From 511011ff6557fe21036894a85e984b60c942f394 Mon Sep 17 00:00:00 2001 From: David Khachatryan Johansson Date: Thu, 30 Jul 2026 13:28:51 +0200 Subject: [PATCH] Phase 4l: certificates on firm footing (multi-seed guarantees + graph certification) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Puts the project's core contribution — the distribution-free certificates — on a 5-seed footing, and shows the graph certificate is valid at its honest operating point. Additive: no change to geometry features or the falsification test; include_ood defaults False so existing sweeps are byte-identical. Halting (CRC), 5 seeds (T6): new experiments/run_halting_sweep.py + aggregate.{halting_rows, aggregate_halting_cell}. Arithmetic saves 57.1%±1.7% (single-seed H1 was 57.8%), graph 80.8%±5.2%; both halting risk <= alpha, within budget 5/5. OOD stress, 5 seeds (T7): [sweep] include_ood=true in run_sweep + an OOD aggregation block. The selective guarantee breaks under shift over 5 seeds (arith ID sel-risk 0.077 -> OOD 0.764, 0/5; graph 0.010 -> 0.168, 4/5). Graph certification: a full-fold run (n_calib=1500) certifies only 1.1% coverage at alpha=0.1 but 22.5% at 0.15 and 48.5% at 0.20, with achieved risk <= target at every alpha. The old "certifies zero coverage" was an operating-point artifact of graph's ~30% error floor, not a broken certificate. Bugs fixed: halting aggregation dedups by seed (the stale Phase-4b H1 seed-0 row survived under a drifted config_hash -> n=6); plotting._selective_row/_halting_row prefer arithmetic so the new full-fold (n_test=1500) graph rows don't flip the arithmetic figures F2/F4/F6 to graph. 115 tests green, ruff clean, paper rebuilds via tectonic with T6/T7 wired in. Co-Authored-By: Claude Opus 4.8 --- .claude/CLAUDE.md | 9 +++ analysis/aggregate.py | 75 ++++++++++++++++++++ analysis/figures/F2_risk_coverage.png | Bin 112741 -> 116398 bytes analysis/figures/F3_coverage_validity.png | Bin 43788 -> 42631 bytes analysis/figures/F4_halting_pareto.png | Bin 44200 -> 44333 bytes analysis/figures/F5_feature_diagnostics.png | Bin 85326 -> 89252 bytes analysis/figures/F6_ood_stress.png | Bin 47197 -> 45788 bytes analysis/figures/S1_k_restart_lift.png | Bin 47732 -> 47537 bytes analysis/make_tables.py | 61 ++++++++++++++++ configs/experiments/graph_halting.toml | 36 ++++++++++ configs/sweeps/arith_halting_seeds.toml | 9 +++ configs/sweeps/arith_ood_seeds.toml | 9 +++ configs/sweeps/graph_cert_seeds.toml | 10 +++ configs/sweeps/graph_halting_seeds.toml | 8 +++ docs/DECISIONS.md | 18 +++++ docs/EXPERIMENTS.md | 46 ++++++++++++ docs/SESSION_HANDOFF.md | 20 ++++-- experiments/run_halting_sweep.py | 71 ++++++++++++++++++ experiments/run_sweep.py | 18 +++-- paper/main.tex | 3 +- paper/sections/experiments.tex | 41 +++++++---- paper/tables/T1_main.tex | 6 +- paper/tables/T2_kablation.tex | 4 +- paper/tables/T2b_feature_ablation.tex | 10 +-- paper/tables/T3_repro.tex | 12 +++- paper/tables/T4_complementarity.tex | 6 +- paper/tables/T5_richer_geometry.tex | 6 +- paper/tables/T6_halting.tex | 15 ++++ paper/tables/T7_ood_stress.tex | 15 ++++ pyproject.toml | 2 +- results/ledger.jsonl | 20 ++++++ src/edc/plotting.py | 22 ++++-- tests/test_aggregate.py | 68 +++++++++++++++++- tests/test_make_tables.py | 22 +++++- tests/test_run_halting_sweep.py | 41 +++++++++++ tests/test_run_sweep.py | 12 +++- 36 files changed, 643 insertions(+), 52 deletions(-) create mode 100644 configs/experiments/graph_halting.toml create mode 100644 configs/sweeps/arith_halting_seeds.toml create mode 100644 configs/sweeps/arith_ood_seeds.toml create mode 100644 configs/sweeps/graph_cert_seeds.toml create mode 100644 configs/sweeps/graph_halting_seeds.toml create mode 100644 experiments/run_halting_sweep.py create mode 100644 paper/tables/T6_halting.tex create mode 100644 paper/tables/T7_ood_stress.tex create mode 100644 tests/test_run_halting_sweep.py diff --git a/.claude/CLAUDE.md b/.claude/CLAUDE.md index fc5edd2..99355c8 100644 --- a/.claude/CLAUDE.md +++ b/.claude/CLAUDE.md @@ -58,4 +58,13 @@ inference-time descent** into distribution-free reliability certificates. Read a wash where geometry already wins (arith-IRED +0.007 4/5) and slightly worse on fixed-bowl cells. Aggregation is feature-set-aware (`feature_set_of`, canonical tables default `feature_set="base"`) so richer never pollutes T1/T2/T4. JAX for spectrum/connectivity lives in `curvature.py` (inv. 1). +- Phase 4l certificates on firm footing ✅ (T6/T7): put the guarantees on a 5-seed footing. + `run_halting_sweep.py` + `aggregate.{halting_rows,aggregate_halting_cell}` → **T6**: arith halting + saves 57.1%±1.7% (was single-seed 57.8%), graph 80.8%±5.2%, both risk ≤ α, within-budget 5/5. + `[sweep] include_ood=true` + OOD aggregation block → **T7**: guarantee breaks under shift over 5 + seeds (arith ID risk 0.077→OOD 0.764, 0/5; graph 0.010→0.168, 4/5). Full-fold graph cert run: LTT + certifies 1.1% cov at α=0.1 but **22.5% at α=0.15 / 48.5% at α=0.20, validity holds at every α** — + the "certifies zero" was an operating-point artifact of graph's ~30% error floor, not a broken + certificate. Aggregation dedups halting rows by seed (drops the stale Phase-4b H1 config-hash row); + `plotting._selective_row`/`_halting_row` prefer arithmetic so full-fold graph rows don't flip F2/F4/F6. - Next: E3/E4 (a 3rd/4th task to characterize *when* geometry beats softmax) · Modal · scale-up. diff --git a/analysis/aggregate.py b/analysis/aggregate.py index 696c0b2..04a60c9 100644 --- a/analysis/aggregate.py +++ b/analysis/aggregate.py @@ -93,6 +93,18 @@ def aggregate_cell(rows_for_k: list[dict]) -> dict: dadd = [m["delta_aurc_geom_adds"] for m in ms] if has_add else [] add_mean, add_std = _ms([d[0] for d in dadd]) if has_add else (float("nan"), float("nan")) add_ci = sum(1 for d in dadd if d[1] > 0.0) + # OOD stress (Phase 4l): the ID-calibrated threshold applied to the shifted fold. Only rows run + # with include_ood carry these; pre-4l / selective-sweep rows fall through (has_ood=False). + has_ood = all("ood_ltt" in m and "accuracy_ood" in m for m in ms) + _nan = (float("nan"), float("nan")) + if has_ood: + acc_ood = _ms([m["accuracy_ood"] for m in ms]) + risk_ood = _ms([m["ood_ltt"]["selective_risk"] for m in ms]) + cov_ood = _ms([m["ood_ltt"]["coverage"] for m in ms]) + ood_within = sum(1 for m in ms if m["ood_ltt"].get("risk_within_budget")) + else: + acc_ood = risk_ood = cov_ood = _nan + ood_within = 0 return { "task": rows_for_k[0].get("task"), "k_restarts": _k(rows_for_k[0]), @@ -126,6 +138,12 @@ def aggregate_cell(rows_for_k: list[dict]) -> dict: "ltt_coverage": _ms([m["ltt"]["coverage"] for m in ms]), "ltt_selective_risk": _ms([m["ltt"]["selective_risk"] for m in ms]), "ltt_abstain_rate": _ms([m["ltt"]["abstain_rate"] for m in ms]), + # OOD stress (Phase 4l); NaN + has_ood=False when the cell has no include_ood rows. + "has_ood": has_ood, + "accuracy_ood": acc_ood, + "ood_selective_risk": risk_ood, + "ood_coverage": cov_ood, + "seeds_ood_within_budget": ood_within, } @@ -172,3 +190,60 @@ def _key(cell: list[dict]) -> tuple: return (has_add, len(cell), max(r.get("timestamp", "") for r in cell)) return aggregate_cell(max(cells, key=_key)) + + +def halting_rows(rows: list[dict] | None = None, task: str | None = None) -> list[dict]: + """Latest ``split=='halting'`` row per ``(config_hash, seed)``, optionally one ``task``. + + The halting rows are excluded from ``selective_rows`` (different split); this is their + multi-seed counterpart for the adaptive-halting (CRC) guarantee (Phase 4l). + """ + if rows is None: + rows = read_all() + latest: dict[tuple[str, int], dict] = {} + for r in rows: + if r.get("split") == "halting" and (task is None or r.get("task") == task): + latest[(r["config_hash"], r["seed"])] = r # later rows win + return list(latest.values()) + + +def halting_tasks_present(rows: list[dict] | None = None) -> list[str]: + """Sorted distinct task names among halting rows.""" + return sorted({r.get("task") for r in halting_rows(rows)} - {None}) + + +def aggregate_halting_cell(rows_for_task: list[dict]) -> dict: + """Aggregate per-seed halting rows into the multi-seed CRC verdict (compute saved + risk). + + Dedups to the latest row per seed by timestamp — so a stale single-seed run under an older + config schema (e.g. the original Phase-4b H1 seed-0 row, whose config_hash has since drifted) + does not double-count against the current seed sweep. + """ + by_seed: dict[int, dict] = {} + for r in sorted(rows_for_task, key=lambda r: r.get("timestamp", "")): + by_seed[r["seed"]] = r + rows_for_task = list(by_seed.values()) + ms = [r["metrics"] for r in rows_for_task] + n = len(ms) + taus = [m["tau_hat"] for m in ms if m.get("tau_hat") is not None] + return { + "task": rows_for_task[0].get("task"), + "n_seeds": n, + "seeds": sorted(r["seed"] for r in rows_for_task), + "run_ids": [r["run_id"] for r in rows_for_task], + "alpha": ms[0].get("alpha"), + "tau_hat": _ms(taus) if taus else (float("nan"), float("nan")), + "n_tau_feasible": len(taus), + "compute_saved": _ms([m["compute_saved"] for m in ms]), + "halting_risk": _ms([m["halting_risk"] for m in ms]), + "full_accuracy": _ms([m["full_accuracy"] for m in ms]), + "accuracy_drop": _ms([m["accuracy_drop"] for m in ms]), + "seeds_within_budget": sum(1 for m in ms if m.get("risk_within_budget")), + "seeds_risk_monotone": sum(1 for m in ms if m.get("risk_monotone_in_tau")), + } + + +def aggregate_halting(rows: list[dict] | None = None) -> dict[str, dict]: + """``{task: aggregate_halting_cell(...)}`` for every task with halting rows.""" + return {t: aggregate_halting_cell(halting_rows(rows, task=t)) + for t in halting_tasks_present(rows)} diff --git a/analysis/figures/F2_risk_coverage.png b/analysis/figures/F2_risk_coverage.png index f98c466942946f2223130b0c83ea94155f0f8522..e618929bc144b557764d6d175f8b7372f8d5e9d5 100644 GIT binary patch literal 116398 zcmd43^;?wP7xt|ZA|297N_R;PjUWgJ(jC%DHv>pF3@RWh54^|w1L}cu4Rh^l@3qeLS?40`t&%JjIvM(d2M@4b%So#~c<^Wqe7!+K1>dRM zuKEl95^|Q&a(-uL?(AmlX!bzS*xCMro%07P6Dn6TM<**gTOM{UE_Nx z8jY`wtUO=zqLmVVO6rfvQ2(IEFC58Xf2m|yqq~KaIlLdXF8)4Xr{$cnDW~c7Zd-IE zki%+}-b!9U@Z2Rw@n|=8O>p01=sH*{-Yl%ch?4{pHK4<&pAda*$Q%ptPsDeSNF;3k z_dlF#G_PRf|GWr%ry&$??Y|dEv6~<_{Pzy9i2rXc30(i%i-p>CIVg62L%X@TIa+P? zj!C=neeGmiQWEq1-F2DgVxc+}=HNAZ747|G?etb&hdYuxjlGZKo~FhfX&a z4%#)=abMZ>b;$T_KMZA4zmsGgsN_!TG2(PL9m$mW{O#Mf^(Tc|6`A8Dx>H$b5ch+0 z{0A8s8LBnb>P+u)e|b1{($dn>F)_XNxxaO5_!G`jSyM{Dpd=TDPd9#?qmUrKJy{Y^ zoq!r(*WKGIpCaV4cCtBEY1)TPmo4JG2=PvZcoPSPgeW*WS3l)2&|}stWu>O3{;T@F z<7{_s{db1cR`NutzHUmFQOD^_rCCI>kc)g~W@h%QD6*mb_ z>!$UA1f0=Wk|R1a$AAu8l8Hj~EQ#;yrW?OAhRn`(W=vadjs`T~Ln$Q(2M3tPFRezq zueOTHh9VU3@$fRvF3*>IhtC?LDaH0g9xEiUWBt4%#k@PXSdA&V-rvki8#+0*y%%Xx zK0XVLj@C%K#1Xw{7joVC*1%~Cz1qyvh$7?D8`1{1Uff6F^-Eq(u7C1Ur}jhQM3E*3 z3o9#Ymy6XzVch9zjF_h$JIT_-e#?1kl3PEF`fn0BcaL;GR`7?nEuJnjH8u%q`ENT+ zk@d#v>S`G`(B-2saWVx5z?(2Try#e~u{w66%quW0F zj~_p{9oOWn#&Q*FC(UTqg6H24r$U&8l%(U~(v~gWiy_$0nJIm59O9QyoW%*E=>B*u zw?Bk8_k=+a;g_Mw%$|t64&t?!}Alt);fM8l#R!ED*?o;K!*_eU1Ia zmK3~iT&BHxoW$*5eHwlcD2rr>It4t&ITMQEz;m=)Lb*QMtpdBjERIcg^74{G&@tO3 z<9L9hO<=gM+_)=w#J-id?dC{(;vP+3p zg(RdeM!U`DHvKvO6kd#FhGfvV%c%)`yY}8J$eXX-zMTrbat;MlTi3(QS#v)axZ-J#?Wd2S zkqOWDjtC94Z!iSc#TU3urLAJ1o*bzasLFiYj}g0rvFkUwCl4o}d1x%5Bs+P=Ym^z3 z46!X9i+d$hdZj#S*{`Zf7KXan57*gDDG{-1!B&os54vy$=Iq)^_Nb$)z&nYSd4Y{K zw2ww}&=rg|z=<UpP}oF8FT0RbA9~7&SAliv%GC0Ytkf#AGli_ zR(mHGTfC~gi}RFHDpP&$Jc>+u!bxv3c4uprZz zmBsK`J(Q}C7*oKu_Fd)s*DP#oY_c?Kd*?A{JjdE-?|rMB5UYx$KSWOc@`F9QnC-kk z%<;}*%begaJk57+>MG1++f2|RwS72kLKJzmsnKrkhau^I#gLMc`0Ywj`A>IdcoZaBp_h&<%*^tNQn0ClcYW4Ntu0qE^?$r4 z*Ndp&A-WO$(a(hoJ*qJLTR|Ea`bfm4qZanVd@zYA_NP~k^~BHeXPNOG=AME+mUy(G zg13w735K?G^z=drua%T)2n2MWT|3v(cj@Qe=>uk{f7Hr2s)DU=uTCdK z>K0tTuX9}K84DkFgGg1DOU$mP#dqbj)s%!vckHl$g{q9m!|#!9Y3K8lDc6-xP7e4@ z(f%k#4BM18{7p8bZ2Gh+@j`%}zIU~G5&5!a%AjZw-sjmD{hWO)RD9?g7j1y=BcPY^o#T!iR^4@((sGFK#Z6Q^vGXBQdW? zD7;Tp63oKzk40&E$wN_`|G9qp+u^>h!L^7i>|U; ztP>#=&!xwW$z>yEWc7(7WrkvWl-J7~$d50Iw}Ov|4WSvWzP-`UZLr9A<@kIqDOPLR zE{^qYM`S4rE`P?cz|UPViBS=@EG%_ib@p9F@szMSR}aX4Ra z6Ap@iv^X;GxP+e;Sl7@hIr^to|QuzHvK#(_No_qZE;f)u3Xo%1NMDIS(kN3Y5; z%L<8ZBH>^L&y$T6b(jVGq7ozO;YZ5Ly{3I|-)a1IsmI@{W^K#=+Y>1M-&{tge_{?5 z>_wpm(#l5Xe~rhScfR-8o2x4$nZ3yw7?o-c9)JUp@$cEp@^TyJ z%+DKaU<`N2onD^!{^{Q~~ zrO@@Kwv+uOL$h&rpkdU4{kF^649t$wpEYn~dXDF-907?_*xYK`8yVZQ-#WJ)v)dVj zQKd&YCoOH`oEs0a^L*K=PLpfN)qZMRH#R9lzsoNi5z=R+O?yJ{YCojMKf#igAYX!0 zvi4c!4|tDv_kx;xk5D>LLeWyVWQetE?-`21hS>4?Po)oXM%su9Y~JE5NE~XlgIJj*Fc)6tDXL2DO1BM zC@*gpOGPuxJf2Z3%Rzy)%{vW4m+pMr<_gDx1=56$a2IsNC2`ps}noKMHR2ks4hE_+NF0jU6Nwp9+XmV&w1b=kgF-S!>FydT1`{RZdYu~Uq4S*$O6ZZpmp%wf^E$^c z<`FOL(KK9H>{bnA3)qV2xM0S@$g!ELL8!kc{wjRyqS<(O?qf_moHv4)13RCAS$K-+ z_qFKn^-BN67t3CLUR}ETq*xm1D=RpSu>LJ=W8K@ay2gcLgf!y=XcV8(OZP%W#3Q0qmKCean%1wz_Jw{FA-=p=^-BH<2MlR~A zcUVn^M+f&e)GPTdp8K&U5JqDBP9y!mmD)SUsd$chVUw|%`p6SUB(7#^BUKvWs|3~D z_KJ8VTkb`Md5Of6KW5ZVx1p3X@XpQq2xXrMJvq(tce)?+dbe%}Efylbp#+MB_mjoi zKmYopbrf0cpABL+;?p8k&3~jz%xGD44)Cht=RA0~&(OC#gvg-GXU0uN2A&VA8D_t? zZ%vkbzfxBAt7CL>9f5az*J{^jKRrOc(_MDQ{hpdGu^#2*08Ml_>>)0?hZi!lL&Vj} zfC38><>GWWXkqms7Y4CmKqdpnQ)-a;{WtS6x}ED_&lw`2#^l5J(P+;JEJdM z@a?WiOR?O_iV;mR_PM^6z88_XIq2AWLnz(r;+*_2)@6J0w4M^MhO+SNv93?gJHBtA z43!KStl$VG?Y~wd$ax_Z!C%Iq;`2&A=U3De@^;O^d>$IO& z8GWAa%<$BA|82uZTs{MAb@}i}{*|jaN*{s}%YzC*Lfb3snPP26HoYR!_c9Ja7#9- zLz)t(kX-#VX>Ft*?6{1WzsF|iyB2QKxcyT*^^OCNA?3|S{+;(J-g^R$SKu8 zC3d@~q$@^r6ee>rarZO-*tBOf(R-xAk07l^?Jm79{q;k)^TtrGjyA7{bB2q} z-Gc3|Mkt%>wl)4EJSuANtU{GdCBgOlB>hE$@Kc7&tz35a-klNU!f;5Q%f^;pWBSlY zmTcAhQkw6bcHG9$&O8x|#;=?MSlQHaQg0L&Z zW)(!Hj^n6=z2@<*+IomiAC3xZ7LfF!7r*~#?=8+u5=l5(*k$)i@x}Vq~6pE~rDx%#&KfA54CCoj<`8!UIWkh(0r52^ROpDZ+ zi}@O|x_$aW=#90&QMz)nfC?67c3u`JgMONH35B*FEsZS;{q1=rv5#^s8~BAcf=#zp ztp#nGz`51RZW9klu+>OnAGwy174xa*;rVfC-pmMu>f|k_kAJ7oJnhVDtE7ePryJgB zDInE^3XE7y4p<~VgcT<|5x{I>tTVXU571FiyS}=Do!?>UJa9i-WY{x`3)#d{DQ7W1TDCP+*iEp%_{^~A#T2wbyBqOks_XS)ekNkpBv9P%(mSf$9 zpPF%6{z_LuK+E^!&7+Il*)Z-H^qk*ihHWAYy?)j1?(r)#7HNaK3tVdi^$8o#p{r5+ zktF=@!$Ir>NRxWbh$~~7Km&b;RO?Ou8wu+HkgoNx>bvM+2|es*fh%1hMT~-z!A3+> z+XUpVUj z2nqAYSfpnB233FnDeScbohmhV{}=+Zl5zoI0scn3ykF*jH}cQtBL08=2T=h355PL& z(f0pGX@m8;5w!2$cc2EBh{PbU+3Ek@k#kM?P0oP#M=#?40IpzY*#D1c?*GG# zjT;&q02qTTdY|hzT>Od10o-Bh6c7UTAPnMoa=T_O!{pf|pKC>c2j~E?Z1YU{`t=1u zV2OZmrJDOD@zd0I7>o*0Ev{AcQG8L&k#Be-Y`^@e`E*iuG+%|bVYhZpFSnQgofzPEGu|tS^5>Q=h#CJEx%A~@s#m6qv(t-HDhDDCw)J|RlK-wQp z$if}!Z~-{)S1R9wj>mBi>bh)lD~g3fa40p#z_$1c-3Q74vrGih>&v5cIw2vQ=Qt+E zx93Z4+Gl-~zF)T{iUvSxKwwUJG~c5i3<&yNBQ|5@uIIU#Q-~L^aSPT$-P9NWY6<;# zn{M9dIFGiAKNzGU?&VW1WBDFHBu-pPCeJY$G}mL2a=+-h274V=R)Oh}1e9c9$jZ$! zh<0_TI22H&+KCLQ#|R+h;87+ULoqqXzw}|9m;!nrQS@=HmVLiOziAR33#-O;XBzLf zPzkP?c7z}ud7g~OPF#X?=8~8iA8)Q+ zqBC)IGN$Y;iU?2wOI-1UQl(PFw;XLZnKX{Ob9D(|_Y~wWgA$V6*w~m|{|dA;sy-|< zs>{6G6fM(hs2peny@0&Gf}r`3Pc~F8 z0jM7r3O!q}@4`|||763UYJUAm|NTHboAq3b$Wf?g%;xm#6W1_vO~qq0MqdfPc?U04Bg<+Z&ZVD=v3#A7ilU3f<#74?w5Dl z7cG%}6|ZX>3mBp{V1n6sc`QYeV12lE`oI=7J6ikUxbd6Xtgl_;9XJFD9-*E1GW7V1 zO>TBEeF1TSrvwr{yERkYhOL)>$(d9#1D!WV>AI=`3e(gl@FW9`h|;0`bcui$->NM~ zBoTzuB@ORk-^+j=B=TtC58sao89^FGz={Sa>K@3HF1Ax8x_Lupt_jZp2ydjAGW2a3 zI7n5Uek(HE!~;A(4*eE2pT$!DuK_XFplC_G?_>A01w$5%ZtJHb9B+UY78kWu$=JnM_v3bvEoIZcm(UH`XTwE~a*&>Z8t!twC&KO2)h zO1_5m0n#~9F)CiDk`;^`0SG{W>EGlrcSOO$-7={s{0%^M&-bWs#W|01=!fZYYN-Ip zB^%zY<|mH6@wb47j5l=t>}+YbQ1C@v$jH#CFp+71r+TT;4wLH{?x8?8_o5IFqdx}Z zf_}b(AKY3vT))KG#eX&2!w6o}Cdu>0qUA=HmrL|Q@>4bS{81+kF>qhWIrJ2YC_X~F$81Greoo=;lu zu51YTTw;dijJ;Abf8V9@FF!l{ZI6YUy+d%l^6_2}E?!OkuylolnPGB2W zykBvL^l->9{3Npit{>V7WD>>*Zm)UT+kP_0C{B0CCRbYgea`#4Tv0(`fRjGcxJz25} z=IJ1eSkdliqSY(N$13qpSCu$?Hj}?j+-FMlBW)Za)&0fr z=5aFv2ZTHSp+_hSTeW7HLCX3a66BeyCF5`54C-Ul&fUX4O1Fj~J-5uY(At?Qiz4rC z`s@s9zkImX!pL=p%c(4khmpW5DsNT%t>d0cz^7YMPNPrwu_yEI>1bLcx=E-@>IS{@ ztNv`bjLyLbwTj#?BCZ=xZ49IPrp%_w3^@=@tEo5pENP~t7DXB*UvERal8&|NY!_JT z6tdG6NtCA+LGC=F1<1+~tJ3Tf(ND7TsViBJtCP)2&?ZQbW2M@v5w}6O2KKv zYwIWdn1<&aON`8fIX4A$W({`nOg*Iy=##0<|8oua%uZKob0dzKrCSuWS_Bb9M9 ztl`r&7pfIx2J%ZTio)E0tf80+t6$Z#Fk7)_;O15hQ#^N)wd+9i>2X93J`D(Y4Dogg z!9y%XT3BYSO+rwy-~aV6*De;+3MmJXvgjq?B+1F|SK+qG9_hEfrE+7FS*+Vi%a zsmz%=y{OueA=b~PxpzJJq_`kp?Bmwfwx~(=Gt$(a_)g3G1o8ps-pkZ^rDGE`gV`=A zavUu~Lx>LKjCC#%uFSSMNXBQq_BnSSwBa=ZP_fj6ssjU&PYpah8z(^PGw#uF3JC4O zQ!+xeEM?83iAwCCdDaM2Xo*saP(wEQGjNawgg}EQXLM`^(U6}F0jl(>8RvJKuWylI zn-A&V!)pEOT1iRCVvQ1=g5wp89j8QYGwx&on9hh=|_X~ z&&KCPn6u}e6^`-6!OGV3Bfr%~0D|=wz0PJVN`Mm80lje2zaW>Mt0#4W6qY|#x9C}< zUZ}>jjp$l=;1du`2>w}IZ?=d(gI#|)9s^nVG|7uqL}cD`r=l`C|)4}7>%&-ct1kPmivkqu{H;1_d)OlJsJ?;lzOaqqN^2R;d`FczU(59LCTeGqbrJWALR)gH?zCVBm-F6bk z+^7EHzQ}1XY;xN#^%RuL__1unUH^VULDKHwDk_PX_?37m#?0$_yLy8Jeb>ZU4p9vE zGM<4~p9wJ|=`hgKIHJE_2RYz1TRgUT(igN4WPt1&zrJYvmLX0-Fa)TQ1k!e!{gR`{ zr4V?}5(O?onixKjh=ZRgku+yLCO+y)A!K^O#J%@-c>FlNuKEl8l>7eYAq69Kao7ub z`a(76^vnLi5`}0RTqQJ0RM^OZqGkiTN~tkv_X7j-ujT@4#KuG_D0zdNteR^!>G*4l*Z-Dnzk$ix|*yTE7?Ri!F4d+o`yMw+~*}^eZUa!~MO}zl-g`lBOtL$ioeq%{dRVD?&~~SVhUm zpqNmZ=`U8nPNtM17+SF;-IjGo56^`pzc*idrqd%CR#ezO$6YJ_;?&*H&yHq3#5M)n z;c~T!Xf^-v{TJij+<>+AwaZG?0HoW;!i=CxYRFsGv~Lmj4|G(#tFZ?@3{nF-awOHT ztp77V-m~7}7i@y+WM@|N#0!%WM?x1)uXB21>WP?vVMoB|;<6Tg%kHl;eB@R&gaZqIab(Nm1 z7kl=_FH(PfL!sf2OcD%aOeE=_lyDQkTigjPIey~kLtxdz+?g)wyPRx%NnPlZf##Tb zCjGU-?R^MHKgzda01gK0%>~Z~bre1Q%xwaQfu0=VWQD`Js{0sz#CO*!cv?QJo?GjQt)Sh=2^7dfyON(pG-!;z)6)*~i%$=e)op z?+gl!9w+;U5Y*dqsS_WOs#tKByhvt zDT6+=BWO~@Gp^L9-@nZ@nf`|}9w9DJ=^6ezuOxsIjzsz|f6&_i8L6zug^col#xV-|8o-xdzWmsGhJ~6%KU!N9bpD`i!aAD{Wx@O`;FVB z&z2@Zw9b_OCU(${B%uqJ!hvLuDuW7fo_7qeChn*E>%oUSvg_Own_2!{%X67=V>#-c2xN08We#RTJYtLfOSxzCP)OJ z7EFP@PARhk4hs5HnB)4-*Pyae54!|-u94MiYn&;O)7YQcN*QGyQ8v>h=+WpmxeiV` zI6iy2cL~(;#LMfG%`A!zy?F0D6F3N8>K)||kx=u|TN4kv36%2-Yo1PLTE+*5w-v)d z5&UG?5^$cCc_W&&%gE5oSbSBgwdOF`@#=iNgXv7Q)t5r82IoyxLBmUgP;A(D>S*>j ze*fb$orErB<~4vO8l7rYK<;dgWQ8zm1yTV=f`Rz;-@7ZlR0hLZM9M>`@E5>ymJc6Z z%tv&Z)v}yG&x17gco8%$A*F7hh^zEGwT5p4Kv3Z1(S^m|56oC)rGp?Lm4;j8G|+Q2+*?7sBPc;LVl5O)`KcLY%H`gG>wdWF=_&enG_|>;KC1 zPHi<8*k}rq_5t*0E=^dy*~p3+=Mw;|W}q7Txln(;O4iW!J`p%c_gr{KP*f8U8X;9e zR3XZ6B-eYVx(!r8J}LngrKmE(v(zW8=6$j;GISZ5Srrol zIn!#NO9i_729)uID(U2uW&`nYz~Yln@s#Z0jDGxe)K(3zn)7&@WDt7cW4~8&o1;1U zlT8-G6r`fidPIAc!fWd1S0-BiMfa@clTn2uK$uXVJBSGJ1)Kz&FlG(8u)|~O0d!aY z<_G$lptqP)iIkm_!?cY_)3IQ%jKO36>#fP7?;&}oRG&lVA8BCI47)9QOO$X^8nYaKLWM{^a)YYe&-&42S$fYN}l zWSlA-m!nuVLah5mY=80DE!Y#&4NiGbfNLtv28dj+%R}rbL^^7`SAvMg9ipBIs?JaY zVe?gs?h$z?)w%N>FN0F@+xRDJk=kNJ2+1K|p08dMf7Rx;Z*ZEnijISm05mjDS2KVK zap2^NEPqr4xOr>5V0Dz>lU(a+wD7pYX!(!5O3g=UHPqfrNb>0t*BM4;?KI!PAM;*N`IbsQ7d2M| zHOKg+c)bbFgUmMV-RlMsWu5WWxmV>YGOu=Mj>=({8HHgH5%2mWkrBm*e58#fZZjy& z<;O=^mT^*t(#&rz4!7ED$LE0oksFwzPQ|qB4J_5FZ5MP}Qi0VlykAx(<+-v+9x#n< z-8Z^yuYnfwu;VSI&qc;p4nqS7eBP9yqRaIlkB5y1st>Adotcrc-LvN{cHjB!G~h@|@_W_8()MduV9-^!EA0lrwKDNs7^=)sC`NMh6;`mFUub7F;7~vS0 zP*WIbtj2S1nR(6&tPZ*C8O?*u%FefXWce0X>?O@G+D4T9?t zi`AUb0jr;`{8z?|o7oegX{*`lf~(=fE0%spiWpW4y7%>&`b&a#`A)AHyftZ$272j` zI=iLP=qMNKiUAwL4_wnS&ZaCp8!5k^iTUWsG(Z---V9(GCgkv=4fD51VUfR7>$_a^ z+83P4kO<|0ed$f{ri4gvs35#lr+f(dz9s|cdU747Ty(#d`7-2oIRJ+{#bUPFYBWcI zn3x3_PqUj`)Y=!hfn!@j?oW={r$x{dh;#4GtTgUI`KI#+bf`kx36jTRy(edy@+^>?#_{;K;{&!dK=M-8pu@Yu8W6Z~hwX#}UwjP;vXoAo_OpUJaPVW=oj4BGv;GYj-v^uO7 z6XiMBqGw>Zc830YHm0@OyymGiE(0E9g!|%TRRH5I*%_(A;lj5}Ex(?M;W@KmN zEzNnyeJ;Y@SI)s+^P>3#`@B<$sgKCsOSgm}Uyq8TcUMXxB^+;J?>FRmIRhln5EJI< zwA*klm`ksil)pOp@`Y%X``{&X+sSVNlVm=Oucw7tsVp+@dI&)U9>IaNLg$r_F)A{3 z@2-BTbks?VoJKjT*qaNb%h<}U7$h-Q%&(5bZqrEfSx&B@{V6sQ(Z1J^OX^9p^t*+YAFc=Q+CXyaY>H>-M0KR$S zpzIAI&{k8xKCLj?o7~NgglB!gHpWw&;MDF}iOzYj5Y*oHwxkyYZ#ekByO`avFX#H*MgLgni3iv;KGn0FqifsTU< zD`Xz>3RXaPc|6({p`)Ygu^!9CVxA8V4J`)&$NY3-q$@l!Qf;5_QyO`m9}9 zps}I>df_Fl-q$)H!Mn5~s#hH=D=YQ^Z%~9<>j0;2F%UQJ-4XF_Ms(;O*e|xDT|0Jl zwKcNkV7dBm8=7Iz;yDQf*6^jPu=9oyA#foh9J|LYK)oz7t7eC~94t$;O?KVFW;}t1 zJ?n3xfJ3HJDu{j(EpF>^nuSv=nx)l<{zrlf4yEXP;k&%(W4`k%fbn-KcLBCX8^3G29q<9g3(No3cCbBN%e;lpTl*nrCwXf2h1 za@#_H1NhKtGPANK8l1&6CK1@br?=PKvlI&PZM(ZFnw>&8TfX@M;b1)rvRttNus&}UCB{J>{ax77$*hEfJ=&4g>fjizGQ3Qe?IY@WFk6%!A&&IM+7 zUC|PZj9U;jbbv8)fO8kn1~tRAVmpweDNqDWX!6jcVSOJ2*PgC55q+#W;x5uG(-4~% z!F>uG%TBFIN=l*(*PkpAZ7vrerL}>vuoht|Aah;vy>D6Cme{e!Of0eJ4#m^x@JCUn z{d~!ymY40=3|)O?{^HP7S<6`xAt3J2Wx#oeFkVlbPlJhomC9s(iRe%>N`z(l z^lG}obP^#0{0V_5(3~qYXlWD*0q*|VN~Ea*`<2eEW+yvGQL7#I#&ZyG*3K(=#p{Ei z6b)_{2I7$xghoH!z?8pnO7JE9`SYjJ!#{&(9(z zH=`{YpdNYrNwO(*xJq}zzbTq$aj&hATUInT$`G@<`BY4eSx;;D^89%`v)wke6!M+c z%^&A19x?qRHNt9LqJu|OO01}^F#-XX2SnTK?kyjoqhzW?pRr@Fr*zRpu!79W>8G94 zct}3~;!%5jVzIPvJ~-Ekx-DWghn{MxD)Wb*9xJ`>xN{l@HSp-UsV9^z77ja(g9W2ACa8kpG-+ysl>VQ8mHCy-o*S92~z&W(eR!)#c0XO$b0)QE%?bq zYuSq?*;=) zwdQzX06mnMusl$kT-0P=J5HD~UWP=;m#R>q6az7}q6a%(>3jC6Lq60YJlOqbdAlVP zRc8}T*HwJ}-ScQm>>yc@j(i_fyjYw>Yx^An6uo{Jj69GC^UHx`^as2!xI~ST)Si@& zrKcZsu=aBrm8z2Z_>MhwT3a(VM1i*!I2Ax#PCVuwKu^S{s^ZYf9a%4fV<8LSFw#go zrO*v0y4-^$`my5tis|{(+#o%+Nt8D?nV@DZgujQqaIi=C*qgdg>uYv5gF@#I_qK$V z-|gXnB^~6`FBt3guqxdyK9|sxNx6n&#DX66V>1Zx*5$x9a?<_+$dp5L@EoeC1}M)l z@piRN>(l-NnsBrYYwW?=n6{u(E4;n&{!ncPVG;VKwr0UrVEnK`g$p7{zS zVwPvYCV!;>$C(1CK7P9Y89d*8LbInfL3>M@sGgsjSHf|Z86<%8#tG>;wQ{(^=M%M% zp1yzYqyu&?z=0yD2h8n5U(H<81V$WE2<8{e+aJ89SzOz83HchN-{`C|R{jk(%E=QW zWZ*Cfe=^zQv;J3{Lu4(f`(0Q=B@W`#fOQFmCWSl$n(+m&7-(v%k*IlWm(&$zk~~E* z;iU6uXCg9*fqk*KO7+@PVM5=HJ*}@7P+Gof-LF8*=dDj`(99_ur)%uL-$1N$Jr}PBw{z4K@5AT+T@LWDvEb|JN-38n zpyV%1q79;Sx5VufyEzCLb|s{l4Rz20!*WN_UVnhVb3;G+{rhd4*SG3mk@*7srdl%0 zq~}}|)o2YrWd1iY2(u8hLlJx%7*aWlXwrfqup+e}HN9rgE>kndRHjM9p8~?%*(1$T zXx}0T(n5^{nR#i26IT1>I62=V2%T{JlM-_m%9a>TZpgDYxpDIrL&<_$cL&HhyDkOi z$$vq9)`SQN?E`jcmKHMgv^CRBzAWJ3xLW=%%wAPkoFxzo%=>j54VSlq*jT2rEJ0nIoRBKfTZmc?V= zd)*{Br`vqnwsuY5U_#vQ>=A>5pxr%7e15!wJq1g`_z)o9m%8i~y*=nW>fLDAhj2)P zr!vvN4c4{`#Yns1E!ueHXQ0kZLQ`BP=}p=v(g*n8z7# zZk8-30X_}{CfG8N$+V_B;z8E&)XHlX6Met8(STEnu3mFo|(aVnbE%+iJhO-n#>9 zKo@>~mR$`$H(p>k%*ms^8hq?n?NJ+E#!s{NU^K4E3U~Y&6wJFABdRj+2FG9{|p-OHhHW-j25cq zgr9=u^;L^u>R*rB)2Wh2DCk<-Q)R=P_kcqE;{aK01P(Sc8Ch9f2IW*tpe?~uZ<|aAq#v6>vsO!VL_`(B$Oz6U$AFo|uW!atlwylu;PLGM zCJATzDij!pD*^B^jUwUdlvhxAw+|$>wrO9qOX7Y;2t{Q$F^BE$VvBBzlA>ZwY2y}i z5g_J zLYN_I*pCKH`IAJw8>^1C0n%6pvu$rx+Cf*#j9|GTxZ+VW+{&w%fy3=(?9VT&7#e=G zlTx`EU{LhBW**EpmG-(r!GcPNMT6>l5-G`QG%_{d-jboCycMT?3G}Kj3-Sp8c zb}-NwZ+>Stz5jFo>WK*|y;{EnA}?`A$4XY6YB{!RuvM5lLelFHRENb+8T+~!{W^N^ z(IFvbHQMUU6;dV+3&u4_LaW7y2B!KRNkh*XJMTLS2+M ztXt>HM7RS?O_uz6hX5p4FUU)EOnkhJe$El~qo}M^Xs)lzd#TDo*(Gi4Li#OFZ!o5+ ztMzz(#s|m;-59s&R_~_5q#fI|6(?~W9Wtdo3SUGHaPI|lE0SV5F@bgs=`pri;z!&D zOw_SXSJ2{Peit2V$!|gu!_N7fvRof_-N>Hljk#gOANWRtd8^$*{E-G!5Ux<+IxD=_6a4Y42qxhvIkQacA2ulFI9Ll5X0e*QM`On!o zxvwi4=8AHP3E!r1K&3QvEJ}rUzyxtq)R$+$EkWM>rV4 z$owCyNcKMVotyvzD0w80Ss#F+Al@R)d-9Uv7&na0CL~zGbFI8jJ{?7{6D77=yrL^XBt&ao+_QF;nL?w8^L{rSsE5RxRS zM-+_3Lhi($e)l$6TBu5S|H{12PbMt*Q55$=`Z9&Yg@r6^d-bm- zQ%lDBIdN_&VqquB>Ow;&_DYOxsz0~K!(s!)+Ev-Qb+%jHsPouhbS=l-jI&fU4N6gt z8LMc%)auP_Lx|(H!ABi#TUUul-qgg%f;|o!i`jFAZlq%+Gl)($^8l=hkt_9cc4|m6 zDIF~S&%tXrUG0aN%sT?68d=K*3taR!Q#7+Jx_Y`qxbkR@zLV>zX%t+B6(&8Hv7$%6 z2psC9oZ}f+!0=Jg-ZpTnsqVF2+oq}N(cl6}AQdylXeo0N$4VS(PZOa3cr(+Pssi>h|a*4*Ux$e&z|_mc+ilRUGU#iCSW39H^QbWcwd< zA>m%5$w~7?G3~~ly)ZxZ+MBW6=~Fx-$!t>=z07XD{QG??^^i|z7nlKx9%sfoN4lPD z8%^8SW6vu?8Em+)=Vk3$PUFdxC3?xy8PGA%HK@oU#gXEB+2K8RiGNbpkC*pP!I(fW zMq>ICM`$e?vPc}TKAs6tW`Y9Ld}g zJCAH}WVVl~mwg;i!+xH;Bz=!dzDMHG2ChPM1d?YzRkdPs(L!hXThf9m62NC+cc zj1H(kLBIrSs(RV0(;7+f)$87S#lWfOY=xVC2dFKm7d?Uoqt8YOq*@&uzdN;vNCdrOJDE)_MpZE6an z7WVg*C8HNyG$z$1J?IwWL5%of2GZ=8j|#794t+N+4UeRE*vPy(aRp4fLey=_i4{ z&KeTEQ0KPFIhN{JbpE&yVd!sAw*2O>1XF@h5bWU;lg;C3yBt2Y(RVk) zQ@}JaZKrU3*iDo<_6|kyIq3`T-C^%-tAiJ?b-KKc0A&es-*W{FiEEKi{>l`p{@1&3 z4$S--VyvHg4)zO|?2g2%?R+x0JguL42{bU9e`~P;htNfRFq-}caqY{F&9IIyLl@xp z4Y*5qlvIesZpCY^Wf<$ykw?B%?WRjxkIN6%0zDfDuwO+pBJe3noe*~K?-U(*T6ap0Wc?2x41bV&cfE0U(iGinvjB8XkHE%+TP z?fKO?8_&2P>x;PB)IzIiTACGF)~|Mq$WJfyQky1;72mOgmclTxUAN zpH9UHnS+R)l@-k~knJ4PB*EqLAVX28mpN`lp2QDvGcs{4oqlf9;TxnA)dTJJ=&Y62 zi75pOzmYPlNK1*K(;SOXzn^HX$dI|LydT(1%nb}b=(}5S1F0*c?cHL25=@-CLJ&pg zaw2Xa((6w!D{2ytsDPT(o6xgxR1`7rwSmVeZ&lxdvSXvwaIrs}ihij{({SO|a$BF=EYR6w0t{14c zNzV|!p~XI$W~~E%T3JV*h8Bs9?(1f+7+Vp?!6zqNJC`uX9AuN6Hn#;Po9!QnsF0=+}JK28Y z&JsiU(E{m$xsr_Pw!VpND*+7cHyhTn_NJ!G&8z>1s<#fSYK^+LRk}eMq(!6^LAp1R zf~0_Sr*xNu64EKs-5t`1ba!`2clYmEp7(vP@B7c|crMwjz1Fkl9CM6&;#Z6JAEAyq zgc+_c`%}E7=ld_Vu+~JG^9ANespw5*jBvGvMAsF_-SCtW z{duwiXKmx)#vwijcnLNUU}SCNykwu|RV8lXG2L%cV{-;>5+E&-W!ks}PVEhrcLi==k&493MJ74yj6snp;zEa`X zyTeqbhb>)h4~sBqo7Ay3JjwS`uXFzQ12m%@gPObGVIjX#eh3~YmIS-+wGZ)-zhn1G zecYhpXkR7{7*GSL(JX2oQQq}n3^D#D!Mv|@Y9tQ)T#Q%jsy0pn{ zNdJ z@ON9_KTk%Nzt1gjj>mf4Ux{PP%~cqO_2eq#DbPaT6R1IuD)jgB>jJ4$|1r4BdO&-Tq%X94^B&Hp z!y%4EGZVCi#OHwyz?}FA*wkJ^1fK#GX2>9|nV*CxqOSh4XR6Y1PaU*{!5Vh`9NWz$ z1a;tjEbW2C@p(@5`r_cgL2H0?_J6_~-G*_Re~*Z~ZuU5ptEzzAV%sfvwekd%HwBhu znE!5G+jdb*SUZ?fvVaDq`YC)S&zfL~i`P)9Fn8HB8o4XE$RMn{%8-_cTH_BjS%m7F z+&34_ppq~EA_s%W{{2!Nf`XFyLW2u&8`lsn6twOFI2XvM;hSnO0~(GuO#YO{GvG$z z+vi*r@B|(g+6jO~(2p|&{Vx-T*;qK_Z;=`TL>O4$6@zkq?lqjZNNa}M#Xg%}8*d94Tg-N^%J zKn{gB7Y9yF(6C6oNWH4Zb{dEiYLFXxFadP&82?&!FxWVEE>>L_0E3i5FcnfqK|vX~ z_=ZLC8H$GhFXg}m2$dH~#W1L8d%B#zF@f%MQ6uv5p$KqRl3gmZY7<|_i zDDe~ugcbx{Y5_aH0tgIjoI(lv#?N531Vl{=D5(J;MIj%hy;*nu47>(&&MFoQnzgPJ zL9eFrx7Zl_U?dV0z=o*@*iB~BpwTbLuK?p}yXhCJ)B!4QJgLJ?el@>Rtpi0WxurH5 zA&XLM1V{~a0|g+)qi?=CQj?`_lNvH3?JhQJi~Xi>+kMR;k+ABt0Sqq+CkUv+XJBk6 z1!T6w0~>zlu+=eBCH`SR(MqX)4FQw@#o*h?S`*)e5=anec08`MTGWCPSsd@OL6UTv ze^>n1obC5IZJEdH(EWZX5{hhFXSOx+SDf^_6zI6O|4e8ufo4lZbLg@gqebkKio&3( z%5EOkiC!$edkeqYxzZC{oCbsLr?k7JaGNrG|CGuLYWq?_-PYeEM`Ma|-4%lZI~R4O zMoY{c;y^?-rA^lcSt-NG8XzHNIv=grw&Q<%x${^+m$|Ff|46$6OIO!ekOvC6l-net zy~cF@!PwWdFJEe_g^u2r|-mV@9#TSIBSJpPV{ zbNZIc_uqqT-3Lqph9y>Y{X3qz$^2X1IFYDIOKL{w928B3Z`XJ zT%*X#8VErPTEMcwqjEc|?SAyttaNv-RfW-}$;J!5b)`gRfA{=TBw15rjH5K)sHu8o1WPiUvJeHjw#f!aY z*uyWLNCn9-r+TCF!{~6_W)Ov{G3&YB9S)7!ULNYw7T#&b&3F7+5@ce8Npkef0>z4U zf_>^m*P(DMC3CsJ^(KvB7FGIBd;@YD!c51<+h_rw5%FAm*p5LbF0G-Fc*3R1sU}9W zCMhS9isRn6h`!zS)BCb%EmE3~0KOMJ)<%1PaBgz1xB6D|0-PI4zH-K4TVr`qBxBtU z6A`xiRQqL5X{B;cTD+d1vn$N)FuV#mi6yK5=bU7D=EKtE@S4&+K2yR@-0Agbzh&37 zS&~@qgbENglVcT;Wb-!+cjDrn=yFGDbP`it;4`a8Nc&T$X1wvW0D>ibdRf}kFpgw; z^?h-#W?wneIhqq1@{>81F7G~CFpP}_+&f13Fy`;UbWH^Ps4Dj!a_P~?f*}NKK_&sz zt6w`eLi@#1I4^b?&oAaO8qe=Mp^=OJ4e&%oOGdrmrdd@TE&rq48$?hA%i4R!>nL+~ z!EvT*6y6hIsdrMAWBXhI@IrnSB;zx8xnP+;<$wOOOusXqJp$Ix&AOk+NE!vHnFChf z*9>HP?7PpGCdQLDjLuC+x!kTrMU7WC*0n?putLgRxA<(fd*fJ3?s12! z#Y0_W^LcikZRlyvMZspu#3>(fzpk+X_c^f3jqh@zmyk&@9f%A zO_<59I8bYLm>#Q&`FTC|{P~#BV-G>#t7b^R#lYcZA24}VCg&(Do8f)aec1^;)8ULh ziJvWhKY28>SkRJ!qIUO+^a;&ru)oH~9VR0DeT8Q)Sq;^xz9S$gVaM-B=(o|jqz$&$ zm1gr@F8TKCr%O-c9wh;vLuW3X{4rgMy8;HXSa7}gfgK6#6&%T)jVHV&RhM}6auD17j|I+gWwdidu1O$I zlp$@2p}KGl9s=ANJ-A-lev{mQGRfqfDTN6CocD_ORxBUbL^ z59TM{Zr@$F-~L^li$op46d@}J`!6*OT@b-T2wxcVHj;XPoJySXpZLm6zk$w8N+l=l z7+~e*$G5DsfV1|6j?nun<)3{Re4NFL^Ju#U#|t!l!j>Q(=YjpYNA(EtZvray5-3rn zdSHm5b-rKP-^lzW)V9;C@I=&b{iB^y_8I$wa-<9c@)qPc?FgU?V4?{st9#TbCyLM_ zlP~DU{}xZ--3(IO_TBkd>AR^nHF!J3t0EEzk|jRphcsRF4m;TvY}1lfQgWoeZQLrw z%o_=aDEGF!7XCLaO%I75nd@-Q7Y4WoO^3owzm_p4(hVTRr|w#NVVnE_p4X?Vuy#`- z`Jz!w6@0|vlRM*o1wO0a2QmpCG*FS&yQef(XI>?CZd)9S3us@Ta zLbP9|ejZFTI4zYV+Zz;$NANw;Pe11*+MkcL<&jv!uyZ2ul0UxpOD<0vjA|sl8N9zO z>0dmkeo@NtM5wZb{}@zQ@bONC&4oatc0}KjHHk~aFj|4*qt(uW&pHYv!x+H_k;3Xq zf8MO2r!5mRGh#cu3#Uu84zTxSWG6ZjSrzR3BTRt9pDHKX`pu2-HJ{G!M`8EJb_gn% zh_NKczyL18e99aNL<0PB%B45!x?M?pYkBm>2({-C3COihha$!*K}ij(mp%Wr z$VA7Ll#|nlJ{IpeFOgk3MN$0%W0Dz6yNxrJ<=e?3?Ea4ca7-Y4U^h99)qp?qDzW}0 zw@cSN%^C_Fu8e;7|MaecXttWjq91qJ*OJ5d`h5*)`Dx!PtVH!2LX-vY!MaL(64`AG z2vZLczjC41#+lNNe^%?x80Yz)Pr_pFvrn}M+~#xkqW~TIYf%bhe3jODgKC_6B+`1L zNo}*$bEN%l!QnTnlFrBz)b`)M3!HZGl+D1ElQ_s}F_p@&;L?hGoEQ>5J4!xN;P4`u zJrC=Df7c>PPSfS%NW=D6|2c9hL0u+B#6_1SZ=|&##^P|`5JRw^fh=Azi+2Dgqy5bk zv^p6KmG+=xGU~b+u!S7c!Q}kvph)}6X1UT9$ObBE4zjma3%etiP;ObNL`(B?)0I&k z2vm29d;mvlJyaOutC#`g&vt+B=MZ>TWO`J4P%6 zU~tOa`&|aiTKdmGu$`gH$batzfCD1UaGH+5#=isv@yxyw^ti5Ma2Fc~bJSj=otl5| ziB7&XAfJ~HVu=o58ELQSuy`5qipV zkXA7K@vVy-jDkCZ-vftRG^p;mwQ0V@6S_SM@dsnz9?(o@f;PS!)X;=DQuw-mxj`dE z;-{4@L+k?1rc974&;Q8!;qO2aBd7vCio`RNu)odiv<6p)DK1M(6!@0cAi^vc_(#@l zng9nunWg#G)fJQ9`p*j4aa!7DeuC+CU=#Z_I zAUijrUYYE=O%M36eLNuH-g^pgr@325n9sUOcfr)kQVa&;(jH9UlRUCX&ISH zkc-pN1dY|e#pi~jcLnZO#}gnTuQ?I-^CvJ~F8P}Adn+`1dhBE;e70>04GF30iDfqD z2l)y$mio{0Jdpcu24wemLIF~NzLgI&ciTVqc2*mJHZZjf`oioTK@hhDCNV?)huagD zgA&j*{$mb+3$-c_9YIud$aRKdQL(Wv04G;@z7%E`tj%|k??8agxOo~u*G{`$i+lwd ztVGSTDWpyHobQ?^Bhsw&vyYY8bHD^2_|}vo#Ot=_kP{pkOCLjnw0EolCFi8NjM8v-cWVVth`vn-Cstj+T;xlxClj!<`fN$(3M3^Zd$cl^8;)y1B*8el_gL-s z#EJmL28DS)2)LTLGJ{8pN>EMZQQC)JpEXUv5cDa+xpy)cO_xC+Wfx`)ZqdznCPrdJ z$K<03s6Y|o6np|FJew|tW+?ZYx7WPCFaFIs1O#kShkw~LEl%N~1kx`}Unb6=9F7Ck z1mFo5w@vYbx|lqGpzRv;Pcd9==cz57eV#R+UP>IHEVOZnU@!Z=%%+O?`7_@FmrWsO!;nFkmQ9b>!TJ=5OjJ}T-f>!%*Z7*~FV|&1Vp!X)1!CpaL zpD6moTf*;r^2hG3axr12ugUN{)jLa_6&d(L6(t(noN%S z!p5jjt5H4BM()H$a7hw++llVuoFa2aaCn|x+0dbOh1Dm4`<7vm=6(>26R4&VI4K`_ z%`zfQswi$y8;xtrdVA`HiK>##JE|J--I#|a7DSFqODD|oUbD0Qi4I`(K^ZPYU`|IC z4h|TKP)5waxf5%69KwT03Yk!(KHOBWWP{U4PVAK~=)$FbP)pS~c{|jK%lo4fr>EB< zOU|!~!jigP9=O2C%Ng8(RFh+q|5EFqg+JnxlkK}TBG8=b(*EA_Jvzh$KHKT5pPP&u zh)2@-@-+ZV8B+@RL83QmfF>sen`v>A(f4^Aa|k)ok*`4Cy}`$cEOoK&{`Ky$8`jtP zIC4_bg7l4RcP}98_8h#=paeqdI{ww|$>4aMY3lIi68}`aQ_1y#z@rgdG9LS`ksHe} zB?^!+xO1X#&)0s__#v$Vt^n>UIrYA1@f&_72#UHfFzU1g+^^y zaR!9R{NxxP1hPg`=$*?o2qsIR*x=*d|BDC_!@Gc-Mhwxie>-?bE{`?A2Zq_8PX&bu zzvOE3dBM#69u$L$&U=DxCKPBnP^p&Dv6L>aiH=#o2tt}pZEs86Q=*=v1g(nJZUUq`OnhN zH?>qf~kAr^K%~O#qfV@Nwm(jzt@xovzrLh_?80P52XOv}&y;dRny|CIXqMrkV z1y9S5C1G0Vyw_C=j*h}QC7SgrYJsI%Hj9ztU)_|ojWMd9@0%90;LYh z5aZ&mD(XXT9`p@w%E|LbWDa)zFH;GSz9pu?sv@6y%~3t4$>^1K&deYc0Ab6-&_LzE z-nWHI<9ZSE$h9SucwfdGE^9fEErr_5a4f^HXA!VhdKMib5C zlJJsE{s0`mGIzMH2% zT&|dY&{-AE{5)MQHq90AqJp_PV~>o)FVk&o@}qs_c%vRGF%0_wS@cH0?){PTYHRjU zR@}wfj!_D~m|b9ie#;{mL#HS!DgkZM7DD~AiLgFOlO@l+dEwpdyBvAsMA4hGWsX%` zzIp9?(6YV8W-kX;jf@4+ANnW;f|$>^me>R*7|E@nxY9S8Mk(09-sX;e`YP0>hQZf{ z2H_+~LGHTp8 zv!JbP%hJ0&MaN(MI+n+Vsh?g=h#3&P_9dJpuX;CBCyGbH3-{+|harHZ1-TlFg^}}* z!C{{uGr%?gIdjXw%{dHU&m4By`F3$Um#PSY(76l+1GCL`$xRx`Er@sKZ)<$&7~Q2#Xzj5%}!NdHP5d!&kk6hm}2%sMW_1VOeS{!ERtl-j#fg z3WFMJAO}pplinoeHt@I8luTp6TpXAAoXa2C$m#i~Pquy{cOckyO{$_8vsp?JQ!GU_^ z`z-!78-txRKAUS^hiwClXIw*4l9Dz}ohJYU+jCVX>W?`AN+`|PQPX{3H;meIi1=Ir zUTG8HW%k$I>S(21#rp|DA%H9yO>KZ?;uFwY6}BD$HsLT+m`z78@dUu+B$Y^7IYH)m zr#TycI|}YtNQqRujaVrLVNDefFCD)Hr84`~uARtwbwez&A@48X@oK>?~mSfV}y=Ws|w` zt#u%v$;UF<=;l;lK3Tl&qM#6|4Cqd0KOIcDUMQS39(9n$XY-G3!CzsL@Vi#J_S;wo z12-q|=k5S`F~K5>A9xdUAn$(6XL+lH)dvuf)ci}+pK+&E5v4ZM3SQ^AZpR=m>eST9 zy$t|4y-x5a%xdLe@;X9toa6?ghQ_J~U~W?caP08bZP3YxOduQ(fC?yGC7WbO%c{K7HX0n_WAf(w33Vd~Fhqg8-1VL&roARG z5Gwl41f=UBQU2_;g%Eb+24K~|@sKDMeM-3-DooD$K|kfUuk0zl92Y+fggk%ry2Fm9 zUkt0Dj$s#N{WMUSpl;Z>+)xYcV}$VpWiGz_V z2iCC_xTKB~Ll+eM)%P_MEQHuP?@dgN_{yEPcgv|ysVw=_1=r_%qnVZ_C7^vfyx(oa zhm}{Y~Y zAn2N~A_N@-y?&`ZzO^Ers$F-%TtYuav%;EV`ooXBgK+>(ULmjkmW<1RN+cARhp$_M zB;_QfR_N&!cl7dxJ0jGU$6BJ;wzR}a8Zj>%B4#wVrL`VE4*Jlp_&#GmtH%ulhNbXk zNNp~gjs)!lYmG||ois_{-~K_EMTMhwsFSUy(`53#y(wY|n-G4QAK{F^@5n)HQ!L^iIh}e_!p>AGWh~*k2(OI^PmtL|vCJ{wB9V~q z3KFaBsvw=?dU4+ba^*z$5ZMzRHvM*blwvo)Ji9tY6dgd@KKfe zKRXxxVUT&GL>uj4OHtVEkf}hAfr6+za(T*bf>+XT|IGSsZjg9LbwxegyTh{t8lx`r z5v~=6z9cm5#}=6J@5Xwvwfv|>w#iU*0OfdmR4cMg(Ld85hCrcyb#J>d+7FM|sezXf4R#-b*BKtm(=8y9fd6xml1&Z{MySfR zMrLw3lzW!-3sP6nH{^9Gscr33A~Ak1%!d()r$$}(Cx=akzuBByRoWi&6Z!kfG09fHM`(4<;xX)9Ah z!|P*rLzxJ&;9YzI!;Qb(4#&q(8rP90676bb2^Hx6$0p!rjj=a~%$5-s#S6?Ul^#bz zTUOL6eOI1;OQfk>n%3h!AM=-YDOJz;-(OKC)KHCzN7jLldR1wNhDVkJrqEtxwJ0!Z zLGW!!O~u|%BZFp5#~reRKzGa4hj8$&Ru8n6GV1K;$iFa~D5Qt7?PKG}zcvNq~?P6p6 z(A88Zk21}r+UCA&huHn?lc!w;&7pFf>1q0h{lo`oG*yAq$(_vOke`oO4?Z7<^N(i> zc*lXk!F{EllYF8^mM#4C1l87CjY;tyr%Tb%?)swWGk}c~x%;n!0BPp#ERt}!K^q|W zj)6Mj!u0rn5ebW2$+w=K=v}-kFL-|`r{MQAHv?V<%P2sg(gTD^RxsFPq8wFO;jp7p z3({_eYh4kx5OetgjLhgj8}4vEAg_|__HAd=v$30V&_SmUP1pxCKEDA;DP#Tn<1~-2n-bG7yA_(47Z9fl>e> zGVV{}oxBF2=RyF9{^W6e+HKq1biK`30MxJT1>_gIXcuS~7=sskv&IhlbJLL9JcbbR z{uY5Q-`JcC;83dpcyYyM26V~e9GGZm760~QFu!>5PSdS8ffuyKA4UP1bHyDH4Q(gD z)*JwK09ueKRAFpaxPetOHB2`kzn}`yjlirwHx|^gi$#DxsnLG}`Y|N;M4$kOl*Wr_se}4bt6_t2wBWV7LfGeWl1e=s}hYhSjG2G^xq5z%dFG%~LrK98B zSiJ*tI92h`XDFBd33DD`Dg{-5Snz@`kV7Eg%^ep&7$Q0Ff(0cX1XnC}s%Jk}{2ffO zV%De`A^~PcMeGLi1fc{5jha#iFi-xJ;&C3&VLb4x3iR#zMIa5xB?gHfb`}&q z%w18qvfKz)E$_W}7vK&f)6jOsc{LC9SYr2E&Qz9x@sp7UxItJ(0V58#o>I{D zG`Fr3h-ERlngFlRRk{!UT-$H~Ev-1Sk#qHRs{(r61dw%%5fT7iVHAoc_PQb4lRR?4 zAwl0=1Xlr0-4iT%+O{h-Q|{=X_)`jQ-w8CN-kJ zK)&QExW2-6U8;)J`xuI-0{N#bGBPspU>>y1QYOaZr6*3)R!vv$Jd8}DU7xNc!wXIx zftGgQ6-L@UkhZRzqoMPKixr+*`xe}FMGk=OtqLqpPVa&P*PtPll=C4?iYE5&Lc*56 z1Hj=iUI{oMHw{gfm(F1aDPTcHzIs)@dWu^c@6-*pjBs-7&t1H=n~D2Qa11LnwtC>Fl~2r`Qnk;<ne7Tc&x*mnP?4#>G$aTlBMx;N}&zEw=#k4av^Mz}efkZnE4 zcPBrIiUxFifev@uuvBBUmcj3F7tqd#_Aj9L&vQ=(PeB|Q=5~YqS=kLqfN9x3;Rv3^ z^m{>&7$VNzx`0AOAhs<~H26vObaT@rC+wsx5SP0St0hy&q{dY!|5MdK?i+z)zbF^F z-X36(F#yJ-a=Q)LGZHV>1(k;O{mE4z|0#N&^s>~;Kw-(4md?pSu;_{6;{!lbZaI*5 z;Dp!VQk1-4VUgvr-%Q;%bzcOr^tf8P4Tiyzzd+551FDl0SY)&gZuA()mf^Mmgg%8) ze+<$WfUWyMU>nH4n*3U>TeR!C9pc4{`aB`)AhOpBMLjMN-6!y7S58L3cCK|eBa#IR zz6F554eou-cYnERrU_+!k3ZOj9uDrfm4VnV#7e1!zOk}0{^G|Q&@vY%HNt!M0ZGcz zXE!^6pu3a0IypM8uvAE2*~26V)?J>n;|GV&OOW!8f>Pnzy)3Fj5#*;(Uiv0!x^>?? z40zFYia-4-?!7)WZ`fWqG*&Cdy?C_X>4n&#m>BY!Il1GBwFmg7GX|QW@9J1B79SY} zB?5TU@{j~?qv+&RsFy)ASq}cgm&x9x9B7}5#@~G(4D6c#40v|IKl_6A_t5S*z*gxh z@;F5BER>_*rM&5IXGV~HE`52V2$|`21--hfH!B3I@h>Q16eJ~|Ht9uKxZO6V?u#n* znTIFR*>M7*jv9CwTmW_B2@<- z7sZtK12%x@IF@R-zX!%Vq9EtaBo>5cqZaF6c0CsT;Dqcq<@P^k z)!!-N+YpInuqy2N2ck`B>@Q4m;313zula(kLrk%n6L3%9%d&8?oG%VCE--eg|BImqi7K=(R86CT+)NBN) z+IXrEjYq%Do}lK~L|bzII9?t`d}}*B$rJv)|5}*^A4cWf0QW z{_M%K7MW8l`9e2JKH+#HnB_KrsEl;JUQcoO%N$^rm-g6{GxiN;d&nn7QN3jh$c@d+Xho zBJF0aB#KUFbNP)&!o`CA$h_1KiBaHeR5wJn6;EOGLf~F}%dNZr~1gRU~>>H@%&0-&sPwS`9Q)j_F2pe`F5%AcoI>k(eNl3tk1tnDcKGX zx*bM?>^vQC^k!0RS?E9K`P)oWqRsiz9=~1AJhIR;{)qz|HL}R zU*pBOL0W8-jS~Z_HQP|1c zR?NS1c5(f0c7=szyJwhnqlRY9m$2Z8{zMqi)e~UAVMLQOP|;K_aC}j0F8;X?5aUvn*;{Aw!ib-`_;5e%6QXi37sC4Z^zvS(LD2 z^2T0FUP22yfyHLCFwh&B!8^l!$4ZaL1;>WxjtzPG5X{4LX?oZtd%groq%*;Pnsm+N zJi0Z0^Hnt-IdIrYZALV^0BdMZPGJeqk(N2q*(+|ZBYkXUIg}#5P!h}@+7=w0~sP22pH!KWcc-0gE zkMhPv4XYsjX{1#aux02ha#MmmglXS=-duz|#d6hnQu86;8Wf9)SHI^yrX8{C7J_Qv zpZJhi@h;eU3~V|KRpVp$%HV7x>sxRDE$r%?+Yvs)&0t~DF7T04SSZbXz(PG%w84$@ zbl@@EIb|FBh2MOrZxxF7PvsIgID)N~p?lZZjNx}}QwtxK?1FXZc(6`04LEW~=Xbz9 zrHcX)rOQ)qMC{3wvl)*a_P1&a3wD2d%TMhALe?8=iO3;i2y+$Q1l>sGFknY|RT3y1 zQ1>K&y>>BY7j_~}owJ05Xd_f<4~}mDrd#a^V;-ih%ey-Z-VwIC6>|Mv{C7S<}la@EV-ITuAhp&8Ux_GPYSx@CtLR?1i>YBGI=Oy+5g?*IiJ=_-)1wyNT zL2y$+O$EDL_06SRYhWK!GUC)VT8IRUR3B+qz!DLu+yhaBdg@0p677=^__kMiQNoch zRTu1*K6Ji8m@SU;vwHHB@T9vO8zCc_^BB9J%Zi(3oOa}+e}I~>al~cmc5%PdZ|}MR zau%#3tngXCzuKVK_$}s2Q2N2y+`4R$9UVJTNc8&wUK?jhE$AO!3JN|PDcn4Hi%I?9 zo?IKsj=jmf>=tweLC{g$_zsd0Q?F5nov3d?X{thOqblf#`vljKNKWWm*c>C(I#Wtcq>4=F)u3sm_#*be0H;r$(_lMPfNY=w%v(KMYHYOMuU|<-^jB# zG6oHW=$=teCas34_zN!pIUMjn_-$k+MKi#L@@gdTG= zQKFrqe-l3HabGr zlWBH19v-YLXL|c6dwb-3n!9TeD&5@|l~r8n`z_1mv0G^PoO!+QBj2%%E~Nw)j(aaE zv4-AT?%R{`*~v($ezL9A5yjE5HY=3ipo>wd}5d$OxZczteuNFAz$8o?F&)EEY z>Bx|B3N#QYc~&Dpr~_ld4Dw`aIcb7E4`hN&-r!&s+Q6JpDTQ50dL_0I?$Qb!D)If5 zp~DX-tV$n%3(Brdw8`ygx+QU(r5JSeF(8zJ5KY-Bu9bCm717_KsN9N%(jjq~kk&Ej zzWuG?@x2SAqpgVw7b9_P?;^oE0;P6Dgm~U~52S~a5#4^>q_quEeevhL-^`~_JH#Im z{O$ZsT*Dva^2XOMuC{zabj0_xm_B~ylzCfSe8-| zTfV$Pww4Qb<>zK8uYXAc@H4xFg=HPlQ3G|RDdkwFF|hOMb{h!tlY_p(qsGf%k2fKk zS*4Rm_tZ}L+daPO!dX+9Q4>!7`7DKMv?uS5&WD^TOwdss)5`kvxsthd@3iQLGBteK zF1bBkDjR?O7*$Su_z;n|tc3H}>=qe=0|M02w;#nuTw?WP{OKTCN zaH)GAo>RL)Ef{tE>Imk3xW`~TqDj-m5u>gYfxr6fDmO54a?eqBR`?iK(k9T>ee$)W zK+^MW8_7s}9!UjRI!OhM$W{K?n_D@7B&IY3(Q*Ev&3uBVYV4&en$S;wwMGcE86Y

ncfKhcTg{{j>n&=9 zI$5=i8(Q<|2!KqD#7oF^Uwapx^6+e)fQHmd8zfpSBTl0$#G>7S3adOJt4@_*R2YNU zm>`+a)Ss!dToJ?i&z4pbj$B7cM+1kz^-`imN47^r(M%)byG!EEE)C(+v}GKHLOo{g zvPt>8Va2L}t~cX6fqL0F*^=I+yKW1(vTeO-XE$%AlqoZX0+dWt)Ew{Yl`#t!3eqnS zM0}IiI+wgh#r0@8UMr_F;ty88_j5CxOy4Wlogp!PK?e3L6Gf28KNq=URjoiN40Xm>H>LqOQikp0^tpCb;a}V3FJ? z`?g;rCwZ1sU701HrHZA1^{X_#Bk}(|C3S#zqC_N4~q2KrL2O(i=N{Hf*!Bc$mi)}DuFLOB}4``T^X5` z9a&SD>mL$m@JTA8c#=v!VhrxuZXc>}S8lPQgayTlv@=L49R)pAF7)xkx9{&TfC@%Y z*@%qf$Ebh9QN#c6CX;_l%z%x%vcT~wzH#mv3<}Y?>{3HgQ?j>D?8@T>S2bcqSvr5? zMulsH0@{TEEsO^k`<_=`AgM5xuPpsO1>{<%Eooo$AW^ud+RiANpnk$z#hW{6+i^I_ z1qz#n`%(KbXv4a|D~Yz*SI}*l$e#%7J9^Y#PL1cvvlG@wrk&(`w7?D!S8Fv8N%8qc zDx2>%gXzY|d-ulbm&J{FhK!pD3vlz^q#E?nt5d1v zJMvHuk55W^`;=|-Gct=j#o~-qoMnwVZF{Fsr(&+up#Qix5^jgEM^OJ=*X>|RbnB;Q z@rh5CyFzK%woZ&QC;R1k_N%;l1`GwezHuLPc?k->VD$8OC*X0XmG_}I8vWUjoGisG z^HyH>fZ;F8+3bDhLYL(2*RS4(#SaUkdI;KT+bxCyqQ(56BZzG&dH90G*#(b}fcO;y zHXy$PF$ZxD5D8E%6#{Xl^*Mo^zC6Q^Me13I2cX8BU9}Gq8&>Qw{^5R+h<33epP3JB z`tev02WJzPc6S8w6>xM?moyv)L1{`_!fIac`v8*i><@v*PvvWB>~xH!d_*^EjMY$7 zvhm9cRJT;V?K13#-ygQ^>)RUiQu#ZqZ_0B!Y)8ugMjY68$ST^5(dSj(a%lfQyVx%=Wt} zBNHFfpI?h%*+EBPP9;&paD5xWcGi2dD>}G)iK`#hyAb2my}xev{Jb>7QP^7|BAvrg zswXcf(z&uYla7#pkwZatuaL#mMl5p+!-Mv|&3(FY1q+|qpJw)B0h0cUZ4KK-B@6d5 zlA+yR7oJkj*j-J2v4wl1V*7&+A5rmHcByL=pUQUCq!{OZTMmoL?oKT>%v6nd8Te9N zU2H^7;$H?;>g4D0(%--5zwy#tog=(W*=x5$t3so_Kut?W=Qpx%KkSFEr!#`{V?@ue0)b3`9rW@U3 zTV>a&{5{82G_yo9Ld%Vh1k}z(*#dd(rDQ1rfmw(p_PSABA`S_ut2j&1uybWGNu@CA z=2v$)N(mSY<41QkNcHy8vfu75ec%Oi(_M}8j5dn|)(6Wex}9HSv5**;KiMo1DMt}? z8HMzGxy^IB!Y-pvyY^FZ&7FeF&Q2|xc?&^plN+vNP>6h3(#_v%GSz_4P{)NZch08{ zFG%jeRzLcqYmgPpbBodea}tik-Xr4L;~vI!@KwhOrF|ie;YNvl6tM${-FcQH&OL=` z>j&lRFMz`)o1`1iiQauXJBrUnT7p8#C4rG8pMJ?g{72$9M#1?QkMV4Q@j8)5@O&?@l@@Ce1`zXmW7|1@$Z+|*exID_GizXsa`T+`jrhKCT%LWL`j@?CSaI;*ctTIH(Qa>*~!5L(yz zImNsC;jH9OsP@53lNkUlVgMYY8z?R;V4UGljP(5p1~T<67@@{J{k3}l3|jORo+%6z zUdQpzxa?K`ExRjF&WOFeS$G!Xc_7z*TcNR=iNN*2(a9CTXUlOX{sd3 zZ{MDA2EGrDnp8-M;xBOQ3JS1f+c_Y!HF?mY*JhWw{#C-_z(DX3PUz$N`wwsLj~b(k zr!uqz!gvMrykfP1Ao;R-V8 zrF@|MD&K3fWOKXXaXCz5e16~N$?afGqoBcT=RE5!DOazB>UaI%&~uh0L4srGYEx=# zO)=G8gE*|7BTI$?Pd%-HRGdHtdp)bs?Agd}Humym??_i_i;NQ4ryx2mt$kA723bkT z?x6yy)`Z+zuuvHomvK)8ZUC+~|KawM^mTImtEKl3HCZ`HdqGIl>Y(+T##?iX-L-zF zT%f%>U;IQDM^~XZjmAe-z!u{PW;3(kzR*}H$S@N#K`qF(&!H_z9#>)ut!JJ{(dFX>AApglk7b$u1q%OZLB zHcIEkx1VI&XeHeh@|8^N%PPrE8|=MyN$E#RGj;{7hYTO4&cRTW4h2V&?Yy(>_`Ik@ zTGji5%fA1|F`N0h)fZPD=P4t5Gs4GnSiS48>uIf&%o^Nst24oJn}LgR`?-SSjrxnO z)41X&+EsC`c4ub;8?~n8aQn@r(yukGM2diC415|6)5@N=O;-;eORaXwKm1cnnXrKN8Qix zEOu;#Ug5CKi2Ho+&zD~xbvAHJ6?r95{qVy@+=2qwd$@!~F(W473eKhGmQ^JsrR5JG zP@;=cVB~0fh-Uj|?XnZFmZ68^qc%;+{C#@iY?GGf{Hym(+GiOW6<&evlco*d$vQD3 zRv?4`SOrB5rg%|0txKI_u+AfUwJPjB~F6*ZR9&JA11Dch>3GDQH=I=jP%WjK8cM!LD~&TNW?HS5ACFCW z8P1P{C{$aU1~SAm$kPIC-W-DfyK zfrHr#-*>}csJN{s{d=&0zcD~?f*+8fufW2O@g+X$BVF6V8_wL9i!vL>S>|t} z-g2fe6aB@wsh-t#Eo->zl9+zwQo-kb^oqWWN(%F?S~}TdgHG39p5)6vv<%Wkn7_UU=>OKW=u=S>+?9K*HNpg0BP%h^PP`K-p$6AdoUoPa1rROsVjy|HqPcZ`YJq9TdqtTu& z)Av8l9{IW#=tC!xrI=T}Enh8(;Bs?wVX(5@J#zI`ZmsE;KktTqIFV0y@{?Txz+hC4 z;4XZ!AoL8^dz_LhK*Q{SRfisIk`CC7%=q+?zG$?{yN5xs%+`?ja2R`>wA6SD^Gg5P zLQ>{^%a3|ddRcnJ(x|aoEGTqXk|8TNY{qLk^vXqP(fJD!44fu84K*-buyx=CcQFVf zEEtlam(sX)Z^Z4Ae=)JWK*k+@1~yFwdE$0=Oh8=F^$17f&h$J6g(4s?!SP$TEwx3fWPtDj+GmyXG6Td9k43jB6 zQqmUgYFK+uw}<%i^U93LTx-~mM1eZ(YStfx#XRS`=DmjBfhc2sR5HB@|UeZ_bpxf*>-?-F6`!;?{SbRABF!YG_TH4B|iS7+< zCkc!ROM5tkk>rWvX{sVWKX0#IQTj&E0+A)}F(QsVq9ZUCt7z&o8MOLieY~I{j@jLK zPeB(V1}Rei?iueyu%rSay!Yt;9F1o}Cq`lY*nh>xVic#k)jhy+-KA%kI4ZlxTO8D| zw1vq|$JJXw{!+fNvbVZXcw01RjIu*HBL0mx=Ps_2D{aM-^azRN7GL55VQSCG+uM{; zG_NcD#^?03YJVHve*D|_gPD`x`g?TY{-M-pf9ONZO{v9nxmmce3U*_wNYr$y?8Vu$ zX;)Q@S6_KFN$c?1=~Z)6%a0Ewa#L8?%ro6<9TU44qzl>o#aek*#qUoF)NFSz?#w*= zi&n}F2VChVb0hIs6jk-A3-Ru3^4+?V$~{L=B;^*Cr{0yR(rdV6_~BWrM*k5-lQ3H` z<*qc%bG7(!m*j_!JziS+oqiYLTz@zg#{yM7TwlISbeo^LBk@xkJqG??D!rZL>iyF) z-*VTU=}HG>bqyArW{VSU#s5dqSvEwqMq%^{N=kQk4W)EQ_Y4w4Hxkm_NJw{=(%sz+ zk^?B+Dc#+DkAHwKoO9-V_kNzWA|4yHe8r9YBk!Iq)J)Y)L&aDM=^&-~0Q4}mb`h6J zO<4#9A0;6HkpTEjS`bJ9o5Yo36V%Cs+b&REoUg@Ow}cP=$ROF8v+xBwTBatl>xqk_ zO3V6<+q;}#2E8S1)^O?LYgff7na(df(&79(3B@+OT>e;lY$zA=Xh$&}Mrq3OmO3LC z8|}mB$Rr(Vf!?Mgwt96nE~|zHiaU%rz_eeuu?Oc$J2vCpJbl)i$}+NyH*|Hdx~h@* zo!G~a=}=+>E2lr2h*@R-TI>a6ruK}t=ux(0?~42;fn2$1!~sCaa3X8<+c@C|7U{uH zl0YPL;f-vF6ft}5c4RHP%1${_y%PWdZ*gN+ev~poxJ|c2Gy!_O=*Hz}_9I0H*jU_NE>Kv$OvJbyOYu7lhL^7U%7}o+# z$9wbqN0GW~YhfHRGg0OV2TJgsuxN$c6g_y^>tVYHr0sjqjhti?%yIEi2VKr2wO3>$ zdrsw1ao9hGRErx0A1$F7KBynOi)1yFNtMt+!(SoyuM1}f6#@fJI4V@EZ#@GdiNS^-}yk8r9|ToOysz=tOqGJkK#1k z3MF}T+aEOF)IYoG0=y+?-Z3k*Rhncl{GcgLn~as|pV$8M={Llk?OL|NZI| zj>|&G)?jH`u4xtWNLn^KCK8>Xz{)(bO-4MzK+(aTq81ZZ46{c@i;2fhN&|<1{uGF= zgE>yu-dTPq0Gb@c2tXK48NPr5K+WzxngA>n*VCw)&BIzQ7Uy8rHmQ(cL?NaJ&H_Z+ z*4x6pxf72CqOSi|FMg6-2z#mDvMtdKy7tM|4AbSCL~%gtb)5^5|9jGM+X#Eh-*_0> zM!Wsq3o6RHywFZWc2ihTzLeX{LsSO7Bl<%qoFS4eH@m)U3@OY=72Z zzcMs-OO|ULZq)#-q2!C%y1cEwt=`z@Yvy?SF19p0YLoHdVM*llgFGDL+^a-UDGSlg z!EsU$Z~=WrzWvZetz$2b7Qu`(g2Bq?+26#>SzZryiHz7(c*YuFM=$H(XG$GZ>ab^b z;0@!5>f{i9TO$+m29C%t=yin+7IUo4zFvd@tITtVPqvj?NR2AEj#4a^r@gx!n)hF8 z24FZKfg`I*SL%l&3gPR2gh95Jl89-mDN$*Whf3@Dc4e?6p2oXqOP8qF*y4iH&)CG| zObvpoO#Q=iBjX?X;>Axq-`Y-vH{?&?NtkREWaMkDltMdnKbi$dPbv4q)h4LF9-8ALM>s4GOfhLf{wb?wCtf&Rt{FL=rZl!2u}A zT%|gIKK%yk{-XcTYyMEK*^QG0N^J2g2BAZw~1h=EjxV4D5 zu{u60Nb{u>D{F&i7%SciNDYQ-Fy)k6tjc}{$Kia79y!;O(Ib~D(j{@oJfoea$)i!* zVYg0P9Uy-PC*(5i-W{)*EgBj~;K+s2gG%81#NPLcgv^f1wys=QV+f-X8^x%jI#tS_ zUZCjdo2QA7$7EPnj;8p-s8iz1Q30?;hbnWK`M#>Zhv6hz&CzC^UmW-uc~<{m5Z@Hv zidKi;J|%k{KC@69nQAal(_1t_x%WPB4kK#PdH#>((bJFf5~3i{>yN}5+oThqNfkf* zS0)g-CyLE8@cn&q30F889F_b4%F|_mWs-7?5e-DPLi##?4CmRjH?qrQu7{hNs#1+u zXLVVfI!!tv{rJk>-#nGqHz(5bkj%vgl*GcKsJ{fBG>U;;P)LK@e!Y%3i*N*N<{(CH zZvHsKPxulf%l;wTaf0_XVYQ-k_)vsB#qV@94!b~os5qB(`h1gOU@2zC?^4`TCVb?M zdbozQSoZh~F&|)>V&f4}0r&06PBr+JLMOMlfG6{!;X;cNXugqYiK9b1*PehX3SQNg zgZI@=oMZfk@YOi!CsLYAd{&hJ{cLkwv_xJiDPb))r@9gw6CJHFRu3v&%J}tykJ5^FDP;TzVW$R*RqSobw zU|1YSuhNfHUqia_55G>1>*lG|bO42yB*I2JbfgqpnHIVdL7IY|iWyd@9Q@pm4$ra@ zaevoZUNzn(UY`+6ndNq(Ol+yyyG*mowc4!UclDRZkpPb8wevNkkHYD58!Dg@b##Az zb3GnQnC(6TMUT_3t1o5rc0B#a-+5U~y5=o^&6w$MdylrXDfSs&P-n+l$wD78$Httb zt*(-lGTAWvK-MARa+ZH&P@}^gi;Pl0FW`gYMa9CQi|j&-jB8NlZ6VYZUQKj?NR`Z# z^0R0{nJPnJ)BxKU=oU;9Okd)BGs?f8iGeH#u+AJlm)*dQ2nvgLh4c>?wn;qjSy%5i zXZCB|78JsOPuX)nl4D9!ZHUr>0qJB1iXQr&9>>Z^P!T zb@TXFT8{NcVc|Y|WmH#HQh(NSUgq49iowB+Wqz*hf*8iZl}X3OCa1^UTmF{aW?WVZ zRt;Zwr(kx+Pj&A%rUnf|{0fnnZ`dyhaIGEqcd1|)|gGl1(TIhEIaLw6Dj;;drUo!wtHAig0yy7X};7KPOqQ;E~o zkp>7D$DE{Gi*(S-)kHhI+j_C(_)8=gH~xOG>iI~M+I>`bo~Fce>D_ic=}L<{_GYv! zR5vo5rKo^b*3vSyZ-4=Eo*OrMNa0ppePjTYn~d*CSFTYHScI(DMtyqecJv87n3ByN zM9&%exG=`>F4Cy6B7SA$>yY0y>ek-S6gECFwN#!iDn>@Q8y8E2bP)3w{c_oMU|k`0UJOV51KMiN9e?4TfDat2!ZpLX2EMCo7>C5BMSsEls> z$Tzv71ln!q_`d!=x#6E<6&bKS#*+Ema8+Bi&!+D#am&GS^2$9VZ$`79=iwRXen@IC z(sQ#{Ff_1LFh~vmIL9G?1NMhbVhWPB{p``uB?cf*3;OdY`#eHcP`Q6VBk?ZLMpH~6 zX79vFYGCWmGu(K%)Wa;NhIR>gitZ zx?L&E(eXZ1v5rO5m|)A*l_ir6IC1|WcbbS&1ozXE1vFNL)eM@fGQ@zNJ`Rd5!}g$) z7JG~hYkuD~1>0RRk6um*Y9={jc!H#G`BkQi5^<53sFrqUhhOl_>Kqeol*AU-rnJH} zT;Df3SFy=gbpy7pTHK#Q=S8Y)*Uwf2ay>%<@%9&?AfBx#>TR&>L@f~sXWYbgfl1QN z0hU^O{8-$Fr+u9lP0Eo@Ert@aVMIw>NFw<71}y1Z8;H-syiP?hbe@q@d)l=^-j5Ja zn#5iF|2->kS2e(UU-f6{+q|BWQ zk?8sSzygKv^0lI2)uu%5(c`6>2q8J1PG-DWQscr)Q*%Lb9GPg)$}`?kLz;VT?YWnl zxwDi>TqBFF`4bz>ap>Ds!X^PlarNr{QUF7CE|qB2wldh!?qh6I`s#C;wV6?MsLw;0 zZ5f+-VB<*2$FfBPB~;3#R|d2y!n)JjCf;o$Jm>{Sgv2?TeYbYE?g2t$8d1 zGV|J|DKIcD&h=6khjinsmgr4YIkhQCecVmJqFkKg7i`~+>e`0l*Fj_2^{p>qBIGpZ zgBTFY$uD7Sifj+6a-Tf3u@(9uCULY(de=~qe)raL!Y=vQ7so~i{o@T3Xp47yx`4YZ zG$DbG^}afPh>JCWa+Dgm2`JYMb>e$RkOLY2g8QGy zgl{nrUHS3{;aXJm0aO#7?)g}^>^B__55!t8E8=E1G)Hb3h6g1H%A4xCTs@m_ftv_ZAL=KSsT&USkbUviz z*Ex};vJ_d!qgkpnviJ&+-Wf2wuJE;S=CWi_rp}O=H@loZwK-*6GDhx@$|=azLKY4S6d_@qh#9otSmZKiy3 zreaJ=Ea&|tA#L4Vok#8U7^k?9Wd8kk9&oeC&6n59>$Rki^j{z0;i}8b{W!mz5rFyT zI@>#)31|Bc+ubpJJ0AsSkcyXEQ*~dalYUm}oKjqQK5OHxBk|mw__kh>G$?Wl%qTAi z`43f^a++?lNB3|Afs<24C?5J4#4>;Wp=G7hdU-x_;IDD3^VDlLdDmhxF9`hrI%Aud z-jdY1o2j#;WC#qFf%P|af#}K*W<|q9Iu-Bpag512kI#xXpl!W2(E7+%G?7nv%mb{+ zXj2-39Yw&TEAi%5A7ly~Cgi3g#|9f)(mtV+=RsQfKijCic(~npewK?&0MaXE(+NfB zt?1?SyZrESkU8u4ec^x3v)6W^B_@=!^v15Xy6N`5kD)xQ+*(GBe^TH{zCyq~3-5Wb zf*^81(n9&F4G~*gs|g@G!I(T9vfWv-zCa00)uh_%5i>$UVtGE^kXZ^PwBwEuGBv*Q z&biL=Od#Q*X}b9#og((!;$=@nt?K>R(!^bA@?^X3;ddXB3ph_h`dj)zsXS$seBJto z^fOA&C~h`}?4tj^bZ%y`#=8$UCu$WiT(9{q(lM?Bb-JLPUD_!#Dbcsz2qv}8 z2brg)3G{QILEEj0h#oc_?2oQ~<&Q1}o~xHLWonYtWp8*ACa{Q!}b+&#K<(H*GB< z+e1^gn-jUy1ABeN$S_S(y2AqN6&UmV9M@hRz>xnhN9k>bkjD}9qtX#Xxtk8_YdSbds!oB)96{aulVp=*I7+l!9G0kL^Qp2z{bu+Vc_2LjF7}*&%1H!Z>B3c`qJ}dE$!GQyCi7*OJ3?C?9QA(4ZO!N(LPlV zbN8aINs#g~2|JBIjWJ^ah*P|U40y&IuOib#1cG4qHd<1Y+0+~Ude<&6viu^% zh7JWEF`|FOjZC5+1$oH8KmE!qpDwSJ%A)3o2;Q*!`x{Wy%+vj)S`D(qDC)(v6M*LW zVH1y=LOJM4s;a;7Rbyq>L0>xwRV1@a_^dmma+~!yU%6{%)arf?Mg$H#qcZ#O)wQ)j z@kEfVYxVR@&cIm*(PEy#*e`wtCxz(?T#ATO!7$qfD(Oi6m(4^5_EM8CEe`={{<96R z?92sr0gmZd{_I~#(Hp$K!lTDL9u>@R8~R-^^Un_6OP4&$p;E=5KS?8`wHO=Z<8sa^T80 z@8U=QEJHB0kaY>}b{$LnZfIuxpMOnp3ex#6m&c?JHLrZ}DlOL)WclJar#~a&ag0ny zma|MA{nw_s=WvlwQ-(i3z+Oxz98DmRQ@M`k9VuV=M|%=J6vauD_HW(Q>7C4f4_7nK z&fROza(-yeFA|_mFOkc^_HC~2BV);}oz187eEBUY$pFr&Eh?~In1U9|w+G@oM=b`) zNI@FZ;qzsO9%?NHw6#QR8bE(@f1OBwse0&0Erz8(KYZYUPmA?7q%g|UULuB8nHkfe zk4p6^Xhm-%kf{ZYsp6|+JRWk6g<%?-QhJlGTD+u`0}B$Oz>sdS@-I=cpOP5m%w8@R zukRN;oO_ek;D{T!1@GN_iECy4EqNhwY8_thPzh9pTwc7`FxzSn6*IB+7^dB_AG%)i z>tsf8xSEG~7^}RFLY+@84HuT6%Wv=EDn1f%$lRkudHY=Qi)6kq43NeD=w&t)Z$Vnn z9sUwGqTx@v2C}Gw8AD2$y`5s&B6b0gZwi1ZD-$Qi{rO?5opSBdtucR%Z$&;W?ND6B zBVDUFTFcQ3M#tzFYmLXO0A9y_LoaB z(3n&8e9N7i7Wch`*@{as z@|4UNik6Lsg>%2n8#HvGTBnjD-SCDz`Qh$bmqbE#LcEO44S2!fT049bm}Cznt2MS{ zFEdx!QG%ZTwau$p9kHZ!j6mF<{F%QILW>Z-qUE`dI<7;&G;(O2RghMwCDt>a4iB$} zRQU`@bizzyu2uDh{KO>=V54;CUwj0N>5*dsabwS5M)Ao?P~^*mZ;}efpYg_!?O4KC z;`T6iKY<4!+K$DfOHi=F2@5O01@--Uw#a+5B!8g0`e6!z1549dBA=rlWLTt`gbgy% zueVmFLsxx!CNCwOY3A~W?>4_HLL|6YxB`0KPZ6Ur$M);~wH|xX^Scox=D48}^@25c zjx7&(uJi?DI__V3+*r^kJ=u)O`w-~ zRPa{DRqJfRpp*}uVv(eaEceBVN8W%2BAiu}neb*{4T(zJVsR5^rsj;!e&&%EB%`h?hyU2#;4itS+vHhvmrD_Cy0fITj{a-4s>Os96x(=v%MT5N!QKCrIwCo) zHRJSL^UY+{@kiL@)S&7O-pYS_pN{m9V?J4-@il%6D1Us0>}G%dT!X@0pgguzAb-5$ zW0wLuK6&*??$sf4r7ND_Ln)bVHWSsqIno{J&l*#SVm^X4wDH7hr%aDm*#}$5UwB?&8kwMEO>`G?NK}?#AfOXH) zGZh#vNut|6si3{EX#RMQA1(9tb>=WW`ZhP0Qx_GXb$w8z`D+!aW}9NPweZ%&7%}$2 z-BFlJE$9@vzso-^Vt6bWX%av>4kg*=BGyGsysiDhe;gDVe9!QJUr%^JL8xz7sbV7S9{@Z(43##Y$*A-^bv4lgn0w(!rN z5ZirG8;6#8%}RGyUgzet`f)Ih*>9uOCY-`gbgzGXe&c;kvFfbk1Eiy`(-W0bgcoUa z9@`EEO=}(fd&Q>?qpS+#Gya;|FVvW)mcJa+T(_GC4MUp%>X(OoK)Nva25v6K@*Esh zpitCSdt5U>e0jTvn1P?@c)cXMq?4)EA38hsB#mC4cFN(aTEj7!{YY7|d;~RXB=?O_ zi<6>zY3^l?JZ=8^_zT)@Br(KF_`XkwjD_T_c-9e{d#-mS3!!*Gai}Ezu8d_&eu#DA z8MT?shn#mb1=4M+0G4P57%ZQedsIu2{bZCKay9%ToCH~1S<)*0ZNa4W8j4$`E^6sU zN}%jlEOPcT3-w$BdWKsz*|~DdVm}K<)o3MH6r#WC!y-Tk_hA z3^x@? z&xc>-3I;S}T@C-ElJdv*8loI?=C5}M`~CWa%-{Ge4>FNf4eNp4PQF0vybXYa9E*_p z(RxL#$7)f0^3fhKJG8a1wpbDu-8NiM5{rjlafafjG(vH8b=_^%sMC=}YVh?(y(*Ql zvKsc@zY?AbXEK}YBF(d3F4u(_P~TY$p5-T0_#M$nF-t-wY9~g=62npK3|P}%NZDoF z%9T$G=op4)ed&jzgH-Z8m-L}+Q`|r$XOnWzu=iu@piOfVKgL}iv+~;l4>fKI`0DfN zOW=eybD*k&ZS)#PBj_q;84n82 z*a8E&$BNXd~10ew`U`)8Ggm8wNUTLE;E zSv9ZC8QiL26n^IooLJ{m_lJ4xnEB-~Gu8MMz&hN4fj};#a8(CH5Tf6Nc>vrDg4vA*v3oaY~mLk&j!4 zo=2;zy*fKX=N-&+JYU;znKC9oSZd|D09Rc$T+NzR3OuLebLdhh4Tf4r{tJ5mE$!6f zJr0lRLl+8a5}#Ljz0*9wu^U+IaJBRjv)Y&3&>p{{&R+IT<8w{3nLiiTt+LxY&-tS0 znyUYL1$6$flP&b1@6EzqGApi2Hf&S>@Tf^-C}|mOj9q}Yb>_L}j_~idW1#2kisF3a z`IZdeY`?N3HOX`_JME4o0YPLr0L!Dstv0*Rq-Ky3Ol6N|{^63Z z);r4j!K$5fVdw6Iw(L)%FGI;{BAB&x$j!GRr3S|*5eTcv{^O0SjQ|EBQ;+Cs{C`cb zi$A<;Cao+&Y2R*+9I(c55C2h;$1d?$RaCC0Gx)HW^_H#^YAA%+I&6KY4M*_SjE6n( zpDy6|;dUjreJYiHTeL~U9_h5mAUi-21i4WPlK{EQkYFU7fcRcAlp9)G~$WZW~633@Nn&Kb0ZQB3oNY5pE z(ZPUj2>9AK#EsPR=7{MZ`RgH0r`5USAxh_a7PJt?{O>#gr^ox+5U&g`s0S@6RvX}3w7la26*O+(2E7zycI z1Xh_2%{rL&Erx1s*5@Xh2aQYpO|vOs2X39xs~%pV(@6Zx^&p;;;tOK}wj__$+uus? z;&Ws19c5$z&F6F)U3=2N1vz8ut<`LK+&}+rqqT!jRd+Lc2=J}8|2fN}YJuFj*5_BChEV_<;Y$5@tY%qCG^#Q|tN+ZP5Fl4?3Q)%Ya`&@5 z3FK4siFOsJiF|c~uve`2Xdy*rIF~hKYcoY!c97XFu}S1Xz7l}%XmvV=P225Nw|Pk1 zYK{Vauj45Qdq;k~H2-7pQ|2BE^on zkMMDchT#saEm$?c0zrAzdLDo(443~GnOwP)7w06*>g;hm6dg}1SK)e|+xov&Fu z{NcM95Ss2O&9(k|*jESUcP(D*42F%NpGQhbtv#L`ei2zPyA4+CDe0{}ulg8|eR_sO zfE1Oe0=%G5>vr5o5RHwU7xmPyD8M|FNNi&lZ-d&YeuEL{-CIeMFRJ0$rKmyNB_H$gv>vO)Nk=4lOjA%Dm=X*~rZOP@T zmMSIN3{lbTVWOZ_0Hsw`3E+Ob^E-i96h2D3G4&0< z3hjqLlY8NVriS8T`<|PewO?1H0 zd#Ozg>a9$-JJr%;_o@BRb};aF%b$LpjB*=2E#u((+|Wm``SKL3oUf$OHtjFB)PZQP z=MNQ!&xjOlK z0uSjbi(HA^;1x!{xbNLX?{Y^x4F+Z+akj3Wu%#*p*v;S{u~q=s#}Q zwVWrFOm0q$6*wdm5=HlBKxecezZ@&JlN4VQAqee^?`ZNkNJU)V+#tqDeXImhMNA}b zl2JIQ-t5-s^2@BfHq70o%+4({kRb@up`TdRUQOk{2nfwGqtI)!jK%eTxuz`R!Lxrcxh zmGzsvJWU!3gi{i(5WHzuYV)x5{y3{%zF>B?Nm)mr|j^*+vv@p~Ay0XHi3D2(~G z1G}HX_cC%2NJ{*my4>W?U+pX-LL6q%G-zL4`D%RANHx30>AD0Am0QYR6DHdognkOR z<(0x(jfXTBI^D}TueXEHH%L!CcCl7wy6+(rhg36y)9No5qCxJvA#IQ@H(MoFB4fpm z>0_Gv`;LyUJPQw}v|3kxUbk*j5Mk{x6v=3HoobE3BjBWX?qTMSdbV%GoiR0buz|0 zwz-5lSo}ir?HAcj8339C5O=-%=N(72BdT#spl$%>2zKeEY2I_4gNrg}x<7Y2UZh)U zv`YXTTf)NO%VwOVgA(u^E=lK{;$ma7C7=Ge&sfk0W5^TIh;^A!Z&{Lt(aHo=T&dd| zG{uivIZa@g_kevcohS(S{IfxdBICPKtdoT^NQ%LULnU-uHHu$h1093gGdbQ{TVXiB zlZJzIAlAE<+^`fqycDC)Mi)PQGaK?Hg!_A|Jv`aON_dlMIH2h`ka(X`DUu$wgl**elat)`1}+B1@RMn+SVu5`e^}k2-;coZ&IRsZ&SY|uv zM3YmjvOU)-csmnV|Lsrqug6#|HC9{!P)oJxYy1|M-G2Z)a-q?}h`!bCzg>w2ey;e<2&0!jHHbTMd^+S1%e$o6g5_ zDeR&8Z}T)0p6w@Y9E0%RI4t+eG~kkSi6RGV4Qi%8KZ>CdbJ^QA9Rl%B(-KK09Af4W z1zMPfuXXI>>Z6);Mr0g5K|{)b1`rzs33$SrYuc@Jst_BTH(TqSC6&^W`WjW0r=16@ zTDfEZHNTg1)s0VCl(iMLZ!(6EOg`u`rVG?VlB2(U-Y^=y=@vj{qU>j$jo(l!lXHBs zt(jDz<(`R8B2kfKK(>O?N~KB{4!;O7akTv0y*ys2 zC%7FI1eE%~fVVK+h?LO&H;3f0XZyG!u;`AaC3lr1N%7$}5NO(?1(@zC8=p}nK`EuC zs4~TgCJxBa^2}O*f$AyR;}y-~$({Flb-E>OtF%)C{AXvyWXXTc`QkvW6Z1yfnfwuHlw3P3 z66ZCPb~|tdRXNs&G!EX@Ut41;JN(O!8<~qv7ru7K%DbreoXT-7nmwEqoYnP}*tB|7 zCd46nOZ+aGHU*7ex>V1}(tIR=gmWpxHIw}q@1h$yL&h*DyocVuXvH>(F@kLZh!{&N zw{o0rq@|SKl!9nc-$|8op(|809n(`rV3v-^H*}PI=21f}&iDk2#XYZ-k@_lh>&Y_ ze$-#gAQ<_%{o`YbimU>~f9ATFBw1k|YeH=8_Q*>b;lZE#x7C^tN~+$pPCJpJW$-R- z7LA5Z3R^W0rx9=LJ_e{fE{`lV(O+JDRaN$O`0_@%o3ozx_q=|Qhmy?~NK?rHw$!x& zFeae7c(^%cMzH)7^!9^Pw1tv!1|wKKy}kX@M;r+PD|16}v=LcR6(f2vHKk(g2yTIuYz$v>zjqF@YNbV>kpsMrvm2Mh%%DypMN&SxUXUOecR2UV?1b}&}!ki9-(Z`4Yz07ubvkw=()5E zSQ7zu4YdDpzMmRp0Wt4VoBDoMKy?9~xNRn!MeLjXd{$hpBjiVJ%V%JmnU7UMx5+V) zxN`pRrC1{AZj3*{u>0U+W0IPGz5WkECkF_bcCUXs8Z^t)wFLI)u>dffe_Jg@0M^EHvSxHjr41=U6#+WCLDSSq7~vPrRV0mfl@xCqb_ zYyx`mDdAmkcC>R|W;?cfP+?Gkz;WdQx-3|}C$1wz*`dCW#_)GwB@f(8>qmJT5X8~m zaq5#Zsyb!&;5*+_g@6P*0;;dl(WFenbIQEGqs7SQeK>IV?}Nt8(^d&{ZqBD-6rM&Y z;?kOBK4m(ZgR^kl{q4hBtp?7<+DSq@lfN$Y=r z0c-sILL~R&+hi83Er%jTeknr$Ppwk_7DOSt>nB;_a*Z4AuOin4e0N56?zfJM0u946 zr?S5zz+DJKI~<9^C8x@fxAC|Tya!Cms;zkk8APyy$d0*>;KSJiw6tefzdY z^Zgu|p$hpMt(I1o_k+PP6u5XECSJMUa!#)X*Y{3nGE?H?B8aA{Bs5?3#Um3Kix6Z& z+>T{pQxar9pq=o8kJ7#w5;%7~Xwe>Nhx$+V7K7Nk!_cy*vM>1puhQQZguhl2778gP z{*3I1Se0~8kGh@}%q4mFU$URLg}xG^>P~zy5nX)13zjB?6pHBxI7*cHNjY3#2;b}i{36L z(b&ftcCx-M&R56M(MycthiObbQAt7n18JREQofX<2_rHD5Kpm_z;9imA+RCTx#zkdL~w@(^2#%|lFtpVwyrTFmnWKW0fP8bG8AB!jo zM+wgfg->Akyw76XcM?PXy&|z(V>i?`&v7a@LEtyOr`u=9X4EausCKn$fuFUxdAKU$ zp}bj#_`X;7C7<+;Ag#;&w{)Z%sSAAY>T{t~WXev#-@8RORwLbg{O?ly{q~Bnr2-as zUn8AnB=X)uU6dJz*!maI8-E4ao`~%c5Jp0f@wblOmcGgpmJP2HX>-0?PLIy0c!&Hh z{I4=@QsMOw{uw};S-dl1O4V2*kXP2TY!#on#Qj4nCP#N)+*r?&F1!5ve7h!Y2vtp? z&Ni{4q%!FcZpn%H7Rx~2@eN)e=G5;s%KiPsN!*>uWeW9>^Zq&>4_XCqd1s)mq_y3u z4Y}}~jDC3|QGlz3vG<8Qrt7Q*GyjT|m^6{Vgl02P)k2aX+E^uTP?_SI?6uyiN(Kwg z;jsZKDRAFycg+LJHEl4%BNohDtSX{i%81cE+PpoNso!LghK{ZL;jzboIz?kn_V(3;N3nM)tV~hfzU2lG*R=U9YZje5p&kj^GITos){K?6X5N*A zl9LZrT_@Vb;n4F{n*huWRxNE^FNw2xL6PiJ7_$R_h~DsUet}&`z+<-TV=Zg~|7!#| z>9ruvpFxCkoP8H0lBOd!Y(9whOwjM$RlWAgb?K?dSf=+()a#tgwuk)QH^B3fiJt`D zDw@Jh(=`<{@blsd#>F(At; zq^tVfX#SdU-j&$b6-6Rw;NW%ieY>*W)?cXjmmPy92yyBty-CeDbPrxP-5a6Ohit-h zIZyL(luR5K8z=LHJeoW{V^f?H7;eU4${3wNX?&)bj_w>2%!k+io`mijtRN&6o=?hZ znKjZl+=}l7V$uKr+>{*tWJHxx|wrHjOJbpRdmvcWL`u)k3=TBii?-laIoOb!puFrW2@?O`AG2kwSX^T zVr4{~;dn~K9D*}3d=xxyk?{{m&_+4#5mf}ZaN0fQ4COTZh4a-aEIdtKI-l$r$71j$ zIWt?SJEs@P%h8d9IwrBFMxe?nHiDc?Z^pju<_*iMBiSeD&3YG3aJL~|EFm6Y6nkaz zgM}`*nyD(o<&f*3K(WHi+$@5!Xw>%?fn2P?uw+!YpJHSe;~=2`<3+mBCteQ`Y&Ul7 zrX)*5XF?x>PjGZ&E!?!`@i#AeG?oYk#<$+LNGH+RCmJ-zdj`dfqkTV*MJNH#%s4-U zj$H@Nqn}A!&c`7^EZi`*I_KIm=*HsQb7)o+IfcFRE;rODD~6C^@7m;J&Sa#4uT2{5 zNID4vhRFM_TQbt8XPkWURoY}_R)*Le9w8(9UfeTAZ2GTYtfTqQv!vZDp~(1UNy)1eYu4Jom5jsU z9uD|d1s?e4`e|32M6gdSSUzMSsCKQhHMXb!1%F{9WkTfrJAN@HX&WXV5;qrNDet%{ ztM#JQcdAPX&cg;p!Kb$>_I*RtpgL0hDHZAF_=dS~6oX?DVm3KGfkI zuDk0U406FL%r3U7M7?rv&vpu$;!q`Ka!4&_#IOj%wiK**l&Dm-?8FMOW>d;x>m7py3O@n0NnLVc=^J|7-D%X#rQ-&E7M^Dxpt;AG0JnHtSP_W3@_ z9pL#%9oyD2bF1&v@W=Y5S@7eY7iy8j{52=lxfE&`toyXHRKxzD|4PBv%_f3Y=<-rQ)dncb*x|6+wDjRo` zw!BX-E$pcJ7tZ-w);Z5*s+3BxI1hZ9d7-5?hjXtIw*HCM3r+_Cp`=_jGowODdY?M( znQ_eq7g&nM{50GR?xfNE>fGD|=q`$#jUK{$<2+5cf74r3m%H3xa{~|ugsGb$fd-66 z-$$h?e(wJIEL&I5HQk@r%&*b2a?h#De% z&G%{skC3@apX5N-hcf2TPeIehR!X{h2+2x}YDV@%(=Kq%Y?D`fy&o!MY$qK^@2=sM zP2~qT%j>F($KR7FCuf0%v!TxI4-ufwyx_m&?wf0hSU*+&({st9)JQ^BLAdo_bGens znKc@bcnxCP=j%LLLlcyMXIWBWPSh0Hw7nD1OMONgW$yd4anDg-uBsLqM^C3Y;T$_; zj;XJlmor@GfjY&tH@Gof*aDnxi&#DPGNQsAK)bXN4xWlIhNaZlqPCH{K}mH_g&u`1 z+7qlKSH>1zLEE%uzSei4#Vu}qqUK~P*3qm!b7Ryw3gT~Gf5ai}^S>*VUaaM`K`p%( zT;HeY`zv2yvBDNdd*z+$848IRx}`as*3T!M|Nc9njdk^UGTkeY0AK9z=)VFeYSo!O z(#b~@AF>D01(+Qnnm|b*zn*y^!|@(5?$^HX&0*m>vW0WOb)1j?`ROrqPyFdpi%Iyh zNR!X0P;T(OnYC-_cdHAb16iwlv)&}Fqw3QJQW&>S1QK34dJtMo3_Ufk;0;g2e za)3Z63gVTFMk2YLigyq{Zz7+o^bHi`W-gEX|>skK?g`b*!gF{(IRvS+p)`@57P2zmhD_c&M z9bP61D4p*-_GfH@g2jI442m#x?!$hot)YQXXFq`Ho%uVZ9;wThQk^cpJG}kHp*))3 zyY^B-kl`J9%-lI4p2KMDNvNDBW5vFg_W9-qrK2^jO9&zu&1K@dYFd5dCv$Vj zRX3RBY~!|;wvK!W*09nE$z1E-?9!)-!AfO&LmlFfr1j4JhbRm(1jrxPP0yGCH4wPp zKG&mh7YUsfzT+!FfLjF{&rf47p10$uXX28^BN(nS^|pr_tPJusKXTQZ(ozs!qhbyB z!h+aB!vt=RCR_T3x??>NEo4_Cr=J?6GPEMqiHmn>&)yLvtdnnEoo&u|?P*1rAO8Nt z6C)c&|8h@c^-(4MgM8kPeL>s!ebi-r?d1h#%=?}2JGYa_45GYX?r(M}S;c3Y=NIpA zv{(nRN=gK7daT3c*KjuzwgQ?3*=}Kn*w`jBRgxx{`p}k!V$}sbt)<V}amTH-T>>B=Z!Ia%JpVKQ>bo#^4lteNon{cs)^OjOv{ zDYUiRJzgps#e)QS0!oORTE$v>z`g2x_c`T#Db4f0wIP*N`Y>CWev zr*n2M-VwI;Dci*suXPmvV8R9HDqW?bVfm*&2cGLgv@1ds17dWvpU^LQ6eaVy`~~r= z{ZLwdm#o_q5Vs*8JOb5cN?4v-5$!Gb=G$J-2v>^Dto}01pq2O&sY#g0j%Fm1x3Hw@ zXX+tox;y_Y(Z3igT%Xf4lHT%CoPh_n77%}f~3*wK;}7SLxU zvSvIGeXW?rCTQNX-Nr~Y5A7@pPMwK;1ziAB+I_@+z#D%$-r?PF{ycn;Rq4>76oFKU zdr+(zhjC}<#wH0SK^~6t$?-c(dSK^ccOmqN(ou_gh`NA0Vn6LpFUtV;Ze6hd^!gIL z3y1btx#j@J2^we!t>b0jg7=D5B>Tv7U(v6ne8$S-$s?=WYdha)eUtS%p`2&` z$CnGGR+y z5H($%7xbkpUz$|91R@cYN219!vF=)^}rqYM=8pnVpW$_*omi#~ILI?UV>SJ-ZTSC&$5u!uu)}?~H#r1wv6n z)pgQm$5tBAz6C0_8I*M`|6t_$XEpcnXW_3MmmuJ|e}!uO7W&qE`jeG)nbG#p1%Dd? zffQb#HU*TWo!o@Uo?ABuMiwgD8ZOh63sa-4p@$dR&iut=g`$yg6!h8QprN6mVq)Ue z97O|qC})u`UmQN}Qq$5Rsj8};-`wEe#>U2K=<72Ww&Y>^beUFnc#L`#78fHD5fPo= z-bPkcIYi3ZIppQ$#()3LVbj!%i?_)BATlLnJE^0i6JA`Lsp{fJ3I_*Qxa#xqF*P+6 z2^$-`wRLvQpq@4*Jw4WE)lHc=u5`{+$)5`J?%l7wA3tuR+J6wo$&NlUNssPxlf8dG zx9?rguPwytbJw&f1s@!+ z-+&y`N0IAUNoLBQ8ihK4M!)t?f8#JE-rN(vR^Z+29q2_?Ei zSF(cdCo|U@VmY!DrUPHIG&9dP%9bG&W^y)X_N*E}W-K$(For2a-6>VsCauHvRP z1II}DI#R&?%7WGNTPEIIvgP*FKgVbommVP_l_hPzbzlA6*8QCA!D)KPnmICVL7_3&f|Vr$f~5X%gamXdMJG zZ-c|C9D$34hhFjtX=+Khyl|&MdyzV&Mep3T*cs4M@2hD6OcSk{nHxa7|9jSbK)n># z+`B!T>R7w^vtJQ$S46lwubEh_RZU0rXG|kTmu27dj(-f2MVwKYm`o10%WD zdazJjB+#Bp7<+nC^BfqdORD~L_#4x4NFN_P%}$ru=X!#Vd{^XLCB-jHh22p(jW8!W0Qw{BcwMLSwF6wpCvbNGsHhUX`k(a;$hj2x#ymG?*Wgd>h*bjWVI_cOe zY#<4;6$EgB1@#YVv#+xo#k&F54jFNL6Vf@q?N3KHe-Cr(^Kcg;o2g(Sb0{kfsVqhg zbSbgm|J~)8slu+jRa5b2P|)gzLLjGuj9LM1!k2pEp|KjoJau)G1;vZ}V*4Wss3#23 zr4Jqbl##Gm@jYeF<5V{D2=Xcy9Cg|Jzlvq;{B%PkYXDZo$5Qi)ur= zG2NLFyUZKhN!y?d!#YP{@XO2N+AI3TyeTD|Ct&;C3VeOc^P({72fVZwKc}_EXIYq%#*#oYFoxtWBFB>8YcvFYUv%b+m9V7r!Zr0$*P= z!PWRYZ-{|(SHg23HVh)ObinJDh$IMU{7+iuybKdFL(g*nz3G=eeTS*-iI`Gw86#D` zSB$Qtv^y*+)i@f}`|Dt9vw40P?u%r=q`-7@v!4tkk%i^1OE2;;bu#j+a3 z{HXU8Sv2XE{e9ED5&E;Sx)vLMj}294&i*FutW619!&K0wyLge`gyAHA4EkxU=8?K4 zQ}o69J|SisR2ZMBmtm|K159_N2r3hv;M6(+lVSBrZ$qQfj8g2eUnn=hefL34rXVQPWdbW{EX~Z zG7&EvQ&ZD^VhIAGc2>EppFe+wf5?Bt+W^BsW{!Di)zWb6v*01ijMJ)<2_Sa?+RG7l zrxS15B+co=reBjcL*$Y(S7{%)zkRnpsbgSgicHA*7&@E{q&EHp6dW>Pos8;ky3J&> z_!*z7D1DKl3nGq%=l6n-N!`eYzDR^-4h_kdYcn~Ez;cIfzC8?$j=nonQCC;*N<%bc zH;`gbT!vbP7QVuKCFi4}qIUIi?*Tc`xr7-~TCP!n4Z8)9}8m&l7Eg2Z~EU)=p}79)xfCU*kyu zn=fTAP~rFm|fcs@$z{fdYWMgINz@>uu1Nc~Mje7F_OVJNaE>|4pi&z-0SPfDQT z&<7PZ`!zSg-q&lCCwCWAC37#+uWk_DKTG3y{EV?AoT63VsXxmV!TWqD(XO>Q<@QvUd-IlUYlmzS%L@bcJTjE{j^ZFLy z5)I-`$pK`0LdV6%gX}41kjU2W1cKkPg$ac(9iC_SqZSqxEPOFQ5;$Lb9#_xP zWvqaykVnWkn&gzJ&;*`I8Brauv|?!#_axN@vD8#yNMmRzS#Qx ztF3JRs3<{<>E;>G3Vcy?xS~c|;ysxJQ3O$-YypB7DB6jkgn{WSZo5I!+Lm{8rVp*2&E?SgJz8Qh`tN zZ4C~Yr#oBlIqD%&-cW#((=}cRNyR<7iP0C5gXYFR;4~75BmCvP%e&$mglk%bf7jfX z6XZ$w40x!gVz`nfb&XGE?~8MAS*YjwO1Qy4$UgNEiAc4DuHKVqtG z@Dk0$LGTZh6sB}SUJz~2v)K?KD2xvZA&RxRl2f{hQZ$@^2;!{KW>OB<#lQ~=rM*zr zdZ?!?b)SB-(WGaJC=X=@e3LDhXxR03>R`YxmPhbnpop1*f`cbo`bZ@i>@>&dJP}q~ zSgEMr!+f|x0pbQTP3YK$f$(iVNa4eSwh&K*?_Cd!heA3T9~IP-KKl13g4ircuotlR z1;J02lB)1|)cxSaPq9+~O-YZ;10y`az)EAFN=-sVtDhS}D1V^T&@-D}Z_8)mk@4S) z7vqPSj!e0>(=B1LmW7R`T>>G?Tfiy=rBTHQCD%xk-z8c`=NC`7O6HyHN?4?on_$T{ zvkWKXP1sjx;*LpZ)LTn$=0dH@_Zy^u$+05vxGjE2=O&(L-Sf*xs*??VDXT<}-f1vU zBjEOlC~r}2nq)L@WDzfVX!grtd=UAUTv@M(m!?z5Ac`GN&L)*@FY|24&7?F9^^Rc+ zN8e!u?PnrGV1x@$8>q~v{69@w zHb=za;Egoe|9MRh@N?vaxIjc|Een)q@Ut- z1zgy`)0MglJs(x<1~fUwPKLxBW75It|E?(j`LiuS+05(?lw^23($X%q<^(?RdV4x zK>_^*(0U*5JN@x|A)?ZDv3=C|O zJt@WWkbi&&cWi{QQWPdwmd>7q~HwvVzhx%iCOIAx7TjLYi;c2&O#v zUscbe2~j5|P1P&d)e>E)Di}uIk$uy8`x09S3)5Xd?y8&ZRomNyt4Z!S*TG0+j|~9b zL>wAm|5!GB`uitd;IaMZ+}h0LqppT_Ovsjk{DKx4V;u@Q3*>D(GMU^8W%HV*BLC|b z+^t{4yxlNYiA+_bxYe&2;PCTsd-u1kyYYgBtF2=E1oXLqhgl0wefOO-O0<9(A#)gcSVOW%=vTgDs zgou6s(x!(}H@Pa6-C?@Dt%1A3OZ=g@+(jo9r+`j6wfwErBP92Au>@xLF0>o|;I}8` zPH)dpyzswrhhcK{f}oKkcxPKQ#U}LJM7YT|`PVv_Aj8PeJB0R#e94bqugtjI53(Av zFh7QZp}XC+k|knZF)SyEcv)I8N6`tPUSr1S(+h6L(8z>5oDF(2FF2p75G+e{Ilmcx zWJSnWV5V$E|CRKh>AVMXN1z>}5ViEo5xAN-1gY(r>Xpwx_K@cJ&TyOVG+6T_sPEN$?EpDw|CjEbgfZN(@2mf{mAi`XFj6iofc-)drepY)1u z6RE4}Swjl%uyhCuJ~|G6M!|X`zdRox>fL_8DQMw{HP$=1167n8cP#MRxSycEtrv_QGpJlY?eC;aMUzA?MNVFpQ2#KV0~4)+!#osr&xc3Fa33 zqw)>iM_T0wMU974HbZ(f^Nk{~xLk`X{rFT+zJM^%F9XKHgQOpd?9uy73NoV>i=QTc z+T`y%7>+79jo#A2vH8Xyg(^-b#SPIE7nP@N;V#3miBau2Tz!uyoLsnw1Uf{>#bir# z?LG4m$FzN-vQb3uI57&fqw52CW;D3pQBMSw4^*R`6crTWrP79%v!u%KV>O4pX=yLr z;~_9?86QJB4bD}sEO9lj<|+{BL02phUWV_qzfoZv8ONxyAj4F0Xs zz2c-eEJ+LKxZFqRN=Op7fytmTg)H@%b!D zmmD}_7~n7o6u~8rZdssD2iMz?WpH8H`-_BloTaxRNl^;WWZH;THASA1HR*c&sbi#* zcMi!i&PoLilY%9ZNONz>zR434PEn2#D}JC|`kMV|G5asy`;qtgSCBNa%(w!Bcc3fn zf|6$ch*b(9eu8=ixFp-!lBu%L7Gx?(=A%^abZR1Op^@Cijr6 zJ9cP*__gIdb*nUiT*8Oe2id+~!XQI>3hlJCLa;CrtpOuGu4;5O0SI(<&j|b&^2JTQ zX85Y3aDKQzjSrTSM-lKw<~b1W110kBL)dy3LP^3|OS-qHl~8n8F;2L3wP)>W(>6g(hCLA{LeMNLet=y-@6iBoxdM+x56hi7%;gP^mV9Lx}1D(N5!qDPV?|% z);ZJZ?BaALkP* zOqHW#=rb^u6qZqdN`i)VWH^n-S|@6vHns4d@xeAiEBR_Xn3izjj_T zOqS|Vs`&^-_$0~_&Di;2x5l1%_kjvGQZ%v43ScDCm?`O$-(CDbt-Sk6VlOD|Z*%=5 zUp38Tqqw#~7>MM6bmmQvsaz;hTP|-G{U?lqY>*J25B5PDOw91Fe|pg%-k5`}zZ3D~ zRQvFvWGGK1m-JH6a|L#cu~t|%G8Y(J;f#f~Uj9YEzj>cqA=T3cD)E;q1xMkxf$*9{ ze9{bM6bmJcR565BL#>2Vp>?`9cvY{mRJyr3O==CBxN%w{t@4{X`J0*s4;QvB>!un$ z6@LeRVEWX9KE#S}-xuOf3}7-o|5hLi}~$tVb!4b9Zxde1b2ep1m6wW- zC#_OwshOSEZc~OU^QapYn)c@9OmZ5s^*EFB>J}wD(oO^Q%3J&Bp7q#_DRWcaU3gvT zG}pQs4};W~VV<9`!)}OZT*5Pr9ExcteUd@)Lb@xV4y=qqAYSouP$f?(k zm8o-=pZJ2TX|nvYxE?z;35o*^T>B_ecw0CA=RdTe$HPski}s8F*Z`5`owC1FDqJ#h z0D067Zlsz;ElI9Kn)g;`mQtL+JqVFLA7SBAm8pIdt@Z9+$ZeEn?v~hOtDy7TURT%` z+iqvY<>$6gT&GZil#vm6`5i(@DugbIk|cC^|3jx-yRh1G}`PZFI`M zO@5f1VgS#=@)tC=ZYpXJ`Lw-q*Gf7&Pio0MHF|D(Cw7UqE`EYtc31zeuXW|QTV5O0 zWFh_5A)18^0m%Rq#{M_C_%(F(hh;-nZMK>RVdT{QN&P%m=&^FA_Vj zMWss;EFMPhOTX$bDNjlbZ}3*EBm-OVF5#9UzFDdY)g`kKyjH%9+AlN}pCmRs`cFTat<+uEeg1EaRz#+} zmz*1cman!R|IxN7;2zQVgAA|1+;C^48sZ%3aZqKSDz+fl(wQPl*hK3KR^O7p+Xqn+2%i?I2 zJDYAx6}zGJl;UdN7)7QmGU8aDxhWBuB^tqtvNwu-1sj&(8k8|`u1ugJVuS= z!HtgQ48`SFqiAQLwU+A+xaZ}M=QAPczqU?S#ab-}9fpj5z}xo7!<$M4(~{Np_8?OT zAQMPcG&PMsdrr!G2{h&)r+5%-r{F+_ty5TJ)M+tCVDvrAv;m0;X-}!!>^rKt1o)|` z*4fnCSdwA_3-;+AZblur+uB9E{2G?aoQCbTv;uhz1=UK3pVD7z;?H+fuil)P zKO1RDmpgH0Yc{?_fh9ykeWPNHD-$~ZJY4qMnX|^)*>9-K6DWLdLe>t<8fYt!0`24X z)v`Akn)(|7S+p^0Pd=Pk#d|t@HML?n*or~!m#87ha^G7|_IKIiAB|0TVjDKY?%2jf z;<%`KkJ`UhT*m3?i&G8ch)A>fYe!x7o_9>~x1t2x!%R4tB_D-= z=CnqmKT1;1^Sm7^`sXuWth8-!2?BM@nuHoFB0`)HMzZ#N*8kdETED@Awj=%<{o#ew zVVF9J`NTWhw8dfHcbkW;#<81J!a84U;BfyrpvFIOs{PaDQa!~_yu?V2_UB}j7p4oy zM$0Sj+id4yN3d>YmFrv3D!d}ji7l0w>H73)P8eZ(2DQ!(qc&l~*Y<#R&Zu0g%^a5v zt80*1riZoTTuSxonH{#NnBFcjSCH6hQk^p3RjP}`kD@0c_z)oDE4Tiy)!C)@AU zSEPiAPbD>(9Z>ks_-;S;abV7v;;7W~#3#))M8DW83>>j;c2-I;6&UG4>2!)2N6Mqy zvux*shy_Y$uph^Nu!DPlZXEu_{SrV15IOD&9;&P>_wA+4c{j5^thFnh$0;5dZ<*fsI{1m^MbX$B~IWH2>bW)%2d!RTvUr+@^*Pm=AcR3Qe9}KOZO`VQ|ML zPt7Va*`-=0&p*K%D)1svwKVMV-GH}*riRR|$ACmG)Y6ZG6Xpf4mtZ5aIc98^Wq^qQ?d*0qn3<)|- zDZD0U>Agh6Xfke!TAVWs^t=fD(Y4eOEmi%k5`7xv3P<`AqnJrVS47;6r8a6msCg*- zh}|fzt1tqus}z3|ub^~6O-xnNPA55 zkjeuT+dvs4`(NE^(53qyKaN9<3My_!@tmeQ=Y4E+uWmHwN99`rHpsA&Y zQ7MmhMG$ZfvG8NHekBx+HHwMW3Qp{+{N?q*2-*$%ulqu{_AU zr03{I)QZL$bI|J6XVhcr`^qJ1!HXmlNWmMe?&U`*NN`l>3*S@t*F-pY;hyDxswGFd_xhk;Oa!Q-r0 zmW{2R7!gy2y!w)ry|Y@$R{2M2$(6gzksvsPbD(MZ&rRX;Xmk!%s%T>bF~V>$nQceO z=*Xq*qPiW0NlLD{foe@c!ucA!Zr;-&6Yu0T0ziOuN4+HxBJjGr5JVpTe*Dx}N!$4Qj8p*KyImdf1U z->0Q*)GWNQ1$>zOGpei|k2fb6)_Y1bdMxxFxuf+`h#nLXMS%e{U#tgCK__8Lg_bl& zfA~gEF8zj=@E5H~eavp7Y=(hw6)6wEtoO+p=q`VQk^Vb3N=jdiDnk&&?vfJvr0JRR@u2_t@BL*D=T*S-#cjZVC~#TKsk@#{C1T6{*_YF{({YDo5H(${A_n8HM8 zv##}QJ`uONU;x+4)rO&f+}>OT@9Hm_Hx%(%sFoxITQwjmPZt+ZpbXGw>}ge@&K_Vg znR!Mb_mX((%TdE}QAFkarVLPREl&jo(t+9A2=vwOc#JUY?nMFBtTCYU?H^TrgaIK- zMUC3TyrwZLLKBt-RM{d=>qJb?KQxcFj|D24{fr6=1+c%-^T#fw((Z8R5=}bK#_e-I zi?f2Zv@eqpA-;v*zvGDd-r$9P%*-k5;~IzsC$l80|u2XD6dFv){4}WL)$D<+iYlwdNYD@A8vp`^5^$ICywrfQTLk zTNDPxc>>$j{QV0W@{mKQ05qJ2|6cZ6zJ>+b1H2WPXymxGcbkLL2Ot4zQ)DuHUT!lt zSB}i2<{pTPa{f)`*?4Y46k*GF0QshWHyXR2&gM_g&7=7A9zTeHPDDfQE6KIbvF|ml zn`qWYnn4{QIk$i<6|WVJcej7mkX@bdwd0#q*b>Bcv+E}_p6uHA2Orwt@thJauheiZGd-oCMAgSkG3jQ^?J zTPMNv^hQ>3Y&<+PX^Dx!|NiM$oAsEYx;`8ye`wF{QMZw0m z)MOlT>Yn^Vo+FOV5!MWS^Jeg^f{diL&7=_HwBL*hzX5rfWhgQhX!r;gaJJH+7j@oX zJ%JvqTJmXPPt@F@V1D?*>wx=o%=C(n4XWRIT>iOV%Uly)zu+zZR>1+z7ZRR%%Q z7mtb-^ohSuUJ*vLabyq0{M=(+8yNVc@Kp=6ml>bvb(;A~)j#TC)ksZ)fI{I-SbH(d zH=5H=J$2RLxUtwpl55d0fPOW#kr(lP$_A_*M@|bsvJxt4xrFhPv`B1AYsVJ#Jwp0Cq_30mNUuG ztU-r;neyO84hPZa9-bQ?@d|&T=QjNJ@6nbJo`r&T^2{!1;I7}5nsTT&AL_B5pY+b(^1t{+5H6o99w>H<_8b8HKwFh3G9%$F=$4z#VnCg= zU^!GP)gzcK$X{8s5JC!9Cu; z9}&&c+!%R&mkQkRDdrhz^`g$$V#bfSR;Alg))DfX%ZnB7$a60-J99qQ^tBl^ukX_& z^ky(Uqm0~{d1JNv-*EEO*4}-Yv8^5zLyzll_$eB$5gxAP7}l~?V*BjCdi(`iLScTX z)=bC7#m*w+GQ^5jy+rPd!EjrLuVAax`(=g3=t$Fq8Ty-k{&u}0_6sMc&-H7u zg|<99YWunJ<}RJ2 z{BV0CTMQp4r6Zr6T~bksG@ERefjdK8Uo)6Lul;=;o1b2k5C8yE-R&p507z-W@lo4N zaI^HI+LT4+7kc>&by!lfQXZC&Dse_qO>`yB*f(yu>feSRG?1w3S%)mTc?))p-GFxm zO``IeiDE{Y_|l$G88GLbjC0821x@%CpFF^8fNhfxdHN5rAXI2g+zCPCx(__?(~3tK zE7urE9!Z(L%^bg`2z)ZXBFhr?tQBZML=#=)KfF`Y#rE#_YwBAg@Rz;kseb1+N`~Y_ z$+bZ8?KN&WJ4M)dqMrzisI;8zd3W~aTOKy7Ex#B3d3EZI<<3Z|h4A&lvyH4A{(y^2 z8|S_+Ezlq3;Nf^$dwsFKBCQ{re3VBhww)=R>&V^hPJ_pTtr#}iB;G0Wy!WZI_q{-S zILU{HyAitu4igeh<2^nFkVpzhIi%h^FetV@P(j9?5zELG{4VxYeJ_Qj2MjJyUy3`na}r`uYNUl zuq?0o-+0h+?|CKuASdy?6Uucqe$mIQeB;JIk4srfAw3cj#8xk$CH8j@DogCdHKMX0BO zLArPbWv#V-gx1qs+BI+$T2`l50TDg0Rv<7AdoDfG#j7JtU4J4MZ~H}pmt4N_^*Zp z>9H|$Lxly!=J9JQA|cT3MaijeEoCTZVxU_JWobN;R+T9(@4}4MZ{ue5kC<>*HR1y5 zo{sm%Q@lDH1^$T>SPvqB59@x2f>?^SK{7oiZdv0fsqVD$86Xu~Hz1j{C0a=< z5k;}d|MK#A&`Koh!7!$HX8kL&A&;~R+y0~SUdVtL@M3H$KWoVj6ZbdWe;nrc>d#ok z^qbzOyDt%@4`Zn54}%{zEcCpRDl;lpCZpzZ7ZV+6({F|2qJ*Et;^|g2 z+G{uOTKzBnmeq$SEjWwll2(>2_U}p%fa?E}lnea_#vsxer-dIV5R1;OO^dm%_C$WOU+c0Nj+d!-1Q>HOv zBw|(5IyegvO%PP%a``!(&E&Hw^O|^jApIqOAhTUwyRiKR*#Z$$9BgazJr0uE#5=x1 zsb-`hv7OT6pKZW*MQsmgeL>_r3{)C&&6qL9xmBag5G=-N25I=j&V_o1uzXGDI5Zcb zGsn)0hzul3bLpiV5%2oD@uVTDDM$aemfMdo|HHW|=SNL{{Uz6JR}-UuGxdw)&M~z~ zP2tFxZnX!`)O~uwAxe8e)jZ&B8;L3XS_SDLc>QK&1L^*(X0=fG`v!?T1LE(AO_7yF zLrc}{kr4cfrrX%@j%C)2P8e+UdF}N%FT*>LX1w0Aj5E{TATboVeoRfa`|R{|b=@cP z!LWX5vNt+OaK}{^K`kVN2_|g!ELa5B{v=YV3mL8fj~w{gJm*egBBe%Gn^c6GXI#Ggy^`*{N;>o394tiZ-z@akXX zU^SZzuT_*l@?XnE!ueNKMmuA5EU0G7e6=P$4#4IFxK+GBffNSWRmlfHPbTXh7?UzP zubB&+EjF4;DikwwOZPFNQ6FeZEgM%ByJa&$W&5PG-{$cFR>@YBmPxKvp?B0G_$5r? z=^gkDuvT8c`w7Vd_%)%WkG4WtFVLabzj>m7J?yjfyFAJw#=#htzFTj^R} zv&W`-jwrKNvTUqrho*5NsxA~NW>qH7{bTo2gOvZCclnvH#0@^k0zB{tJD>rf9bCDx z{TE{FoLJDu=pT?}iO*l-~jqj-)tklqWa<^q$~pok+PQOZ}(U{ay!5Od7Z z=kLh;9;}p^3i@iLG9(LWz~w>rGSg%;{PA5 zP#&y9DX(g==Ld{@O5*!X@5Dog94)!+JdhOdaC_$H{k8KNX5?4f$MKUxrL12)`j&6} zy~T3hr@@?WmQLj_->y4vsLK6A1{}B-Lifg`ZvaTTPR-p_s;~uP+3%U-JjPyK!*XCt z{H(wIm}c^Ro=n8vwd;q|4S1jDq^NTP>D_iy@;k96h4EtXG4{)M`!95;l_>0#JNoaXoN84} z3}(JIZj10BkI8B@&IwVgrN-I)eJHbG9-B^F%g$B1LCq{F{4}vEg;2bh>SiM$+u-!c zC=n%v3-+QR_GU~7N?3L>ZT*dUdES!8b`o;dAoAS?^^jCK6a!SfJy&);6BLb1uV?1p zH0H|aB(X-5N%pzcGCY4BjrK+L8hDsfw>snF%97RCOO=|kGjMb9=eLAe8XEjRSRitrs`cKi$fmLq{ z4{Fc&H%#%CwA)f!=6@zo{xa>Qd#N~D1}^Dg%Sdbr{6wzG0Lfw9rm|m_iONLx*APGX zl*x~KAn5!=nS<7QI1n6|P#!O7FC|PE;(AzRvYe@Ine0ldrJE_TUJC6t{=&g0ct6c7 zfG+t$#W1WIF|8g_Sxkd^Y{-9KF9>A~7OCfnmC#@V?cn8;_V%6Z zE&nY&4rN(pWK0%W2eGfcs$#U-CZ9g>0S!)3rRV@`BYo z08F5YLAFS$=2Z8BK_&_*Qx}t-A>G`p;RB@pssQ*iz}tm_7u^gD@=w_iYbPgUfmqQp zX^V{UL6`OX-4&Cb`l`rj1*if#x)rR1ma)r6bvYcAvLw%sumfD99{H0QAKNn=%|y{@rb@k1>ZmbrSS#@vR&ag01)Y{ zVNKV)S;#^_cf9j1Cw)Rv%WO%TAxU*F0a*s-qgea`n^F7qXy%I}|2}}>d|KTt$2?>A zVrgUUUotf>TsG7!`io`KNqkuAl|1A3f{C$T!+T~D-Z47<&N2C(XW^|bwz}W7)`BNH zB0o$&+9&QMqQ@Col5ws=aO03}O|axxJqPHH1MnH9C131(=Sal~@cQiSo-H2`ij@2( z^FAsO+p8NiKa5bARu1I|p-IoRb+{Y8`{&MZIj$>NGq*k&gYD#_McDrkC+9SgI4RIr zZOQrQM ze;vug#;G)6?Z3AlIUC)`vQ?(BBQ#5b_4FyiBFkd>a(KbH@`#3y*1!?%(<>#eq*Q#h zfXjlkkLkzvfeez+%3x?&CoR-6=GbB^cz_zsUI-tg?pB?mtc0)_G0uqgnV7>sufIUX zxUCY}=fLDo3~o4pB?h*?;*-2kPSaP&`2k@u{H3?ApVk1+pB}$6R76lk5l}?CDCJshAP{T!;7J)t~iGngQxLcqM|e`t1FA^dTel zlu9@YO4E@@GG%d?B#4cbmihFJN5CumpE(ta!U$!_gke`8J!hl0-rAgxXq{}8VoI7l zZhnbJy*P?z42L8AE%;^|4hbL_sB?}@c!b@lu80Ur0XjE1=OFuiRU;R{ZMUZY6;j2!nfUi?gtANiPRrQ%b*bN!BHT#mhjlcR1lHz<(r6BkZK)W%hg#Z$J_dD~HwR;PIwXl{xKw61Cyv&fi-275 zL+RRE{4F45WjpPEJbeX2Ro~OK3W#)fhs343L0Y;?8bqYIbR#7l(v5&1-Q9WV?(Xge zsrSI||Gazv?mlPlvu9?_nzgpq{N(gIz7};1X>ILlf6;~OpTZ#K5?~*FVddS$D&-(@r zF~oM73uZvaV{*zjHLH6llV+kOT4eGRj^?wJsE7;UhI^ z*@9wTwc|if-!H@Derh!N_&4Ww(XiF!{m9{&OmrP3B>EC+0kMZW`#3)V#Uq3qX&6tf zx5LEo_|M-71y;9hwDr!6KKc5Rj_}NVNt8jTFjh2G-yRn&85;)lDf06Q+T2sNGb@Tv zZ@$H583$lKs=LCrUu-{jC>lb>=1{Wu2WsXbSEe*5JcbDgXDi+Y)>@$t^b+6wl&dgvHdx>3=f*+vR2NeTzl?z*lv3O8Ws;M)(i%*135dwAH?uR% zsXxS<8uq4@{%&H1rM-U^?|-_Y8H$l^K;wav>=LhDw|5C}{KdGYyNWlDhYw4gvdQD_ zFvo^hP`0zfrI<`Nl{3v-Z#Zhj1n=G0_f>51`othFR+>Iq?Jv)C^0Y6=8feEb4|eC} zu{z(Q(DcefMDcPyjLTH^Cr3rtww{48Ete_!@$u!Rr<5>1G@5e0%as*=DDRxD@(9-a zQexZo&e)W@Yg05oZki&S)9iF(b8i>RK>K5`88DZT^Clnv%||6AKf-e;6=iwPx$7O! zeRmZyD4g;q(ok6urfFJ@Si69)`)|^#50^RhG{Yf&Dc*>@;c3B%ENoFJE{J{i(_}}x z>bO_q35g|@8^}SSpE7}Fw!irAiS9m?{kj=t8q>AYf_kGjpJ>#~0pY}F`kJK5K=mxj z`>?tr3=ZfPB@UB^mu^{ZbNS^ch_d^nhXOC>l=buTNpgheD)>nB^Sv9{_)L+%mJ-IE z`uFPH(@B+6LVn!2d|J(`$AtCP+Ww-h(uQM2v4l_gqccUBPt4QC2zm!2D#@R~5<2RbqAo+ zOhj0V_P03p%7w&@7nlrw!!)4`l3BOzL;h`lmWREoyM{Hbr76cAA!jo{+-k)Wef?ln zzj81yxk^O#v1>brPy`YTOl?y~qKaKMX7(MJ01ud01DW>M<*u{gAnv2_tCT|D=uhul z^_p7x?6$+Uc{q<~?-=}5(&aVH~CMiz0LwW9oV<0f-bBZPbDgm~#UO5vxYy24rr zj7(_?oX}40bHab$;UbQ`nW$8nn0+*_$#Y$(0%etaIv~;E(-aHUtHATY_?aMceAcat zgO}R>rcp$v0MsAleNA~DZS*UFF!bS-x`q56Bm*Di$vbA3#Z}sd2uGe>S2&DTGv?r4OjaV+SG385AkesM(&w|DX%D zKOv=5{@OGzk1^`_p+Setk(M3*lDqjz8@G`1Kz4Ffdu_F!p#~4wvTG?PDMrqc+c&j; zas1xRQx;qV%6^i4HsgN0ea__O7?ucYbT|T$Z``x?pLN_=pe5Ddv4n`k=1<*H@Z)zg zI+}dtDt$AU;liL(g_GN-tL?KuQ7t<8rBBqFRTUmA`>r;BN^P^B^7w69CfEUkh^Z38 zi4)XoA-Z6X5QLi0D~!=y5wh;F`8V~G-Z&WZPr9(wuof`eHhu3lzcZey3ND8TFvCAu zyNFO)L=%C`PS9gp#spU;K8 zO1*MccH;XKo5WSGin6j;Er6#2xtSPCO~OGO|2+uZc5wlL&+naK$&{f=q3T*drZLzmEVi**huk++jJvw>FBOxj z@jS23#deB~gt<8go8gOpTU*=oEdNB;2Oi;$8Wv&slN3TS=c|M77;-nkg!Irn07q@& zI4hy6TGAF4oZqZ~DUi_R2D;1<|F9>3A8{mlyAvn`6}Bw!D-V^;2Sqy?f$DK+OtY0>JfzSIZF{hF<40pD^>#URUfW5g`&O6eczCoW2N(x~-9b?a z9DPFaaH+`F%9!)`rDZLOdF<;ohjT5OQJ%S^YX5Xxp>M=KC7j_ndcVZ0ZKGcr>OZOr z1;zeWfWY2=L*HxQ^!#>dH%i1DXYKeiwYIc+cH2@%^{&JpO&4mtTmbaOt-82zPAj{+ z$C!Gp{HgPI12nH*E!LqI6)ms)q^Sm3Q*(2-qmt6EX09RjN)>7)KAv&>eEv0B8OOMP z?=?3i6ABYH(*!8#DRBI<235G-UE=DB?au(ML+#~+f!~fb9TPI(mRw-6El9798Ix>zie0E`R_D5){^dxt^|MC#4J8x_Nb9t2feJO3McOrKsa%Jo9Yni6dY0GZ5@vi2L)&uR`D}w9E4DN zKq&NftRBt%1OAfFR9YjQ(UJ2SoW-V~Zml^AM`Skx2XZ+_28TdK+Df>^t64QJTccF( zdYOGfEY>{DWO6}M*;Hv8ZosOZOc5C4pW$^4S^$X9N9V@llc9ZIyOOLfl+?&R<^nMo zZ8)85ry1q-6!BD?1EN;JOTj)LqZvL9{qZ?o)k?(T2*5$FPIu|Ic{k2stzODbK|_xe zpjU&9l_TOq{T{I=(6|K632U2`*8Ou|D!`R%2N3E&V19a@ADe_E5@@o?Ev9o=#J_AE z8a|~GgI#t@(b%8i4#d8a_0@fp;uzD;0~OGk97kVdskoUnc|QW17)<=O$UDBRHrB7{ z%k*keW>kVoqRCUggI>PBdGz~XV#qlZ-PzZO%N0FC=1$w+Z4zN5;M1N3uWZHKD(snW zw*g~<#+a#O4ugJ+JL*c1TW9kbotz9(q?442{rmhFz5IpKo_7XZ@noutGP^LQALrgWWBzEHmCvFOT}$%6dqy-6$Dq>0Nj+9YoY#X%CQn&cP2? z4v5qW({Dkula!Qv05mY_03)3fdgi3(HGX-sGL(j|2n`Z*31b zaNB>KAxISF88I<2(d#x~1_uYv)>x%{$KG69BcrD$;(0EECytK>Wqe z;97rFnJ4m+lZ)si%5aG+D=M2BR+BnkBU;#XqnDNmexARflZAjbyG|^lNtycOI+CoE z12A)>{StrOl;gfo&l0IVG5V7Lnq9L+6S!J=lHUW~@|gB3DQQwbMhccrLt2aFKK4$) zZAfw(i1)LkTq)HDa!Kk1!=VxKH;>wi-#CR7eEB+)Flm-YFRrybM|I!B0AT$cY*mpT z_vH@E!a|}6Q|H-HsIv7Cr?LZ=;c!TZ+W^)uv#g4yvpL&S_yYV02&1H*PSV-~rvKdH zyv$Bj-i5`0KST{V-h_hpZX+V!%dw05Y(I$3R?fjx@Gc<*aI7D~pf;D@sP+e#E#<`9 zvzh6wo#3lT6vG}JMNw5@PWdG&#dMkRRs^zk>7yn-L<-6`=A07+C zxHVfTHTvo5=7cAWL8q9ky#d3dw12A82EAEm@vZ)GiHPM}B7&e$Rrc^dJwjgO87Ir+ zso451=(}VNi&4PxO*TW22tbxa)Z(m^KZ%ZBsF&!_>dgQeQ^1PW9S+#>F0MmZ6KN`hU*;({qJm=5$p<6Dh4C;^6o?l35e}2V;rqUGCSj(ELj&H9(=p0}4Q6 z1#{FsaO;Ld7vUr2YUG{Cwn2Zj(6Qv-;F3Ls^uX3kYpqoJNp3)f1sG?a*qpATdb+#U zwi5IqO|BQGfP_Tw*WtC!K;zXmpIyM$tL|ifKO`%Qd~<6n`1`-(rN(l{1NF_FozRkW z6rl&^bnmCTi#eoi{zk{>NS*w`yK!uF1UroknzaLKWcUDd!C!t+PM1uv+aI%wNRi}; zh2wjMzF`a>KJ+OokI@y8h`7nE-9)`yM#-Nnx6M2%wk%@z2Ufqbul0+<&`2jVWMfbT zppU%Q@lh?dlDsUGbTJTKOy@mtw`=>bxLD1VaX+9H5{}Q6s>OS(4nx8U=bg{kc`Gm{6od%VZGJ_Rm+#%)c3_REirF3PpD|(paQ@n7O zwAvvHJ||T}Nw8%#UiDAOf5_cw0{_m;Z0OX~J32eh09k?1?{`m+_vNk^-v|XgaJ96w zXz1xNKcYN7Jd749|WDd3_ivJ5X$9>6&pw{6iocbS1pl=0#lFWNRP9F2% zk9^*TM35s7$TcqHeB}pG}H#->)HC{ zVO{J5A?hl9xGP!>z zBC8GbUc}GHBje0Lp@3b6j%+L0Y1VsP|A5!oi8B)<*Ja_$nz0UF=7>Uy0sCYuX!`ro zDao@E^RC)@%eXMg$gU-BRSnzXmmFz_)g};V`yCO?gbtwIsS;~m9HiN4SwZR8?}Swx z152p$IP6YP(lTay4KHb?hwkgH1GSO|Bq8#zzgM^!za`^!wn#a^S`oiU$U*#;gh@@y&;V{?OIceT%^6muQq&(-&OC?a{PlLC{(kI*o^Nl@2#y zagtf_r=vp0M4h5`0d4b5t=f3jl{UoHmsAo8)1{2w?yR1zm8~uc6L24 z>-*Bk#5ZZlCd*sCODlloBj`K(j`ekPax#e5i4_OM(b0%nyf>6pUK9WNJCW7G(F<;Y zXsq9VxpnJsP|Ka_wFudFRXH)ME9m@?>UymdOKCF0|4m)Q!VcnN{RnO-fF>gXB>t7T z{(0JrymhHtQKYFaS~Ce)S}<7#lmT}<f{% z49VBM7q1AhKV4rrnGXU>hC`*}Kfa-Ygo&)vDP*pGO`f8|98FC9MMCI5v`tUL8%J6j zkE=n8>Eb%;!awhg8+tZ?+9=t(Gm;Sr7}~A@bEPFWXVKBr?J&`jWACf|~LsOya z5jmiZUW3n7G?&og!$)b548wqjZqwQ3E-dQdVfbe4Q{bYga+K_E#ZM*<5IiB+!>gjU zweQ}sGj`s{irXTPgu4J@iigc++x!iV=#0y2j`1&v2mpB}4pPzc)V=d2 z)Hkj45Suhu6gZ{Ill%6^`;2wGcD?~NnqTMF`d*NjGnXYIf{&8f0lSkLx8dDYS(Df94nA)(bTURLMyivL-W(NL|}XBsXFp-`yGM z|MpJ1!bcz~yZyO$>g>2(TNplS4g>0|d9g!Yoz}MLUT$`00SpM~wQCT+6xUR9_OWB^ zPtsL&wTeHW7m3_SuCVgZq+J5O`b)C5u?ms5l6NI*kPS5uYKa+Z{T z0}KXpM&2AR-FtA3%;vM8uI(=f2id)RGKRyEX;SyhHa(Gftm<0v;;w-kXG%Bi6Jta_ zn`AK`9;l>M#+6d@Yx)0v{@bqQtU-8+(%$}Qg!iEMvvd2O(c?rO%8->TPKYa<9gkrp zZIo^f=QsPIOGNuMC2cMB*^-Wz?c0IfOBOPoF<$dA;T(Vsxa#&4wf0jl4j`iw7lR-L z-1YLS7`jF%y^XzvU0YEU>iSnd^@T|hJdL!r|EY;5Tc}meD(V7#b4JHL{PEhqeicx0 z50&_&WUzVx2F0nSHf6>0(1&BRlVrn_DC(D{LzY?q@z37RfnTwpKs!wR z6DXo@tVSEM=U@7+Icn1^WkozV-9(C5F`>cWI7W2!j|E39i#?iBB9E-6)C7>Sf;|Ca z&Cq7LmlN1I>9_p5XJ?_cb2(Nt-UO5be}mit(@6mvd=Dt6Db;O4B^H9v5{!MRTDr~` z56e)S7I@yx4fCd4C5OYuLfktvvpO3q->*{D6jW)sBzC_Z>j{w%I_cA=+yBSh0OMQI$3F_ipQz)=D&XWkVaRYP3pj5QJPi zq!I$uh$KG>VD6w(hZ!iM)B-&3uY+mR+Wlbd_e5X!n9<_pUcCM}5~Y-Cq9V1GKZZg+ z$fmi-Bvb_>UfAM0_q8L_j-g*g2JLGUucI|!*y8Nj(Du860oJNWHIsc9;afBQm{*C5 zy@ZgVtKmGO6VnIz@E^HyDL(;sp5BV0swzxi(5VX;GhnuyWmd{)s;C$TfT)*S@@GA1 zYM)$JsKrV(+D0a^u*+WmUwWe^%bOLNFFXoCXP#TUsY4v(d3l7Jg!=Ow`1&^|>Kb7Q z`ATm6+CktniGX@%q=3qZ!r#@7Kxa(o{*bLmv;3W*p&{o0z}5+nilJ)n?1W0;wCVt2 zgeedjUiYes_kM)9{op`QL`l5dL<^7xa#;(ZUOn2REhvt))cd*K>h0f$cu4OorUN(1 z(uf)d`q+Xy3suNBb@ZsDmM7Y0W|N2n0zbzc-2b*HgKf<|deew-F7*7MqN9+?Jyv14 zH=dKoW{M1$9|yQzYzLblVAGe~G`l|W3v3H>tKHnt zEY)G=q{RCs+gLN$i+JFn{?hiftA1}2u^Z^w?&UbGowOBJ;iAw+DOH*O*uiW;Gecm* zG`l34o|s5elAZMjMM+#d;5+GM2LjXVBn2R)8wP+P{~aWd;^h5|XkpphD222EU^)!O&pE}`;14}LJy#wD` z(`(Ze7ru^T-_&SMxNIe9O>VbK8pQX0Il)nYvUA<=$Vlz~J>+3T`}ek)Mr;F%4(kC{ z^i}7}44*R=3WkCHNK4VP6(OJgP#bzj5vLqDPz+qo7g!I@-)xuWeIew=oHUBbq;VXv zWBGB3&NGU2lLZYORaS@AT7bAT+Kj4FfM%^Vk*}X0!vFw;Tc-ECKA(NQh#@czPtlG}3v(vf}i!CiB|n$zi{0Bnv`;6MKLummdM>A|O6 z>~+_-5n*N%Xn_`` zE;xUVb>0t8Kd|V)=wf#X6?C01)&B=MHK5fCkU4@Q2z8+0qBT&Yv1KOx5csekVj-Rc zCB43A$T<%YCuzt+u*U7Wf-F!tolGXh#RUMM)))Tu%*@QJ4rv2}-*pP)o`ABhW|Irc zx4}fi5xH8aI+g#xM$OOoO6S`O_wgMDteWw#b^n4l<-9fY2Uta&K<8dAFs29tW!0Jp zep)_u)$rOPol)!R8wd$tdV5KTl=M?oF`Zz!^M~yv9!%+vz!{316%aYz*$I3Ra0UdI z*Pjwuj8~73tzA>ie*v~e7yz@m{Qov~7@=7>^3EOGSmr@KWy{mIapNh|xlQZftJXLA zLGh=KS}<=bVWUhHv#5_HwWK97f~auu{2FDYUx6?9fpTkhF63CxF#Yy+T9_WPz~U4c z#daM)vH9{!?bmp1qQyVbJ{OrJy@S_Tu}9B!YDJe85AW_$RS#P{(T6;3S`yODd}zI; z71BSvyz$Fak6md0DnYVytXU~gI?M?rKzP;@qtUja3rgy(QUGS(UZ(*80F67*4N>3n z89C=4>i{cGF&ih85=p_U!5FjyXo6l4TCSE zpPHU+oYHzytIQvmO`$`<+YiupFJnhnfL-?iy#@SqsqugJ*yu~z!*`AoPjrW&tcw}^ zxs;iocb)jQ!DWGe#gd)WzAx~0M+}~|aFS-!tek}xzCO7j8bn1+Qt8RN@r1T}~U!Q=Pj}nK@$Hqj#_doDP z3oJcXzg5EJ%Lqb}mn0Rxs**1+AuV)>W|m6c>v)3T*it=gEtIHZ0oQD;BGg2Fj{ATwOTox@BLs~B68j zEYW7HYw>yM4O(Y_>! zdW`c!L$stTA~u?L1aGX5L}*zE9Y1c#osBht#YthG9T$-~Np{q{`dW|UB?L3eh64O{ z&AMl<2hVZNm16f#ru^)OjCH(|a{Wc)8b&=GrDZ2=!l{nncF3usYTiNr-!&<-x~o;G zx_|%H150AcH#C<*KtvC^EGH$I^pK=!oK5`jZQp;LH%AMBK-w8Hzf&~?pgkMFyh$fu z%A4F=0thr0Y1hh6Dy~$TM&jx;UC&!20&2A|3$U-D^-w`coJDZ%ACeK9!UBE9{1Hg@ zga=MK$Eqv&X++_0A)H|VC%di=7xwY7Y-L5EA5tY^%)TBkO3lJbOJvT)RYc;25r0^i z(zUl&)O}+w(XkW^6+Z}rx>-_EvhJs?re=b+Q``-x+JXVvcE|T3srcDOXC|+QGf9v0 zK57?$1e?-S5=W{ zMJD`y{;$8kfA#1n9FT)v1wgj$I`9VLt>EjGZmeHJecV1^u~Lb)@wUXvWvk%te$ZJ~ zlTOjo8Xf)9sT>(3_)+-jD&$PiY|h1OgP0701;RoOOAD?=p1KOdI2p$O;WO6qFwKg? zDDpIG9BFueeRK;H6#?Pd)h@t-{Bt&c5?G@unI{3M5CZrS&_{jIw(kIFpUG7%Kp=o; zVN9@kLC~50Z%HIH%r&>Gc101oHkD1zid_86hw|JO7SXI%aagk#Blw4pH#WG1f5a5x zl;Nra@g+Sh>GJd5=s0h(mqf_?wA2eGawXD>DAt~jx*eHXvVHvV4?`R;cM94S&{qZn zcIU>+O|Bg=0v6Ns{(*rrXW!78-EXYVoPadqV+k-0*2fz;#){rqs8NbNW)TYYA2_#3e4JVuOXi40t1Ws{GU zuTl~tRUM7ytANfj^$g|8@YcOaYtXIaylNw+~& z2R|z(I3qtlU*zz>f%o!crFD1%m^=j*8hGj*<48pXd8!s<3h;dCy5d6OV}02@>h@&5 z#!pri%;Z-kvTbrA!k<`Z#k*~@b24g&4aNch44|8dNFsn8anRarnlAiVg>Sa#{$IAK zWtWTFh@lJr5<$FTJ(fMhsiRWGjkC~@e8Q`|krdD>v>2U_!le=s5!nxK!T^#z5yZba zIn~LewL5~nR}c$PBSBu?d1Uv&mxE-LBG3oR>GoK{l~PE_}q> zo7w>G3eU3Q^AhJE_>~UtfkCH_sG9KhwtHA$^IAwDkPpf9Dtqhzk#1E*$Yh zj|1Bd4hXxz8(TrrNK;+<(^Lgrxb5vz3I5yV83zoa8<%$c0BZn*%-^R8k&4L~0nU#N z?Uo;_IH=f;7aeDn;j+1DlpWTpydnnEvU%6L58NtoO&Lo_TTy%E0`sbjM0c;;z(=dA zwpxds<;Fe9SUj)s?>$XvLEB6F=xLII^RKZ{0{EpyR+Xj=q~oGM8fbmO4B&~5`$)=< z9&=oWVy(Bl?$sB+mB(L=vQi(gZpl#CU){5gs!0yB4ifMHx4OMS3~8u0tc~3Lp;2AZdySi4-y&iTQka+-Zd zh98Vr*qRa8e{KhERmpgp$b2CGy}-{R4jic10H%#F*Daguo-RO7(+Hy5i0-!X^J62n z#q(h`S=tt?MGS=$PTW`En$Aa(QY&zX=K+2czhYx43+gmQ$&)-5Hjj=c^HoxoUN2#@ z(w7ngmS89IcUEu3;6fk9Gae7@uVuq{6Ijk?x@fNaEO?g90$N;;sjs27oSUYZwsSkA zGk1;Yzc@KDtH**GuSsy!v0|jddzvAn`nod(ZWB>7O<-donz+=qJ{|IKA4X{NYJ5OE z6}vW{;+s|8BB`1!9_!bj?w%#=|Gv(7i}P)Z1Hme^jj?4km#czjK!|(ndz(<}Qy5Y! z1Ex{x!}`pk7t>E2fo^ihkXjtg!xaTkl%6$fGb@{GspWQ4k>=GQexo@p7@a+{n> z!1JRxyB>T3qSQ5YZ!pl`)7NP|gAimLH4g$EmD6zo|BFap z-orGg6KRJ=)_5Q_ms15=*t!Tktlajj9`8~eYsAJG&NW4XzaP8@ugnb*aS1+P@?H^r zuPn)$x-jgO?Er_fc{w6g&8dIYBP%JWDHy!M4-$>2iNxSS3~OuTLNZ_(PUV8PmkIUJ ze>a%QGcgV`F*2esEBbZi8W_QP;Z6`R;c(EA`|tx}@Ro~pUpEuI$pG0{tZfG`M{ zyVyEA;|%<}ItT=eNbZ`hGwXY$(vUuSzQBdD?rjF_22bhYvku<^htAEGN+DMQr9R4J z-MAN(Y~>Gy3coTwKH}U5_#5sEptMhQa%L1-j(ou^y?g6M!%SVr6kTK#Yn~0qWHlSU zx~@#`eBevc{BLS56$1;-G)`*<_AWO<3IY?4X5(B0CT=6 zAN}*Yj{!XDfSwx6fMA*N<;ec%zutD|HiJGH=S7uC(wBV5p)i-Rn5x|QxBfj0 zO`c~2RrLg_Gp%3UI zZxD=GhG*jjctxxkawjGxx-V4>3=I6N_VcAsf-MS&w@_ ztz3Z^rPzQQPA{KJBidc@y4_Uq^KXD;NxX@qlO;r`^#L)akLDPq$%y=?TdG=EDPeAV@X1i z1}H`3<0E5ZF_j_#!Sm(f8ls9|Xjs^c3)jj1bjB;;UzXmhDg$TnKeGy-{ z6A#4>{>=JcR%pf9bJE4t==HOlcU{imDze>};y9pE9weX{n-DrBs{9ZHwG z`1r;*n|In>gwDu0d%ag3e7{q+{J9K$qbn8z?v!zABk2d6V?P1%(Qe=i@bHsjLzd3q z<}Mz{0uf0Bbtu^&U0su|sad@M;!r^;@``bH_P#^_JKK*fi)Y#4xFkdZn9oEPQQ~3L zghIYBCoM*38?jA12WLC@m>ZDjcbLUb>WhL~hv5NK;X@mSFyeLRs;tWwcs^f?{|*^9 zcQ!LWFR$5-&h|v~GZ!=Vu6SI?0rCXK0M`eo#=oca{FO~2FG*m~{`Nnx!#+tvIjQzh zvWU9|QX%^m#(Lkr`17i2;UqI&_-u_EUcBcW3#gxxPNvi_o8_Y1boE%b_SQG2YmH>3 zLr0HXOYL1_`}a5zN%gq7d!w^vQ*T0B_zG96zn`2R9G*YdanHiL{H*`OKxZd#Y?|(x zy9`h9>({UK2Ep>-*NpuBPgB7=5}P}G$#G+bfhrIZfR<9f_m_74o+BQh`Piw!O+bxh zIQrcYS7?LNVNYMhEJqN8Y~}lO>Gh|u%7vj0XE-)tfnV@Wj2xL1+PiUX_%QIk&S1?i zdv(uJJ14-9(N$P}44fs{2)}l8#846%V%}i!3PewM(b6+@G!nFa5=s9<_JJ!c^z$-M zn%hZr3VA<#39{?pBmEF_(mOpLI_dlVHZF<{dpAUwIqeXACg7`GWGo8rR}^uevRaIa8vV*JV78cLP^+ z9}f6S*l=yP)>J*#+1};uLqz-&LXks>3Bwg_OdE;mUNAg)IlJq_ISP7OVAQ&hqIafL zpQ*(CYTxAlHXIKvBgNmD%OfJIAAO+Wr4HrT7Ovx~%L4KZ@5f6dbmNaHh0@vWvj?V` z_p57i=1tCS1E7am9UqmeAn)^aQJ_%o>SdN;js0J*OmzM!0@B6yBBW}EHUj#ep zZA#&%7Y`tyJ-`{+-P5xH*ezhxnhquZ{}m+J(j4IeBn%`F5=RcK`+KMEY>IE4!W%!T zy4W-yml9|gdodziZ*Sp3Lu_on$4s|wkIFaMf{=6O82$>==3_PX=>K^lSq`^;r8Au$i?4bUDi&w+=nwQPf!go=w}Wf=p!M%xLZ zQJ&O=kzy|V)8mpMO_z1sv>tW|3Tr_R_$6idQ=>MQKNjX$tE&2`(gYEE{Q}B~oTRltqT{=7m+2tu7!8k^ z`Kk39<(!V*e6HUky%mBcV0x~?VdJ;mIW7h_O5E|ury^}wy`yzsh*dzC zx9h`XiIacL#+-lF(?i$G8X1uTwkTA|FiHb%A*s%Xn%?bRxn7!4FF2>w+n7H8jBBRu z`4|E#^RSh18{U$AehEsOr~3tR!ubYa{_V~EpZS#}A_iirG1Je^9cE2sjzrQ0myW~) zsnI?Q$mySgwWt9Zv^ZSq1?y?E04w z(BRUOfD<)I8NCYKRGs(k5ia2}ibVEX3+%z3zQ@LHV{1LZx+kCKwuxG6)^ApbR2a+4 z)o=*LC{s2{JPz;_G2>C$b`XTO8&+N1FiZMPB`Bfb+#|U2wzf};4ij>T0`8pe5pfMp zre0xOdpxCFrXJXoZ8FXSAn0a8-bO|?py)mtE9R7Tk*o}fcAQbf|689~dM2rJo2@Ce z337I_DbWZUz>XWuLjRhddgcMV*y%b=;uDg+aMRq&j^D^FPCW9!bw*DIC}WUI%U>SO zseHK5wyuXsPnWwlVyP)J33vgmb!cgrx)uloCGDQH23S!fogbVB{XSeE3X5yuse3%q zDsOx&Cq^X>w%Fl^m)yZ$Tf|LKMn7%i)fVDq;pQQ#CEDG*S_6In-S8X=m($L!7y1V* zEXlS5nHadgVv(cg@yv0^Aoi2PLD%-(F*6Q59L=JcsuifF^GeD)KqcYN{0#Iv#f+%Y zh={@)Pg)27BeB4;TITRh-_)1au?XNs?R>!NFckk~qE+P#osNI;M#qtnrhlM&d}KvZ zI6@UEM|7km(dAx1v5-UVi+8~CkArwd31#u{S!!!O+jqw3D&1D-8(&qxr&rw@7>#TH z-l8cMnm2%1WdY$tR<~wiI`tw{V4l6fk$fWAJW4}Nt@=8b$^E~Ro`6xF^!E(tlK_=Y znbj@ZIWc#hWAD)CP8)rXpy&>dGtFAzPH=aaws>%-y&-E-xt%ww-clV5mR^k!aQSf( z&$()fK_(_kZtFTJ7dg$|d_xNNmA5~3WRp$7*<__`1E76-{>RIS^=p{LOA`cy7gTH8 z&d{cx%%kQ?TQvzW-sGiG-a=L@T&zq%LHoAV8;GjXGF`@x5WV1ED!E*J{S6xq*m~Mt zsFD>GtK+n=@P!7o?&2Ty?o4qOFV%WS8^|`|1Hl7bPqyP=D3R^+RbLNf06-zj+7NoW zrKqU*vzMPoQ*zVyTGIxEo7QZk=n5@BFHkw)hil)<1uAEXo$q0g`&1TGDD!iAj632o zKkeIxoT=;uxY%zQ7V#xZ7>eCG4hx$w?DQ?&M|?TF#88Nx_{G+WBDg-?`O1tA!?ihBbJkl z7@7&0=>(r-OWBLi&IRI{q4%N)AuNxxc;JO*Z|^#w^(igc&zz7%O2TdXJBRpA*Lmq< z)*28_%5VPHrV7H9l`vGALvNkPz2AoFJZ}44*0Sz(2=q`ZR|yP_mCV=^?Pj2Y4(GGm z9k&zI*Qe`Kz|2x1E_@(Fm&rTs5?h|o#^%PEy$&AJ6PE9?RG;$2X49N&=F3?JO{jbG z`KW_!_ow&vzf7dZzo0pZHNrW(~GFRd> zI#}q7;v(V=8AXL#0C35@G}?@OAoA&BzuYU0&>~tpKJqe4cEA(Jm%iU(>@ftdXY z9j%kRsl>xCmUmb|+(VsM{l^MFz;ZEx6f{8p6P}9y5Is~ojgqb(-c1t^+F6CBpNU6~C;0|ZKsNtRpS zUc~sURc3}PPTVoum52yD9_N~r2X7rHqV1$tv+_E$t=(C~5*Ud2^6ely$2BeBT6lru zd2gLb0%~9DCNz8$?e<~jBk3rc#X2RONMeD8;fh>?N8VR3mXS$h5$bHE9HnzW3cU-q z_94_eff2K|8ablw_^AgK6K!erV52gG<)SkD^8k}*PysYa{fWe%F9k|oi@QFx+4pUV zWJ?M47W(vybXVh?4-czCu6A_9WSu(ieLV&fS=Q$pw)2{bXl;#2@5hH)F$gD-o%CMC z7bib<3B1Y93halElNgC%XD6Ro1;EpgEG+qa|hU@E&nkY3^26pSzCYY zrQ<|5KR<;7brn{vqr80LQya!93X;v~outY+%R#7IFI|3@Y9J;5fg7%Q0Oq20-U)Q; zK}okh>{Gd`+-F0)T#`bx-QhIeb_?$39TqF7fTYgJ1a{!o#h3vw%EsBo*B^15Z9Z18 zeB5vbFSO6WaPi+ld-#bG@T=To*nvtfNlIH#F zbL1^e#7m^B!;;qJqWHOUByynR=%VJ$KRb`z7L-m$EhE=-sZg%)&a<||U>EymHv$&< zLH=kXfMLfv3;wj5#0@dYnv+ZASq%c(AwxmF(0F zmN|#W9PvOEOWh&a>m?}%v1Bz|*x#CMeo}Y=QgnmBOwVn=zrOwqI!d& zk5i6|Wr%ya{lW$BtVulavkH?TO>8Kn4+b{DZ%FRc@!xN_+>VE*Ym<Ffw-^5Egc**F#?Ts1uH%z6iJMN{u~pa;Q+mC3)LDZo6^bX>j5}0 z%6(B+Qu0~xx|dsw2sW!T#(90b1Z*wSm~!(c9n5WMIrqh8Pw*quvrk~8zF+Cxw@#_1 z8cQK!+mP;~VdWh;pa;nn>Q25~iu3Q1YvUlhalffVTOEyf0sNlWu!Tm7-%LxMh-Yee zxY*7E(S8o$9F0PP-7JX%;L(-|aK5&$W$El4cGMp#FHHe(op)_)$$g_g<-}JMs&a`c z2CtM+D99uQc?03bbNBBM)&PPmKEE#ao9!~>(__UTA&l|;CMQ`D&f|H&_IziC!y0+qLtg56T5uhza=wib0EIeAhU@n1a5z$s zpG-cDw7Na(Aj=;QALoaR?3tz1qz4NA?7!rQg9%bABxdTYO5)wTH2GNP8{)aj`_c*> z?EXzkP2laprJ@&KZa+-XOmx0Buz-cD}w=Zq#s=yu8zayOPh?wa1ke_%iNVs z!C*P6|AKk~$(ba|S}Dv*Rz(A}q2gCq9s3@_IyV>2+vG~u9GBhDR{H%{&RtH5KXfF$^9B_EZXZRgWEK2zOb-soqKZ8W$Ckd*=f`UjIB zSahN1#dE?OKPX?D`hXRc+#zFU;ioZ4lC>A=SY)I^b0omS4()mIFa;(*UavIlsCd*s z0Qd>nR4HE@p};P2e9}4O_U7FmM&;eppI7z0Wo-2Ylnh)T=2};bGEASl=r+9f!th2>I4mpnuE3 zJ9Yllp1u^nLtFjD!Q)!%HiBTLpl5{HYFd<-gh@_DoOAUiI$>bM_%LwUpzCb%-J!x@ z_u8gM6PLxCNl;^C{Y&1%`AOalG);dytt>8{uWWc(&Gwp0V$dK-V0b<6TZ9cRzpIu; zU$6lAp1FaUVw0wV;A}e;ZX(5gw!C?!lo|T*_4sfMg^kxq^x?VLub_x9(ZF|3G@zf7 zHZw-i-Gy}fZfC6n2|CHie+hHpr5kl`Ij0M8ZL zvF28ZeuCxPV}rGp6@3d7mABXFYqs{2t7mF-R(c0}t3h)Du%t&R0W<<0gjBc&wt2q! ziUokx@bli~dgv$&J7Jn>Vogv*_5Qa~{nnP{o~6X@`3uE}Ke8OflcMQ>ej%alwGRBnzY50&if4r`V5JRL@^ZUQ&#P>8yhNFjjPwMa{nf;wzs$ex8CVT*ZUvAsgA3^ad z=%$}Pe;7^|2nFoM;P_-R1U*mNpils=&QWs9FGm<2$zI@vC9Xg47wKLKM9UM`riLw3 zK#aOJ55&znb5E}eC_eC1pc4`1YoHC?P7*m&H4K&MgC1t>onUEEfg zO`HO0TF4H|b+&?wt}tF^xL z`#YL9SUDX3!w6cbz%Ff&>dUylT)r)+im7`9+CGUK?DOY~zWwkXu6em%;Kv5KG1fAe zM5Kg4vifP6pwBonAGB7xQESa@`YUX>Xh^c%^Qbb3}m=)YdU8$~o}< z?DZ;maKEOWahyj$(&7nw~WH?Vz4C)1#YQLhR-vKo0<>r(d(KhxBL9}JfiV4nN z*Kahwyle>0pL8CELs~msurIGwMB~~w?uI%}3*Bd8nH~zE<1RKv$uVK81A_ zxf|ZnJKOvBxUw7_+E(mq4I6(gyXuU^{82aN-#56j$-3vLLGuKk0`@_;K;!;rW&?!Q zbuK8yrUeh^rE6AwM|%e~CvB+kZz^V1DjqAQ4>6}=jp}7ya&RlHs@`==IdxHgV~MKt zHqXKa)3>SqFk}GIykgynmHbuKzUnGme?f?_RQ|*w=7$*#Jp`if)%iuuXUGlmF}J<8 zsOVcpgR^AM^Sw}BTAH@YAH)eKpo2cr*VhhAGXF|V4J*-YT)#a_F0v-)=Eg<9VeA6f z52pZGV$S{I2!B6fyGSLws)EWaw=N4FoFpCu=g^vW4*?*E*#uBC@jd55_|n0%uP~Du zBTxWmm2m5_&u>#-lF0Spgm1r-g>ji)n*M)WePckL-~a!zZQj|oTW-~sZEM-)o$XdF zEiBtuwzsfs+x}nsd>{RP&+1X9bFOpVdVyD@$p>Df#XOJtmnjpY%w%16!tHbJgNYK~ zhAI}=V_wr(RBI}Ls#HwKvhkIQ&iacN&zXj`713Y|a_2({2(a$Kyqb+?pE$uxy#h zIqtYb0fsRtbF@rjKy}3a)RqZQOn`==V9ib{qL=Rp9aUE1dJ~`DK|>h zk=Vb#-A)Z_cS#G`E78@ou(};p^fWgRn{H1a$!y1Y8hb;Q~FY+ttk`&R)>IrA~v z4wop#P<+KP*1kV&*m}CBMC;Y}z1=KO7ss!K8%L!h8CWxGdwpNI-6me*o*!LUScs1^CQ2&l?cLVm?qR9u zVmneb>7VezXqN*=s$du1>tu;)G-x@9kfk$A^77^`QIbqBcC?|qUkv0Pbma1D253K7R zIpk{G6XCrb@W6EzaWHN@F}~>O`f?ty$?baJbq?xlparnY^qI-fkH8>-R=`CE>;Dj& zwNgbJaChqLz%cy(s)HgnULxB2bFr1gkl`=#fFohq%)Y8~v7!>2FQf$eY9}5=`)xxs z$j9`nz`tUX8V9!2xOc8fhNaCc;sBTZgeFknbh5&sJm z#2_?%$dFAUVn`2Gkg;+tEmG6qW~7rc&u6x3t^7LkKRGflx<4%dB}az5R6?@D3xFZ$>AYWy zUJrO*r5MIzHDIc$Za)(k&lLsPxRvuDNcrIHGy+y4=U`!qe+Wy3?G#*`t($y5@laK` ziNBtYa1<4)cn`2(M2s^Mbm-Ws$0EcmDDFrfzg_Z*MaK{Ed;_N6Bsla^xcwMsKTlrd zUebBXJojm7g!SXC%N*7wZ?W&68I%S~Jv(P^g4B3}sZTWQm4lGUDzC1ux$g zL^A)yDsSqPmcC~&HFy+LXI~NHOr+~z#e>Y`2^k&Z0*Qc(eu3y>p0^x5wGE1~V=hnQ z_la&t56m&ws^h+2Gx_gzsIVTOetJ)LAN${h!;Sk#tcz^vC)wDf8ulM)W2-2*S_ z3vtMZ8Wwz_?&8o5N#LX(+%HzDm*6RtZxb|Tpue$LYfu)Xcu zV@rLTXFtICHSm{Ho!}_XpEvo-8(r`|d&WEVe?xCY-y;-ZU1btO>;|fqI$?XqW)6c% z@%kFL7kOu;53x4&WfDNL;{!KALLY(mj~`S@dcVY3;^;f|C0nEg?@1?B8b~A1=SoBc zZV3YOBCOhdeq}Ypt9z$_)`ElW0zRr|z_gkt^i=%%J4*JVhZ~Cpi#Co=9LZ~W15i=K z)$K*La{aJU>8-a~&nRKZZ!UeD+C<++X6hB zbpXxiW#1jaWTKB3Fa9Jq__Dr3>Uj~xLSa)qH7@p2 z@9;5~#m)PUSQIkoel2+Jyek@nSYKmebT6-t#FhoS~9YY0u+HK>3N(Tz~qOnoa@)8AuLMZ zf_%OOC(~w2dP=KwOqj~0A|~ql6dVWk6cFd@RJI-@{%d;{x{f)w(WVskaN{x5S}~u|e#wpiMq?WATl0GB-S2PF0HL|0 z=Qfsz_pAWn>Vt<2w2uHeclUVRxE05?qZ}!$$3;8c^S>9p)pb-s#7j?01v3s5Hez** zga#ikz5pRH{qsEV`YgqoHHI2Yr}M^tgD}+p%8`tCp6WUT!pUbVpwUs(=iOiDB<}Ph zNocj=ch;x^*VYVQwM@BTZ3+~C+(n*|MGlC=IM+zsnz##H6#ohL{Kj*_4FpEe`_#k0 zDt+{rkYS;!TK5Ps^5+j-@%f#9WKgU$SYy4tTr28%?WU#;>G=3zOP#NEp1B$T-D}5M zJI49o?8Wqiv40g-F(G0X_UnrK>H;L(@cO%b7LmmD2 z+Ywr&hvgEsR9TEOPFPJL{Y6lu^m3U_%J28rSmBKtf0Qp3OJoJx9DhF;K1^i8CWk;C zuY00^?#`q1_P+^ihpOcNdWVhetzhYF31#V9%p99U@s)l}0q{jH} z2wMbBN~nz0*e6TkY)8fqPAHATqvo$?x0+UJ+D1!xc}99v@2~rVL+)8vO3PsL+G&WD z*Q5H4Dn+CbnmX)Tgs9uvO}jDXN*Kz=r|Ryl>VUVwmHbYn{93=xdgsYnX=%HmKLpt+FNg7@B1c)z7PBY?*OOr?{A*?n3#Sb4Ub&5Z%0`NwBnOtT00j! z=#_Hgvl%KK|64Ya8U&bP{pV##A|WU|W(OMo=E*y^eAWsb9XE1g7!^M1C~!J1NAmbd z9z)fOK1yR!V)y6kR@YiZzu>@XZ3RX4Uw@tN3#}Ar@f0YV%a}&OhDR&)<>*^#rYl-1 z!4GJn{j8STwT`l^8Mdhw$j(j!``aZH00%Re1V{>ab$}2CAMmglUh92pV+=#SbJ+d= zl?_n34R29Vh$l<+L$iP##iFv&wWV-+2QjzYF7M<2*(Rx_r6X+m@tpmj`<#)PM3qEx+y|fl z#i+U{3&9`mkx)a9I7;zVQojcV7lHcJ$WohxZC!fG9(BB8*J?d!~SP4{J`Mu|Po+<_^Bf8sFApeS% z@>4QNW9aZC8om5@hk;B#*3j`KU!%wL#PQ)vH!Lc-Sn2qv_|{76ut76waT&un$ux4j#aXiW}%QD<>p(hV(xG!c3Cb9q!EY^s2=hfWz=Wo%xs!(=ZZZnLMZ0 zLZjL6v?*2I?tatWlemP$44}kwX-Ysu6b*R9Yyn0yegA|LjQe1%8k_VdjzTacj<*32 zOhnw<``sj@P?U?`^L<;t@JCE4a&)p5i5Zm+rF`+&@7{_&M^6qulh%szw}RX@B~`z5 z>CVp1aA~FW-0KC_Z#zaOUn!o*6)eGA) z&&@wMaJLM~y_es*h)q00;~1twn2S!jY+k{dmftneubQK99Ha^z&xBQXB#W`O?&SuT z*~{T%nrhs|I5Q5ui&#;e5%Cp#7OnvfZ*V)<7iiX-yo_^BwFoDU&}Na{$A_7ew3En*`GZu%P{w3CtLTtb>= zRBHItwv~;Dz~ga^k)C4cW2*(d5q`1%Vugko6*bx8jZ=(MRlpO%Rg;5P6XBn;L0+l&sX^!OHXske!~BTq<2 zO;j}{%IG_wjsp%JL;Gnc@7bpztvJn1BcL6AP1cy_4Sd1gQI_OYO1>D*imhgYDOL+Rox{dc|ee! ze#Crf?`u3hr$r2Y^0S=0_&|`jUn51v90ZIP>F2Sbgfn7xL&FpIPaVcuPJAfksJSw; z;3UFyW-gtJ1CRDvvj9Ycc4N9WM>6KVmH{P|n9#y-AQB1tiTirU$fnCm4h-hoR*M|W z7SgJ@RdCLv(JwVYr==usU8|T<88DH;3z{96j$>(D(#elVfgbr)G2F)**$jsVBQ0~K zE)yyL{7t4$h0hQ|07=3EVmZs^q^+ejKndbbF+7s~jUz2V9x-yZEZ7&I)|m(y{R)Jw z4yiZ+zD+4-XAU0w6}qT9H47l`2k@&D+XyhGJRUx&U_O}8CaQ+_5j${2OX^Z?##nGi zTt+3&SJC9y)bL5u+r(=YG+sC3?MMX$gBHxG34*}K!b0+DvqY1OI0fS~m=pbOLKzP` z@Px@vo^@I>rHFGnLhrk_&EF0tm0LWF zP7%<7H$fseJDC5|f9iAi!rl#I9YDAEWnc6iBNtFjSTDUD&*D|A*dD+$#HKzX5m_^v zaK}E(4;EI3qOx8teV77_#|ZzBK1Qh~l(YxRk8(+zqN}Nu7Lhln+5`*HHxD$GR4Q52 zZl2-{UHnO_r(-E+!99}rl!>GG-hw87SF#oR*f-7H;+C>(k&(91`>1xJ-Hcg+dg62m zw{UP9aYR5{v7y|#wE-gjo4?X2_Zj!ajSNThOtycfU*QG41p^Tb9Zl6&FD1cNx&wJO z#vf&*+q&4qn0nwk_XobyBCIAM8au4}?P#{RcMQq6-h%w=-Ph8Wml|qr43H+qwz0QE z2ai7?c4fr{MxQH{S<>>063UPj7vg_lMT`?!$P5jqJ``Rfpa}~|-Sn3#7fP<#hacF% z1KKwStL@Fh-r~pFHAX=#jsXDth9VUgQZLAP0d4<3X%Uo3YmM&gsXu z-U=CPVlUlh#wCG@qjgt*Gz}4QT>=6Ei20qNF)%Pt-b6`6cDNri7Qi`{^-iaPL}jgY zC*Dr(A*)qnk;9kb|KUf@(GBO9~ z;YTOzn8eU`dJ5MK4U(Nj`s5#CqRJ~{Om<$vMxm0rS_6?bJ$MN| zHx`|#R1W^G3@2vTa5wwncfhU2qBbFU{h`h6STZdU7(q`QR@Hf4npke;{F7N%25eQc zHvx`7iB-JZ+T2E9DBCW7b-B+ciNA&7VrnzVvxKusseo)oqFoWY$%-m+1Y)Q-OUQ)( zn7 zKxER|kTB!L9c%blt9ENGW*)(L($9FeBnJ~e4GM#sH4s>1O_h3d3quVGLz^{rvkh$O z0I=dG@d#RtNU6*M{&Vc@5@eAKPYw+KJLi0RHS6iaI3f&wpSs2H6tUfL#_C_$elsAw zDviSB5Ht$UUg&+U`iZb19wrZr?S9I>L4Y2Uac4zdjSM56^v>Ptm{3YB@4Q=g^PP;B zR}-61!1tjB{r&an7VC^{Crt(C%J3afsh8?0z%%g7pE<=yxl5=x-RTn7a48d!>$;vR z{%m_A5njTy(|@y|xh;D%kfPO3>k-ov6XfP&#P+SS<*XBBGJDT#_w}yp5+4IA_zt(> zQ&2|t2@Bem8=Sf*2@z44HB{S1)B_9rom1{o{I)1i;oE_ z9S+GD$8dwgS02VMr|Xn@s?MA&ZR5(+FjLsL#H1$SDbJG$x~b9sq>^IKR6E{#aSdA?d^t3WD&!90H^6r8{R{tCLm zp${WrT4}{tC-rE|GhE{)bWr^ECx0avzgPm%8Mab7;eB24Ih57a{48gMXlkNT*YR{L zfV4N?Ogn^vtug?b=I#(0q`&%ZS=S)? z*hhM)Q9uJL31hT*D={O=P}8$K#a0;vxJ_pBH=qF%GfCyO0WF+Hjc^E4TBD1HF{zQK zChyoqw=>ADzrMJPs6{03&F>1_$aCcS8aBNq zwMOe*ep{()!-?L2ig%eVu;}>$uc!asDv60e=2Ltv{#C$gX z6v?baIc>eK3K1`$I%wIiT{>-Q-JfvPR@>OTE6IsibkxFXqQ>=O)o;xR@60}46BIW_ z{La4N%I5C?S)zf^a~UNvm0=Y#f+0)?B?@h`QaMKS&zpruwqArY4HnIgTXj8V=?76V zcTHA2c8n4=_0)}Vc=z|WS8S{ETcQ$+=Q_M1S`KN3v8!&WlMt^jU*@k)EK;N2Up!_N z0L}He?=W88-U4Uc@BW*DpLe$-XJ=LP094~*t4qnl7qL`8WUegjJj+@jtc4!km5$r+ zLA1=qDK&3naHxA7#Rr8|rEA3(w;oo?#7tq??^dCs`G^ml63Rw}~bnbEbk0j>K2C>pc*)mVDz{fyNi z#rxRJda-W^LhbZI89DkSE*b!FZ3f1`&$T3%eQ{NnRB>Q0+Pi=XC z(ou+I_*1B=;|`z{w8s(aTM2;k!9dw=0$kz;z&z6tAPakY-zuQEBHc*`rV~CES>V!_`&DD-5qzrv9hzZ*e=J?xXaMqps; z9XWM7C?RyenEyHC#pY5JTZL+ejWYTiGQyye6IDAxQOqCCfh0Gfk&xr^f(Wsbf`E@C&NB%~ozB!rqHp*~jY2KS)9!~9F9U9giT2IMyCg)8 zKpFJ;AUlIk4{NyvaT@46L4dA;py0Ot8BDH%U?edC%0(aW%jIS`DC#S93S=TA=uKR8)Jl0MtVFMGXb z{Do7TjB==RJC`5tF6PFpKL2x|F@ zC1v+-=zvZ_2m``PvVD9wjhNSR1+=UAkCiv$O*7PyP1}t{GLA34VzIz3Od!Khhu8Am z(>1)K3`h-kCxBm0Dx#cdW+%Ugjc8G$iqlA>pRwam_@ciAiyt+T9vOO5Hr#0m%w8c= z3}-?PHZqtGXpxiN9M)G|S-N=@CoMV}_RGNV96`jAzQpBB4Lz9>EIJB7Jttq;_)==6 zqJx-gci;=tN^^|L{Af!AKW0|fn$ryTURIVvqbqWfp~LAj2`-3Lywk$g!|<-r-{{$kINQj*0H z|M2h-fgJ%gB4F-Y$0B zVn0&G&`I`GCv3h@9LCGg@U}f!0-~Grt<+o63nPZ>KfK!LPd8o-PaJ)&UzZ=gKiqSN zd|T<2VUZC1s`x8FleiDO^_MYF8paFZ{fa-NP^iu@$wjW_Osfr@LU7Wo&#Y4`uc#|w z8B6=UnqUoQ4o;Hm!EWSxD3Yh!2nGh9CSE9CJZjGgq7&)SSlV@)_?E!CxI6e@5_>xs z5yQSJiLjc(84Q^y>{GZO*b{NaTVnw6jmjU5Y9A5fi&!FkD?rIQk}s{seWidz2jK+1 zB?d*rBctDbun`CtLF{^zEJoi@`*g3eZ=fD_3qNR=J`iDAMUGZ5zolaK{y^*OZwT=@ zkL0a(cP>@m1jF-ziNm zmJof}?sKCZC_npqK(rYzQ+z8myOw&^2i7lHc$V#6X+9q}z6Hz4-I5&poK?G*&sdKF*3{6Q- z4l>j6`>&jy-dU zq^dE6mua!+wV+4iGK=fzcjGJqbw-eA6odd7GmKGwjaN zIB{*?a`$@cX0}ZFYu)u@8)G~H)mz8a{-(`w%lp-ur~8*x6?%a@i3E-WUnT0wts*9q z)f~uGIDN>@P^p-TBd+SE?6u`7v+XoAUL?>{SwePe%a~! z-#l?De*tG|Wu(Nj5?0XMMxpwqtd0mz?s&(K#d|J{Aa6a+iY)Z~n~%Gba*AmF-RGa( zj}_PPoxYrLc{hp+k@4vf-yB7WVcX<;G1SIXnSASG?q#3YBo#x;jC)K8?_z$E-nS?= z-98tkc>M8f8yO!!1_W4ibWA{q0LK)(hPFJlN7i1d@N&MxBV}5tLM~#f=4{0v68$)O zb(xhiBBxvE6VY4~Hbpd}pRIX2U|(Bc^=z>7(nuI)TEaMyh|Z;kG@=u= z^%V(u0C5rd45Z<@+D7jB0x~G&IoPc)IY! zUd!fp2_*vn*6osFIe-8a9`BlGp^HdrF74KK)+BGb6Pk9aY2T|OLmxN(@keJ;mh*)r z>k03Z<<&p^WH8K~ZgYv;L@R}N^E;A{(uH%MX8c);j zhKb`j)E~xuV3$eF3ubay5GDNtJ?IpJ|Xdo!Ps=If8XFZ?P1)y_Q1-+Js5p-2NX%Bc-i z3AZVv7cqN(rAS+Mgt!?E%&HwDFxvETWLdG_-sSI0)EN2KaPz|Zl-cv~|8PpeyVm{U z0e@^>{>s;2ht*X507WD57(!S{>wm)+^PL^f+hKthx544(xASVBq>KG7>>cc^>EF|R z6T=zk7#?h`4zx}adE32KwQ2E;b&2@QySXHD2R`0iV$k2`E|qedKb_?nsT@T|`t=T6 zDmu6j&eBql-7V7ZFzJW%p8EXMvwn1yP4KeLKb%2Gn-#)WnUj@jw+rDY4pqns+xV26 zcXQXZBS2rYS)Zk8Z0Gza+{_`-*B&944WScPX_SGg{d3NU?LQm&>;6cr&3zB{gMDCf zj|Gtnn9!rU3UtYGHItR78Ke}~aG07%TyBS8InmV1K7?hJKZGT->0IvM*!%UZYkMI( zFqNO#zD+lXc$!#DIEc}e9yx#YZy~D$h#ejWg{J^$mw|JnTzEEN#iFDXdxwSeRRqDy}*C6 z!+!A4O(Kk9N^EuL{qT{_qsl7yZT~OkR)u#r<-}d}gp+(KH@g1+>vdUX0`SODKfE{4eQyCWC)N^OYtc&W; z2>ZP47GKF!+_-B}cG zZr+Y1815Z;e6jaZjKz0deO(uWn^E57lSdG896Ijn4bg6DEcc^JGl zJiVCX6FS4cI&XC!Ax}w;UNqecDY#Q3K?9q$Jhf{|eS*&--%s}ABRQie+C&m$`;W85 z_$rgf!H+#XR3@JuvY6acwCyNRLn@|=qoz{aA}1#}qV;`g*S(#g$UU$zdnzYLD!LY9 z%<1?&72q8ShhS`Re8(_F!43Zr9{1Q_c-s+zwKrJwfd3O#+yZ|!`2 zEkunkX;5kQx8~!UB8=M05@de_jxL1|1yGV>q^fJt^m{!TNm}G%5IpYY(hrce+6eJ_ z7-ivY+d+iie$&&Tll`-qrO>t092bkhjvVWzYxPAo0cqVL`tt*Cv=dm%!~#l{QJmMo-WM5eM{J;ZQGNATi91a*kEL8n@tX z*LzQks-CsbCyIV<=uh4D^u*iZ7iO%3in~Lil(0bcol!zMYn}9asIl;ohF3h3DfWi; z9~5xt-4$)ghPX1p*~>bQaiXgw`RhqlZEtmo^?CySs^3{KAL+0|bD{bY6}cjR%};bJ zl$1?bYsS70KMJ~;P}b@Oi{OMe#J4 z44--Wy>NMIj-~+ktL0!AFF5QSwam-;c%FqeWc4!N%fCIhHt8y429NW9ZQfR~g;Kve zxY`o@NIG6$;YEEdZ1EPH3ot}P?%HBn%&8vKh?Th%>w!>Pn=PV0gO-0k= zaZ*L*5>P-?KpH&14imfUQ@$l)Q7#mEA`a>u!89c;W^-(U-wgUCC`GMTib2tO{^TH^A9kX8k;-?Uj z*D={B2KjUt#T9WIwDR#^_b_^151EcK$z>&dC=U6i*CPOQ-ed%H7i8#gHkT;joQ$q@ zvPC@x%?+$1v7h2HBQW1d;bj>yt!bIv+lLXpcr70WrHSZ%RT)@h+2SFdld!{tBpF+# z-aj9n2=X^ATwi1y5Gx+2yEFKgd6VPg zes$P+;2V6WjfktTO}vjaJWo3bkt){3L~Y1ovbKR%HAjl@?f`qbsk=h<(82_p!gXJ8 z2b4I0RB0PEj8xk;Fk>OghCE3;^pi60k}py$JfT@%6Yc72Wnv@2Vwn8L(;{A}BOQ3e zQ`Jm@k0W5i4;lO{olYv8Q}9qHqOap|t(h&55sc0~T2RKxbebBpT6xdA~4&o`biRBZTxSF?Q?&*bv+S z*Iz83E9?yVvC_1Vz0K09a5k%xKLEe#>HLXMbuBgmp=aMC)=r7Q<{u~vK!TNbJwxJN z4;Wd*eu;ljPLH(7>PzH?p>Kgb+tVOr@Q%=P0}%_&xfeZ*`t_;XT&RfWpKC#qLSEbfX~ZhekD=u*}lVo zWsy$oA@6!n##@MBMYX#;;$?+{ztm+{-=uXCvXUE`l(jeJ4@^zl2F(OS2>r$jLN>xD zc1>}dy~pe|nsOZZshs}iA+!yJE?f4dvKs`%0cJlqW1ai&O1wel-;FcoAuP2I1GP1X zmBe!5=X0>djCo^D$Sd}~|5&u@0g6jX=4Q1Wyb=aKCh3g0UtK8WwTJP!{g#c=(dB1e z@cIS9pogK~F_(HA9Nn)l1MR4K*`-Rqd_0BZI53scj=qT~<)@LCwv$3^)AZVorz*j6 z#|GRDra*Vpxlm3nY=F*66NFlAx$F-);Y0cauHBN21wJqh3;xqufc@?vruB-~#QLu4 z3yc$6_SR(3*BS-Sk1Pt9_))ai-=cc$e6r`vgWH2GZwFi33slQJYzoJr>j3~4+fz2i zLH~*Iu0of->$ua;apEX)D|bF31>e~|JnY`<;j$O%54UWRxAWf8tD{QL=^s+skK+@H$sbfMx@BiQFE z`Jajc7v_llP}n-&_Qu}rR(rGIhpN}-LW+pw*8Z;grG#b+&X|%(F&s>20IWYP87i^L zPmGvzm()I$k{Eje@|Ssgs{DXMoh+~93DK6}sHGrn=AaAl9L7L>X8!S_rYqv*xljp) zCsrz(FC`}^z`-zF)r;SLKUr{e^)=jBL6Uc~xR*^~=s$$B0%gtesn+xM0>u>HCJ}y& z7n%@kcOLp$b$XuF6uFu2Qx2W-`5i0{_J_~u7tV%NavVz4`3D(xyQABeEws>6R0yV;i!u7-W z*_}>6lWZ}8|4`NnP_;`HT92pue2oFk52WaXVYcsY_`WZ;)nUss!ThbC3-z2CA}!)E*N%HKy*FgqyBreQykoulU3yQl1FC!`w(4ICU184#rAqV-m{ z5pKFvUwK+TkJ`%v&u#2#R-&ebYx{U(?d!#GP3^^k-ruJwU^9aLw~!hgX05wjT0V~- zLeVlxB`hs>Z!(*^V?CbSXZ5`%7>=3kCY_u$Dpd>QH(~tn-#KxT{=m9;<0|ZP8VTgm zzw}O6AF>r${<7fuNW>&Y1UbADnE#80X|s5X;fyX->EVIQaU367G2_E|3*Ye0=X7W8 zd@?jvd&BdlabXN1^cXy8#ooM{$fo>io1lOxqi>) zswX_d>>iENDz8a-94>0a${k^yq!+x9;IAYDrZ~T_H>)t9kXu~-OtWbHm8ac8A`zlk zFR$jVVKw3BA9#EbVlgb5Uk(>W!|LCERw`jVnoIlon^n`DNkC+)&(N!T8coZt{PrMI zb7qMcWQ+Nq^2cu|7j1&KhxK+@K_4C8^A%xO&DCo(D_8HeczZp*DEAYor;_!)luU^B zaT-`FoSLjLqrcrGi|0YBH0gaFzdjfvpyBglBKpnUONnJ-7?Q5WpT~hJYJb1(lv)W1 zI|~W;Jo7x>0dc0u_#ehuRR+r$A32i4`w_X9(0*u60|la$o3b|)_0 zn+}5N7app`z|leDdLIZ~!w&B|?6iwtV=N&8WO>?8dB4=cs^dIt0m78&{OP^eQa=lZUF;3WzyEPS^kbdo zgS#O`R_^RNXsxz(`u+f-RUFo3R#|IqN}Pr99w+^$+Vm|5Er>F2e>>PX!}~uP-yWtt9vn~QK(S~5?&Dx&esZzN z$%!qQLoZoR6ZE66s@V4i)^15{G56T9krcUu!otnf^7I;eAGdRcy4+tXVRfd0%55(E zuDDp}1(-H^yjGAMgN9JxW>FnP%q>IEDlkCjMsOM>7b%xHco{^OpIbhE44dAHNL;fyA80`Lgb1)hkY0Lsa*b{!#j z`M%elvyXxmOx3Oh8Lsy^*!y4uNA^sJ%3>lA%yih;TzT>vNZWAL*g|5I|#XG z#GFhui}%wm)Ev_ldyYb*IaKogG0da^uB>t7=-$Q;aVnG&vWmi7JrD`Ueq;+)Nv|Mpsrc(oPh z*#AjGZIfzCCcf{7?Vd#biD$ZeR!`t8Wr>pS20S9q*YL^~Nrm(s_g`bk(LIfI#q*j$ zlgV&>mM2W5KBzY>Emdf^?I*~)qn4{vMaGourmVBw6oug<*|dA})Suuzj908Ll_cSRM#hqP2; zy1QQRe_DQt^2@)z>p59ImylHW8ht`EVu5EdYeI>=fesAPh(f0Bm~*Xc0a469?hDvM0J&5Kzu22WT8u6~0uO@nS%4{rS(w)IsL z4~Ib=^tzL*^w{L9{h;cFckwWYBTF}2&cix8gyx`-5%-XAfxE;5@Q2#|y-FFV$%7aD zYdI7pp7s`PV?sd2mpEr(I^565V(n(H!~VuYRgS_@1R11C3oqZqe+)UEiD$$}@1|3u zz7{C(D}rO2Af;FzD7;hEop{L^k zooFQEo%xflTcts1Ho9AbyVsYJgCcK0Iu0XMFfk}jk-ehwZ-%l}oV9_26_WPLsFtjH zJg|d1uja<-Mx5Q`$@7V?cao>2q&1e~=eJBUSoP3D=oyiU zZ@4lzIE~sI#h40FQ9~HWKFEoR0@%NPt;1ua)b4Kd%^UryKRhFEb$?IAjrIvqhS}5N zug7ABnHfR*1Y-lu=$q1%qYPKQ3xjrkaNC>th4~samH*;BjA0`HoaD}kfvtGF1nSco zOsbQHPjUnJ!McVVH$gN{abth*A`6b`5Lv%RrdUh1=O`g|=^r4ytoqe5oHlg_=>5#0FN&{?8YA(~ zDGw`r!en6px9-mCGN_7f&$pv5)qCXlE8-e6xG6%eYU#zW%n0k+{d@*S0*M_L!p=uw zdW9b!Ry-F~(3DLa*>gunj>8nf+7^_sNfjkWX>)%W&{sE!SL1c91ep#Ma2qvg*YaKq zHT|HDB(G)bUYZv!bHT=q%Zftp)U2;e1}F-1HDKU1zy*iiyv5qqm>*WxXqt^4;4Yxe z*>^#CJLC#9OW+zihz77XSO9%j$3D(3P^9Zkz~TqdRJej?@a=kfG_ zNzoxv!tnfgwuZNsz7DPKshI+Z71UREqpF749^PM+SBdvCH-D_M@xLdMWAr2w>|89q|_0$_;+@2;;{b(mmFtzAwu=X=$OgH?*}|L6n7UBm|(Ob&1UcNE}#W z%uePEURA$`Cj~p-!><1_8t*u@cRiXf{G)rr%?3pq4>V)#oM&?E=)|Nyv2(~r zCo1cRh#2aHLM;){HL9`77-DtV#MYR=9pAvudm#jUhVtuvSVDRuWX3+d*aU2?i$(t1 zF74Xo)5T5;e)o!pxlFu{{Xp^z=RUjXqua_R-mgWl*k4!TR&a=v(h~GUlv%<#exa3p z5T%29854z@T?EMm7|L=WxVsQUp`OSb2H6Gvja$JMte_D*nT!c7`USBdE|q$DxPUej zFN*4>lpymK&ry!!!=aL>l@dRmgjsSxgYg9-0|ljlDvMjauY;)|3;zi=yp|CGc*RNn zd&TP$fn2?b+3`%kc&Nr<*he(vLDl_)c4sI=p3_h@`snBkhvuH7 zEF{%47+DS?szvgVlU-AWUvn7&pIBt}kE>T>w3wKza7MBH_GIqiP7awO@Q zo2O(90m55Cx$N0Iu?Md5|MJYaAAfk7S>K`d6T_R0k21R!wWjYfS3d9kwX|jgyY+XfB&i{F~dr|-#{*( zn-LnlKKth?rM&lpOMDS?>GDsXt(vur<;Ff9CSvW*U;I$RvuzRjg><3Oo1F;jBCT9$ zr~K&c76{sD!lYB@q2qQg?f88`B#PCgx+k^<4UVLP=d5f&FS&NE;dx}@DW;jHe^YO7O(gaMqr&wot^(n2~`JV8z7f!ft&a`hJHgqmX@ zJWGGhLRtSf1x+)Lrz|bP3mWWKBX1i^Qc+6FKvdOfD@{49s~}dA2LIAafl@Vw-H;Bq zrKFS^#Ij3gnRY*dF8UH#m($&sB+0^0)uFC*9}zxvvz(8NZ2P-unfep7J`K(u<E!3a4vW=^nAh^1Q$j8H4||0=_tNOq$MtM9FzMNeo1~ zF0QWe{;+iVy$)D97>e$5b;eiMXA7$juQXZ^#C=^x&YC#-TD`Ebi7)&r2OY{|h(Xjo z!X(+u&R(SF8|_5+b%6UhGl5$VV+Sg;8#nAXRfY;uIkQAiDny|)x$7Wu;$tW&7gae- z(ek_arC|VMlk-@^${3o6<_hHJ^~Bpc4pq6H&|o30iNkM;Xdd~ig)%u(|TQ({rf`l3=I&+C+zaD|L?j`IsU@KDP+V2l|KbTsK zyAjpUFMS32)x8dFwM!DRzR4Fn-#>+3el5UHFUsyV%PXSzg?2Pap5SFm)&sl!sD{rm z2?g^fT^$aN((Er+s6r;B`Nt0`(MV2g%m^(BuuE0D(>P{5x=9-_{4Ze&q4z(I$rtAJ z20MSLX^^+MfG^zP9qWR9ATc+AQM8dw^Y@kx-I(Fqf0|| ziXvXHb;%Ff?*m7mE^z~mFjcGnSKC=eRT;eP9wn4Uxj5wt~#6c#J4&5ZhEF(I7o8(3Lt43~fO-P_z^Mp<9TXxvqW7AME6H;XBB+isG zCqr2|zeV%1bKh|c%D@TKb$d}^B!)952j(MxVvOksJA~3)O~#eH=^ru$ehs>8;p=YS zc*d9#pD**F9_i;j$_x(w14DcIYolWP-x$qNv-atMaY`-hQvcXJ-Qn{SRC+V~l1;`l zIgjX(kddTT4=|gmZ|S}cTN13b<`*Lr1O+h`aCpCDrDD-w#vtb-Q4#U{sjD0Q1&<-{ zl3myh$CRW(gMln+Rz3C7DVi>okwIB9vpT6*GT;LjiAqxHqZc9w+LqYPtdqrYYuiz0 zG!lJ#zY+VF1Zk5A)#y(&0~+L`DtsR8Z0*NXaF3aFAe^L9K8BN_>e10pHGxtuWRm6g zcxvT)Kdn>$K1bDXvX>{(A(OLiI$w@8>3+7wiomZf=tGDg&5VXWQGq z6B6Zp_)CdTC4%NO5~R6kfs%+Db}uT6am+^L|2Vet#0w=MxH(Pw`Z6cPm%3T&>7i}= zqn*9&80zv&t}OVJP@4@&Hx$GPUV9F3a2()l`awG3Gbp-RMF+vOcZkD90a3=ZqrbjF z867wN&xFNBm-xofh^WqiMoL+Amh2>iwneOr9<=g)3>zI1?q8hdn5JW-QDelBV_zXg z?IV@})0RQgmI1#krQ8{J3pfThWh-dg<0T)*m1yv_OBU!N7U=suuR$gByhCJ9bZo+z z8^8CH^b-2Tt*CaQ=?^saF8^_27N}U`Gdp^Xx>{e~F~G}5+<G~XK960{cq!PHGVFIddSWwY~1W|)Kv-!oY1FKV#OO#=;o$iaObc=s3^=( zi*5M zd0rwY9nJ?$axnx2rDxx+##gkVjyv4No*&rtUE81zsXtkI9XTpK<9SpP-@#e%sy1ot z$n@>ib@PxAD0*wT+IEkpv)1yeI|{7%OTMm?=Lif$S7^+AUMUFY;&C(l>x2)yiK`27 zQebK10~`gHExb=SVjkq|Y2IvqEJ*mzR-70W6a!u#GXDx-P?VhU+@t@6I`BftFC($( z17nZxU)!E6Gv&rR%qGhPK!V8ix((8M(&Uk#J=8Uo$N2CVjXggv0qwAl;8hh6%hJOkM#1cpJr0!BDDJ!KjJ6p2b6mGZm0&3`2SU8R%Y$B<q9>M%gvXvu4F@}sIomfJr@Hob7gM@9 zUh+#{UAI6QOz3k=7O`J3kh>_w-yr?)?5Z2Ppn4KP+xd1i;+YQ(FndRd6ICA3SRffd&JoW0vJYpzn z&F|InXr%SnlDe^&Ow7tSD2-a;_6$8(B6e0PZ5m_v;nSy9bIzIa6kLrW2PX~rtTo;b z4{kcP-W;}E-Yh>XZ5+6Jkd=11_oXE-3^`k0uQRNkuKtLrwsZ~)>^Cnaj%4IY>ayk* z^sdn(4Y!IsoyzviCK9~45ic1RFI1PRo->s1KrohMrfmU597&ZNRL%9*==EBE8J#=s zJo;Qb`ucJ!)m4lf`|#z(bMTE%baqn;KSOf-J8gTX&-yP9U>qE-Pjm{8sJL;@)=?T8 zLuUsu7#Hi2f)`o(hYUnj=tFG#Xx(C3NF85l@qsY|s(7?kN5bcHL$15G_orpu+QqzG zZ^ox3Ch6m31D~)Mv3qT=CE#_2f=N1Q0)M!LPXAW-S4sHBHv@G1tv2!Y$~1ikCL5DF z#Qe6n3AC!;#0OJ{=~FN^CEJE*W-(D#M{U?y@Qfq+H*Rg22c&dt35SAJwP}Aa_jhaK z?n>^3kmEKNAM6xX<<`uxOS{|5+Td0zlYL-rzdSBJ(WI_TrgVJ0IBoD;?L*NtcAAUd zZ25VBj%JPLAV$lhPYNhVaX3WJ-=p>o_HNa3#bWdjHRR2_byIGp99S7~nZx?`gvzx# z$%(+b;N81~ z?~cq!V6GGQP%y4GM$h-y?4nUq6zxcq6bt<~UF{gtXnsfK8ZVe<~H)2s$|0TAV=hSE5z?QM4#_S4YVzvtxBgi zAz|a^Q;(=-9<|ozo|m$79JhOkPO7w7+WdPD=eTU9B`I-n{efM|pSe18;5YdxRi=Fb z5e4IOax$DElv5mFU@ZlI;Kl&94MjwX3)EgLLy+?3t#tL&dZ|=75c^J@KiwrCzjVIk?(w_q;e*TEB1@S?4LMc*t$pgjWIk7DU#8}=&z)%*DYAc?;Oc-w zgAXdy1rvCtgLHD*Wpok!B*og%7@PwaIjT~ws4Qm%-}P8yA%wC?dlzP@@~El0m`%+K z!@raszMUAYFf>J)OG;T{tklZks)nQK*SFH+bCSGq4)0E|ETE9v9>@g>n{<^~Y^Chsmt5p{u3P-8&rdZlE zENOl;44TYhtt_{C!fA@U5NBuZBvze?=lizt{^_g}f$u9T9Y`DrXyl8-Kj77F@7ffj zwk%>rkg=F+Y;dMB`gGNbjARIHx8l4yz&l*O!2auS1H4>-XC6S&M3i?ZqB+(zxAw`{ zNKRR^EAJOhS*GQSDBDRXvNQdWTDo5qwSBOsYVd80nHO8=JxFU$K9)AOB;D?tlSD#J z)bp>YW=C1qpyT3Fi$p|3{P$lG|KVmapd1#Uf$tCEg9OFIsQUY-(7(jW(s$UL9vn`6 z(~5_;^q*^8u%RHmjcTkKeIzO#edwrx@10BCRl-Bsm)=NHo`K!ogI)Ma0eTVMk-hn# zy=%Y-5^r7Wy?YPcbJxv*#LJnE`_1Y|MCLJQ%O|@9kNt8=i%Oaf_#y;fQF*B*<%TkZ zh`=k~hV$ELYe>S}yx1Pw;nD_>@5#=#M?YO*HYLAR0>mZznA(}xW%xJ2zpMpk1FIpn zv5}uuxAy7p-@iu?1_1Ep$pg4uz5njfcbb2}-Lb@995N5@(VyN6|J7Y~(RJUKEaMXe z7dQ7t`ucE0=Q9LdN~<6u1iLi*JeycWxgt?El~wg;t_4OuS zJ)CAyfb&OSVq!WzKkxZ&_>4 z%Y@2k8XtC4=f@gDaK}YfS(%8(EEt;#D?{l?jJO4M>Q+v1v1~K&n4)XaS1O^;_<4D< zUeMj$O=uTmAq+-lB2nx8?gSM#Fkw>w=9Z;SLcD2;9*0faz}~FU(#E1rch~O;Wh|KS zjf1+M!Bbl-G&-<*OKz{QH7qbdeA^SUvPKu7L)MGU8hhXhNS)Qb*u#}_ziq?JH@3WS zMv<$H_%96&pX&C{0gf~u^dj%?ht5W`leL}+a7BUEz`(%F0HCi+IDi2?Bs?_Ke%FAC zifR`uw{gOc&`&>ZgtJuGzYkjY&g(TPFo$kgN%xl=v1Z zWiYd!a@&C6a%oS6A}={~L6KLg2ifr?J5Di8;3(pDhgVSnfocFV)pK9uWV1DtIswjm z_Xx;Bi#y1E24IUamP}dpFE3{yyQ9Z}@dyeI4!@%!CT1K4(}xu3aOknaetE-m%P@F} z7TuRnkdeczctr96GA+UahVX-%3M6<*t*x!y?MIU(@i?|pAy_UKjzV${s+n{3?kd$DWCLwn4^co1o5|XkMh(U^}tUiS-v^&<7!2mj#LKHWl@-bpw1);ByL7u?zD#Z19r-@Qn{cL@+kv35;4<8&rYu5t;R$A3!O#$7 z8wPwo1*w7oT5krWuFYv%iRSwLywX<=yhcV6OOgw&^EO7yG2{XhJ;hZYhe@|FQ_H@& zm<}efDRDfENntk$C+L-H?gHd!?8L-`@V)OV9;sg#d)OYc&(}i&05)ryX^q+7r&T}& zV)Q=8T|4JwW%QFva5p8OB|r737l5i-)U~x4DT7H_`aq)0FsBW$1pIast>@EvN{EU| zN=kY4smShJ5=q>9wICU5nrq!0Opb8R5c*V6(*uPba76VQQat>r87&>Y0S2lWnvTjV zZbGMkyX28CvNbW*wBhz$x6V?d?&@XxWiC{tquK@G0`v~#zT2HWmRHr%oC~C`LB69x&oG@%{Ew8d@4@%Fqo*0W3=4(A?bIH%EuxzrW%r{6SOb z&JP^cpJo@XR02yttu)_Zo|(qzyQIg;g?o+h5D*HWFjC;<{k>*v5W z^_M7cd3DN|8yp_)fnzAQi?V#aw&W9tZ!?g*kclGQX$FEw?{IZ zQf2nRdmVELd6Wm4H*xBzvhN|#`Y|~8ZM+wv_?_Agg_Dab=xLrcMpaeS)D}ykJnV2$ z&#Un1!Io{zNu+jOgdO;lo}MOV+H{KXi&|@AMn=XT06LUCZVoJ7hThcmOj7hV;S&)t zcC81rxTL454yEz2*zb9Zitb@$9OX#RI z0CbWuu$E{9p2-=xQ5K%*7sniGc4tW0D}_3Ya$%p6{idxVycT_K??|J|WEGl`UHYZd^|>F?+g# z(3=!=-Km5MwQb&|KiQgiB=Q698`ua8G(|2AUS;mAU+Wja zNs%D3uUzj}_WItRQEfYWh!@>^!^UNL=G6*qRL78dBZahM`4F)bKFo(}dnNV40~jsSGZ%XU6qG0yc=QClFiu1+^|sdf*v0pmK5 zpCJ#m4cYznq_;H7zzkXk|Th1T?QTfveSes1Ya^oqKwI$i}u{2WME@}GO zDfM9uAu*|6KS(TLj9JNfxV-lB<+(}{^ncmnvOoO?9w}79;@L+CBMRoczP{mG35Ln( z=%gNt-GW|s7-VZ$0nVmo3JT~K4pTqap5PzL@tF$p?UJ|Q0}5|sV=Q3gkwy@`bR{7t zx(B&L_z_Qeyv35JtGjp zQ7v8Dd7@jv4m2?Os$%0A-^~8`8sIgx1zG3@V^Qq2(zncSpC1_a2B<5NRFXrl{JtGf zaRK8FC!vn^|AYp5@_Rr0MpJImcUx`8UFb%T%5n-KJAd0-6AHa1#{!u-;z;{`i070x#f`jn#jm2Mf;ai+T)d(p|IH3PQW3ji)hkeL@@iq zyO4Q_Dbd3A0`RxOV~D@3(7@~Db=~>+JUea>?cCm3_6RpQuHqtKiDK)61*?8g*h6h~ zr>N8$GGo5CUaC{g_?o`Tq#xHIQaxxvgZX7|?R(7%eWvS_S}oHZLC7}axky_{1xAnm z?fuOWJ85<&H*wxO5KX+~N$wD{hJGvjJ>zCpeOu>Mokz-$xONWWKU7jf zwZySBZ(td$po}Qj5U7`zSEow{AXRR8|jLUbqHQQazk z4dOhaa;_mmz$eH6DC3KofAi5|U4qaN>&VG?t?f16w_fjkd#zFl$6D`==Y^!a=}BWH zjF{icoSe}Ra99h>pR#})87x>LATs|SMrxdP( zykq2sHp6c6iSt2)*8@aJ$@L1;f=xHNR!f7NK%gXi9}!!a+WTrVd9C&Q>9@D^j5*Yuwr@0D zK&>2jNr?Pi__;VYbPJY)5zFYr5MZe*3rydO>xI!z>wBM2d98RImkPrj_5?p2>i_%D z?ui1048FtCpTK`>Q#Cxzf>IXJzj z_{(2q+)WpP4fT>MAjh2MM*v8wl*EYi@%D$=NPw5)EO&S+ zC-de__{{Hh4LJrNQ_psX%Rl^#N|{_dnHo@Lj(DC`ROer)2+$K(?pw6Ndjj#6sfV;l_VL_J8`( z5SFc`?aTmRxD$rQ0J2E1SOGjJHHhJTCY*2yZ4;EAPg^yNM-MUIRW#HkF*P+^n+**M zb4c}>w`u6??R^5UT5R<;7*O?p64jyr=p@Pqka)u(T-5CBBc@Bim}K3sfG5~pCY*J% zyqpjxcy{vBoO{jS;hN~kxp1-bA9OYG6d+UJRk>8N%o5%sFfstBo8?1FM@JV+i|TuO z@)Ur1}g7RvZ0?{%e{%c`DUl@397^MN-^fb`ferJqE z__Xf@F~2?a?d@$KkASG?`^%g!D=ts&9^kZshx=mGy5iu}yMk=R6WgDWd!5<#h+( z+n>bv-Xa30d)vPb^77&QNnKrC-R{$1ac?~v5wfim!`gWu5prd%uBoXq8_gt5HcsIU z=?14#*W;B%;}YcQ7S9*3%tq23bng2ZvdSEE#KgobfU(+$sSR%$KJZdy!bfA`;i;Qz zw(v1A$)`;oe7EI$@!PD<7TWP}XP4{>H~uY*p%Qx^4yL-!1(=Zz>{@D`+;I14W)_xj z^-*wZ&iT!NfPey)%2rh`xOjT!qqcldi8&`zYv3Z7VN#{7zXc58ZFnlyfA}FPHXH1D zEC^rB8uPGA%_{>4xnM&L^SQthGZi`roC((#2m2SF8kRh)GXDBXT_{V{`(2pX?C0D& zk5w9oIgEg+4S&(*_IAWsWYUyn-IvY}_9F^%a?hYL(d4_;BZd5<;Smu9fJ!?v`_av< z(fQ(!PO|H+HYz&0Q_Jo0{RJOKbab@qJRdNhz&|qp3}g=&BiMciAp`Qn7(TbsE`h%= z3xFpoFf%g70!n0lB8!$hm)S@HyqKu(1FXM|bZ`-sg--$-<0dfOfDE1hh@PHSFrVc= z*8ILo13)tgi0^Jea%s6Q$Y70@fYwik0ztnDRmGQ=tPIY;`FbDj+H*MXwu)XB=?aB(M zF%X{lUhguLR87|Ul2&8V_LXEYi|J1yh1pz-gc&>mX;cIMC@2Q7Zi#>ZHR4XIX)yO;!K-E2_*BY-tJP4|EQ;a0zQ@Tt64j6L+Hp5?eh&fF{k8O(XU59Gs z57t`GK*2Qh*^93*;rS}#kV`5+V`je{r#KLt%lCfc{TfBQnNnS}m+?)uLNfc*Tq}^` zZ^>Z3&osQ(NebEWtamX1Ki9o~bd+cvIBR)?NueP)g&2EGrM*C2i4(sBg`xDO;*R@k ztCZo7&-z~T**?LFljOUY%U*8A>*sL*^-g`)-=-rEIfbR46CGIhl^7?v|2-@KTyn58 zGfV#W<9f6TJl0>7cogiioAv8YRaHxM-SXj8|W|n=CwzMprzlz?~u!}Ycdf_0M)B(J}lZ|*a*gixr ztJ)r*X$dw2fQv^aLElgZ)+G7tnCYv5vQ&JFO4PLS9+kAbmG`yH{M zb1|`+OKWLtsCKJc)u1C5BO7Ux_gkI6cI{UvLZ1%tlSDku58V!qKC2cy1FSeo;dg9b z+(6vmu%3NIgS6u!s1nalPlVuK-^9F8_#NzBT0%wI?o6WcZA3sWAHSh%+XtVHQg;Sv zbibj!dJ0Ebza$Cmud^tH*I#+g84k%TW_4CgS}4!XfTk26&Py~CICoxraKr++BB-d+pE=j4-{Px;|N1jScU7eOwLS4yBOQ5)hX5N>nn5tOBvV}$> zBXjBHxrpr!agMkssIAAfS6W{G`sy!8=Ggjf&(}4-KhK)MC5Pswh-O|~#~5CQsIp^< zKras&NxeGTjxeu63up#=x?4G}|3_>;OWdK9UFj9lm$#xNnY!glqf#3rZh@GB6xzB#0UT^@K&DM$_9 zX+A5|N*3OTPOv&e!9_{r0v!Y_NJhB*6Qr6QD>eqafHsswnyIBE<+YxD`!YxTH?1sw zv(-LWp2|-O4b&>^2BegB?{>~WU6HizE_p34MijTh9l&fw*!##pn&G(s`N~});oi6J z+3Kzv%V+YyZ9>LPsu=aaP{2@@{_A_`Wx*=g0#4|-IEgz|0!;i zCAeTd13_~YB)mNv{qa5fz;-2jrMctoz)s0`=4eZ?`BGdR}?PEoZ* z00rphh{#uXLXT&Pn+?@$#m=J|8u)-F6>|F;dVj9atSKUfaQANw^|n1P!1E# zGY+`6Tao|Mm7Yb%&#+q6mO;4|(%4fY+!|JF54#tLB}e^~-)i&6t)umyIXY+?iK z4bWh<$c)*sff3{8_^Hoj#d+vJ*-%XM_E;+JJb)65xG+x&+YR>pjbF*(gaRi-FNh7k z)u%xnu>?|^Xs~evbPzDtltMa-L#}RYB-VQ8q7fhA;LmBhbMDc4I>53Z0}0s%8y}aT zuqGYtLAVa?n*ZlB+<=pT)eya;z?TwYdA03z*JGfz&?S_xpQz`9omXabLROq<2i&a< zl%7`ypy~$y-fVqn!l^C1>UfD(U>gbE;`Sb8vAbF><$p(B+#dhmJCWG`z5V$ALz4dQ ioz4IEj}Ay$e?W|5*zmMq%I`)1f8<^(NmodI@c&=M%6Dai>F#EOba#g|(w$1V2`LGYMp3#`x*G&Z=|(!F>wfon&hx+H zzP{t$7mnk>4SWA;tu;S$E+W-bUSXh;pgwu>1Vdg9qVePj{08`=gp35fQngpT1^y%C zA*1J^>1^%cZSH3EMA_WK#nIWr(awU>%gW8&&e@5Zjgym&lZDdO!^6d0h@Jh-|Ga|D z+0BOC?f26w@KaD+kyf$i0eoaqQQxhH(`||bD&8SO6`hZemno^icYppe% zrFYp#=Hvd^QOnSrBVRfZI#Otd83z%11a3$9Yq4qAU;p{aTbzx5{{=}J#R8%6pBI6z zsD%(oI{o)WNtCkjFaG-z=%D{!zC}(#KB%F?OrTo7#x~)#jg4YoELpz&R8hErj4Qo)m@8tgtzloF8<5qfurg=+1a*ezer$c6a4kJNB91 zx$GjheC2BYcdg>u`Z^sgE#xzg-CD!ebA%AAbbjZQPoIc9+uXND?L`+8GBQ|3KJ$L; z?uMK<);i2w40sU@G9tT$cC3(nAmO!;NI!aWx-k%YNh;*gf3aUZ_8SSE$P5h~efa!0 zCYjLeH|2CMS<=MQ&7oq?EPiK`+*Hrg4f#GXJ#Z|l(_?1GVeiTP%qq_I=T}+YPO7yjO2XQ8yP45FgyKK{v+vai9#|{-7|`SpDBrn8U^>Ye&_b{ z{RuR0;NajIV-@S279E!ZFEi&aon|XKF8=;fosYnH@hvv@ye0VY-n%vXe%jEva7N`b z&j6{rbk-;2#tt=P5`HI^XPBh%LY}(~F>f7uW678=m;B3bf^Yu*aa;795zL}Nef2d0 znNwXbZ_DL+Gj&MUG>|SRD4&1Ia<;-eF)NFWz3r~*@BUo1+7K6xo3y;!&QB_QG6w3Q z$1>{dn_Ds8)7RI>U07!=1RQY6JSotlO3Oad@=~qB#OCX*bhm@r8G^2q-wlF_EJ^+-lObtIq3G_bjkmy@UfAAJGwFWG05}rsO>T3-P31V*Vn_8ZcZ5j zu2R$GrtD-w9>+c}&}~qB_dG*8;t73ah8HCNHW@X#)hV*KOx8KhaT$7f)vY%^eTF$J z8%KTsZ4Wr^cxE#K#{oOiQ*c|&M(Y>6h>`J4O$4NZZrvAy%;mH8(`DT1y#D#e|e4)|X2WbCckWCvQGU0&*r;k>*|8Xi_M1jjLXTdZB2!kpJ)*@uZw zwe2e|EqB`KYw?9_QuuK)&64iBeLLv|53~CIsbx7Xy3eCi?X1{L`uZY{6*$BD|!U z7l`T#b?*yu7*Jv%e2v-#YLcaf*QcBA8SI%to`b)d_69Dl;jMj1cx-dTj>_8}5<5bl zH3b+;Qg2)vqs?v)v%FboP(Lti-d-AI!@m&l^7uUdJr<##;rjae8_jP@l~bU4rqvekNo#86HS>W3-0lqZ8`m@O7KY1)k+S6vc<>K4>c3*Db7&-Qy`F8Q~^)Lh%cgmgC zx6u+*j95uU2m}(l9ZM#>MHIi|9vV2vw$W zdh)`#dDU%^kVS7rfI^QV1++BwTGMmaUqh~R{Vj`PrY?PC-UXY6mon<8Zl5^JL)P_Z z>FNDyTEhiZma;CFTG`pLzx(<6VrUuGJC&NagDN^FB6Gyd5>n=Kdr&`hW^Qh7|I<_} zBm<7_Xtm)5dKpK^f*B*i%`2bbPn^4J`@36qJ=opp1lfcpKIWAa6vc1XWoZStxF$9o z7Ew;4*?!20j>esX`^-(`6#}mKx9Z^OOv*l+@}9%1lXWy>SIP3H7711n5fMz;jN^99 zL#gb#;}XmR0|N?A{@7kkcl)ovjNU`pGZqtXjg0 z@7Y0vCwD2VhNO8`EA8)0C{*jb4;LC6FAy!{v2bw$N2~E5DxQ{I1U;$b5S4fe-n@Kc&yRq9o}j_fEc*DiFB@Bh{=Fl_vAbD5cIcQ?sGd6}@k3-}3KLUkorFxY!KhW7D2x_W zA98oO9PHgZ=6f+@_IAQhb;L{{lv3Zd{Vg1ua~4kn=)9&I!H>7uH{g*z@=m`>qhPW8 znIp;Nv>@;joqx&ai|z+OaH8@b%8Z+)_r~Zuc0v!uAL zY2rlkaHZu#==47yQOh^e#=DKQ{P~tnN=Fcl>cD0H`t|FCpp?~N+MKVt?PiQ`KC^_1d~#ybj!KrmaKk+&JST#)Ve>@l@n*2#s&r8<`^UnN-yFK2i8E z2^a;O^u%t{t7I^OsqGW)#+b;NQfw5iY0!24%}}cw0V6V-1Y{@b<&S}?b9QZI4^I^5 zScZ`89zs*C+S5YYRZJVW!b?Syr*hs;iA;X}LizZ}G8UOOCxL~(H{ficaiL(8+oS`#yC%`<|sOHIOlA%+hO!Awrxk&jP z4RGVjxzg{8CR~z&hS;h5_N7d$`k_3Wc<~fTFKT3GYLL%rl6l9qC|AT##NaoQ9spXOXK~Qw!MqEkUx{@^El9{+6f_#xpVw zKVVRWn_)O&XfB8zQ&m+J<#^lN!4y$S-B^&@o!9%ZKfpZpWBqnZ{m~GXF;l{pyl+B$ zeEj-3qE$x-qC(6bF@IU5j_m}u^ODHaQr6L4ad~SI(x41n-^e`Nr`7#-Fq7woNkTcN z#{RMf2Vbo|OZEIA3K#vXJ~5=9Z`{=$`SEH!P9Z?UIn9=QzV8F3&r$PP?9k;|mw?sa zbA8a>cdPjQFC1U7Ixvew2n=9tQ%EeH#345^?0Duld7A8O3V9RI!YeIWATdI^YbhBl zR1g}}zsSb+mF#p7pEne)XHgvPrfws8e#Y@FIRjmq1&>~--;;dz^Uu+pnax^LXvh!T z^`6?m!;0yHI@zJ<&-lzbKD>C^pVl`He2)%pPebQuSAMiNt~y=Bz%g?iA0vqNZ{Cj8 zPGu&C{WvrF{&z1Ley?%UuGFV%?i~Zd9Z#E>sxg@oyG^7X-(|b8m<|g=n$!)HzZAt%dMiStMXwQ zZA)q3oEngMTf-5>S1A83lk!;5wxqhC$&S8aDnib>u;TV#Wm3T6;detMAfs;Ok#~D# zGKSxG(HMI4VjHKLb-KV)J%sfD6cw+1VI=Av*Z&Lxu+IKti`vaHe3frKO-rpb>27(HF7ZIa(j|8*i-QSpr)Ed7C z)UmUV3T~60*EjrF<9Z6YG-z0l6^=T94V?bBt#|EC--Ijd(?4_TDb!f?%(*2}_)#QW zeNJ^w7a`mOS18blbB7}ZKa}~O=kQ4oM_&x_m#`P&Tvp04XHP%r=A~W4Y;C(T_&scA z2g5Y#l<5WW(@&jJiHr&AK|44dhuyVjaSa#Mc*=pW*4(mmBQP$miz6OT@k0YK!Ig}R zfvU=CD1~ii9Y{It1P^PX3YrGImiE@EDP|GLS7*kN_IXKKwyxXfbi3tWR)8{dE8AWa zHOZJyCH!$OSXY#XFx021DwA&^4T76Hp z%piJPUFEr_!>d^5@n^zlO#e(vpxh&*0v{eh2wptk%H<`~-^Tng-p~$WOMOnFC}OUg zmkxb#Bm)pgKbRi2F3utPEro1qi513I0oU)kE+g4L^)u*T4H8xzDA5qou1!_jsPd*u zJWu8d8K%e1rC0xDc|(?` zXmd$KbZM@DB@EUpx)EjVG#4Enle3^f9z63t6Vx6;h=!Ujemf-dXWze#H>YrS;+TrX zN_9sQMqlgx=$7s>XO&ubKiDV-N!rCrDt~RdqAgp)Ve7F;vqfYs{+ZJ<0yjNr{>u-< zd09mK3p8pHVa&Y#xfD%Oa4OON9=2SXcm`RF{Iq)7LKb8H0)IXV`T+`6yM3n*&uU;; zYoDLjmX+1_--)oCQsn;cuhX+J@nTCQI$ETffUzs`lQd+fQMi5MvVEf@pL_6}CKE3B zO0?JFF;iLz;w+Jxj&e5dk>b^b>l3$pKu6Csr&rWXgWJ0hOig~>VywMO1#MR%(kpHz zLKo!NoW(#oI;kN>{yD=cYYMkHwrK7z1qHCwI}}JpW{@3M@E);bTgVI zYFL{z=CBNS%*$8nxgq4LBpXVAB^_koc} z%znGyEGTmeOh9(HOndU$>76wfB}F{r|oBfJriFpiMgFK+z*+W zD;7db$`_|9XXA;JmjB_N^CNm%GfFu!{XJc2&UsIBEBP z$zxO907)x$`j_`cR7L)o)E}G!#ZJCFg=E!`RP(ahGb^l6d;oe+HcD! zj{>(dy*Jw5J!M-C9%8L8IQUs%3}j_JY>J%#4`1zLCTzg?YCZDV4M;}+=29eS{SFB1 zoabT#JuWSFyI*ZF_x=UHQ_}5Wld3_z(}B-EU@0E~Sd-UZZF_tGMC7{Lb$M(|6-BqM zi35fRxxl1q2jp4VcI)-l3%ajIjoajcu4|C8ZC|IkDyv3vRXI7eF{%RIU*D89Ls22T zAzpxoDU@&%RfJ%_%AwxT)X?|}IOOS+uBrS-dgYA|{7c7v!ra{4R^T()8v<-@1<>?F znXz!KN~@nU;9Zj!`py6&#lfg5E-C50&mvIzQ!6KPHrdHXt+y(RM(W>H$QnUQoO<9P z8Z~f?&`_SQdowj+#*w5#uk^|61+TqUWdPtH>yG@(0q(JVWzDAw^MJv=u?n}z^6dpg zshTv73=qz(qa$a-Ca+T6AthzyT0modU5xNAHw3i82p71F!#a&_R`U+!Eh&4xK}_G3 z(h~Phq~Bu}ZUZ8|L{CvDx-H z06NaGX*J@7f8T_060R4LM8}B(QRGSIaG>CnT7|sOg(-D2^cca;-z^o8{KP z-A5|vh=O!sA@4&+e{e<7YCeD~+O4Mp9`nj-Fi2bj0iCR!qnOGz0r05rQOo5XqjdnO|wA#Ldene4qaetmy)o|ur3>xjW&Kgnpt zFrE$inCUe*DD92&4ENCl;4Luuhv8@W1C>AJ5d^bG4y!|-VGRNl%A}aWGT{BVRlmG7 zoc4wKGd2Oix4ciy^gKK~GeOm5k3cL_gO~lNBB;nFU^ij0v$NxR-H%-xoM~=)wnXzP zmNC5_u!F4it4`o7Xl;BRBLU1}=%C}G!=@Rqgu$xyml+hmn*k)D!Dy^94r(EOx!{gV zPF_a60J8$Fa7SG|?w4Bpry4JJzQh&n2xQB@UKj+=r}5!X$SHW>ivwl88i!^^>OULJ zgnNo~b=Ly91^lwiQAS6g9CZ-f!ll&vSfYR9g$U6jsxd%4i{Q?<@eW|lf7f-S#Q1aX zQ?D2hWM6|HA10c-oqYi$OWyVZ?GJW~2qV0>JlO~Skx8VCBb?GLxW&gX+1b|!;Kp;M zpV5U>ICME)W7ARA91*@N7EgT{9L5}qkY$t8F7eiLirhuwm=YoFw|Cz{qkFR1At>2dxAl~_ z7^{Ud{G6QXD2qVc8R?Xpu=3hZx->n^5*=f4LnU|Oh&ZJ%!t}(%%n<4$#@5C9K3RTX z-2enicN$C7%n#;kaApHsi@S&pUFdMN=2*c%e@%q0fAE~-LU^ysx4n+~Y7Xl+-XR)M zgG&~r2QkZPVK2d&iqCYB_Rxv3xx`1pk~1{+=9PxAC*yP9M=qZNvn~ER<>f^S`cjc@ zdD%YOQs;>kWf<#F^74`yfyg;_?8&tpD=3d|(Vo+rSLTD3f@QrzHo zI>3-@7Py(LtMyf^{QP~9c5xB3aA~^)&q5#V6nJiQRXC?tdFNh<75yy8)SsYCp$tpj zpmG<@Q1%gFRMnlGCE(JJ5h5fvHxElApShbtnOHGL5I;loJnBT+x6M09MdR}-o8@=&F9 zyZc^gy#lB|K8m*&2UB(N(tu#EI{wpc)f~HNxM=#;WV z^yfseXOJ1Fmy!J)z9!5BfLpi9eOrajELSyK43iU6Bv$*0|J!rU*|y*|+zmwi_jh?y zdRQ4f`)}$Ozahcl;NU1Y;#~K}cDkK~%;1F&qp41(Jrc4Se1)*V;R$gN5j`7LU3U_ zPFpBxh(_LfbJ=UU@3hRxK?xtTDb?lc4op?6x$$ZSM~c0Jn4n?7qa5vGomgVL3to%6 zDCRP?(0t&Dz06hpDk3gYTT+ops%Xtu!oeuP;~<>K+#GlW#CVxCUQKO2hf7RM@PbR7 z1%fZ}+VSV7r$-LKa@O_{F@csLT6S8cUv&q}`rln(()a-vd&}nz^hQCG_t~=sCL^eI zQ2@XQ4bI>g)1Rny1u{bED8bE-8LvAzj*qNW!UsIS+$taJd{((Rspa*gP;1SRuyzIlO?4x{p`c`NB5Y;$V&H2msy)MdjR08x8 zf?4g7+o;DhI?0l~`!3%-r?+3Sy`RnBOenG!HlqZbulL32%2y)4A@?RY^J^T2_p#T9 zYt!sWO|FxOeF4IzC?*TP?hN?lYL9*0-345ss-ovf>(>Bs&v$lqx+RN-0bw_- z(-st%&Tr&C!nBI#FQ#Fgqx!tOe8=Cve@V+NvnyMdxGa0oJ)PW@cvFqsB1?fOA~+fe^WMCG zAwQRwmnRZbzV+@Ic?IILXTpPmTQWgbST7L=d>ehww&0zMJJeD*a89`(;Rg6*j|^E* z_Hj>6q6JAS45l2jOiF{nn&!G#z3;{Hbu^zrsNjWme&`>^MN#Z!UTiCR`9uvj+6|mi z8mZB<)b(jKX$$=XX?6ds=gC0noTBmW5eBZKQbd6xVt^!Ywt#eosn z-7hf^17Lg3ZwjC7KMRd!9q^4?pB+dS1uM-s6+6)3s)?66OVDl;+>gOHt}`&9CD$?I zoH;@@p7c$2c$(wO+JL_&F5@?+_Y4S4grGTIo4X;)2f=Z_S`PMqBYS& zZhi=U$9V~bZ*FZ)L6Fit=&NioOOW8umm%PT3u~>VeX;y~IrvfZRf}CB2@Vt)9amOV z|GjcL-%E9QV|NHgD28sfLwL~(t#9q9>ObxLa4~UYp+*?g=^B1l9isX+0b3>fHr$CA1?n>y3UO8=5koDd-AdiR4=*-(GjsMni8^oZ8a}C< zuGzYutoas3^Atl>QiM97@VL`W(56CPZTF`!FTTB7wTg}*)+TnLNmpYCUDUhqTCb*B z1j?tp%vxzwa|GqJ zSB|_5->wZDB4%Mc{4pyO%+{CO(P8YpxVcFH0up?(@u^gjnBt3qnroQZNl2n3Lo2Uj z2~K_!qi2O)(NmY;37*7K=iwwH+WBb`NN{Lozes={Gc!VUyLrbAOl4~ng|#9j5ez@F zi-t?0fADEYxv{wDUJ*Y2rJ0(U_{BFFD3U?0-YSjW0ySD6_7PI?Hnd5BMP9Xtr1=XE11;yxcEkW_h{hFrp{aey(**4@)nh{~ZQ z#F<3?={%R=EkF#_#CTNfmN=)InuG(fzUg!vv}$aRyGU`&5=7`daMRCsV>PEj3xh?9 zb&3`Aw_sM}Q|!AT9fug0&>l5W3TuwRG5K|vIU+X@U(?Nlg313YIAY>itJUYBj1PJ( z$??@7Zkp>tZxd&Yp>7x)YedVuz zrTkw@&~d&5gqxwD*YX#_5b{4FKVl*dXTvJ~Sd&axIqY9Nko(VT>q==eIQ~Oh05r%a z)4ZraXYj(Ms<)c|nvq>72mjh)U^lOZP z1dZ<__aC+y3INlYcAwhsYzNy-F;K~-0(L;%PJ}7Bbx-$J(~R3f%Ix#s4(0M7A+ibV z)XVFKUkiSFrMP5dtU$qj{Q2`I>6mgFm&|?O&7be?El07!hw-4I4}mdftH4kksNr+* zKsx)$0xA&rNCZ}cknX|5#>Q)y3HQ1u3cqMNxGMr99}G=g;6+lGU(FhUux{GrQpC6i zQJflyqoKh;jiXy2-KK6%TG)s8c6M$D7_$3s&iCjT8Rgg-R^W;G9Q9&nSd1FfSoZ)I zI<-*5h)4uVeuxrZMPH|iyNZ^MukovHpILtgOjnf})Fy&4br7`j*8_=kHv8lTb&hGc z6+Ljb(U{q-A((|g8D{-Gn4dYB1yPyTz=_o;)S#Op{^(S+RFZynzqX)?cwGYe(ZE6d zG6*7aiY*3z22ySiB(x~2ZN_x?Z61K5=;@BQ=08=W9eW$_$~Ld39T`Wdej4;49lwkH z!tO%P?0VaAsU_cSvU3thb#``kk^geJsrXCu`USV>-=^pYL=tD>Ph;-{T-U;O83BPf zQUHQ__<;OmACyuOoxD9u z4!u9qrRlgq5DYZt=cDw8Kla$BR)6VY=V$WRCE_mYaK0cSVgw{NBgsnOsOAan;6d^I zcPgf()$K#(S2Co~PHTn_>%K$uC&Wiv|JxRIzh>4>C$4D%A_Fv1#m z*Mfd2^ZVY<73-F#V2R(UZH;C!)5ylC#ql~!D+Q8UWFJ)vcXoAI0qSAT2R|+X{A8(? z*li+6seZLAlTOXEEO0dnt%b@O)=*z<&sS7%Gy+#(3cVZ5TrEppp}7K_AoV6ZqRL6`yQ39>P~^B$zUXPm!3ciaP~Nu8RSS_|Q0KOH<9yVw3K z)jxm*3@@>LCIKfhR*Gh(wkPk;!|p#W{~+8Bf?rZa@YtI!)hO0srSqBd@-{P`1ytg< z4VXu@<4(vAK}=^|5W!K>rS_5M&!0aq@jN7rCHl35fwT)uOoec7%O$Y8R=eK2>xB`S z2{FC9(0Zi=PE>PkK%`zrmPwBsli+SpFxsGAnUyzg zk?8;MC|qzK6IBR1T~~gM*|CFirF~Mp1h}h$zz+k=Uw6M>#nclT!drZ}q?**r24OlT zb*V{waaL;7rjQF_OL=Hi)F94CFTn%|(%=_50!o8B1vNqNGdaH2F47t>m;bZe)`lAk zRX%vZ;Fk8!`I>l&hd?D2zrzt^S@#G+3WIc*Y3gRdZW+P12mJ4eo?Y`UHoz2;3SuS-0pVh;vpOq)7HLT(kGaW!(AMNKj5eK#iktYX zB3cE%8%gIcwlq7>ITIfb&SF-X#T-Rrz7gviN8Ri5?XBbrS#Y5@0|l6tLqIKrIb6=( zDC*C&FTq=);GOym^1-#OeDyfR`@wvM(@Y8<#Wn!+<{I6<;{vznCnX*7HQbEm86uqI zTN{D$P0EFmwF6KnwJ$_hxi2i5IX(`qtLoSkS?nQs6;&w^_%14@||~ zHss#zo6b6*CVq9mzsOK5a})&kQV+^NH98|vQbGx;bMr}rrD*1=QPs@}XudI^_srJi zAzYt?C7OVa^c^J^^5P7>SURYj%0L049xW}Ago95zH;N}|62FOk(i_!bc@-GhPa)va zenK9`1Ao}@j6y499S|<{aXTjajmx>|KmK<$1~r>>{qlL$ve9SMZJz9-_mD$4oB-+u z9oM1uuWp-zvA5_h6_lWms?5F0D9(7|MX}V?di>n`_6W?ir+~i{Fj4C4A?V;DsQ^Ys z;B2;qKsEBv@oQw|!5u>W%{ul93+4D}5nhzy2WGW9gzuIng9^9w#byr3ZW31g~8*I~w7%kFO^*l1oh z1{v27p=d=14TkwJ&!?_~EKeh2^b~-?OgZxNx70VU*s-5O|3q~BoMFFjOzc)^gs(yc z>gp&w9AsPH4H6)uJ>Z*T0tKKf2KIj?}T5q|eM86VO|!Ths;+Q%y;-P-vyVjqyJ(N2Tp zOrO+ z^+IY0GXvjK_GX*nx{$m4ip<*Q)?GHl%{*`KG-Esw6s6re_ zx~f+^^`fWqus>d1qnV-n1nnk(ROod`CCs>C8$rrWuI=SItEc5;ODw*Oo8`AF#+uQt z{+e0gU#hZ-HV86#{(3RmU7YGUPP3MBM@k;mb@HaLGYuaOM6A`_W4tKJNr`gH=jmFv5fk#NLiuJh!R zOX?JCPg#E8`Z?u$KL*pDaG|N z^4+k~mBk}Yp|$?iW9Di`uSusv#X98?yCuYbPG_G9dJ&>k-ci-r9u_FKt-c$^ zP1L^I)!N(JafvqW2ZpSudNsa;doKV7iw+Q3v_VlKna<>bPM<<0L|vxed7A4$f@kdb zAOljWCht?TG?E>$bVw3~=gQ06WGlp|20_YV{k-l4dXK!Ip?a7px|p>8Uudt7)Lo^( z^+_G^?xVZY3tHBQ8qQZeYE12)bAo3xbLZqFcX0>_ikz_%@|F2F5#%Q<6_$UI{6U6C z;O@jq2;nD_+^`p%<-Q{4K8TB_Wr~d>lFaSM>U8tL7@8!2BOC9*D=xO%8S8Z@>rjw@ z5X0>jaYa<;(2(2%0U<~}-{f6>a|e>{el>@qds|M%fwkFhf2=|IAski*$AnGu?2Zg^ zXV*R-h_TItbN!-u2lr4$Z{u)dhaL6AXHXLyko0HcYi1~dMdUxDw;%U#U%6bw;ce2P z9u6kHdxOLrsf|DrQ&~{qSeMWoeYZP7h>+s2SU>&RVz04}Pmh(~=TD)>dQ5%!(;_%d z>CT+t#g%Y4gyoYv+-6YNLqK8IinpaaB0DR8z1 z=1?CD65$FB%M@9&Sg|vE(?`vQe|?kqT_t?<*)H4QP!2-Azd6ZIxuSw|+QM{d zQ+v3Qfz1-itw#d2-R(b}BiLytST7|KpT9Jvy}ZJ5%+ z0q@U|B>ESvf?})@EMp1Bp*Pl?ysFOVFnof#I_~F)O~-JlTL)ff1lGASOrt3Vt2>VeYY2f zi_BpB`G`X!E6`07;sX5K(?8#{2iJO{H$iOj?;kH)>^sZ8SnYf!JDV|y_Q6k_mahks zUY<^Bs&3-`65)UIo2S-sJ23R8O*%Ih*T#CBxK4R()YQCFlk&senJV25PSxH_xvo_5 zfe0AiTb6W|9>Ep>GiTsRI~)oc+DEW2!n$1caX{wTP}-xe$&*^`aErC}%PcNsPV zNE0^E_PMicgFwq@EvsX_c0A2*K|){x85fLxN9xj#8&NBQ?Lz(V9qQFDe)EgnM|E-H zZC)A$8brGE7g#8nx(7ga z{w>^7M<6{$E?Yosv`#r&Y!R$L*}~HFKQ?8QlwwMq7JtAzU8TUSe-h{7AQ51wr>EDN zJtSrk9B@9Rd)*3l+#34OZH5HQzx~b4O>(F!J<9Cm^cVx~M3Y^TOe0+@Dzh1OA<;sp4dTh4WIQ!F_$4v2;nh#3|%Za7wbnqdY3 zWTwwd0s%1Ya{x)&C@ymV)Hw|GKw$#ew2VA zmg!{J9bEU6!^I}%1a9+A1glaBy%S%sh@*3X7S@@T;O0^gr3JK8g}@CF2Y zQ-43b0c$DpKmmRQqotTW`7L${bxjV;qbtQqK{a*i19NFU3lmgf6fc%x&#qSMIYDg8 z#W?pY45VxVZzMhnU-iqDR+r02p;y7-+Isof(hto7;Cq4FEe1>n>L11EqNHU1ebB_mff&3@bP@j;L4LQCQ7k8NO^44-d*6YCH4ma5Y?lZ z!lIuivanQ^dd@qn(hQIRli(Q$!XNt>LvO^uL?uVXyO|)QZyBW9j8LjZIq=#Yi2ojwQJ>(62Ldz0tvMDB@4?mjGc3^449X>!ic3B= z3L^z&nqAsA0$*b5rRr2g9a)3OfVyBaU*eOiVfYq4u=8j={3(}u7}@qbC^$H|yEl;9 zYxYHs^KvV&6v3vXh7_7Qqz7F!MFwIN?J7}XDNJP%O*B(3>0wy>o>BzbYC-oigcJuV zL9OtaVss!*ViPlrUETv55AVyj`$9@U?MlN)_d@zNg7ZEiyWRUCApjT=+g~UQy%LUY z3Mok~o-v!FL`e)>QC>3vz>f4Ox+-F0(?NSg*oV+}@Eo_UfAUj{h%sttr3+)?yJ~j8m}#9qz%j9D^sU|* z6zFef*>$k@wZg`$8NtnygzOfCoMzIy+!sf0swKPj`pX`Lob*Y1oQB?rOQf)e1}=@V zdrlRqMe|pLMh3HIJ17r`8p0 z0~FLa`AE8eDkic3FP~kGSX|`Ca=_V#q3xiDYa8Fg#_c5hC*zT`tJQNV2#+8e!~jaR zCXU6D6PA zepnvw>U<@)1n{c%lYqK*P*R0@Ig0bV{n4Kz{*ECG#ddV7?& zZu+C+8UgFGI@nfFQw%r+8|Z|kx&#ypOj{`TAmL*Rgb6VSW|c|k0PEyw_e zpM7_Kxm<1%XB@W2vL~O!%`Q3mHUipIF0r2lLRo|5M}?F#d#iRH3yWTq6E>q6g%^0Y zU8vP0P*2wWlPpU_SJ~*@W|0dF5+1c`MyML@s{zgSEJ=9@YYU@}x9Vd~J{J0LNM^JZ z22={A*(oMTB>uf50-GPHBfVRJpgVTQci=*to z6{37>=WWFII4J6`63FODxLf6h?8QASCehEx{~QC-U;948Hm_c@F=9edl9h%I1IN}0 z3SUC6(;#b`U_3cQP1Jt*#5_-Q9O{lx!94VY++v?=GM_%Ah?Of`hW#0p^jD*qpY*O8 zV+eXwAt(^V1N{31X)=3AE<4q3diK#QJ|wCTuj~LTYsFsp6YB~QH(s=D-6_bo6)P}; z$W<9`&Q+FR8Pvz&z`N)xhpnaAN5Tyf%GLooO79f_ueAo6vjZ>EK#2Bbp98Vp$$Z3z zUdhT>lk zxBavImp>at67gk6`>T2yC3 zEiIBbZusc_>SE|if@E0sZ{&&F0TW|cB7^${;hKVmJxXjjoa7Z8kLK8Z1-PfH3wsQVBlf{wnUat%&Rft+}o zsc#4Rpy=fS4Jz@#j=MY?TdnONeX%SC+^AQdKW~p7&mjk!M6WO->(TGj>*5zlrd(~Y2qpD$Kt%EV zF#5~RpHokACCgnds)%m@k|dAlk-uW;PRBKtvO98P;v6?lY&!zKZeIJ95=)B5 zi+nBm`9T}aDa7<=)Y=pj0kh$i7H*@|`^1dc6>EBwH?{Y&T|!mtCA@uUoIm>%QyC^N zBaNEyHul+5Ht$5Xg+f)k`EFy5kdniVwAHV{?KZX2XaC4Wgr21ydnmC+XGA1=5K*cSXs2KxXPL1~3Zxc{ zg1p?FHW;E*?@K;a5K*m#mp)Nn)v~bhl1=&AO|NJ>^=q^0AKMXpGu`{_azeRubFRHV z?SG6uCjGAT5MK8+v0)VJRQ24OBoqE`&=}x>a^S<<_mr)nF1mSxZQ9lcTWRvJ%&HQJ zHCN4>pa1%n!>De?G6{Z_Wk2Ka=q?e(?~X%vZ{+7q3@9&;mX}NGMcI$*@LIm#l6gA) zF_;DlGL6GrsuhDZpcYdmbFz^0e{{7~xAb8Yn0|{dKv2`Rc{S$H?G^KKr_^N~NS@O7 zSpt=$&QE=eSPCz#iLHwhXu$6NMdyUMyqOwQ5YvQGX>~jzQpS#CDOz5(9uW16u&@C8 zszulFfM5v~l{Wn&#$`yTg7+)?V7Xc**N?VmfB*c_UdMRAm7#YVhNVV)a>cfR>!>O9 z$qJMmci3-?h>VHR>YM|~UX$Q^&ug&Vgi1Q)JXe)3a@tQ*v^NWuD6|a^fG94o=MYhq z+6A*>lp|SG?L<^%2{4!8eR(=MI$~j~?ju>Se7W?A$L`mS)zbQ8ccsg+gWioxR2az8x&Bs#P_g@@Nt^l`On3S_@iK=7Aud5jv`*_ty) zmTf}^;z8Qz^DX}MlSSH!rDzurW`xf^qrQ(M;DkiV^uMEs0UPQ}k_u#Q3S_6Lw<0Eb zoTG?I2HP$muf5@CqKd+lYf?st^zblaC;+N>_#1N`0plCi0R?vZ^{@eHEPE zvTKk}-AFYJ8o;5N;kTX=u>u>DTn^JLQ!^cDun6?XgzF+mEe@WGjIkQjFwU$))w1R* z%tMFHk;>e`FN09ygD?-%0~ur#ISYt`ypd5xGjLp)E+g8OLc-(W;Gd!)kst@LgqKY( zr)>_U7I(sx;8j=sZS9W(fp9|%y;-cv%1XhkdC-edqd08MO#^{gu{Y~{(v39VnEll5oJE|z1 zefH$m`ah|xuyt24{cjy+F3P})a0FY1#{-A*s;c!{VA>Fn&Rx~I3d==OgxbyiVTb!{7dNs;c74rx$I zI;6Y1Tj}l&>29P$x&@@WOS(fsx~1!%`#bti9Ud^&-g~XN?&rF%2jnA708Lr}XJHZ@ zb#?Wa1z+&HY6RrDM|&mYjKcKc!|z4_HV&FB=sHDLg1@tIY;25gqZOpjE4_ou|i|8YMC1X8Wb**gD19gp+AAp~?rYWrWP*1so} z)gO>8UkOGHD+}jDzNg?NAN#Og;Q615+DQD@ML-l|FY$b_ieQ7_oHJb=|Kdb?E`o_} zQNUqEi!#3kk}CkTnf{&s+_wt|j?9N#fw_|lI{&msi=edJrjN^}*A{cO4MLFe{0#4= zTF$%2$I%+2UezQwaMVY&0#{@v>Fg1_4-Z|WgL|Gzz&lOh?CT3>xY^or%ZtO#Ec*E|L1gYAi=H93#WC&ge-qG~^aFuK-NWnMEo{tj@ z{meqB?1~LKeQh_6s#dEQSlQU-tWQ8vad7!CdPTM86jghw*@V`*1$c=fK;Om;YT?d{ zFRL0Rh2$xAJ%OCwP@XoG;6xSF1BGC&bTVRMmm>y_so?UiL zSZHOFx6lS#_xY~1bdMCV*LV~@y>S#w`+y%20huq$cW7dP7iktyW>9@Q=k#|>3C6&t zTeH)*9rN>!5t;qD!pxsraCf@2)MA8r)CCu3?sbv=wBym1x~jxCP!8HbZaoZl$J-5~ z3H(0B|J?s*!3@~jLf`HZ$_MO5j8y~^Lx7uDkK2V?2tcXcms(dsr z(Cd&*b@QIrsRr2WhbM6%_1i(8`R8x%Y=v!RFfxh*dCBAVBT6tK3|2#zyEC5#S+ zQ#0zixa4_`9wnjGb?5c1*0O~=HSjs<>BrOIy%5z)4 zaz`n8CVZ-Rg_1J+{x=E?cia@fP`BzKQBr_?i6UTAO;SgL4ilg8BNIfon7}Fe^+PkU z7YVxOW-HR^%%qQe*DQP!t5g6ybdtY_7vhP>9bKDmunv(c|1Tu+>8&6ZPI)hvvC~R) z(|MEM+DqtT?gg^ebOySlO%xbtjTd*jO*aJL7_@ag#gJ=b^uA-*y>BGmJ^)8PWU0?l zKVP~cfoFl3rM_X}?Z_^X5AeF3R^+Qryp;^IhvmlOu!HjevF_XkNS#}J#=*14v27kJ ztV9n~Lg}qYKFUk=ZRv;RcM1MB=qXc+D=ZOJiEN7sz*pd^)sg0ThuFW~v_OEqliry{f^_4!Excz9h4Nm^KnSKU@k3HW#G;^)_%SgM)n?R`-g%M-1C_rsSXjOz(De(7W2EA2+yT2-2WraVR$k69eIH=l+ydY zU#+x?N>w6Cfp!2*+Xkl#Gr0U&iyjNB+Me4DLhgCv$o+C$`!FL0Z3H8_3umnf4!nZst{#ds zUr0Cpg0&<3KTcLG7eSuvOvv+z|0e|65gPp;I=9k%~J?bv%4#y8%w|@ zuJ?FnZJEOx>9JXlUXiV>?k^nMfJ_+&lM6C}v81Onzt>R?+TZnf4}K^?oaK69NBq^j zB&Kj)XoG!;o(JG8wAuC71f=NIBNEg!H06(=Yprt20jjBLx4pf+NY~tvxJ=5UdDsi! ztYRJ+8`}g#V2cn4OGik^-_v%z&4YXW3f$S3K;WuI$H1_(wA2VmEZak;)c~|F)`&{2 zb%Q=XuKS19CoSEzB`LLAQ5PBiLzgQcHg-L|LC33cc2@QJ71E{xRdl78Jz%l<(V+t9 zAclW+6qJ;69Q(-{`VM>pw;jqw(v6^gwyV7kELv*monVYpei_)WcViEZ49(Ja?cfa3 z2pmlkbe|ave6EvQ=0OQdbgdKovrS%JUL*MS@aT&)_gribmODmx8tgW#Tl4PX7;O(k zH-f6m3V`rZ%CY|dYRd@Hh;Cn>2e6E;o*+199J(Kar(O*tDU%5A2fv495-^_80pG>Z zUF^c+JabSwVjd+?_KZJj`3v}{M)7FE(&vLQFqxWTUHk;O&~yT^aacHc$awL18d+=S zE`be&$)6NaqoZ6$Y^QJ%zvFlpy1!Oz3Q|XOk8@`fFno&BJ*Z0h10K}X`II4Ap~`As zsab{T=gO4<|77F{yk<1px$~r#j5>V<}=7v332WhzQ&3j8o_4Jl~me8Zaba3D_Rgf@ce;HXV~aZ4zDiiu;vyx-%)IemNBT6B^_Oh)`s)>ffL8&ZxOW8 zun*x;J2&SNIT=e?HC%3ZV6Id!cD%~ry{QVWJ!#(?!#)f>mt8X#qfRZdKD*r{LA zP(3p;?;@CJ0-ZN$&?$jHYe0;&}u^Zix|DRkXp<+_CLzH`$jUu;AqJRje%gP-Em z|8;@*^BCCP5Zj2#KdGetbPTDFyWQo%#+zT>R@DYAQAN!<);iU_g|5v^mFH8TFbdRC z`v49oGorA{#ofoF{R4#C-Me6gDb?Vc_H zD+MaJzxB|d6?_H|9Q&xUcrbF^J24`Xc%0#+sx+;C62t1Cu74HfT9GiOg#O-gIv=dr zzFyU~oD@oxhx;9^Mi4StNhvQEJ@7>VTrq464z&{=vu;TY6Le~KsFNdKqvXUJD>2-g z^2Q-_A-a8mVtT~UGG%HDe`rDO8!<%eO%sdXf0?C{c@8o~21LLdwk?my$nk>Q{X_Gw zslc=$Y^bXl_>+ zjIF=dE1_1d(zM%=kX`faRS79AlR;&%hKf29l&B;Jmr287vV)2&eTk5VFg**mgkB_; zWewYh+|N-ou6QC-A|8@FJ%k@^e|LA@71TN!uY4Sq99CotTT0hDbkkU?xSe|~eod)X z)c;)eP3lYXZxz2@s-Ko=N!9p_zAxrv`Y9J4!MY-$t?6tQWpv8k?mtThLI<5$ccD?J zvB*P7ZJ(tvDYlBHxW`R5pL@siKDdh=|G~fC%|7c(0q{`3OG1{z^7NdwBEx~Njlo;g z$<`imDUkz#!x*Rbd@2T9^zz;Ryx?*9pb<1Vr77aTlU7=gfbPK2dIp2L?lhN3&+Yf# z`0dDja^mJlo^%Q2?*`3k)y_IqJAIiROVtmqrEEkwNRE6dnyRbG{On0>WUt$$20M!b zy~>qZCZ!k>KzM6o z>q`q>E7M!yjpumlZgMgI&@_^&V4ObD?w!%RK>32>(3~&S3;ks!$rS_5*S7%Z&uO&~ z@`cqvjVENXl)A=kC=?&nR<87rdyB}#Po0W#zwML*+<*MD8x)9b@4Pj?hR_2-;M1`| zi%z%$8Csrpea4}BCxEhEraBKL=sB3haNA#skM(}acy2gX1*=Ui1rLG?GC?(ssiO7) z3Ch=Wja({RY6RfEu9dDXMf$1Vl$%sw-d5uNMPK97h$}R}+*R35pyvLjlTer`tL)c2 z`Fis>oRAWC=&oRC9YLhsV6cY0D@s3Cou9wQp*^qqxm~)za^S|1^Q~ju?AI6}oc{== z@uf-saGXRdPH(9scQSq08I$Z`OUf=?QkzNT&fnCl;^Kd&o+thb*~2>NSf@@$-M2+E z%Oq6eEH$TzD835oQL^0VL=}5zlQ*V9kXFZ6pCz4+K)KmS?Ycmx*@&Zbx~7ar0Hk5| zra#Pvma9`nZ#=&Dr)2Ri;_DYDjC?O$C<>UhG^qX^e)1CVf~n50yLPElo#vy2*?bf% zdcw!7)Ebl1II<|Zg_@<>#P4U<#PW$&=+L>M81tNfF)c#KOgC=EINh_BN)`+U#}^_$ z-L>G}hr3_H=&SZg^hu3nOMYD}50AjapYWddYaPKLIP`TPeTz7%)teFNJ@dj;Xj}YH zXSuAr^mr!EajtW5L9x_inwse}+Qm0&Q?{?@9T8E3bxp1uERnVUn6BJi#1*zgN^R9~ zrw~&mxZ{qz5_iS!iqqGOd7zdoxo0PIpT87K)XM*Q9LFSBnYfcEC0)qQ-4rr=wk#PA z1vo-V&&ju1Pk*7qc`=XN8FRh;XHOL;N#M>_73t@P*N}3#HobJ~U5H&~AekICw9kS= z_-crxAUVPC%{rpiLAs;+HIz&FY;MF)f+eL46l*Qh1sIXHUPE8&gngDFP<0vHgiRO< z^%?n|gx6_r{oHF12?;rh7bjgp?`Nt2SY`3!wdx~=lL2S?gvpQCq%>%vqt&-Q#W6PH zoGv?07K3~5%(^=+H$q<}l~wawVt?7Pl?B1V!Yzuq=|&XxSzTmN97ug-P7Nv{dfJ!7 z2@(7*6(OH=&89J8o3CG7wGP!CVENid65;T4*HV-z>3$+TVzKNjBvU$*=bKkt%QnMm zwh&lZM?g-JimXnuYKke=ZgAk8mAwgP>a^p(~{${Auj5S@+kC6%PhwM;qsxq6A_#k58vg7f39EhO$1ezOX$@xMs~U%IO!yn5Ez{qU%&%LFyaZyM zrOQ!Ij@;?BhA9er3X4n>U_D;++<2VCM77HIB+$ z_?o$@gxcEz+QrKwcrAXLQc8Nyd%T5mtPOkfy1fnJx}D0-*;gBJx1{VuK0!PAuj`~HjV)={)t*@l}atbBXF{vy#qtZ&|s=E^Mh`Ory-NgtW zRBqF03;H-SY=nOr7^FjyTBT>8DwEJstj7LuyfHn?&vG<93D6CKZfiogdjG7GR@MCt zGjv0|@38Ev3>DiP^V6O0b*EKHnY4P!W`@0ezrfPXintofn)Ed;N74Ac{PZt{8RLcw za|?5+f7z1YN{WIjsdl7>52cTeA7;o+fz$Z;#V=WoqK%^H_ zp{Tn)kt!8YUJDHoh&r9zT+bRpMUteD`;{%gK1BM(`BedN$#ZP3#kV+kS54-*6GPc5 z!{G50>RrOhM62~DS!u+a95Zp#9D?A5iV|3lt&2N~cRr*Wj_iXb>~bFhpTi%y3ia-1CbfaABN`1dKw~>s$dcI~E&P&MPWA{$Cgj5z*0PkyCI30aF z_LL3Yy0m|eta@y*4n;R1px<^&%fFX={rzo@s!+N_Lr~Y-4*k&~RAJ=7PMyvx?#JhG z`vXrkyRY8zPC2S&SBh!Hhg3C=9w7-};5Q+~FYP132WU-NvDz^YXlUK9-L`NsJ%LFNhK{`O-qfthP#&_^`V03Z6S?(o4uY8q-5eG8!8Q%D^S8Hh1UQU z8kLw^ovUB_lH>pF0Q5-%dc*JyM#ph`C-HSBM*g!cDWePTCoTq8jKJ?&|Cz74K6Qxs zOqRxpUB3J!UDzFh>;4hORlPj>(&YF|?*Xs=t3$aio`46_&@t_p(e(o?nG+q+0 zOQiR3zt{M5$2TJBdb=8W99}@^{gu+-*YDyS6I#zO!!eIdFPA@5c!rBJ+tRbQk(y>V zA-45GK6>UJ=-t`=`|o+$t|lu5NYkeOYHNHdokr-^mwT(0)fe$wV;G#Lw1}I6lXOWQ za}jrow)@u#uftI34^@)RMM9FzHut-yTa2BW3s1OGt0|@DE)q_aJ9P?l_j*7;P!Vhg zBJ`^1CeRU3SK0xWSQRo)qOLnqt*9KSQ)+SO516}LA5qwCKQc%SpOjY~F3ter7!}A6 zVU9mJ>st?pKNtwc0VGJ zEst}41*>4=-^WHI@&`OV@kq;3vykU4X{iCTHU`+!aKxIq6v`>!8I{V`dGw^*>U+%X)ma5W8a$1iiIqPsY z-&QIV7Sh&vq*YWfa@a^&Su!iX52$xm%2WTf9T@D>wZJstGJ{X1wA1D1HTx4k{A-jZ ziUoYw`7SAovY(Xoq04r&0LFHFe7rOjxilE;(1Iyfb3thut^rjb7_}O_sjj68ln?W^T@jG996YWU|>W? zL?pDHb*ywS3JT&lKrAsynm>EI!h2ZvL89DJCyyR&$M!4eA5V2B;m>_q*UCnhxy@Yp z;s@JFm|$w?B{fpS{OO88z~om}5Qa${)9FB}z7vj$7VoVEa!*^JrCQ4}|frbTBAgb~1m5 zYq2nZYKU25?3Yser;A8NGaUDu$W+I>!0KOa`E^v8PTO5X7;XvgO;ANgZ(LS(V(>^3 z=I3;zTMsL#1(@^&w62O|zxW0GES*wr_uY+esC&6HFV(4xK|bM>jG+-7`TpulhPGH6`s}gwGgT$hmBhwXh@(8MgyFgT|?7ImW<1Vq{GHT|+!Ud+4 z(49YI@w%1W3jX^oG9rRZKtNC?%Y=)XUoV=2VWmlWr}UnFGFlQZsO||B@#o&dCrM@X3#%o@u}&#SP!BFC^KY7B>Yk#^6QgQmSMo8fqx(5s&sYf!oj zLnJT3B8Rn(P$=>m64B$Ud@x5G2|@vct!tPG!)k1FwAzb(pPiA=rjs{VZ}7<_b4dX? zmup}-!Nd&E_9JyY$e%$Um-um+;6}k0!&~D+q=;ciCd|tZnxYlT-^n7$U21-Fa)P)S zcH7(&45Fkx8U84KGjkirW|Z|3odvUFn%R{Hqe~2Taqol)_%uo;CFM2t>T#?wo==H% zkMqxE{ryG9)f?uW|2Y?BWuS{Ug5RKmKqkBoRKCyf88 ztFJR%$uW@ZyMa?KmB$H>9Z%||ySXmiK6_~6Hi6>?c_-|zu0&lC(GiGBO0Dt|LZ*pH zAybQlWX%5+dOITWcT{p9+{P87_{qTLu!MySPapGB6xl65gP>l%?=No~gWtoxqE>WJ zoUh?Ij_s)AyHZ6Yu$)TM{gHUQeb7CR;MAa|zq)DoxEY6Vo-Y=u3QYo3bc$m77`%J~ z|E|7Yt*3gr91RC$+2Or9zEBKP;8gz_udV3oy*Lb|_{gl3eeBYy%4ql69WBL6cXj(j zZ6|R1ZS`T(M!HKNo3%fC-~7IkY&m=*1*|zR0`M{YzW-vqF&;&pa_Z%O1|FEWx;!aJSlY`Y z&fS_A8zV9e1#Ek4M}MMcKsC-u-vc!ID4=6b@3J?KZtIg1Q{=QIXSAcKFGZ91+&6+Q z$C6%$E52^XY~x1HHA`WqIfX5~hwmBVyI6CAgP}M^)P*^}rR*ZsBjYC{;}TjDLqA~8 zHEPy&FYlB$`-ntn)(`)~8q+=FySfJM$G%PNvv_jl=7DnSQ%{ z2G$9gTUtvC$}B?{a1;aKJs)rm7>A@jd+fwZ%!X|sTqaPimg*F<{DH1IqFOjmRG^Mk zVvUF~QfG+lb|V|eO8nyUHlA*8T65hV6iSwUITU2jH;JF3hf^(L@DFO|!zJKu=N6w} zVDj1O&HrZZHNE)mM{hK}=`Kgzx8D8&o}tQOn3PzBQpw+>qqbFc-nh$+i4SL_Qh!?Y zDG=GgR2nFcw4aURa|hOM2AMZUgI4K>7OsoAqK{>01uQ#PIZz zC=XCuQHFgX#d}aB=OvM-(Jr2n(x)=EZOB3ImXN$63e-J2g9O2yqMgkc{p(G3Im_@& z`wjSS1e$WyK4<13hUw#3;Lnn!G`PSz4F4A}m_@nRuriMjE3Un6uW?V;`V6YXpP)-c%Osbq-JHgV7OqjRLQPHlrq6SA@M7hb zxR+t@romF#Zl9Rx8m)XDM(CoKE0ZbmJS~5B!QYKL98F~p&~o3Vzurll#;DfGY8!2| zoQ9Mh= zE0oH=A^SM~oIP0-66?Y3Xx9{r!&p_6M(lggC&pv%3Fkb~W5nV_Siv-DXcY1He?c@`6OJEx>nn+u(rNtL$zrgzTd zZC-`k@#BWXE8-s*U7Rad6L}2DN$Oe60;FqhFiccvA+AB3pKkT2)uD|*w+xU6ES0kK{>q39@ z@S>m-K_xXg*OjM#uo!K;ZWxV|B zP+M0DbTXl;l3*s79ady@kfIOxG~};9rWK|zD*A~hqpoL zMl`B)l1s`a2n2o@#)XG<*A<(Q9(S6}g< z<;zgV^+Jie6^6wU^rgNKg9_O)=??Z6gJzr?@_IT1O9du|MFD&vLvrB~t6!6z*rtj> zVRMZi=(W8cH`vu~l$|#3glog)s9{vb1Fc*Yzjz$%EJ*!5dyy%-tzw8KWdHd_lOEkv zcr*J!kX^E)k}7V42m-;ZAL zI6RXD?q|xQJ&lQvl~I0IL(4&j@I>D&<*aBOCLYtq{asCXAAbfJ10$hoI5$PS=m(eJ zjIX|!m>?(9**7zjA4xhO^?skK(4yQ!9~}NVwwlUqGt<>8`|n-J-DNMDpk%w}jTSM( zY=Ai5R{Z-<1Kh?ZV}MvKrzlBkfXQzRW4E->$$Md^$l6Xq&tEdKWxVO0b)>%-@dG* zhF>-2V7`KOtw%@BHKuSG~2I(}C9TWsH~G-VH~yQ_N2#$d6^d(-qqs6!Z>XCV|Hp z<3SVH`xzaP8PtjQ%A~+|5zPKBuHL_AGW4%ZXbiFG^g>fPY`dzZJP6M`&=&cmSq~-* z&*9P!O6yDyMG`Q*M9VnJDSe9>gLLZZOkFbNGO`l*yhC!IvM2g{ z=fVGX(GRw4O9Q5(!|x}_S*zi66hjEg-m_sAPlI7P`F|(Szl!}NUk<#FV)|vj|u4#Pp8^xtmi^I ztRFTrD&8Ov!V`!lT|`q3VMK>HXapq$c3o$mh?}b5h-y&g%@o#4&KZ9C!+-4OD_BG3 zK(b%LaU??ipE=zR-;2_|E*Wk zsmZCRf~Q5kuI>c4!c0PaPfoU*Ic_WLD*eI_)cd^G$ zCgm;odjF9Z{e%gG`trW`05TTq+79l`sW zP*hG}SSqBNRfcw$4rAiT4YbC%3p=-7HesRkOAG&NBC}r;;UoVUZq900^WHuoEn zc(%LUrBLsxVtb;0|G^+3z1Ub2g;{Gf3=@c=m;B-9aY)@jf~6ic%rQbUGAtm!HJiH~ z-8oV)m#jyOF4F#@Hb_>t_G#vl z%Y6pyNarkCI1d?-a$e=saJh;6RF#j+OMlR-*5+NJ+UsT}>ZWV`A^PLJ`>Uop^NRH6 zZa5TOClIq*)m42n>2?c%%4|hsfbUHo%X2|8o5^%5(pK>bX|nz<_wsYelvjP(X1D0R zlqe%R@8I$(99P+8mCM7_dU#naLNuP$+^z$d-%}`YI@HqXwd z>nGjo*RVYn#m0ml*4sZz^dr*ZPiYIs82EC1DU_FwSc2uRZ&p9aazKL*boToB<>BBk zwA1%ioONeP3C_zv`dh@NQdxyK(q!WF1QmLnroc}oXeEWDM9H%H1p)?Wb6Yb>4I!KH9K24;j?pPx*I zr`UXS4rgE_zj#2hU-@^)&t2Gr7{q*8^O7jWnMmw&Z!fJ*tY}Fps3XvgLPAD} zJh|0dBYudFE|(prG&scNec`IHjxFf0i_GJGm%0O|hDxiZHQ&#*S^EAmV%PsYgJ^24 zt>U_(hyXHl<`4&l?Fc5}nexlcl`i)2?j3_AjL3 z{!%b`Of%uwpiu(1gG(dafx;Ce3(6G6nOe~1rkQUt!7r7Jfl^O>+*{ZMS}dx!v$)Uw zYQ(1K@v0(@!XsFdjvG*KHMsMk9>RQ4@F% z5Tk75_Z0pM@HY?AF7Gp2{_#cwSsfSSWntVgB4Q1xsoe8W*G?+i+_lE2dbQ|13cSc4 z&DR+(OfS=-f{DAEF~<9~>P+XV6@Awy75U|XDy7(_x5IISG|xL#%XmXQgA!8r%F}%% zl5-zB?q!CViN6Iw&RUe=tSbj^nz;6f_CBJ{xrsXBVjhCxmC^>_UKgjBA zMYSYL7}emTo=(gpQ-q3bYEpDfVr~)Ani58fj%&$6I@19y3vTX~ui>Jn`i#kk zATAQKFdv5Qbeyhn!~b)ah5=N^yXm@g^&>PY!u zg7}bXn}KF&#zJb2wRFAE`1$Q+*!Yd3RoPe3k|Rk9*D#lu1=4tP7^jqNR%{ZzLv(HA zMw(#)WDBgP)7i6m$u%&Vr~mG&$sjRLth&H$QSu=PGqKQiqT<=I8Ij^A$&xVPzc0s) zyoGFK%ag|qb*D!-hXP7>d9l&TW@4b=WG5BD{r>x2J4^E6%DeI+Gqd4X-(4z3E?wI* zuf(>y)i=|Hne9fpw;u?hLnNKcZ&(|T4%V^Fl;c!BN4=x28an?0i@lXeSe-qK&xws@ zG}o)vi|t!+!?aGlR?R&)MK*x-#T=uag;l~$UW4ooCJ=_2jRQ2lDg4DLyapaNuF(t>!tfqWGLXcy5#w<0%IB1Y9qqIOv%(XdStzUk%A`_Ws zLxiKW8h+6LYWd@Wl}=P?*|$6RW~m?Xkvph)*+kU+9ur2wO~|sdgRGyMGdTu

O0~3ah_`n!mL{h&-7YVW7HspDuM?SR$0fE&i9p-Vnij#l`_$EuU$95@^ou^g zjMg^;?Q110cTpc|)nd8jy{B|vNBg(+Y+s_jE1n4TmptzRADg znZ9B3rOyEAIQo|oMc_9@gqU~c_!HAb2)`QDQW|%>tf-CP&X}8q>as=L^`$E4?bHkg ztD+v1gKE^R^}SEV-5cs~zufkW>FF_9Zrn|lL@zz^xzUAqN|_Ceh@FST zAuVzd*Hfg#O?|yWyYgv`UtNIxU#?o+ZvicnJj+eXfDYFa?c_{J9TUqEWxsjFOZNRK z(K#fV5jp4tjX>`!j8M|Vlw%L2Vi6zeq5A;Jyjks&enl=;v^+T8P+e%|n(=#c#)COTF*0OO$?xwGH9``2-MG;I&{0rpl zDuI|bm##uy9!9#w${w^4?4^m{P(^#1r#NEf7nAr4Bk|=hk&9rKKT`Hc2~k*P`AgoM zOHgg8-pWb8gHIA%d+NhcmqGfxj^MZd;$?z~^QD=LXL z990`9*QnlToT^Nn4-pSiI}Ya5wK$ewZpQN-56>naYQ$gth;mqp8U>qJvXd4iml`)k z2BS1|wOhte5aXDymq*RA21gsSyu@x6ow(@lEo6@3WVM`X64-e=i+@&X@Kk1s-9*Yc zXP%mg4uA^18xCL*&7PMES4j!&0FQ|PdAFP9j-Vi)VM+PNvZ$( z*tr`I!X#YkP)?WH?Z5YO_FVu;PX~+@aW|~3l+xQelNa2%Bn03$5Yjtpvz#GxTu&^x z`j-w`RA<8zBMBV!`bjm1H>XUekIC$x$KMBSAeF4Hk`)=#r{=xvMBt!@7$(iY`Ui*~!I}}`j;^(il~TUmy+Sr+gPyFm{L={43dUxSebJHgo(_b91i<7F551LLKVxeaHE7L zOTZE`uRH0$4C)`^{6K-WvFC0Jw%IxUZ>bQ!hOR>h5NYw2QMZ+wIq`z#hgz!_3B!KW z>dtWD?q1HxnyJq)5ZveU;Lp@SMxe^tZAP&TxC0hsoBh~A`(iGMtIKh9DwrFi&U zulkTF)!Vz_XLC#BljEJ6huJXi{gAi)Ju`FKi>`~5xy+dZHcNK@aM5xXp7W(hH~LPXYP2h&B4J3|KnBx>IQ z<;ipn1d2CJNX);4fS0Je#fpdJySZr?Y9N%V$Clqnt>d4GNbs&zVuiOglxn@!qE^WQ zg}9u|mP_{K0uVHkUxU^$kO@IezUnhdQtf-?-gdAiusQn?EUlQk*gK0MzKIZ0=el`= zwp>o3DQtR0lp=w`Rt-v`hxk}K2Vj>mW_hZft5(cy-3EA-dkd%KP88-%;)*QRR$5pM5HjkZI zKo%O@Z~#}O2CNpPu?9&LUScOQ4a&wU552POANzOp(t^PjJ-TIXcYRX}prY9jl#WVu;fG7006e zQ?lV88ceDLg+SQAF~=^=t%|D2VN+_EU;9{nQbv@YyptG8#ndfu7{%YBML%Ch<43}a zLtee~qw%r;FL;^wkoRSr5M@y=A;h2up*e2xaaL!wVKzpm5{68e_A3k4AnV1NMm*rAXzs zngi3Y<(f1lphDU$D`UENfe?n@SCB0~Od6i6Dh)VxFuRFNUL+Xt;DZ|k)9vquzG{VpO2j0-Bcp_PG_ib_VGv~yScJ|QRJjRr7-f1!Mp5`#QaKy6CDqM;ct2je*`nd3qv7N({@kMI>#gLB zLf|yotUN5|m&E#xe2-!%mY&ilpHz1dfE!ho32a$IaEsjZ=Cs6$=>-51JoFnO{qcUM zA}|rZdEvs(vEmtjt4Nj$xA>(xQR3Obg8oxUe#;iOjy_O(yIpUVJs-Y`@3FhBB$fUQ z*e8>=v{4ZRm0B|qdy*ZjW`59btTMwk9RT%B{g{M`Q?L_0s6w|{8SD-kA&PaXq~M7i zhx#}uYw*8TDZ00z&i>snhb<^)$LxifQ~U#aj)>+FgE8}y2}Gaqm7IkXeIQK@uqTc- z66qGB36|UU;ef_P&rI2=%mDca)gPj$6oio z)AWM}b!U{3Ya-kTh)Bw0tmAJ}kB1*_%5uYrTDE*9*Pk$Ew!7L02f3@i<8v1mxG`^g z^dlWMe;F-x8*fG4I%fYiRfu$mcG;BdXPr0$&a1!fIm7KA_Y+f$#J}selxD%;#%mn)*)7GLl=kc zq8(S9iMg{}wMBHz{3S4{S0dZSduRPKCD{tyq~9nTZ93M)NNyGaxxG(`V%}%wQG|H@ zU?yBV3NNsks_^OIs9L%U50+`CT;0V0DdUSh$RB6fJK-ad1e{Ms#rify-Vf!HKSZF@ zNXrzG21u|gM`{K0xZ-#+8Ym?fegU9A zTxuowNtC-2gX-lGu`#&=4=8;TKnhwOb_Imos-jE8ntcCiXJH84KJD!q##%11TBI-H zGHI9COj`Z7KkP*y`LSf@6V3`Q3~fxUi_kZl&@)tSZS{XX{G0V81{C)o0GV_m1_JPb zlq{w<3*+$(^&|=YWM1`EWB-uvas0r-MMtqzq*;@}o3rE-8N8BLhDQsT$KSHUlGy2q z%ZjJXMrn7C_qwMKnG}_g10h8!#qx&(Lpi2&%vA3G-N&;f2}m{;dgEPpe+?5J{bD0h zCqBfcyiHBX(QTyh7WwVi92KRY0zL!GIrShVTmG-(^^4F|4QtpNsRI;zC zm=PYtpEEcUS`ukIp9Ek(U>N4_XfEH$%&IC>>X1{6SxF%m|G~#F&HuYddh_xq&nTC|9&9BSSiT#7P%JC8 znQlb1DgV0q%OPfvJ*)~c-hEbJ_W>)MEuSG4(Xb!lqM{`TcVKP9p*!h<76a(Gze9FD zm+l2B^V~k-Bk{}kDdaaaFo+E74!x&B>?$ z?DD0W?f^}3oO5z5toV_SJ6zl^c@{&DWhgeFpl)&;;C)5g?Z8fEaLWuQ3_H9U^3coRv?&t^?UvTxKnlt$tXZqnbT=nSUh^0<6CG=Z!hoOa+iw5UAv)B zHyD+4gEE0EV;%q2k@>G|>p@+p?ji@!Q!6P^iyQ>~#r=P;*qHU7q~D(WEc`; zw-pvQlaX;qm%XVQyNcjA$z99;nYw>|Og#_ z$%+N^Qi)aa;AYPH&hSj$0t+l%waXzijYmxMxC$0FO2Ygdh!&j3++*Bu2Mz|q@$I5I zqXP2r;wS(8;6P0_aK0?Cg$ohJgL#=)lOE6KuOC1^+N=lr%xYYRKAyuNU z_aY&dO!nx}e!jPqKwy`Ha)_M)X06r5%doF|#;r*;~Z z1Wjq*M5Bqc#cc%}c{&?MZs4r7{tAW|?I}WYlj0V7BzzB109YFq3cW;yUhHm#vFzSh z8JM|;-Zz4)*4kM1FRviXwuzjIDi!YAEqVY|U}PEf`WnuCzegO(VDJU>h;Sk#fq9TjhaguztFx(sIg*5PCFS}-B_?8(g@OgbU=Xs zJi;9~pzXS&P^3DgO}d8CoKbI09)Ur9%)32eT>A2^fz%1sG&v)9_#>e9aRU7@SkrFh!XL{-$PyL$z|niB3oEo1~K$1+#e_mqMf)o~m=xH|!(6 zoQc^>I~L;rPcs3j){2D;}> zG?y1D1#^88crzs#jQ0*|fh#`~vVlk63FWqd^cUIu8v~(cNs1j!&YwdR#uZraIKud) zp^XdjqDZFHtdTTzD)#%lG|Z5Y>R7U{ZfjrSeoL8QT!vSir_Eta`t}-}MBNDQH5mMX zY-__XO$q-*(uzv`-oW#KdXBW0z})fT5GdQApyehwR_MkNs?15+JXqM+CK@gX&OfBp zMmQh+C#Dm(*koPrBu3iD6+{w~+^slG)lZbv`)*HiwRWZ%bw8l6gi9(aA~AZ7Vff}7 z2Re$Q*xto-N6eadYIX{}pu($YxWi!?GO_|jrz6GNP_g-@BT6@CD+&ut0S!S_*v+lc zXi_b28}XJHM6}T=qo9`MEqw*39 z1*86^kMJl=$gRrx^Jm&AR|`82;oBl7HMIQg3QAnWJy9t}VPwSfY?x}J|IC@bOU`~@ z(x~9PU-GZ{Ka$QeEULC`qoQ<3NlABuNVjx%NlSNkOLy0UGz{HCNlSM(2nY<)DLue^ zoA>)a2h3&%_r9<5T+9AX|NhT3~CsHP)bPTpcu z4vCL`#ybAJBJ)*j#G&u=9_xuywTAl)>AoAHhqA+n8N}(awv`tN=pu!KOir1U6S#CXDKFVb%!k~X)WUi%OifCs?JE?MCiBzWjtMX#ubU2Jzl`^XQ`$hHk zMGFEC^SIU~%30d`pZz7-E~$YrSgyfd&$n5Mop`VH)xToy1S;%?G&S{@Sn(|6v(fbb z!M!F zlW`{wo^Hnj2}CnrG~#yAQcbtxCYL~pYO&AIL*Fh<>M=&#ag_gIKR&lL#CoA=)*IKd zUCE4csTvFqs7W>8&zO1%a*|V2)I64_$vbtAaC>t({cUdbUH+CI_r&{4!K7uM@Ia%y%HE(J9_fAo<@c`{WedkMYTv zMav2qlX=v*_#c?2b6`mesI5A#3JP)=D*1lu>bRz~8c&curcsq zXc&=o&`le~eu!8bYl?Cd_~77;XtzLcIJQVls^P-}dWlSuM$AQ~U$+)a$BCGcdreAi zIo2p+;T~D{s@KZZKi7~e&hPkh#qVEGxTV*S+N@E#qS5W(4Z_*24^N~25!EV(Q{%V1 zj)VSa^}-0Ju)$UIw^_2hvl9FA^j5fBVy0hqE75GldPT&-GT>6`BYW`YLDGlA4ddU?P&t>C;PKYY*kE1-LvS}yEIvNS%!)C}6YUx)0)Zz-HqR66c~ za)kH}cn-g|*6eouvr{?%o} zV$PuV-}ASYwI~+}HIqgQg_BuU@NWK{#k|w>NVH)riD{?5H%7%~Q_qh*;ckaWiMVBl zEuF`Wt9Hfpjv30K?YSv4x2VgmY+k+rZ}~RWLjuclv}AWli1Y5PRW%`d3KO#;cWCn% z95%gBDov3yN5K(lv$Ka3zQo~c-3X@(8fIP!FK?N$;dHQ1PBW_J8Kgbv+ahkfa{To; z*-3zP?H}wlvY2#ioMcUDo~Qar**uv)O+rja6R8CDmFo_3%2}U-j3^}-H3Km z%kZq>a!`U?l@477i`7d4(H{Kn)F14YO^4lXm=}{ryXXbp2GWGG-QOz6DK*(8%v_nM z)6Z-o(Qbc_a=4&Csmty_KhR|zk^4&N5&^=IXH|^iI`M;BJpy4f5n&;N{-4Zq>;k9&#`v*m^Cx?SP^A!O1si0xwi)>AI=;7XtxK%hW%_? zPUtq+x{O1_na9q`13m|&3mapIzmtIExzQ^)f2;48zuTWU1l#q0^bi$i8aKo{;g7ay zpmLGxopZQ?=byi9JEAm6MSbJWNgbL!p#1rdhp_(B?j}#prw^hiYJMfIsb6VF8aW#+ zwu*9@Vldze4{-=H*ce@gLE_DQQnttL6qU7i4DEWhn<^-|PMMg>a-qyJV5C{?ZF%^8j5 zXjqla_w)5>8Rs^IXse~cS574ZWUNgTb&Iumm4sA0$!6JrnnIbs3HK8GP<8|M+vXXS zpB_uLUw1Giuh6V>9&P&HfqkI!ahtwkr{Nge!28v$ud={L*041%aM`L6Y$`gGoYQf@l!zwSp(&uYVpGnqA}korN0?z{0Z3l-({m>2+}26S|Gg}&1g&1UXc{$@SxBRO@IWj6x{l5(b9twCM* zRcxd-|LMQH2G5q0sNcIkx4Hu9iIz7v4_KL*HU0)%MxPk;=xvCapk=nrsdc|^g-$fi zNu|7xZSsaNF%qcqu+)ldp{~izpDOEXROjye>4u)AzH)%H^f=NSqY~`&&`&A+ULFM_ zzf>=EhM1QQ-%BPP*XAN$l^fF2P}3T|l?Z{@?+hd$uX6d-kQh0@&a96-9LmGIYg+iT zA+JDDz|mb5$xBWm^s%!HI?=K zwNj1d>*V=WF${mmzZwZ7bSQoFoFw{A8~v{_os;w~^j*ka9vp9#`P9S&`h&xo4A`ns z_=-%uIIq*IGdvDmglRc!wfOK5g{mT%YLZ;oFua6C6%<3m&8_9J4@z&W$}TjI7Hm<; z2zV?%`Fh_xuJ2mi#bu#qnsV|4B_16y*!Hdqk_Clnw#e9X&Znj|m(*jyNk{c%be}_}%YxERCg!_OQbOR6kBbI>8h_7k;;RF~p8Kdo&Rg}L8HZlM zS+9utQ*K7(l|6 zzmUM7TQE)jZ_Myqz+PX;$D`RRVAJ{*(7E$H0AnXw>SE{RUO)vm0^plU0N-royHgsJl7v7!>g}8i8xg2uA6uw}ZsdhX& zIZNq?r^iF2_Z&tYY8qMTi(1l%`?tw6+LL1oaO^3_{D!{(tT?|e z03;ks@`(ILFss@*K-S>p=Z>&C9o*|VF6#^4U?%&z9d@6dn0nIWJUU^-+7PMSI-r~& zOeIE{j@m3DRw!0kds~E0Yf8GEy^V}xl^;Ph(~SS;GJ{~P6!Kl5&SOFaJ>2CAvGKbm}GkaaV2^YYFZTH66OW*e}r2YFS~Ef*IQ z6g7)o>LO$7gq+&{QCtli*e+Kq1_9dQwmdV6|Ni@r;kE7l@j}nJT5y+c#*)B4*J^anK02=m3{!8PXPgmpVNmgG*?Er8q1d!yZJ?H>0 zUnxMNXz@50$;Uv)&?EizLzU%y09I;&qkT6)(En=lm4`~F_#Xv(=S9fT!X)&S6m>)T z1ejLaS*cTev z=LCYW3?JjM2G-jx4y99`u9D-yJK1khUBLx()XWl_de;p&drLG^jyHl#953T#=c6wU zw?xo(r5_6FP_?<2yS6K9DL))5X-od-Ez8unrZ+Z$8(GGG>qijyN%;W1rRhx&0JDE= zC?RVryiU%tBv@0e@=&Q*-Dx6V7UoyqQCYDIT(T3@4Dnq8)X!v_^hnz5SfViW)kcU; z#p>E-SE@X7`f-k<(=qGw`9~^P_iZGJuAl0|Xpo7ka*) zPZ(~SzA*z}qWod+NB=w(82gz{9Va{$K_kR~qf_#2A8CUqAd`GsvjuF* zvICIc78WD_b~Pz6ik&0*>Ri?wl=U8qaeSbnlW?-u%)IP|ECo-8y$SH!)TGMG6tmRJ zr_s>R+;S2nM|mB8ms9;ZDFeJfBY-Crgi7%`d=v0IIq)XOegaxqRU>V!ScFv?WeKlo zgCoCLu!X`ZkOWvO9ZjX*=l{Zb;P+7b+3W@|#8yEiM=Pp_{>0&o{LfdZQ=u-k!Nr0( z`vQ3UAXt!zB#y>kcZtE@3*8q#PAfJB^^8aIq-k*)y*?2+`NT}pL8tY0Ad@?jwrU!9 zSrOxB!vy7=4Py6{%hT!1xNEM5z0^C%ulIh=B}R=IpS^b`uZF6~1%X$HjC>lW$&3k* ztNL!=HjZZ1$+OYpZ*U(25nI=_dURBjYO!tU8^aBBS_R9sLs{i`%UqC6{%nH&R}a?j z<8uFAGWp5QU1uHCTMY?r6t$&G&8=3gzWArF$WiSBnqv9!?bDO-c;ytI#1{&grF=(* zKPb@5(`3=orxUDQZ}4!ZZv}kBT99H!H?)=7*2sW{Br!95dOE%OTG3(_c&cAC%E#G7 zeSaNHCS-vDAvvw0_iy=c-wJTJqu`37wcU#W=Hw%J)DWN~{=RZ5v^^zwu6e8i(omTM z!uimIeIJtGL3iPhUBEdr&Y>{E^6M3T7d8W^cR*6H_)J$(U5S-iikJU3fx=dmbr0UA zLWM3>g(SBRd$JN$%hU&R;jj3O*uXebXNi1id(|r6!=9ih7MT;T85R8e8W8=cs`~&y zITq}yJ*m9%<1G$>U!c=S4eY{bK|_$iyjXBnt4mo#c3+~AdH4LH^>7dGZWP(5ax6ZN z_iw#+_HRz=V_YzxNaN|P&YM_;GOBibw-YoCuz*V=#0mJV>I0~a%`Lo;7sZSr#-Eit z0fb0 zxR3QgmT`F(OKJt$4KDY4)X~eFTIP>HqcmZ^9_-!*F+!;5?T;{;(d8td z({8x2zhGs~&@ozzP*#KT(2Gf0@(vboMD%_=9zPsmE=0yG`tPxBdp}lvzINn)(Bfwm3!>P|V>|6TSn^Le+khuW9&~-&CLSeUBK2 zp~Xc-X~@8o?3NO-f$th{&7Mu-LV3ZU=$aekZ6=OT{6&m6^#h`uD2(J|e90v=yzH?{h8 zG_ZfM`c+gYTnrxDP^TWI%NBpaH_bQPiT6_bE3en;6UDa(YoIc0`=yV=GZLSdgkE^Y z#mj6-HgxY8SShq2lA&AwT0{@#J-2u>C^83)KBa7rM-r(!C*j1K@OM$n-TwHitpJ)q zDF0q)+wTQx;4@4~%cEwBNjAyRA#z`PSF#axO#5l2dBYVNy>T^obWU<75h@PJgG(J0 z)rSlxZly?cv{Vffx7btfI<-qAus$13#nIxm<7jy#_u!Od*k(Jav$Klb)#J zP`z6`3S505BHHsroq-TcUU<9PAshZ<8+>2WLnBD7kkJK8q{)#@V#~an;B5R+{v4x% zKR&f}%B}iN;;qH6=l)kdTf#Ee$)e_3{5gw$pdLHGB^tR(B#_%$MnX5w_Kn7PotAru`d!$(Clj|Fmx$7XUYu;|$d!U@h*|;X6-Z z@zdI%i^Y63W6B=c28i>;1MPwrmAF|JYn|l^H8`imC(Li>j{<1NBE9f~i@JU(j=v&5 znPQa8b^WriztGejmD2?7Mw0o;q5vDpQvcx;4n2NzsF&7E5yOCLtvlAsQuMv5eG`{L zk9kc;ZEe0${89=nAGYj$r0Uyn=qck1y8T~!Xjw=a@=?~GW*Z9%e~YVvcm$AKG{~&1 zn(l|dZ-D~5HWUd?7zhy6wMGWx8QCIAM`>A#b^{*V8G!52>jFm zY)NrXZ3RSA=;QKV9_G~mzC$EUHVC8JebEigwcMaJ%bmn67I+=A+E|B-L?Fx+#CDbU z)L8>8opi|0i$R|;Pq59>M>!#~x`D=^Z;qEL&#tFpku$a?OdM&Y5>xJ7;8p-{M>*U| zmJaZCl77@2%Z~p;JrDJwhX<~-R!Msx{tlqgD%%|%Q3t$-f*o8-gHAY&B zWk;(Zj{L|Z8Hz>{c)Fd~=^NUq{4M|jYoS#1Hu|5hhRrZ;)RSpY1)xf(-N)KI5w#@e zLdW8tCT#k8=U~K#d9=Efqkm#15(-BX{R5)*%{VyZ{_No?q&D<@f{j8no&)vlG{J2b z#9HV9lKL4?#&Lf#TMFG3-$dl(SrU62D&4)m5<)n!0YpV@$pwSXiB??R{}JO&Lw6tj z{Tb#U7u||x8RA5bnUnPE%4E*=Kx+=PEPMHHT3@`>s&t$u~y3?x{bJZNL&~?0Hrg@=|=OTYvor%f=6SNK*8_i-CX) zTt;+*V^ja~lMuKbVwx$Fs*u|v2I9+~k7=0ajrv@Nw~XZ$rAcY4E1hcKkoS_=4M5do zm%}@nNaLUbNuui+%UPDY0b5@ON zPm;TaD%F}=^sy81G6qRl^xj`+)`!-^*N=6c>M!7dpPFs{>UNOS)mbj6BDB7J)R@~f zS&MTmTfVF%Dx`}EpUineo-s^;e2g`#VbA5GOWxW0(5UOwlHMDRS5&?7?E-o{;t+Tt za9h&&MjOQ{KdAN=Z`1eKiW-#V^Xo?;^-yhC5nJIw=gSgExjIJ^fog5?O|&kek%}vI zXlds!a33psebhI%x0Wne7fZXS?GWc;>9d|zgaGzN%KtfWIk8U>6L3WW5N%gt|4VMj z8}~InVBl-yoG|4$E;gy(AIo-RAWJc$Hs8(39Gn*2NM?^9bXwT;w-LT(VkFN9lH%^x z$gjTFyB*A#f*sA5cW@~qr#p!TsBy!Bd4wNFd@>DLZK?=lck;^y?^zr6N6L5oN>Ew( zFEh42tG#8y^Z(*Y2k>^4!Pv&1x|1Zwvr;?>vcTGkz@}jq{{J2TeLUM~K|-HEMeyNm z`SN=^YbMtY;ImD3BK@Cw`ziSGW_4lFW@>^Weaj~#J=d!~%Wy#d!F z8n;hfhCKduu;gmhsv>k>?=bkZyHlNM;&EA<;vB)GpR8$9_4iOoK7u92VSOgtJ(TJyK0fNYI zLrlw$SiX`%9v?_^27gwaE;pi0@O|ywHNvWo|FlA(er7wfEmw;t2;~35lDZq(^-e9uYvX}LW+COnafJm! zeF#UfNjMzK77>p5zaonBy|2{366)@;n|rz!?@@5LU^qrZWT`i~@SwrMYP1l;ZEt1i zVQ1fX$^;J3hA=$Si_)YYL7p5TgHNYu7KbnC3D9_^#Z>L#5J)2JJjsLgyCGvH!zog# z(eZuxoshc4-G)^$S2ug|3aos(ecBbI?{chq@{@&U|5(OQMKFyy<(DvoWk)D+`ps$` zBT{HNPp_pu75JB%X;x7~@G5!7mo?(^JN^HAjufKcQ{728OIC6dVM)26{U~BTMriYU zIxx9ZVC;9K)p&+6@5s||pnH@HQ_Sm&_-O~rc7lJ)t5|C&J6x$D2DaNgq2m*G8v?JtSO4E zUi!De&sq>;v`W$p+iEJ^tTP;&j@fFW8XqXgs^>=va>KVf)_{TJ29XBOlfsv{RZe)8e%lS z_?;;*srA=BoQOY*>I%!X$gWQ{1CT(M>VDApT26??fox%{ew$Q)7hSpZEax2SVUbiI zhz#0%BIsoNP5@Ff_$Nk>2sQS4kk`a!=@7lbjdUm^=h)~x4QO_Bj zeVP>g(k|qxkaaYoYme<3h=p$DsBWdF-c)>SY2=B=3?w7ud|l(>H} ztay}9FRzL+e3|Nk3Zc|#sqsu2Rr&8#mF2RPNLt{h&Bl>jKYv%@p2 zN9qt#%1dEpgyswNoLW5~E83$VbHj7Ls|bm570E{CNwzJTk8(vM`I@srrb0;-A7!rE z*DW44H<19*)uG6XhET0lC@G)%IV=_^(yI9hU(8wr{R^EIw|DbDK#sYv(69G$K8b=f zZlFucw%p+3$YL*G3SE4c|8CjCVePCOW=a0%xn;`n!*v%;jjO+0H>4wuG)m~r)VnJ} zE*js@KMP&3^80VGFNcfI?MHXKXA&vc^E-NZ%|Ub{g%~}u3!C#LEENJ7AJS=Zhqu_d zfB0pfc~f!7Z|YuEnEk=apqV>U%^*o^#T>F*aw<>ZW!A2tlN%3AALuzjnW}i9{QMxv zzM8A0<`}oi(#(0PE!cxLEbn~USj^(gpKX(=TS;mAZj&Z;Z#chBF$I!(OFo~h638Ch zTZq%_RB~eY67E}lT8{(_d!C+qfowr4b#|ZzOvZk-u4uezLHI6XS@@%t-`JFB>)2J| zi&p+VIfNZ(A2P*HUv`hG00OXA#-3bMln!8!uhyQ7?y(o3jcBR#5Pt_gH>QbI6&{cyU&Oo`1eR+5i zIv5l;xCi$31Eo+!_XH201X zn&n{{V33=rNiGIzm#5kz1e53Fn`MzG7(bTX+(lTl+b#^h$=n_E`CD12V07Nxcg*+K zXV2l{**w2(T?EnFdx5dchN=341AqTW*LWJ?w7iO^PzE6peM`0Kp~uY!l)CLK3t*k4 zl=t$i#qV8-?z&jcV&nUZu0`Wne$D-yRf67tHCRk6rO7mpI(Qx*hYgq7cci5Wb&lx< z)ahv+4R#`_($~7 zAMZXgPC^X}mKfOqP-}WXte1Q&r31(p&ClVw?~C}sZyC#~wLL%)WHB*u@1XCA zJ}~28!_!Y{f80;LOyP!>xIOh<@GtS}D)ndm1rk+F+&jckk35&2k(-f*J|ba2>^$q*4; z?Oiv78;!=~jFH{P`e!GBh`h)~8Vg*Dy$|$)$|fg|f@tpvw1%c4znfsCLKY0X?M83{ z=_k?pPKkS4nD^J_A%t2+GWEmjCiC6AewuCXFc9sKy!y48E-zH+X0JuskT85$>I4WQ z9Mif^OGj_^@J9w9woK*IVXte;FDsm-*H@kurvYUpEnZ_Wiwq;)Jm&FDs6O|3Vc5<^ z*>H^U_n_dTIC_%A)Mn!v=cvoSq51$FR*BRww2&*9^c|X#7}Et|41s-Jt>p*waC%Nw zx%UoFQdC~-8YMA{W`{&LL0r;vI4LD+ciVF=C#*XJe;)33_F4a>6Jn0k%gt|O;pBaD zyC0M>^lId?pViw?S-m*7;UFWS8%Eru%2#ehx@6;T19<8y>(3;Ip7_z1RCo~cYGCv1 zoJoJpu(a7x*@PkMV{ub|bB54_qDn^EPcyQI;X?m#+p(g)tIz6YB~S~@(K!`z37_!> z)^YRc6ml(IcGh3Tw~lFXY4aX(8`y5&VxOLGPv=aomFDW#eFifY#G0N5Zq2~2-ISv7 zA9WcDiFQx{Nw1w~Z#ke2A^&O3so^M4UDGtdd`5uuSRp?fDF<(qAtEJQ($_7|M_SFx zk=r}chUqI85cW$Ug>}+r*+eHqd8(V|(B#zfR@#YPA=<2Yb?Pi*huz;115HVgy-f?V zMu_sByZ-Y|Qu~8wG;C2M+o;w*O0)e*EbBVHYfU}$8ouoW)Wvnt`$zusQ+Mb1_Y5hZ z$o@(fW_qCQIf9c;$|TVx$_Sg%pvCmC``!8RZCjB6zbGE9O3WtvsEbUAOCrjg=gK1Z z=I-$5f%Tt=7VjO}MJm&cJ}{Zj1xYt5r+jcz?ZFHqC7qPK}w z1OMB)BQ+Ib<zYM^^jqXBsQysWa~%Hw)9~rl z?%@MD8q>Onc1_EE8AHg(9Lp4?a7#cKd3?vQW7RlnNY5i%umZpMU*H3L#a9k5_Aqe5 zG{VC~7N^!_xhyQ3^ATEgWhTu#79!3juwhc! z0GSr``QLpb77nCV;3#t}X1TVTKgt0+$mwtDmOugT^L;S3FQ>q_(B*~C-}hLAysO;1 zFO)0hOAo4SRq6eIz3aD_dfrlyc5MgV-RUd%a0zi3wmq1t1P&kZjapv~1{&OCfEg2r zLeNx|+r##SZwM};>7`cPhQ*ndlZG-vcV0RL7(_=|=kzf+joW~tQq`bUE z<@IKx9m}3#%Sl=aRVHK4eDg(HI(B~{9br+Xz@8?hMz2!?gSziz zff{w1Xs$=dZrK2gk|}in7NJfxqN1yfqrwf3TA@iVvp*GrrM`N$J2fVo3}oFZd3=^* z&(lltMJjDmBeiR426)vjj+6xDoxL~rV8AmqIgq!DrUQR6;}or19RqDHb;o=!bGrO} z;p_fOaG7lx0Uj60M`ze4T>zu6W!AOBxJ{aHQELawA5byF1BSYY5;oM#oJyWA&_WQc zWF<5@xd9mVm5A9L`Yi4~yx@0GvvL|nU#2p;s8dqvR}QT#dLKJ>fz}I5%8Xl`8V3Sl zbRxRgus0R{hg~=eKXW_zgJMZym6CrDjk-EC314`!P-wPqa;QB`r&KZx&Q;)6JCp~P zGxi|_zX1KO3h57AI-k-)*mA{sQ*v`@|*A&N^&*+&*$yJb@VTr6yLg0@!c;JtpBNM?YhO~Wea1+GJw zpW>71V`AILW&3sLI_>THa`eLNm_C!Cds&*(l3ELy2;=#4n^M=^?MOV1X=6{2Zlc(g z`*4J7%g#R&WNvA(<&s6+6M=?C4?Dk7)_5%Rk`e{0D4>Qwv+8dnbPXIHYr1xQjrEhY z^L994t9qLwCz12I{#^AILc(x)em4NP0b9IqjgB?dwf=ALZYB6Z79fSICV3vAX}oLBx}#J+GU-= z^7w~LF3$DzlIQid+P0_o`^QuFcGV2*GNrIXZk{M$1p>*Ln#VMeipWK5-bHgAa)VWQ z#!H~uqbk|^6|iTp{oJ3CiiDhWCWS=G{dG}J)1I8CN2vzJTs38;gj|+h2R}J=y$FN5 zS<~n!rR2)23RjH`P#9gh0ll;TLFk#k&XE_X7G(l%Gq>Ce%zE0jPY%qSp#L+f^DWrlQakz;Ne z?>c9@`B7WD!Lj%zLuZf63rcWR_0y7p1Y*Mq+p*Do+73wh&?wVqw-&IisaCO8kdNMV zu3_VHCUw->{plp;H?u}&(BV&GUU3?VjWg17-p98nOyn$C^zX+MQl^T z+)vnbS65xvulK+Y&Mc&id)>IlAoZ32Pv}0r@ZK5Psotgi(0<5MWz|XG`m_jC@i|w> z2mA;HWJ+=X3vBPT@p7l6ZgnVk)FeI9U&VXldtHs-WBcS1`CP*A>+Sq$3RJV7(jl@C zWwx+62yfNuK}#fMzc1@lSMObto3;C0K`kQQE2A9ZF$338S~k(pEy|PB zTKQA;sPFq!R8+3QcR=W((yU4sY#!~)$_fb$4W;Gc(gcRI)ODR#^0m*h?C(WdP5L8_ zgkKwIK&`q`-aH)pA39DRFt4f5Z>`mHEK|-k2insje&-Pw#0xF_TiyZ7wT5n;p0Jl^ z+gA^j_B8N^5L~vRqo-E@Dtww=nxRgVb+(Jk>?5M*n~cM)b!k_45+$aFeqU+~+TPfR z?oiseZF1uf8r;7{PL7_vdMrw$pLo8hdDTf)xE!)n>u6dqQXU$@}x(=b4Dzxah zqmTzShj}(5P3OIsbg``1V=ftDwd%BkBr9M9IyN+S!hz|U$UI8zK(b8kFAeHzc)tot zEghHinokHQ)B!rMY{&*6$F3V~*fJLTVd8z`d!Xi#++a3527HhmX!PAZHp1?DU=`Cp zxa~4%XlX|Q?xY3KYCF1HBsiA>3Caw40aB{B+?`8f)QLhxFM(I1^p2vbVrXz}uMp{D z2Ofy!9q8&l%>DAiKmf?rz&J5^- z+Kj`5tV$*|aZn0w2Yv%810P2lFab1at)EqTz z`%L1ev3Eye>mJL7nY%7U&-H$HXF|)+HM=^ko!sEKLos!}`viwvQ6*1yk$_GhV^`wK z$&_nRB^@vc-C)40*Qt!7>Qxoob4Q!P!yg%Q`KPt_>vSPQOSGEd2(ce@zJx}-?8 z{b%4Zi029%SyicI8cZo{z(7IRv&{tOTrtc+HqABJmxkv*)h?wsdcCciXDz5CpWe0`sl< zNwc9iwOI{}Wb4h??{-_=U3!+UJz+`JVyZ-i!IyAkggn1r>AUv_vlSlvP_=wfOj(z;=#cO|giC6d& z0@bl(3vhaVm}TwfdgTf5_-vp0@f-Kt@1PNJStoJX>=+KD|4VeMV7Pr=g~vYhXxtj1 z9I|UYqibGo)ifpfc-ILm-yMz{r=KjJRlv6+hjwm_mGK}zk2tIn1bcJn^%7kYj77o=&w^=zRA|`l_Xa8_tL}n%X>LM{RnAco zo_0;c{e#ceQLfEr$+p($+=k&6`D-}nf{p)x=j!7yGuZBO_#p=BONt~LesmiwDQ^I( z)No|j_@hs2+6rKCZiII_ynwtPbMK!g0m=OYQb8JOY6&eD@r$cx$tTKBL|^8gGtYrs zKs_R;q=d0yedS*@1+H z&voXQU40JY-!kVV4h^%93h3F-va+)Bva)peoBJ>KA6eu42QgCzz@E<1CrcH7d#Qb1 z%=7U|Kz_R*Sq@cC^@UA}p3$OrqGl6QO5?C7n$7&+t91AyjXgjZ?A;{ZD7%viN$6Iv z+XNm(uOeNRl3l=UY-$q&jJM6=`(3B&`hJRSf9*Jady*4?&3|jL6c@LNTjH%c(8e=> z!q^ zto3?>XK8Po1q9rT7lQ=G^|5j>Zvmfm&bQW^=)TlXb*#qgzP`YJX4YvHS=n!`?#X5t zhAa341k1j&t1HWMPLU!qkA}R6F=O*iNp4V^Qu>T8of2&(5Db3BkEx(emKNoNAJ2%? zetFav7GJzVqT|!>*W^6H&j3ozEO*Z;g-MPvKV@7vt!RFiE*#nKa$?{ar@+q8-ye(M zCt)gUBXGuqeXW_{L_yajuw@Bd)wX_Yn5PR-{(QIT$1YeggA(!{pG#6PnW%lOj8bB3 zwW=;#!PApR@aG&{it?DdYjNS^b>e5A+_FHvb!IJqF_@s8bU;tqJq@ zC{@XyD!RHhnlOz~Cy@O!oiEz6`Ttkdm_sQxVi90}_>bF8I~4W!z^~xWG(wt6(Bo)8 zL_uI%eIYfWmFD1J>Cvyaz0j!3x51~8FbF1kPC4U_ZxLSc)5+k&UswS5*l!u}Z3;ed z7QLn@l$WEBOH00=NMgGSp_4MVA5qg(kHSzM7_C;W*Ar-dw)5$KTueTEgJ7^!YZ%++ zw!h<9=msBv__rw^C~T(fMc`fw&znt6P2D(xbc`nmL>4uSPEJl(66d3$SdA43O1OTA zQ)|Zuzr?!tO6~NBus;)ArLyT}c$iZ6w=s-j0|0K)$bNR$oJo-uH=@jo z-RQjxC~C!9c61+-jvn2Cvvom%!dNDhb{a}@gV_qqob$k(=UatEfeB;(H@bu3f2K8z z6ZT!F8XjD;-b#piBbXkZBEi;}YP6Gru4iQw<1`ip3_Y?{R@{?HLIrZ)h}{2q2dLmN zZ}iA>M`Y9dypi$706eU57Xp7SXo9|b=`WbdOLc`N=y$uAs)(|fjuCur>4(o!&p+Po z(dB*d36H`&;DIBfd`|K-VRpq$< z&5fQOe3Bh8vBOLvROvFx5}28zY_McSm;F5piwbMvPQ$g&ja7ol$wwD#yV;e+0Ly&6 zIUZ}$?(S}U4&@oSkUX;%vzqu+dX>+GsvB}eg{rm~h6VE%ow`Y>#~=3CQ`aQ}wqzW* zARgB~dJ?9`me>nN?uQ?SPE-$j6d%^DR7i)v2fnKWnau1fs@)uI5uYY<2NZSermBcY z=@wGVA-yLJ4}O<=*`L_+?N>peJxyo`#oz0Kh%aydJK(=%{I4low&!OkDf>al%k(6a zLr4I588ZMA=7{@ee>vE2WX8)3AM*E3Ba_zD!G_*_s?ZHHm(%(~Fq`^8NF=I^p`qdG z@e#Gq@vf^xPtd=h?{ksJW6`Zft^7Mqft@^#`^{XYQf+(glxFy6t+77Y;o;#1Nxa{$ ziVR4e_ej&swLs1>k_RFW9M3x}1gnD^gVGu?X^^#8GLzwho|5?Hx@QE~K}I zy)SW|5wO+o@nCawSh#6l50*Z8jW*%ClyPMP*(Y#OUqol(+L6Q2KY>Q)%Hy9h1Y;1R zB4*s*SXUZm$#65f$ZI2z-5OD|;W3 zaM<@V+{^tr7$tQXeD>|I$^6Scv+*zW2*sZ+r>K2`o)N{xSX?8cqb3zfOG$rtM2(hx zLI(mQCE>@3%D70~Dp5U(OJnP^#i+rH?3R@h+!!B%=_eSa(!@;zri!4~{0n%bJsZm@ z#ApcJAG2#sUjioxwsIcDFwPU6u7f7(zhlXP4!ED47^XzR!EX!E${JPZZ?ns09rf$E z9eD#q4V!;?d0un5d!i0acZ9WLdeI}n?++x~e_!DJP|Il?BFkh2gL*CH3~efiFXa8~ zx%gPL(8nRg?@Q<}bJ=;xLpyT#E9=LPF{Xz|3}WqFmzXKlUTmq*a^wjC^`#H0lnl9F z8H`gRb{KUMo}8WVYpfY_1xA^o3rLhJI$f}_TFC6m(u>?Z>wn_u>fCPK>Ph5qjCH850otVuYi#v@5B9TKW$s>Z~`nDS4hflZZl zsb^((?k!EfFDZp&yiEU%uv{!Wa1Z#Aqneg04ae|VSv*yGv-}O1^1QeBrJnGat1y3mNd2XzjQyI85I__1!d58IXTQQt6gIx>3gkrmu_ zlU+SYHKna+28N5>B%SfF10f^Ww}G*rcbrdWFr&J1!Y;mH-g|Y#mgFUWFEgWQOGv`x zE0Ibn=prS^WixEA*MIDu6i-!OZy{AMnPRrX?+MPir)Dvo5}_`}P9=B>!QA?hNRam< z`VIXyUjXF7BxB7iViPmk)(y_maHcKdxBSRjBB}{vSstnb<6z|9s-7~SU(&EV<5XS9 z$n5S|e4cwma?tuM;%vSJ$ei8ICoTUdJP(g?I|exI@Q;Dk@WCC`i-{o%K&E(Ha3%v! z|1n&03r6CKdn>L%)s*dpb`pHin^^x?Gi1s|)NnQN$i%-p(7AbMY6>B=H@?Mn*Tc}q zXB~_D3NQQn`47)NcNAVU)eu8lQUD;A+44S^^x9Jiz?W?mXCBB56 zYUUZc@^39i$X~uEqM6oNez8iCID?GmC@_R5d52?K!rC0Yx`Gc5$^UGT2ZUaq;P^Km zeYir^lAU50I3|BWWass}c8wi=n1o3i?a1BgJ$8CkLlNkCV3^~)Lw{xtB|C4g&?fi? zNtDXWc%KoiB~wMF*Q5JBUJHah`t%L2qv(zkhzgUaB?<7ZH2^6YR$c z-bqo-$-*oa>&u~(zQ*YP`|M&rvSs8!Rt$~Yg>CxcZ3Wzu)aGtRRJk;q2+y1kpEMg1 z)2zAG=1ZwSNf<1sI^+)|&%x&9 zEVi{;M*&^)j3vB7Cyt7uOk9qv=go;Nw&7U=P%`j0zxWc4=pUAI2 z{ntFHqcPLwX65P1wVtI$$*AkEHp_?->b5_{T`qg5T)3PeR+*$5(Pp1qD?T%tQ3NUB z=RL9rkoDP-{y&o$#T8;#M}N#iuw*tTtRW1Efbq_J(Mv2C00et*yN{sa4& zojY@9=G=4c>9lu1S+H7V2;DFjY>(0n+nu13ROwul&?iyoIgvEwMI_K);@=sKc6P zHnF?@B``}ng}OPF%qwYLb}*cRex$0XIz+%(`jYzVd024S=_EfO|6*~9V!!kTk^1s}Jw0IpzgSdfv1BIKU%kJ-F9KNsg`5_z8Mxl>8k#=21 z*hbC*xe91-^>J%RI?5Z2`EHECknM}H#p1j}hrNrdB^c8oLqQSj=$Z=kfa*BI=S;OOiIrI<%ymY{_z0%C z)$Z5s=^)XU?=+@&?M@aqU34AE(f9ZJzRy=d7IfDg+03#TER^(&cDq}fUyR@wqhzSw zCyEELBN-qWB1?u5^i{c+1v93|YN{(X^5s~eGas&H|!c$Bbq@>;YcVK6( zKRMe-^6aFW-o3B}6v5X;ATxy{_|~M7sQzdJ&#j~)tUan zc%by#2_euwF40Q!1q0mx>ReYL|Z1NfGdz3KTZ<}FbcCwT93~-E(g~1u6lO1{ul)6(5vLn zUsIjQ=oC;3>aY>=xyO0Ts)c?Ii=8aRO}`mV-hHec!(CjrpQ zgCf!cxeTEbw0LFqr5~N^)$?#dgq&qnE;gz_QK@q%%AphE;ruV{-u8#p_dT%Uz7pA=F<^)HZT1fbQg>(?lW;ocshEYe;r8&Z zA=NTq(Nc(sx`92qdnhRE-)|&74cAD7ik6ag`ci&KMyQBQOkkrW zoVgTlaCg+4>GvaDzG1RbRCbgTqeWun_Wlf{Ob^dS(&-0yq-fGR+e8z$ccY4PvlSwp zHY95-jYawPc!1Zzi=(chQ93;#?5 z6LYZJHeHw;tB3R1IMFGCz*g%C2I#%8b$4dNvaWHmI&8b~9j$`l)M-pYl^V`WhlcCo)Qg^?J;6b znGBRM_~p0ssj`g;Azyjr0QoBMQ_edBzusMt+AJZr4HHpoOtS+zf6q7wF$ut?d(A_N zt>Hy0E;J%CZ^BYjCD9I*wM2w;=J-6W(8kWQMD)lTV6(HCBTXQ3joeoLaD68$WGNGy zSY@%4@mM#dj?D_0t&o6FoT@gvbI#2ZX3`S6>K|*Et>_WtWcnR7-Yc z{V@ygB|YFAGlDPshxq-UPZBoL&g-G97tvoM3P0-e@0y1fZtO3&1SEhNVsVK+ zN=`yzG$5NYk0ZhqIoqa+{2^lH7U4ABlJvYlv|Fg+QUpvicBtF$#V#|xachydhg0q%BmmM z-HsiDEP#cPNaz*VKO)^`{sJ;`p)l>!LfHY&pwE{XhN-#!V@*2>=#B0_i+40dmTn}k z9OQy6>iI*CB4Es;{oYsZ-j-^tC*rWk?kYpRMD5ssI;VZwN6iMH!-nuNc=ZTI_H3D2 z5CE+C`@Y>saFX2&s2=~@cUN|~4+(Z(_8%ST>cT$3Zq}Sa&KBqEpF(chHsAeC&+}SQ zw1dn{UC8u@=pLm|)dlJI@T30T*ujj-JAWtwQ561W=gou!Z(Np|%T{>IsW6|eJFWt) z!2kdk4W?T{D;3jDI#&hDB?HLB%_h%`f*7434NjuRMXYAGNjFsy8^ z@nIN9096_{kANARpo0hNyTBR649fo#_zRnoM`KBaVBA7) ze}Gt|?`CVoTj|}+MpnasgN2d6lq}29#3~4`$A{-3z-~#1-Xz$uKNDQm0u?RB&UDZv zNzWEl3j8W0yFsEemN#iju1~v0r8ie0kBtd(w8RaRCV^tlCoR$`4T?qUx1SKiw_Q z^(GW{Za0}`9hMXAZ!qzq*hHh03hNc4-k7C_PCa(f(i5uII(#EamF6kxLkeMd#1>0E zB5@O+=+N?;lKiccTIZAwlP*;*z&zAXI$`H>LR;){la+NA_$Hsq6 zMXML;aRv7+tbT0$V1e{*yASnyBE3FYz*ZF7!%t+#0v5nKiRWqG?M^@(ItoGNAqHrW zOC4CV96j*9xrX>P*G*L4^c@n*C!|kB+-Lc0Y;T|R<=-m%CHQ<$^C0jssyPWW)Fl8 zGXQX*(Sxy6i?@nH*cMT&aP%{76HbFvmFbpE1o(%ddX%uJ4VhkS-R>Y>N;tF4UX_Z0 zA)1{tuavt{w|L5-iB{7p0*0HA1VHbHtxE{xctRB+4)u~}T2AOFRBCb9H{;lHd9Sb@ ziGJ@h;_QO&EKWMGm5_byrfdc;*cu2BW6}^6N_9ZazoRywj*j3%H>!*@C8~ZpA_==u z8EGYD^Q(7y6ia2EZa99tfUBwx{0m1F%Epl->(SSUbfSi_>xy^tck&jFyDI3}s9THi zTUk(*m;DuQ7xH~h{&5{G6Jhzg3}u(f%ZdEKtan>c`fr;!&@Zuk^W6yTX*arpiri@W z#?Bfc*g>@P_|?m(ive)8hrJFIrOaZD@k@B@+E7N=Xfj*c4hkY-eb6KR$L=T#}ZB%#QVpzG*M}x=s=VE^YpSA?ms~GH`kw zH}38FFWp}^q0JNvI3d8I)Ded_R5Qv>U(s6d68?pwBH`G)K=*;5wSaYx_TopZHBwmv zC91%_g8Motvk;`_ux>@d$-bqr zo}f9+LKMLVU!?jfT0}-Nrb%ukiAoyW3CHBFgtUTFL_U`XaLu{>0#S;AdaUhxojP3l zT)I!Jv%$YPdb-Gln!Vf0>6MO1J+%}qPpUcmQd~L9bAiRNc6veZA`jdNj>ay@WF!#y zJ3_Xji#i)Oaw_P!l0qtXe;+9k=j{l6o5LCcI+kO_wYIGyqTpa(P~lq90wR;M`VKa4|GPh+K+FZZJv3vwD-aX8^^@ib}~kMo?d$CxQ3u+h%d9vOsF6{c4!%ao*n zj*d1AaVK8ITgHb6ohi0;i!)b&u+By3g_zBUZ!&-B2sCDsWll!$w%LlfC^HqMV#$u3 zayxE{*NMW#+U(^78sWJ)rxv*tmGGs^kSY1HI)XG5C^2Od8emII$B+Um+8ln7y6DWQA7{%k3L&1vE!f+$lc>t&O;MPxH%t6!ssd zj(5A!R!(1`g8jgGcz-?;{7b+f^C4S?j59t2!v_B)Y<+z zrsr}l)5b2jxrXuJjf$fp*y;mZA`wj)6Z?NQ*mfr#Pqx%Pbkmnh>u}L*t?(@naUVi* zX%J#uLaeCC?Hn^*sgt?x(MZ{A&1v4x^s>F&=Zv(`B7D(g zb*LL>@sMv?fDZhIdq}C~yb(CeujO*I?jLz$a^iDk?3E}fxAvogo9raz3aAtj02AGn z(24POhAlUvyhu69so#ZkgcV^fVaC)_a{9JBpzOV+ngJy?uQ+XEco)E5(${U(-;TTm z_v$k*W;J%ABgyZU8-gB;d)i`lbqelrex$sc6Y1#%ea=F zPHNo9RsrJ&mBOaQg3NO3@=$&9y!~I3ACg>c@vX(mazQ?@agA@Y1oS-|=Rp3@sg%ENBy2s%-?i#-Nf(zEa-;_MAu2N^=h9vWT5tfS=( z=Q%j{e{9EcRB z&xLMJbf_oGc-8@NJ&$J`IA99>>I-XA|0*_`1MMtJQ3P%Vr1pPggTgY-B;+7YBqNTF z&wuvw*j5ywyJm&JVwL<}k96yG0DaA}(1zO+vffbtwz-)IwfsU|d>HQ(aDlqIRD7u+ zb};yCf3+k@Tl!Ay#qGWw%?!VDrp#TR075{mx<3#Gc}S7hw^oqmbziUrwsHb%+=fUM zvt#{Cv-}ekP|jAQs37=x(6cEKk90UQQ5Z+o@<_Tn&fNA(2=r659KDF|+m3A5AaNbq z=@I>L$e8mfI|t!NGQm&3x!;$3RV-sNx5@cNWG|FQ=FoX{hlv09yI7pxiUdw9eA9yEz23Vv~%DPic zV)J*ii zYMK5B<57-S0^Z8@6h~`m4Iz%|3%%%lC0>o!H1=vR!u0HFy(J-)0a`+G>dxBNx{2mH z>i?Da238WX=J>p8kED%NTGCJh7=7J+?uA>lAY{ZKbN87T8zc}n)!t>RIprZvQ$<3=%fu?z%VSDrA5)nygT7)i$|?W?4z`2E!%J z2F}jS<+lLf>!wWKjS@|W9<#r(zRDVhi+m+c#!P`m0-&UUfLW&^j_l_szhs34? z+zAhp*NYayD_Y&-5uzWmh5vAs!BFf1y~wI9XiF0h|M^tJd3$oo{0nX|bC-Ke_X0i{ zvST@Sk-)DdtV5)v7&iT6+uPkj?44z`KBTjiwN!xmY=4u+NP@Sy47ZP!PbuClE$v9_ z>hgncs(wS~Nks8KMkR`xo$dmjtvqZwY8;FBUr zRl)tVtJQXp1xKTw7#-^A5*ToYo!_Dz*)F&lzNt37JL^qaq(qLk2c9E8jY3rMIvC z7vlJvm&4jLJ@!mgkl8mVf%{lm1tG_b7T06w6vkgCG#DPm&dHfV!eQq%$NpwVyNJAM za?9=Ry;J`(58yeASj9o$D!p2K+)8Wxht23k3(an}g9Vke4NEg#34F7^ty^!&-H2yv zd2Ol391W@EAH9{ValF^5?`?8{5|TNT=$Xpy5@q^JNbd%y(iCBBa-Lbyed>N!`Wm!V~3ML=;(l@%cMBtuij=E{}`bw}}RaS{O*Y!&?2UGbKc59ya*sy__G175WM zpZ*H%C`V0>)my`TsvJY?0+W|f_L$a;1VNlW5Y=QY6O0-jLG4EZ=|Ju z$nV$BC}D#gPSzEe`-zmYhKR1zE=6)V0!`X3?S(WDFE?A8!IHH5b5IW%kliDB=CstY zF=V+6vnfm``>q4kUqfhMq`m8QQ(E*jO9=ep7ng+VfwmcUYFWuqbY3asA07Ik>Z#`W zfn4Q%C0b~wV-Z?Q{$ZY&&l_-KYBKG$z{!TlusWuumM{TAN9t7xRG%}LCJA#th=(lC zHDEo+Ps(_UE4X7I@#iVKwJ3S+ny7p5L$d{{u}u|dcfOZOX}P!coN=DHv^+0mUsW-V zir=O`P|Kk6fWb@3XT>8|NkPWi3sUa07afeqm1=!k$B2hPvhA|JMFcj;!DP|d67qd7 zFSKybq@qA@LN|=_xj;eSQHdrqwA9$1cfjd`n%chPY|P4xXsBV9)>ZMvgG*o%R?2{` zQ|=3P_mP+_rHp{JpGLC3heN0L%Ot_9=8-^`GrXBCno)yb4O2iBxi;@!sWfV3Zx?(M z9H4|9jsu%jz`4D?v;K^nJko!4!4*#}u6t`m`?2|m|II@F8%HM#62Nu6`mYN7ZGl9r zAxKVu!|O_KWwV8*HQekbMe?xXY=!1k4OZ1fS zfC7QNi)Rl>kzOX#aD~}|png?{r}B3&V<0*(Pp1ME;FVV?X?@*4+;Nk9xNC+CP9C<6 zi`)$7%v;^-i6r>GHK!uKc9ZA zEX8|!f=k`)S&6cbrVnniLj-hE)(IaMcl@#$mG=i)W;*W0=KZ|jlZLc}!T2cNjw0y+ zhAR$TQm|z-@wkF>s1oc=%AD{Jv3hMN8DVg^d{MsFxe`7UgoJi}zx(`vq)?=B%>n$n zebdXAOOHGIL*~@8<_n(Z=xl1(QDAphoC@k~%Q)iqa)5ol(T**h%4lM94r&Bm*-vxYHd)V( zDQZItTfE4ylA#2Y8otuR+!AeV>@S!kY=sOVq_Idx1>Lks`i63BD3wsMBONNfD1Rby zTi#CK=iI)nh``$9MXGMAoOp_=u!4gVGwF-7lIVIve-AI3Lj8|@MMsM>xiG=D-ufuU zLeKRE{~}m649n*vbiZJa=tSvA+QP;+|rLxYAp< zFc;8fS7EO4bHWU#zk`*&pc<-d9=?dPkr6$?laMa$;V;M9)W$D#)L(PfV{COu;yLV6 zHIj=F^oT!Uk60L0Gz54(c`h$`s)z1eURItzKAPoo$>nJE9~+qZnw2=$!=`_oS6q?j zclh&PMf72QU7y7oh6fgv$(*lnoI7dZ%c?GD5WDY~If1B+o8cOy^4;k7QibKkk^MjN1ab&jp+b2+R)~QF&(AJdQs|;? zIh&o#hM(Utij7GQfqwlZeIBQU+1Q&;7x0bh5l58HgH~=58@+CW7K%cjeqU72d`qDq za{6>irfnkI@3ApEc#`;tm0t7Xy{(&d;flVFNG(6n_xNu5oa6#E6uQU4>%M=T!ga{7 zR;=Rgej$aYm8cX(abazB3B_1=k8Xi_K}(AD9d4o{R%GluT%*bjCHuRF!+;V@Sply0 z>SOY=(Kz6+e0Mi@;N9&DTE`MVB|B4mqe0Q75U9 z!`&#?m2!k8^m1VAXjxL+0)Q&cpgpFwtxA6LteUmfGNO3?8_OvvReSJLN>tdJ0)4lM zCvHBRb!V9}B@uw?#k!EpWVnp52wC0bF}m0nLk~GgC~l#%6GR-K^=n>Sh^^e4f}o`I z8FQSa=9(15+Jy?QbZKRpErrs^{Xw*)T|@)(!I*ExH;s2>&&5FadQm)dX%qmg50s(K zzQ}WSK^bWMH^QL<(H43S1hIK*-7H={=aarm0G|fa)#7p-n?k-gH7;ng1LGWDamL^s z8O90MKW5H0vA1y-*<6F0ddyi2|AU|1bg36O=O;^8v5>qF01>Ov-~Qr;t3RA{#}%Lt zJb=UN28XW0Ixu~}D@N;2w@ej*r?@)?zD-{!mkU{`1F-UHj6JcGch5(^N9a>^>d4;i?k%KxC<^h2Vfw4yyWq+~iltFQSZ zy!bCs{-BsyOYqB9Bn|ob0#wK_$Co`)pacj2R(P<5WcL+b zYWZwbm}KE;;ZD42q>i@HbKKv$vL*T}M7FRJtpEY2j>@!CZBh9X74(FI)|r+*3PObL zI^dopH_rlvO0zx(K`_ML$Zb(xwPM(OsV8G~$Jb0;*{)qn=ke1FDxDiLC@jzpLJ-UU#L#S2ctdrLbv1j>UxjMprdRox8L9yvYqaGo_I7W*H z!DBP#*4Cm_S63fzcJgAtts$N@L0dD7z|S-8o(70ncTq?KnVH_L6CawyO6~8S=lt^6 zIFkumzUPP3<;gl$x1h%*lNHSv>s47BipC}DKF3*1P{z+uOrU3LFL!9S=xrDTui7gu zN!1L0 z5F&hC2HCxQC|oj#_0uGjGE``O6WGE2@?4wuexAtk#z;geg-*l3>E7 z)*&btooqD0!*PLIZinh0rApH6P4Bs{{84&UFBVO8i#TmkfO*UL-=(~NlBO(g+dcEG$$i&TRT!X&?GDLue?FYsJPyoLmJ#9 zBr%cb2hZ0-?BacQ><`ez3`G$#V;GGNo1QbuSS^ljCUeZ8ek?E7*glC7$Vl@#yPpg?YgPdylBO4&}oC5-CpJBE~6|%3VqK+wj1hY&r_80P; zj%V_?oDLD%iSh6va&t)v3kyk@6}SQv`*VY!k#7NiCP_xnV`@YJ@1&IZ3y_LY?vL=9 zm8tXh{S&{L*V5E0i5hKS;3U)_o(RFSDF+j*PPS1aK#bx*nU&f|gVSIc$M)I0|X;O=#jzuFVsVDhiL@&=#QazD209e6*(Hek?;uvBMO=)O-Qdl4|Y zay~(FrX-Ub&Z{|asYz~ypW{YLmLW%|RhvCAQE%L&eEm8W3sfgwpT!@~Ap|S?_cUf< z6quEzwCVe1UR57PhP2>Ny$)DiN!>Vt%=Bje-!Z&%&1puW}f|J@3fHRJR)?-(K) z7<+U7^_RK~#!-D*B|bP_A|%ZN2*p)iG%9BnSS+fJ{>^Cdt7bm6ic$v28AHfWHWmfQ zm&8xkRy#FE6e|owynDkd8m)vcAiO+;Jmes%`R#C8!Bda9I3D;_?9XM!ukkhNp`yjv)|4p?Qc1@BN2Uu5j>&1w3{I7{tsO5v^qRXAT zW6=d8bM-vT#F-CMVY9jYP*z)tUqmCb!%UiJtr9J>$9*UD`>w>Jd&mM26BLq(z@Dgo zCVx{04-yf8fitv=-_S{Dvp@nlOVIBRJCx^jFSHfkC#doDPF^UH;dEqlbmKws8buS< z{t?Fq^l1Wuxw7TRK4yw<7RMGf#loh7IIu^R_g# zoQH`eYz+~d50bglP21U~Yr!UxdsDi0DgGFx_{k4$46)auNBSqixWOMUZws@cJwno? zrgKr%$suM|=Jl;G795eZT(wYM-Hz`3<&b2XA9NRAXIyl7GL9#~hmb zbUx)XA9K`SSkk>CZEyxGt*7?W$30I?s&Xhk_;FJGfjc;0}dubbFAr^GQS00R%y zJM3-)_AzQq=^Oj*bR10PLftcJ!FFP;l#eKiD*FjM_D3R|ZY^+J%$|rMcN%?a#r-~q z&3=V|My4q77UZjel|#tZ&QAWZ@$vZNWJtTKjLp%Nm6hw88-r=3GTN#QE;sSG9L+mC zgf^BW%%q;#PW;y?&uG8<3y9YWrNXp1!I7tRzJ>GS3pgs`yO|o8-p;E~yQ7^MD`k-A zZ4l5?&X;#Q9wkW+)Y3<$92iwvUqnJkPnFvqsQ*&Wf3xk?nen^a%5!#XJ~6qY@0q$-$C@O|ycAA+)$K3Nk__ zB&4ySIRtK_Y`YUgLeUz81W}J^1vZ*tJWPb|i277cqOd~*!((&_PaBbZe6b9>J>lQ( zfcLpGVN=&DUWkBG7J2Hu9A9wzG2I!%7%7@hcUz47K+TeAZ>^)4FGIU#Lat$sSwytG z!=PN{UZ8|m;HQ^zmLxjjQ^JwQ+$bUSB%^URbVN$xj@0eRKS}O(RLaC+DNe<{*#>aF z{?~r~7Pmx~6^f%I96Dxz!i^B}7d}f-=cFL=l0OWG16LN*yW4Ct|L|at zVun>A%LTl=!}z$d=)6U-$f!;V(dIWd`fk$Z97+=7AA9>dAc&;Wz}mRp_EL|RzL?Jp zDEZUB6zUf;RK9}D{^H>=Fn?WYW<-g{KcVx62vA1Q96NC2%91oP%zExE$P|mBGe&!V z-~Y)tw3bM6wOmI&mCZYRJ9spiskY~!Y3oaLG$q{RY??-5_nnRhaOMxtU(iX=hE9;X z5`UjX9Fdy{)(YcnU}nXo3r!S5gX9x=1F1AIk^qQTIIAf2H95WETU!lm+oVi4y|K%> zQBGV0Yh)=S`bWFNl(|$BMHH1@rnYb_dq_e$SZO(DE>z~$IYA&Iqk*+98J58fdC<>Nqj{okZU&h+!T%(;aozWG($`NFu}0#3YZs|nikR^QUG+tGL<=87e@Fnsx6)vdl*oUT>&1S;a&x8}j$qC(2s=b%OdzUt~}gO5K+g{Z0b znnEWNMgI6)kf6y`8{TQ8CeE5ncsmpL+7&E~TWDdJm5?0VGZy~Y7$?*G7e8_V;sYmz zJ2}eL0^xt#_CXT79S|PEiNx9L#0?AT1850A|6o z8mAcVC3bF+6%@|>8x^vRJqqOyum*7X4V&xq%RRq_%A!u*$u#a=MCmS(rAj_BY2K)*-R2A{CMIU@;2;F*p`$SiR^kEi@*pBe zCJ4-^IF55-$LzfFOAn_(I#NAeFp=!GhoLEPBWJo1-nckvM+^~UUoA7UdU@X;=X$;S z(pcS8!k#3J{lbC9lZHg+gY7#khabFPLvwTiLUL&? zU+H$zR@8DckcVd74#WNtnJQyaR>qxkyo!IAKajZ1`CI&n@;p3lNU zCl%~`L_B6Mq8?`wSI{Ja;(sOTAkTl{9B5@Y7W|Xrjyh`Q+K_~$<1s3`HlO&*4W4Y; zkr{5jdT#`z3M#?M7vInSrDJrV9gGea+Q0KxRnotP=M9@<_@cbtTlATPZ`JORf!9v6 zfN4a0Cn8pVWZZui;#{?zM61Xo_Ca!~Ps?P$GxevpoLJVvMnk(`m!CI`G1{g9&IP|P z5dlFIKBp}SIXSqmZ*!{BSRH@Tm$YHlOeXpM`eWQY4E5;+>SDGm|CI-p2Q0Rwjhvop z91uTru`Zm);MMhHy3IHP#4vRo!#y(00*%czV&TX0;f$P01g98^0Z4f&6G5;>lH;rc zGU%v$$0&}tEW|zl8<|3nsR9FfQVu}-^V`3oQ;&_FXpqckjrtnfkEF-*)sSokYL!P zak;RY8 z2sda%>?K$6&Chd*e!uK6rfPy?3$S%@J|O}(zT^KD_;*u9y-^p%4#l!J7{2WMhHa7YYdN#J?gK-m%wY z((J?%9IhM)cs5c=c|r>7n-4l-_3D2$WWfXVk914QS!P@hg@AN`8dy#h*&{I+H*_Ig zKGC&EUS?fJ?B-)9_yc(N-dqKl3)kp+mnxQ_U}Kw|P1@VC3ImAt#XI9C?_ju(03xno z*=H8vBRe3lDgAs-+uLCqngXj(c@(?5d3d9x<{rbb(=b5VXZ}KEE4c_N@oW#t$p1ni`)9bqo#>Wl!{Yqf%QZg9O zff6HPpyiY%t0YoJ>iip~9h_L+T}bd==bM+v*oN{d0tfbfOiau~Xcp%;%XAvL1&pw` zv6pMC|9u9r*PLay27c}4s8#l79f4hdd@~83;5)24FULfiiqiud+>)=z#!~Pb@Yk$L z?yth0JblO{GQucI)TXrj+yIiPtu!>C*{~SLNF8t<39?m&@?F0L*<(DQ3*Kogf?Ieh z^Sc8VD2FQfQR8(Tk1%m1sy2Swh=vNew`|Fb6g@DNvz7iE(QAbnN=i&VlLMDZ4Dey> zZP}+0rgGrW@?FM=L1wn!(yUH($J{sO$( z7E~3tu*C1&22mog!%z}`5MB0Isa|GioW$FDsR{+S|X~pv)YweeniHG^gT-i3NihMTtjBIrwzjl15=9jgP&DTZj+ars$cc1iv zwXeMona&K1qu#7 z*4w5FvDGC(aB*@HmR=f!+NUaad5XZjYOIYzP9>x9BDY2L60Zy8_sr~GY`>t*{|T z95oM*ixf>JDnqmN`Y-imxsDA^(5ZXgA{Sy6scA5xhW`%|2+Kf#7W;sw8U6j^rN1Um zc09PVCXN*lh>wPwMa7TMu1k)ZqZ7!zbuMBl5blf~5?P&BLml4PMZgjaNQ-Pu3|)P# za@2&C1^K?l^8NS$sachRZO?(;c59dx71$6Vsr=${2gv>}vKDsxshEZ|o#L0=1J-Ez!jrjMf;Vfe%2-mXL^_0XD!q;=If>;QT9b+S{Ti zQ85D9il~N)gd&MKI6WRVvcVxJxXUODa&rMNE{ybr(ke zi_7CL?M3!?w1#qV&xmZ9a4)a*Nk!$thY9o=z{v|&q0b#_mN3{`1Wv*7D5CA%X5hj+ zG4FKA&Ph(JZ~^NtDE{IILsT-f?NwI1`Wsp8dmbK%RlS zpcY8%L^(VzwCy_0w%Px^#l%tp1zq)?+WGRg(~558Zf%-je2i z-Yf^iY0D<}HjxfCrW^^McidSbh#)Pz!$aWPyIUHLo)aHfEED&{Knt4YIunlx>K-?m z8{hLBzIK#T2BUD7K|hr1+hL~NqU8!Y)(|bBCFWTcx@9sMDF8EK~H0of~19lZ(-SnE3s%*x*rEdexsE%Wp7E!#LSYkb;Y zDhF%t94j2K9&zvgZ&6xlkIVKl`aku;7u0+j>23MdVP$;oz2WuV9t1YeZ*>!zKK6>SCv0sG-r5VT8tk&`w zV>7ei{QX5jZdMkS<*LZY$p0q&H|X=*V4dRBWTY+Q|6Q!@SK)q~DpvBf6TR6Il-(@* zZ*Qip9qGC+x33s$rFYYZ7VcXY;XGY^Yv!&Iu|5h5N@cyV1>!sj!h6N{&vmS|4v-#;QPV6;JwNNpNB&MU%+PCmK|mOJDE*W=lex_UUvs(M1(cE zg+5iwqz9%UkRQ!SN&~E9);JJ3!anDi4TG%Lf^n%Gd(-E^g8HE21HQ!7`g1(j5P-Iy z<9!PYg-oP%VVa0=#B+Q^8p3p387e~-lqvRn$B_SvW zP}u{V7*|Od|G^@2CV-kqHvEe+e#P3WsP0I<>B(#)8~8ToKV_fu#3RZADfN$s@3CJ9 zBiSN-`Wa$YEE0zafhgFo`x%D9F!Rmaeia227I-ZB9uU1D=HYC~XuBsM<@g!&xUX|L zQ^mlX1N)zK|7q zRcpXJ9dO5}`ov*@7Mhpz=I#B+<0MTO(`U>0KeFDzyUy;57H-qnwr!h@jV5Vq+h!Z9 zL8HcQ(Ac(<6Wg}aSl^Srzx$0l#{CPPefHUVueIi!Yp#pusqMy}Br}%cBnX)p5a~2$ z!1xWg_iG^>tFm@~w%T$Dc+eUFvsf^N+nsi0-1v<9;h+CbE5^z^RDv?z?}(R31qyhR z25qRuRAVmYzp#A94)ZXi;ben@3a4B`?_TE&3=en-@LJ@*sd0IOEJA#SDOR3fBW*yu z@c4U!zJc+=mPg_5rJuU>B5+vEr)9_nbwW_6s4{+=Y1Pmnc z1$};Tz&W|Pn$pk%sJcyryptU+36}70mk?np7pm>Uceny5{*~P2Bf?ZnR+<@qU<0}! zvzt^%@1_j-hGO}qh^-&827|JJMItJAiQg}DFBruz4yn?{e{WpdVlhJZpAE2h`W#)lM(ii$4wTCPDit*l{4E_#;Y=u_ueL<~A>-{p!i1XkUGp@+a+87l$ znIvM(p`YUtC%=8Zt_ESV_S9Es7^Z`r_Y{+^lJ)uJFD|}J5qevY7=4$(Dy4x{0}Lm_ z5^dinFAHuOG4IbUzP2Gd zjwACcw%+RB3S2;uV-D3SbjYx>yls~j;Rw{OJ8$=#GWD-EDBG!E#OL(c2b?7l*z-71 z9*>1&QbQHBn(z{Yq+&ca?N8rGE?{7*7TX<{pngQe+Ar550B)-Ne-EcqtcP{oS0PeP z=KPT}6bA6sf%%XTuE%p-i%s34Va9AxkOVHf;+P|%$IwkP{hn_BGB9?`VFaNpl+dr{ zZzW%9ixX&B_`Gl>-0@+ncyg8@u}5T-n$0nW$4~MEkdqE6sN3zn{r(Kza7l(-dGRF@ zo9eahj6PcDk4>#4XDFGt`eLhGd1vg5j4rqO?ytv%G+Fj1 zazrOP{QX@Hr-E03(!hve`TF`=ll=uS^5u3pK!$~d)leS?y7;zE8`kSr`L>wZOP;Ng zLRSKT;vcB#;_;foH5}atwyMi#Jxc8j6z4VuR@-y&Vi1^9qU3_&vAGhHVd46eiy}|H zxD;bE!y55lG!2x5U1CUXUtYhKJS`9$UZBGqV>A1);_KQpWc^!bQ)nnu`}39T`;_E& zGuxDmO~B}R!G#SB4D1)O|5i8^GB)8ncm4(04?#{beBn0zdA!<$8fFlio@ zZ5Bl+_t8^drbj5rWtX(a6LfU@SCr-%8AArRvxwGTJ_!TnRlgdR8#g(6?}zSnTI-Hf z6J3W=K%1_d=O<_10ZrPXi$2dOR2sT#NKnv*(Axu2WKx!(FD?`!ekWiJ7@Ls+!u=Sx zo72$1<#l@!jzT6u%*%^wv{qFjJ*j31yA%C$)Mdo?f2qbEc}FR-^|*ciE8dKi-T}33 zzw~C9u8Bzb!8qLi@a4(i-u% zWy34d^Nd~%l;KdzeDe(kZC(`LiE72v>LL~n>8ceiA{G3Z=(WUw$RMo(cjB9;)#WXK zHDx=)`4{czXytJ@^_ITds^5yJmyP@uDe25-%ws7&)>s&^Iibl!58aX%>WjL%Jj!xhjrE9p66Z)r1l~%nH-cnr9V&xvKJ|YAt zBHITHTH*t1_%g~8648U4-#@I@gX>XV>s5)OzJ z&0uhbx{p^rcyH7SH4@tgDoKayiicJj#S{0Mss&Sf=_!r8f6^>FP!bx56d?dgY};-( zXUS#;C39CEB3)=WA_i1(xNpP>t57{ks!{n!9TJnT#!2(B*hpSgzp*F-2gr)J3F7TPc1=>zEZ)sTFXyncij?NdSydEl=>nPVm1p?lVty zyb*+;b}{Lk7$Ch3>@!zf9z7WYL49RvTMC55>!!t4sQ;lh`OV~7!S3upz^N6q4p->g z4dIE`&Hq#*1zk09!dk!~J$){r_KR2P>5@@(S)CHt; z@PSP>jd{=t0$=RI2rc%k$YaC|2;$@TXz0j9b6LvWfa%sW1C|xNdsY1Av?W26@umNw z?^`9bOT^i{s(s8fB7h?R5OeHo54I4{Z|i!p!>*gGP$h;~JQWU7zaeSZC*pfQ&RPWh z)u=_$He#Bih@b6mS+J_`9Yay>M2^lrxjNx0cW4AuHx4VLC!}uOE#OOK;2-I@Yqi7r zKiyE%%2bqG?23MJeA(c|Mi3FriDNT!OcPZY!Q27dPucyRF9*^Dsf2`xwg+Mc`XaGx z+dQMc2Bq`5q&jcIFUhPst`*0>-={K|JhdOy+uC2;T-fB)$l9->Cmrx_K9~r0F(3UM z)M`guq4W)c)5K!wadQGQWp*S_=#g;3RnTp#jBU=0y zP$4MW5O)jjKD$~7M3oqgIB7bVb4O>wc*btbDZAZ3D!<(Y77QWvl6tH;WBCxZVj1n0 zwWz#VuWF{t4_8$AXqXep-G1Gypy;T2Qf|T*Z32-LazI~&qfu>vxYmIZCxh8`sg^dz zvQA4bOORk7hRA$;bAfOUUN}*o!oT%5vHC?!=D$#hSUKH4F^Wrug|6J zRq8lmM0vT&3+u&Ilg7~yqsrc!zoUKWnBjZrjh0Uv{msYKOE$u^#R!4DX4_{SZ;xpJ1 zRV1dr+AC(lSBON}0gQ4NYAw+o@6KarRDN#&Nd>@9nm6rQ&y{xh-mQlI!{chwtqy5} za!E~tBb1`NNSLAv|NILQ6uBcZ5{G9P6KmY&nv{TX3F_U?c+nL^HU12C>fJLs5!O23EM!gMAFXtJ$=VmuVI-h%V zfwjx|8i9cKZR@G)B(H@o3k%D{@eMYQ6FC58I!WH;uvl1E@w2NsYHU7i|4Y?*uw~$K zmhL2jgL4cVNq^F9GMmu8$C0>FR*-T=YV?d^hv;G=UN=}3_8MDZ^`M7u-WGVqAAK!1 zFc_YKIoU}pX>7`N#CVB?n4+#87OQ?6sbk>P->j7)a+1*orf0wq`dkR7^LzdjADu7L zhH65P|2sWBt%*AdxO`dJcmpCGXc(B{yXJD8rU0PsL*jEk6H`-5GUd^gT9J!gb|Olp z6|qxLCn1l}#%g%q@X+ZQC2#N9Nd{`Z>dC1iLQ<9!0QcGs!$Ib2R27`J5L*i&7`%7o zki#|XRY%#0$xhA;@ZEphjnKY+$)sDr)lNYr3q^DfXCxI#=sA-$VU23CXKY5oU%KZS z-ej*M7V^G5ktZsYVksyBH1El~QTiQvbr!km6?zSTpi?3c292_EkiB4Nf?s^C!G1lO zOfuR4m?iNp^1{bIy2Y;x$$SE2Ml}ybMbtpqNW#PO5PcTNz<@--$|toOu+Hrcv}yvS zg>07ztBCIz@_US{%Fm7n&YAwT80+=*^h+*iSa|rYJ@iN?z49}8z`R6(L2tgzOU$b9e$ql_4Qp0?kxjvD!c2# zoH?rOR#E%6-(VAzI$|&wHfAt=KQZ?c9Gx})JRRYZ_&*7nJXgZ^I~nGypB)ExVr2yx ztL~iI>f3k9c4U3)qCSXW9PH$tq!<(n5S7)KC;WB4VO$y(-($%U;B&PzWWflO8pC!p zpkDeQ7}RMZpNo%XpwvEaH{U*H@nbyPs)ifE1y<9(^m1i3k+J{Q;%q*-Z^}$k-bTqj z&w^#c8{JGaTx5v+^$)}JE(Y<<2nBU`cpbfS7B{Y63@D`c_0~Eki6~ssDHGNRn?>0L zqfLy)hG*!SB>uNyXQMR7ZE%m}WkC9hYHDD0HER;Ah+GD5UH`WGx-VD0GEKQ&TSRLs zKcFS69r7)*#jdQDHQ&80aXZ`*|Ch%%IUm%h#J-KRBq64$=35Hg+i}CE_eOC20X*o! z&;4M1b0(mrjy6NG_lZJP*85W^5-k7T8!b~v8T1O|)<9#x?8|0oL_cMEKyK>h;nw*f z7P8a=J{`m2g*o&QVl7QgvmWVA^`O24^04J2K~+h0J^1)9{s1KAuK9j~#t6EVw1NT> zvov}etLLRW_@8L-L&uB0Mqv?zcm3dcHAdMk&x6dHZg{h;#isHTbU<@yC^?O_0CU4`XNVfSbhKp+X=uUGkkk40VQLKCGkGdPymm{=I#m>442zaJ!!> zH>lfyfg@CoT#D333eL|)L>O;9m$iy0ZmQ+e!iJb9mOQCR&RtxPqLY2dgW%ijRyo88 zJoXqQRCt}1)4wXA1^v|YQzZHmQuteWX7z@@!WvH zU`SMA%zhB0iV=qmJ1`6hIaeR5{J%VhGGr{S53fCxQQkpa3L<0teKHKiak6W_q;iu zv?V``h4i#fR|(gkixiV?#1VbA7H*j9;np#j->apA|zR_Hic`+ zV7L5fx(E!rxtLK=G~Dk0F-jZD?CrOXXlQ77!pXpZu)CXRG|ekO$-+`K>})G9yAaxp z98Q58PTQgNzud?j_9i$_>GUlgB)j6X;q?ZWRDU;j;B@V{U3Uj~d@Yh-qr5#8%J*?! zu2y$2&p3pGl2c6Xr)TiAiUsu2HQv5qg-#!VH)#gB9T4xB6wg}CeIBHr0Qv1C;Q6*F zp_l?r)3OkYb=K-X257((mv64G{}@SH$ZB;9&S16m)0ze^|A+5{PyE&soeVi&Xq7TV zn%AuOY@^Qxt4o*_bjByriCOv|)N5~7$NpYF?XGR1MMBlEi&4h;WPlK{zF4)a) zp&=5T>+%+vKtz%jRuE5{^PHaHmaLf9PK8cU5rBoIZE;>j1V-9^XnENedK(I=z$sIU z1PsPohMh+N%}KcU0hi6fM3F6*8Byy0f#ZMoEB)$;;5MA7(xOa$$_$;E?&!acFg9}B zp7m#|aP0P-!Um30fy6`Pyo7`i|1IO&N7P?{M7y$U25!+IBv2$Gc-y|B3ORJz%4v`W z7at320RO8*G=4Z>8%|10+=)t|AdMn-kB7(M2h5A9G8sZm2g-Th$G=^zp4S?FO8{x% zaJItJ*xbDJgKMqmD)kc-;zo*j=6hyfT=f{i^;8jJ6vOTEIe-LD-41+V0}XuDAQ9-0L^k5kyMGgOu;iR#^aym79-m{zyZluONfhwdC@b9Hnv0D*P~x z{fQtu)K^~DtNwp9WEDw4OOmkn3w>I$`w7HAm2(Nd-HcV-)sPW70dLottZoL%@RF42J|UIL>w{)!bGyG@#tePbG{}}!^kE(X55fXk?nT<6`$LQ!8I@q zCjj6AcB?AdA0r-q3cWpFSJrIXuGj}J-c~Yd_dZtcUz;oa%oyFHYJK2KP~hN;d@ox= zr5LKn8q(|U3px8Ra?(yZNMv~44%SrB01~n(llURj z{tB)+6wuU5qr!ZfM*V>*zsG}*&hefo-jJ%pSOA45-D{e>L&n?Ndk(*d6O&PA0X=PJ zhRZbgzZc*56Z<8WQuU2GCzG4Q_v6vzKVgo#ek7=C4+~rQTN7n$B5R&7%2ZHk{;ytb zhPGm&dR(Nm(ziKG;ZI*ufe`Uwgi>CT{eBEV z$smA0E%bb>fr4`z#aBo}+j+V@HE!7Xt?aAmTj@tocu2ORI$?P%$fT&W0F5YY?`@U-gnRnA})M(=;lROZ7mgK7k=a|#0AVUDFY3bspGEGec!*h zLMY*iBXd_O_E5ZYUB=`ON87h?=TQxQ#yl&n1 zd+rLpn-Fs@TT+tScf#)AfE0!Hv0rOk3w!V1j3Ukw^7m^?a!f^POYDJd;D5j7fQr1` z36(B2Y1P`DIT8r&06b2kszsd@s&*4}FWj-rzYcr+ny~-m0h?Y;2b6 zJCnS{1NTj4;eI;4#2xp&Iim)g&CUY{H*4{QUx5ogV~tbzup$5b{m41D0OiA%>Qi-# zE$|T$5yB+|g?OUC8QqIp%Ku3KzPoP5Ng@B8QgOiJXp|Pbrz=!6Sgy`8C_;YyL?tKO zap`Wn;@>duVK;8TwK7YVEqxov0FF*7@5!7jqMFPO#fy+r=I$M8QU>?m2eBziUG zXU2$uG8_#tzyhcDg95Af9b{o)1(>{d^#>nqe_Y@PgXQ@9BP#fMFFhOow(UDv3lYV8 zGr8+Cg6SXgL0*MI>^m}sY0UhpyvS-=CJd88d0nnaMHYcsA(g zre9v$lOUNTEVW-gaR7gBa~zCYfip9S*TbtWamkbJMlqO?p01><43?o_jkKpUY!|)7Edf)uZHg zrtONyE1)c>9hLhsu;81L4AuSmf}7A)uZT2RGQ041#E85fCU<#gmS8&53sLV?1Zs@v zPG8*0Z&pNM>j%vje%>+5j{hTh=>jGaP87&uX;)-UfBsM9VM!F5r}l6tphTd9lqXY( zUm7iCI3UJ}b|6>OA}n_;1634l7a*u&Q7XD{Lx?cIQbVG)q^=pt$VCJcPE*PYzUQ3=a^6=w+|7-o9Qwij&z_LW;{~%GtBq%`2WWei z*DnA*^n=)cGZ^a^@!M7;Q>Dcu>HE5VKUIJMDZ`-P65b0)T$3zk39}}C-fVHCUOioTFJ#bjh zRTeV>oy5~Uwn*>NzSA7p7;|~ATA0LOz>(dY8!>B4SzZiv^Y55bzreFyDi>qeEm&BF zfCj8=;($Emj{Rz26p^|{>WkE>F3LrNgf)6eDHpXqk%I+AM1mH{B=gZ zoM?C8T79L_rc4}RahvPmRz3-UdaJ=Gu;8@dJuV9 zJxN3St0mNsjn+RSg%; zb&0>p9532--Q=+OaN{JPJ^rlk3J7;`X8= za4OC!R{`d-$tA^H7TNQ@$CF`-tnuoVj#uw1&;I}A_!+`7-!22b6c9eO}{|Q-s^l7Mp*(&27n-t48f{W^{Vg?|H(1tFI zefJEAJi({4KRqE)=4Hkc@vO9c^2VOELMP*G*fADaI-u9E(ZGX|D*jj6&L6=lo5|(l33FAZXc2 zEl?f;Ln%qW>H9Om)0f7#qCSD(*e6f2i(sk^Yz)h5TRj1^z&303+|vAC4Do{|2SRqN zjDy{mol~;Z({sADI2b$qQ!2Tlw(CA&J(e&I!fVkHycGN*@6Xr`trdFjo%jJNlZd2v za9u&Flu0av#AhIoUQLw8HhRK=>CjroWmUIVLPK~XvXPWamoKd03et$3K_sny2hClZ?R;!^AQA$z%L^6RGxQGq8E_(pzu5dyeXLib z`C9}9su%(;qIfcihusUs%Ai8o^-X)Z8@+(3kLowj-H#s2VqotTl-0zHSbE*5h2q{x z2(+aHI_kB?{Az9$@{pJ8q0%4x1=+QnGUIw;-9m!jr3-85(P+>e2&G#Hj!kDr>HS|Ly%))&W&np3Ava!oW!_x}}V-=%&vd z8fODWtI9!SMBMWbX(V<-%n5XZW6vh9Pd11^Wi$aDbE7vJ#O;4V9JQB#9S5C}_hfj{ zTg;te+BqZ6GDus3Usa|M6oV$1IgRwf>%lQ3=)`@h{C`Q|e1jJ|9;`h7A1$1!!OXFS zBWf4Jzwm(yy3s(le?aNaLngRdV+3lTu=d&uFh znkWR#IrB9 zI{_{JX+*A0&dw|@2TJ5rRC(6(<&7?f%83J?z`!bf9-KR_2#F9OQA#*5kfTU~bq zfRaZ+#XBtdYQ%*4~(KuHHc~Q!omueV+1_k zoI`Jg3(IUa&~O(C0;%K!?H#CZ`%E_y#g}6>HQZHO-HV_EQ&+KcpLNqrgn&eiVNeJd zNJrUkK9)X|yRl`(DXvne&^JKj9|Z^`9EI0<9YWv5g2hpU){F$Am-PkAi7tt5)d=WM^QQf@ag~;(#^3bycC`=7Mm(P79!U>I@shX7$7c z132nptOL1mrk=`;Ep+26~FA5oWL&o)W$pCclJp&<$nC#;9*ulY@bfx|B8qB%sZbBPi~5pa_=xCS$Z zwg)gZ25B@diu%5Ff%LVE4xCVqfSyRIja(Hle=XnWR&4&Lq@nO_^e&I%Sk~!A7s~i)&H&GBG z#MpN@%O{SBiV~BPgO`=YVKtFL5N@}gXLMVxF;@gup@Oi69tLcO`Y;o|@TGnvgS;R9 zz}WYfPH9_O-amj;`v^;@+OsN9AV6RDLtMV00tN36NXNoxw!(z;+5|xE+t`>3sCEtn z9jJBqOO{YdTp@j$cfs11eIX?NIJa^DCJJkrb8twA2=*#jd_Z~w&V%9nllTpgj4^v$ zY|zH^4Ge6z-K`m40~u@8&2Jtqx+B$v>v7=Cul1zf!gq7fa#W?oKgR>}YZ&CEg;jq`xVwi3mgFL;b`R4OXry=w)ZXd6W z+P;z1$4qT6cy-nz1pxx&vtY&jTgf8zW9!o1%~}}YcEBY2eT1q;ho1l>BECg4U2JW6th65dGovH#x!282Kff~Y_!HB6VQKs0 zM2J(wRnLLX5l;xLya#gSiesjh>UE0;6G5?AdG_bR!B$8yqN72olBei{Aqb+CV zyR=W#(`t{=+M;?}_A3Oz0}^fl8l!nz_zE>JPQ!TXZQfj&K3CGEx-E00gX-lvLw4q) zzrF+GreTyVk8N5fQken1{9mNB9QoyDoPvEb^0m&zKySyJlcC|HTa)(eTG58rz)9x=sm7<73LIq{ zCqcCB0xRFLRZjiP83e_cL$Km}*QX|DZd-}iKdQ9c+|aMNIXO8tTkIVpu>Y_{TNF`4 zL|8Q+a8Wzn7AClG6H8hn(~-L-N~-v+XMWB^bm)dYhzE=YPbCbjxa#r;5n7mP^~%NV z4u+B|WEdwJ80RW*(*laGgx7!JK35p#O%|Mg7{s_intB&`r`zA=HR`BuR{_w+{o|(a z0d8lDEf6_mVxJx!s(dbiXBq$H@uueHh^Ik!X6U{2QCJp9P6_MAAnMW38a$(GM*KD+ zF#g%ZbYvd&dZl=Btm-sPNsHPeZy7aE7)eyq6kVG>DciV0kRDedgB|b44Hck|7Xe-S zmz4T`=?>!-*cQWj8DaL?pLzorsxZ5i5CDDSj2lf|?e-^i}?i zowF{xjMbH#3Ea5n^YNT%%a2%+_X2m+QLHpBT0fXAY3=K@5dPvKthD-0+IAvybFb^I z$Na$&|8+Afh+bqBaS2UCRr{5g6x#AqU9#*= zFpT%qpJLTZiZ3;OC6lQ!;6pH(>!tN>F2KeOHCxC0X!5uuMceCmxy7S9A@A<#xm+-g z*#vA!%nljS)27R`xkk(vl@NX3c&|{IDj&X#`@@-MA-_60bEm6m))q(eeY|-PouB{| z2NcSUa@rqE=e4z8Ik0#Rix;Z4uuLeZw36cf3{B9m5ho94gN8ym*|d?@a4;HNgxH`$ zwn}(5g7%}NFZ^3ApFGHS5;oeIZb|X#D8Xg;y7Z1W z1L{`zA`Olti6g_-!O4#1?8%8~3XWD%0|eg0X8^j+L*7`sf-~bupX(=J`2=ZA43Yc3|6KC_mwZ9zGKVq-#y4 zk7Ty6lvbcG7U44b^Yq$uR9IQgtN&iyKM!y@})y}du59V0ZwBs+{y zrrrh26j;8rO<#SbmoMOCIor1i{R$`P#4@&7YbNaXaAk@~px^95Q(aTDSdB=)yAm$n zU|q66c$KRF)bJvc`@i#^{}L845!P}4{*?pBUEGbm-tYjnak1yh?Q77`S3W^U483hG1aIH*A)nzD z;X&IC?p&5Ja3H}^3H+X=s_fz>WOU?Ht~UF=B`quL$VDe`j(=7Flt4@*ftnB(Y*%J9 zsM;Q1!8>~h+*4B%N3BBd2Zd}><=*EYk(|y_%^D5IeW1%d%?b!`8NA}CG`$Q9kWoMY zR*X7gEkZ06*Bp&5p`k&gK}TuJj=}{KAJ9XV-ff5~RRK>Y-BcgT4AT0lwd)w1LBQ*B z43i?#6+wIBCGeM*0N#tBr7n|e-D+kx2~4gTBo$>IkCUSF5rEZzU)xE=oghAYSMm0# z??I$3L5#L9yXAN8_`xM}VlkkUl%)QFu{0|SU^|avk(KA@(Mcb6BUDWIX&R39B|La& zs=PL(f;@RRM%o>@Pa%M=*I&P|DZAB6HNO8u%^@!EzB?<8y?*!g(BK9}n{AaRmVi`Bih>5T5sk>7eS;w`iMpXei+Tkof^9_JJ~8a4CHeTl!flQi}bm zf|6DpZWE5>AV^Bw&I?z$T(^G5A}z-yDRxXZ1v(}yB1FgYX+Tg$^77)wTxI?>kl{AC z7F3!*9;q}gFKcSadg4osj?KBN_ZOe_=c(+_;kDG`1|9ZS5rz7ZiC1W^Lz=Ke(7IIU zRqtEVEqr~teA$g77Zi^gBYwDoU&Fp301%{V#;Uk0pl;5TO%H`5Io5>*C=ikWI}M7e zDg_6$=fGvqi?FXTHIh`H8o7D9f9`}Xi>_3xkiEu&5Qhl+_ql%>2UoG5z5Jl*`BSf)52k3)aoSeqR;=@tn=R=KqEqC{M1NxI9fc8Z|YX&sCj@f2c3=&Z_ z_5gUjB_nl58Z4)DRN4%NB?^GJLCFQxo`dh~?Gc;5G4A9SK%-oYy2xrt>BjO!6#5gB z=6dBgpT7dqCoyN?rY-syeyOmtw7*{Y3|bM-W#06CY3La41cJyQc{v}%H62&jg&e<=sN@h&&9HjYON3BiW&{_-Q~vEXxkvoR!jEO$|Dc3wKqCU z2wK~W@oN|5jnl(6=DhQ=w$=A{AAJYl6O9_?K!Tu=qaFea)Fr#(CG@a=YYHh$1xiJc z#mlm<%d_gDw%}KPjdl)uyUX<8e`xyTxtXI;sk@uu9AIFBXLTZTo*x{|0dL8W!RObp zPlNZ4;-;5-l^>zKE{Dt*)>;$P!Wz^vdWV8qd9y@`?nFh&E{;~eIMj7C)sX{dp)0`W z_4(mOtKV`D->FI94HmebZOm4~IRKwijkat^jhx$N=7M2K)kJ?JzoiYHc)3?hW|ygV zfrx0Fn@Rcjag~GvvSLYUBN>Nes2uh!h<`uMJ9Qdps3n5;(rsQ&D>};4X-T3WF zrNz3|wS$HfWk)!}aNZujHzLXNyAAcQDT3D^L|=ZfO-6b0GRYVtI887ux+;5HF`POV zm&en6XXx)(hkLhk1W!Ev*M#)8Z=?EZdCw(Z^N1DQ?}cuNbwd8O)I2Gd_nbu@E}*7G z2w<^g&uMaKlt54wz*i`ScL$30kXU~Aya%F5-=26zczE~;Cp&xX7eJ<2T@=3`>%8Rf zsq_ThVoTw(FAro$x{RXf&EiXJsB7r-y=SZ!6o;4g<1>B2nHC}_F}~r(>-0tgAZLuu zNN-=LK9{z7Zm01DGoJFQ4%B}}!xHtRP1;3uCqfVT+tseJuausty6u`^w1tNcyS~B9 z=?NA5*>atgiPJ`Wu@2niwHmg-oKYi5I-OwirBl^oMF7+2Aw&9!nts}7|AE~`FfHO! zG8-Vj_Os_Q_F6}76=SqR`lFm0lMzHOM*ztYCJb8=`-VXC7sdsKILr0v=>Cj;0D2qoPLI-}|oIinS0W-Qu zmsr`>-CKCIv9_Vm2IInA$8iz?O&VDImQG+WjiKdkWyTo3ql0q*QG~gzsCK);TNeaFsM*Zln)uxlJ?9g8jh4` zn^;p-H#W40`k;VU%0uee@C(Oof5L-6k_;c4_AGqF0ZxbI9?l3oRo9-RqaHz&Z03kzT|jNZ}F=jsFo3wI1OatP`+lQvM^r=G)BOyd^=z&iMTK888MGK ze9_>WEGH;+LpW&=oAbH>9S&{ zwP);At-VFx&70cNDy6`z!U2~q6=Pme#-q}fHto62100H~qpgYDKPnddTaK0&D$RNR_SXT!u-ho`Vj z6qGA^$J?Ar+o zNV7rTXWEo)a3ijjn!Pb*un@JA!ZTJjxXn0$?gtjbRqUe9gb4Xvg1nqIKQ-jN!^_go zs;>7;34J!9y6cV(*z8TfU~GEW8_l5!$^qeI>;c@X2?(jOCxdf%PY3fsGOzAvt+^5O zK?V8#dixFoh9j78u?S(b!dxJ8FV*$&4zPhod%M9=-QS`gNq>~=aesZOgoCT6rvsWS zCbm9t$7@WhyeMp+miFojNCiO|8U$hG|Gee+*|k{Q^dwm0*= z3vhDcbFC+UF9@dAxh0W0H2Y|=Kg)0}ajIkb&8tLXyke{OM*Ma$){SmIQ9GgK+G1ZU)r*Wg}bAZ&+ zHyQd~fD;dKl~>(U0q96@?pTdE)}#Q{CV%Tl6QI0>@7&Q721x8K0u`V_jRr=?ji#_M zT>nukE2$cdlA{c)s9^f7kl`iUHC9z{3RvDXU<~4EWdU*?*}HOy!*&k2g~b~}2v3i3 zwZ!d_tM2%U?B^9>3r2zpLo({i3p7UF+>p;(TZ;#$;rUM7@F~Miorg7f(EKo3sGX*g zoRfr_nX;D53dUxl$_xnobIgW>lV||HcOnu5b3J2y5+I!H^H4u0jpB>CUdqizj*9BO zu2n{K%{pCWyQnxQgQ-(-(7`AwzD2l6Jrcf9ue>aQ6CZh7Gw*WJTMQi~?X9`AFk`lqrltL7!zC;WfJ+_uDb) zwJv1O6Fi@N-J#Moy3W(rfTzM+Xm?D-l93&Jx#Gc=h1EdOyD6fEVF5e^4Ck%E$;OLYQ*k})vPMhVg zbvdM75bxyZPU25E#$TD^mP38w&%funtLyC2C~JYPF|EQc7~#d>25# zL2)7lCM2-0BNG+^z6%St+0e9T+X>UWmxf71x_)L%Gg#5pVYC7gjD!h4NswX!bpKK( zRg8-1pmD_;9%ACwiGRbZOC6qx(I!c%r0%F5P@*RF5+(<-z2>sGhtQtWIDgoOBdi!$ zGK9&zRM&*>;-fvyHA89uTSUBTh5xg>pX6bfq9~$l<*OKJ5|yCGmO$+A+35ZnD$9`z zP@l+B_J?<3Wfqb258XYzI|!)AZWEaE;rMwuQA0|wpN73B_ALv~a;FCiE_1r+kH#DI zp|;UZGF8O+PDzRkYD;YUNCS@7zp}|485;wBY3Q-a8$p>Tmm=cp>T~ZL)wUWS$R;?lU+$~l(IMhNHa51) zVY-UW@o^*~|7XV^1Ny!`gBGg0X+M1Slv}y}o z?ROTOhph2IMGp@>qsL1{v}WBYl*Nk7=y9L2Pk28l$CONVconMjW|<%7hVacvz3^6U z30H~0#I-KvR~k1M{dqF-HkZ<|JB#fu7D^ck59`1l5%Vxp4k-9MjoUPZC``ad)~KQ$ zsWQgA+q6CKi((Y+YK{?lOSG*hsp)= zW;nE+EdN)YBnIt}cl)+?DrWP$qTmBK2?@#lxc{s1G03^3pfs%D%FSGR4)cpXKNxmU z>sQl?O;|mYQE<1wl!c< zp!od9(Qq+|XySA}1qCfwL|d)MAlZFueo9Q)4J^#)*s(adAeB&gKl-A@?`(kl-bQch z;9D#gmIytXLgY;T{`&FwR%n~L^~;GjWQ*b9bkF8yFJ1*;x`;_V$Yo2~0A?_pKvO5E zFtYA*Eq}|;IG9$>P@hBUpX`2UqkTwggcak3f&a6iR(IUcW@`VpsP7@RGb_uJ*Suu9 z%ElUXk!W92ay;jQlgr)8SE!fL`F8Aqdi%yfaqfvs2(@uDxE zLUZ04r5)V%EkEB-oG$BUE1{h|nP0gogzUw~0sgw4 zn5yFP>dMK))Rct1_IEHQ165;&%5+zYOFhG|@tyWwF^ZCLtQSP*)*kcX0#P+~{|L{Cg)>UR^ja;Ci(0v!CpfPXbsab83u>Ao&cn40kDxE^nbp@jhS{LWo4au zXD1CRttz{L<=APw-ad%hWa73-*)A0Ascq?C&cNITQo_xJ|FFRsZxR58BlV$4I=CD| zCSbxOC5wIbkc0sA9we7P4{yh7xTME6Hxb|8n5SfU8>V1hx39+3NI%rIp8--f z4>uihBue09&5;Zii&DW{0U~ZZg!)uCm{{=*xI$&6Ck>HsX&bWq{4!SAe}L%=CioczvV;d@Oo$ zcU60Fu(88tW>jrf8siFT#3Ut+y8h0_E~ z45RCcb&}J5juBQ#b*{PALhH9KWG>_F+&RCv_{RrSi|ODM2N@N|lSzyw3ae4TXNa*- zn?tVTGb4uaBv~L!j^t;y6BgjPRs;+`=n`d4R+`Xa1Rtq6Z5F=G{!v%=^jWI4ya$Mr z#cOW&Y*{KG715$*EpKuQ0nrtq#HPQQFYz8#vzF@q71}Y&uX=*z^HwWev|g|n+~N1{ zVA-)<_?|0Dwa^L(LN1#9VXGsCQ?NUdXre?k^J1khw$VhHS%1c$1AocyLTN9n`1$+y z)`{c!ifw>;pr)q&2KY(zgN?P)XD(A;2jsMqP6`1kfMH`PS_zXd!S?B zoutFuF(V}?l6&D=10=444o z%rn^kxbs?vur!K4;)J^W(cc?=k+A*O!vkSHhN{;|nxpdY#Z$oM0U9McG{E|$t1%2p z2Qle1BBycLhqUK zOegf7EDhtn{L)oD^L)0C9rl1449Nu;=(UiH!R2NsiLV6;upv)aBej{B&oH)#MceXTMj-amKEo@gfyC1GjIAHhx3#OrmbTt%*bUw5A#Bk^! z@W$T6?eE~%6CirLM1>)XYcv;VM4o^(-ufxOP0;RKr8MWn6~uPlQh2j-+nc|riEzhr z?IP_P5+F#&U#7GrA%bM&QHQ}nn(!Hh8GxCUIXi+jgEg;iwgI)!G&1C!kKr+t?Hy(r z`HLfT?f(Z#V^Jx3e}BJ*f*UP|^yH)}FTX&WA6k}9T0S<+2xI4`gIMe1UTWQ5TM7=CcB6In z;u5$a=DwRTg`&$JM|;OWe=0n~&-72uyc#h7 z1lbe-)C{}M!mund0xS7e?C}YH@kSBhd04M|%Sjz!LAnXY&2g+F3yOcPogc4qF>LTb z9BjO|*UTm4f;sMZMs_X#jb}i{cSAzGV<9;a5ZImt`P8L<0At=Ui-z9${XQ&IZ%VZw zg$1-K9r+9;;KjO0*3e890NEG@S#3PGsV)XM|C$JRC<*wAw6|-&eEtFawg?!V4bq}B zuApL(9|BLe9sx@Sh0>}C{sjwQO7*vr>+e5^%>ht@@v;U0YuFn}_5)@#;#f&tH+&#c z+&$u5uYnj0g%+%BuDNKsLj68JKU!Wb@N_>w9>M&e(B;4kJ@u{q36<;~y2s;77X!lz zjK0k%QbvJ@#~VX2*0=F#Z*@eg^JL|CUB;=Sl>3VM?ERgrMcVXI}6EO!dI;shid>ub8+lr z4&*85!VSa=-9Ey7e%#LOyujm->&wl=gbYjy(~uetL=&-sHgN>x8joj0ERC_PyLVes zCiG~KD3VWnyr))ex9Dy%9$-!7^Yi)^2u*~>K=IFYcw9Pv(%|+|=o=W&mK1qMT$$&) zi4>kV@Y27;zrP5g@pgM?AgwiHo%6T{3%&Ahj`Kk&$l9NXiO?vN&lSEa>4>f|VkSiI zZ1Q5!5{36wT+cvU_ZiE0LPnF2$uJ~t-I?L|z3jm(q(GThF&5q2thAyr8h7FN!FK|b z$Hx$bDiIAk2gZFUIiEmQoZ6bdTtNdehf%A z*zc8wUE<5T4l{aR!?J6bZa=o9Dqc55j0Hz!wbVvfnaqsiWMP7xsiRr z_iEuwSGX9c;f!2rdE)fTM6p|xS zwj3zfA4AmqZz?SP+2TigAc=ZkHaKk=PPl^DOj^j=@S9sf4s5UFT;G^9y&miQvB$q= zpRROZq(R&UD4tV@ouuBe+ir=#x33`u@g(}A4y5v1B_0W)hllOp14HqJuaLH(_~&bF zIDom3%vETcBkk;n35+B$e&Kt)LCKq1Uaws=yD$=tGy0sKm^i<4oIB6O#Z_}Ju_`W` z!5sqNKsGlz{nT2WZQ9lp8!T6kTGb#CD(&G+{tzZont6-?vnR?zW-3+KnHKHdQAQ0s zs!`I$3T8CNJICRxR;vMVP)%wcG%-AdLfTGa#mU2m!WK`C^M>rpgAA=<^e6{kTb{lh z>G$PL3wUf&q$UR4`}c%A6E%Q^i+pvge8WqW#WTKt72x`D)h)~Kj)P>LD-*R+rx{&m z%OizdE)4lR){SS2N=NaqBM&RfJjyLK8rd~b}v9RY1^ z&q7>UaVKkBRAQ_IuoHA_vIAclBsgjVcdZeQzU=#WY+%j?2=}F8`%MOWm2-s=+-tY& z+U+VZo}FTZ8uvN!Ls;oF0vv`K-=u9p^K0R3NyMe3dN#X*l7HkMFz5fVKUr{Q`y6V7 z!dAnoL;*@vqEeb_;jzQ`){CLL(f)kq*U{0DFL@|R;NcIt6@pa+@C+eb%UZ@M6vIU)Tgac1fhr&e$DS+*+l7uwecH z%3|n@XJ^}Crn0;c$*p%`JE`)ncf?cZuh7z=(NjW=5m?acqQ-)IdkXoGQwcen-Qa%& z#&f=sC$rP@I6*O_sJTjx9Yt8=q}gfAA;uH=_(BrQBnt1I|I|R+4diRMS)e`U$4yWe z!9n)v7j^KZmz4&P;pP5h$~qtLs%q9dH09g)25k(ggwRz^c6OG4spe*TA} z(#i$ykSkr!$9x^%7hu>QE;Maj3Ka8pDYc+fONPE@G4BO?SaQ}CxTt-fAs}@wfBD37 zqm7ZwxmP-gLMS&`ktgbFtRsFiL{BjIB}F`kE88d^H(n2jM2Fbmna42t>uK+ zVKl)``ErEx>Q)(6xsVO?P97|XT6MiIz@7$lt0+L;{ z=yC8I+vKsB8%YlVd0!k6o<`6yEMX3Qcl$1NHM7b z4Mk^6ptC%i0%6H{6HwssU=W%LWUbw*rLotb{r1`#?65_$f~P%ss)ig3-vJ&Y@Hx2< zXUC}On7`pnxrZLg6f39j&IdPX9nk6o45x^p$*G>)4cjf zW*?un=j{+PwWo}1=>|QrITcHU`;UG7#m9d*eQV>*hqz{1$4CqTx|P-!5@a~?MWImV zoSFPqO^7j>S*TMmN+M8W$E%2`$B_&6PBhE=kxX%@L>ZA%N$bGeEhOZmL`KC_^Q76# z!*Nh$jO{5?%elBd_cs)#P7Qpr2=Jv^sc*AjA&$uFGx<~D04CzX&6aK_L%^V}FYDsk z=l<0xm-TypPSg5{afN8zWnVnLv#XS`*e^A$(GL%iiH^c%Es@7*xFffs_~2|X-?{F1 ze?=JnO_UY1m@hEqt}WNOkE7_8RTmep8M?5eSy(!Q!Kdmrf z8@L}e#b?{0_ z)g2(8Qxl4_tsm?S*wb}8t)r^U-sig>Va8VAnoUWvlb`qxc~0j zK3q2V`=?WlxY$*zZl%lU$;|nflhHl%I5*hn>d{ZC7Kutn-UIBPYuddEX-e>z_~;~o ziN?2`5w$cXs;#g^KDgz~M*HD@mDM1Vkxd8NNz{Gm%FRdN@X&-SRdTR)HqssA1h&ii z&&!&ZcdK6*SUGFge*MIbs@@`gX+Gh5;@1DGC)A8g3J*ZcE2CcE_yztgBaLqo&Obf-1OCcf$}#oq}4APFF| zi@BP}qCni;UUM3JDZM(PL^N%?jOhc(X@a@JdbOyDoDZ09qNPwg26gIJN6@d+_5YKV zDK^kI549*ZOzYPm5V6MA%Mo&;6ZvS}ZYoTkc}=^UJrGBAu`nOxuQ}SWcjMP#JdK{L ze+0-xelq{<8JuRH9;Q%a{HaT3^%KuntgiFE)9`Ta~P4l6O^PlpHRPd9&SG>RXK zlIV$>bzXjq$G-OBjh!e1bJ15p8YADw8qGAyZISq;RG=%mczLS7R3*s1;u`fmUcjg5 z&ePe%J}D_PRx^j5p_ZBmM^X`HgHV}BOWE7HA4IF@hCk!Rn6niK25cz4p;(O`l51I5 z)_rez`hKtMk`GQ73E$9jAG^F=X6qfR@L7XjUWZPF{uS|29c#-Hq2Q~sTe zRbgjb=b`uWSAtQL_Y@y*BeU#R`)A?S7XHLt0kV=d`udk&vD#5z#Wi+3?#Za)DnX!0 zUkKcAO1?RSz2#I(05V-)so*_Z_|U-&_rj2+4^cuJ&d(~xho-c$bLjZ+CH{(P(g!y% zoQ+X|J~u+IyAN5QoE)oM*F=^630~5yHAervsilw+OwUdp$o&&7p_{+>yLS*XNUK(l z!+PJeu+;GOcUg={2$nm0ho0Wj`Kt40S$Smz!nhm#n!0_U7&U=g!MChn+9 zFIJ#w70W=!rx5F_^tD>kxQkM{-)#hi?utCL^65U@hI6CUkE+b54@NJD(^rO1Q=gOBZ`#`mbl4z(0*s&u1tnUkM0 z-8D)$@k8ahmcMuneA_2&QNeb@@1b?ZTL0r*o>F>G_m*aT)bNV`F0{N>%AU zcxd{r98Z@*{thaas%$pxXVlzpYVz#)a;oBqGjOuyy4DJF@o8rx5AnPmMoNl24iVtY zZ~wNY{uB5RE)d5Nc1k-?w@XT~^jMRm;`-F(F=LT>C+Hgo+J8bBRCyI@TPbv}YeSYHKklO^>y>!c&r8+l#KT0OV_$k2NnxA}lWrnQB$t*2EIy zsX9S~^LY4~`1J;R0Dkn4u9sbq|(;7WaVas$P4{bvs-W#VIft*jtV~-d8!+OaVC5_@p%Ie zA2e&lC7M(#4Wx65HlJv25j^~tE-U!A2jxs1HTpl{zA z?c<9}g?a_6#iry|^(`ps+ZQpy2;h6S&pOjDXL>SB&d)8YCe~f)z$z;^3?1L^gmM#l za59dpT%@x7(UHV@5rsjxtuz`!+y|A_I#=aCvlVSRGcVKw%!jLTl;Bu}AjaWCGdwT7 zoG#nD(L@xse|V@R9BkRd(2y8!?$Z?Ur$-w>bg5gr`GHSL^YVQZx=Nm6R{P`Ap|A*%6^h&O!$H|4$y#?)& zLZ7m*p9w1t{N#TTv<4C9BqZfkO#>(jzpryaIQC4(xz1v%Iel#arLW{QoCC|IBqUlx zF%<}QqaHGcyv$X=(2@rg{5sg7ii}Gdp_A4kSdK$&pJ{5HrCnl zDWM#BfV$z6LByLwT*h7C$?ndos1%t$RQWqp&6Q#(D?1Vb4$IX}fcG#S+4^S`;U=;E zI57`N*F1iQ&OT1qlYprz#`;|8Vq3ZaW?+*cD2IA~$R7r6M<~IUSBJwAXqfjhf&PW4 zl2F^++91Bv0Upo(zmZ)}D(NnQoH$UnS5O`vj_aClvrSSKu0(W&N4_T|b}e|*cnXBZ zlXF%Mys#1w&eA$2s+-@Ro7;3DV2{^CAT<>cFlV(nk!wPGf51bBmz841=c^p(MrwDb z@KuCeKaS*OKq6dD(pS!yTz2rNUStB~o@g&Wdp+DNz3dUL1YCXS9->OJ3f6H8m2u*F z8+j=RiEvFZOT@gRw#I{InIPDd>@kDSUcvQb_P^gi-l4D47dC<<2kZ{eQHiSPH9_9R z_b&$QiqTL6m^+f~6{d^YC(DEzL)7`zUVJV)nGreZUuhHIePn56Zs>5PShD9-Zxzmh zrMdIPZ!Mbp1c#c$^T#|3@gROSV999EK?42aBnMhcn)>fG#qFG^^*1?SC!EdLJP~NSOC_fP(JUY5Ur`Hsh*_J zbB4P@3GH%)!WwG2{JqpN74fa*Xw%8l%b^rSR*9}@7WLbK{uT6@d3yeIdoW8sGX6u0 zRDx>O=%_j(53Wa|n}W(#o)5{#1UCJ5*d17lBx2rpKleq!ZiZJWCcw|diDE4}H5k|G zA_@_^3qq?ANy>gUgbubX>s{{X6cx>E^(K&8la4;Q&60;zI<`{Sx}IrrI4+pkS&)hkf9M$1`RePNs{+lY=D~0o-tewCMU~?nV~V+at`a;$bBkr>CI^xGHF3SgrdG#J2?vJ zhzA_hl$r+U(-cg@{POhv3d7T+v&vDF>&!(cxF72&Ip9%|r9``iIFZy&KRQfb?mxs& zw6>(Y9tm8qbw8Z=*MKoe`TSTd1qyyF$$S!|DFDAfs@aoeDL>E>4-lcc~D< z>XXrf`OkcK6 zRj>J^&3#9Z5NxUIEbYkXIwub&U_XLUR|v8<_k)rrNKot!wH@xS57Bn5QBjuB{&m)_Qj^I;vcbKf; zp+$?Pn1hYX86|{xTBb~2XDL<}ldfiTUY(pCg9MlO7qB+h7EY*PFOO{0NX4hx?sPxS z5KKr#-t`;4@5uos(O(9!0D}fc-9a4@_kHn0;6)_eEscwg_V#)`*41|fFxGdxEC$hc za<U9=Df)IX>$FUf`dMF7nmN#ez$QJ ztIR)B@P;Qa&eh?Hl@bfa?WSJku$2UVQ4E~!Oa}#XOxoFbg1bwdTdv;CoPfBKj4Rn|)7kDD7?3qntf$b81N?i%0iGV4J!c?U~?HZCHPFstOu1nVg?&AQ^#}6i2 zK;Z3j$NL*sAS_l+L*%;MIL`t=ss~WaGXTTsAHYVyMq#VVaGO;8TVqp*Lw!FP!yTC1 zY2BV4kh{CX1<-;y(IrnV>k}ONIt*ux&Ohc-516UT2gbgdZ(=UtjcL;^N7Y#{#JIJ3 z8GE&;?+k3|I@%^CIUeCc?PvH+EQB9o(SZA&(YoGdrfB-&xa5ls(h5uHP%2zY_^TC~ z4pGx}WaW6~$Gf8B3QSbSbqFs~Q~3374I(-Jw7_iV!=#69P$S+E`@_TDjPu6CDxoH9 z$fB$)CLY{dH;fVOev8axNpk3Z31@BZC}J|tPM$}V!HPSNA9smlp zNeg40NQakIMs|Sn1^Jo8-jvWOu9>2^2$f@p@^#}Z%(RQx(yHh?LWuPfT=teUY&XLQ zEg%bIH&suq0j{s!mTFEnv<=jXNL1@dlF7Ejv>iqzY~BD1VrJe?-;9#N=9U z2&c19csP%k8;K3PkX-Kf8kMzV_&yplveVu>w68*(r5;AA);A;;aPIc z?H>(i{9VxY6Eje+ESF!YDl?PdL!|>?PS^z05^~g{qoSg$^$yqC+yrrB88@q|@69UY zG1i%DCDZrK?*2wPyR9x_du_x!{sM|wdV1v%p*@(W=i0OdcrU&itOiK&be0z>cMQ;8 z@30E$jL@F1nC&pS0bSp%7k!oT134e5D)vibG84eS(ukBxj|B1uPx%*YI zK_X>1@^lVB%OIeTG#kP=1N2$e_SLUGJ~SJKRZtY1cAt9-$8&EKaPv}?ec|LEUAuX5 zf-HkiJ}1iHW&PRUui@a|PpmGP_#TiRIz%OWK@nnCoiW_&LgDGHHQ8lZAqfI#{Am6Z ztJ>L21>C&~0nV)841jE7CiOX~6Ul4P%}SXstc`2uO_jn>CCaq-2KJh&;ENY2N1H?2 z$zLohq(3NdbUlyt3WTFn$af2eJXXhX$tNR6<>2v!+WXvQQ_-V5@^v6c%U}-}HPe17 zijEhXZzPsJ4wPWibbQc6S?U&v{ z13CLA5Q#Bu?}EmevS)FJ&u6H(Gdosn*f_R0IVFiK6HSt1lt`XCcpxjcSXO|!Bo*XN zzKBp!Q`=nahy#=)o0^ttopDtN%52&yGdbqwq3|Vr#Y!4vLKL()*@T4x^zn~G6T|G$ z@>5FL8x68I3Dj)`ypa9%tRAyGe<4g%21p+phZIhgF1d$D*0b3}>b8@_vOV&I9U*u@Wu^TAVu-Nfuq@+-RZ~!XgTttT+7wzm z{tWX!kEvuh9Zuo)FFvc()se=ze1*d4cm;V>_v)jHUUxFN#~1K+I-AAe{p@qg8JIUq z?DWjLW3d!+^>9o!FaPKa_4y15?{idpj5YKO8Uc`S=q0I80~n%8%?5a&{jC$>0#FNd z0}?Pc7aRO1%MA!05@g#^e3s|k4Q4-&+2!>n>VSek9n~`AW7HVd3{>V>qmsefP;Lk@ zn$E52+b%{__RRwNs^c%N5N}aC_6N`|w>gx7_s1_{324NuIqTh48u7exc|KhAV7*+P zQ(&r8@fT@NgumAz|9*BDb%Z}d57xzq^2?(ODj>v3d9g5ouC|P<$)=_Bu4gA$+5!khm9$s~-73hH2SL}R5aGxco=j<>5UcfJ6Ufaetl+rDoKj` z?0JF5CjeJ8_5`LMvec#!4G=nblzHPb3IXo5#xq#uJ zc(o!FeL0&=$f2d6&R-iOamQwd>>VCCviV|YOvOj!k42f1L%p{r`nk9ZsI5kn(4ype zO@*5(i7q64P8AwN1qc=oaIYd!7{5(1*ru)csA}wx?&sOEDQSGqd*mUL7@@|9>Uq@U z>f@!?Jwv**($DLIqUZ}~-C<9g7wGL2)b)`qU+%pq!d`2-%HD56OrHO)T3c_o&VL;C zaL<`Z&@#;W@15y^ zKry43q?MNoTUj$Pmr4@j;qcpuW8=+%)urcP$`Q{o<&OI;f(eYBeC;blkuO&_Wo<#@ zH0myq*Dgen#5~VIu`lOe<2Buljr?nrHS48D%!z1AACv9bs~()DpqDCxxHo3B9*$NQ zVEliHic0>DfEfvYjhD4v{C!hZ$>`vT-uTsV$7130xX&bHl1u7EQ+wmM`i;)WQF$;PI;YC0hyd7l>*_0pdiW7{clwBo@U!t z6ntKC?v5u5DEJ2ZYHe2eEf}rm?kYcdvwsZx>hd8`f*VpX$3{Pon!7T}@F&D~XrXWC zU`B@=r8V5w$evwlMo9KY`A<_@j{# z@7tRG+*|9Xbs!Qechhf5^0uw5!W!JAY z;=W_r5^@~EF2~SUQ-H34vxae!4Pb`T5L3uJd?Ph-fJAev+ zd-^|Uo7~QtL_0pZ@t3-AQD5GYtwrT^TfBW)3HX~2G0vB) zaJdV)<+j<4<7{KnN~t9jSG5f#GJoIyDV#O2ws#+2KCRL%b@!CO+Y|;P`&a+i-J{Izu9$8W_~LQ>Aq<3k$kRk` zR9AV6M`S0kM-1f((Lk%_TCL5Mb!%(uq7TO^%j5aZc{f4zW>~di22CyduS}n(T*ysa z6(vCYoi5rdZDpeA%K?|nRvV@h?M6A{y#5S|n;noR2LwgN8yoV|(p?%_5Nl*5zf`oi z!5v;Wb}F9`kNmH@%w(LyGa(wXrz2LVrJ73RHosC?%uF@9UtvPq+E*@xt2*UR%xCnBvC}M0^_Q1{_>#$8yy#Edq$W623^= zKvmCK2+digZt_@sdNG#j3YcNXU8~l~!y{=Hz3bH2PfjDyheM5km_9sV%wYAwW$tpx6Pbz z4YTnDWm6yYTKuC*-{+Lr(7Vq3MXre+9U%&exk+1D9%|}$Wr=3^Itjh5uNP_-Vd9kp z;XAaPx=;c;OvEzi-=7uHFAA9RVJu_@y=JJm>}38~nn@=IcILLbGd3QV=|0d2H0T@JU80NQT4SMcTVs8EB*^nj+&9XK1ghq+5D@AzC(@W3QEJ`nv6KgB zAG^22h{D#j+`uRCUT<6mOBLOZT6MzR;E!M+?weQfsOp5}%xtr4%wX<6_x--?Vq(41 z9rSL#l9gN=c;76X9^cxoD3f;RB5Kd_O23Hz&+l3&C5)-fJd*^x9mARzzz)or1 zqDdtOhN}!gtX+ySudD$TnA~y?&dGtl--}!GT%QgqSy=62((>mM z)^eX4sGGs)R@Ii)OhA1#w0bEUq22CzbO}pWdBMg2lpgJU@KK0n$U+TyOlk~S0YrEU zT%=5$SH}w8b8k2Jq|us`bAbO3wSf^4Q|b?gVyN#t57C4qYPVG5<^X}vO%K$~zzdr< z5)~}2)lI&?8t%_7nL_)dp}JPpS6%zU#QnM|Z&K1!Ajj}yGsI3mAHJp4u{>NIO zTa0>h z?D2r@OOTAvEA5o9GTLH&dzmQbtYy4vC9yS@$uFZ3nsg>RXS2UVz)s=uLphR8Sa0Ih zN#M}_xM|~KmtYNGVQdRzSRU2o$d7vJ&1WFppC-f}nsLI~qUv>e&R{t2Lvyx`Z?AbE zZt5RtjM07@tdYCf@U$4U+Ym4Gqm^c0QLn)-K!c%xk7b1Z8DH~pJJ)%#y`I;3{v`gc z6IkNQ2ZP2+7SFSHP!FfS)H{kd`L%DZ*;82$iAi-`t3%Q4Arfzr;`GC~Fbm%PAaE#> z8yefs*>sV*vT(nyTpo61H(D+7Hf4lm7^~v;&S8*epA--}9j<(q`P@8|QV&|XR%1`U z#T+4QgF}x&)VKWHvYZc*?&mkLn@&2gT909>`wIHY6zl39S75UE+T{KL#Q7$|yPybE zS-5v%Laqqsn0NYxm=XP2a#(2Z_g}8Ut~zt_)TKDZ8JU<5p>&cCy3K$@|)i> z?LI;-y%cct9jWpn6ayh~w=Py!`v}92qW+*c$AXi1Vyadj$2S#S=v8zZH3j()u;IIH zi`X%OW=@@o$cM*bM&eQl!A;DKAeqW`Z{JtZm}(0!mVmm(atc)9^jyD02wZM%Ax=Jt zSRH$X#H*?Y`*2LHW=EmamV$1#9zu$6=$o0HrF_hvB)TZmuOhNVHJnCI_ZP~WIH{C zhj4O=J0m&@xA+zJCrf?f!#*EJ;{X!#|Xoyr)+3fsjHw{I#xF=@kP3 zuKVQuOe9|fCZ8nxU7Fu>WD>zrofKWYMm$(&n69tXTxQ$4%<}oDk)3m+rC&6K=KD-^ zQ7aFU^)pFngB>(9V*7r)4x{n5%+JboI+{}@CSHd{Ct@DGzv1Gu))3IN3@$r@kED+y(zTacuORaP=Vr_%A6`ac5Mshs%SyJ3L-4 zx0E+U1AgSQZ4lNo`EZ_ZM~L5E9f-NL!mpzXG2%BTW?^xgV|@fwnPXyAp+%L5AV~-O zF-Wdn32jOhwo>~=euy>fsm{jqd<(H1+KFwksrV-)@xeh1r6y7_ExM=ys0%9Z+ zsJQ&S%^?l?(=85j)tm=LW?OI$O7n(#$DqL`F7^t=I?&Pv*7@A!o5pBfMaNa6D#Zn) z>Q#!TXfzm+CxP2gix+F2>U`&UnuJ$OVHb&+kp_!?g2I3IV-{#W8W=u}umbdO?U0kKGL#_thQyGWFkr4P@1#naN_x zx8s?5>c%O;#$pw17?%yj+-a*b6U<`8n6 z&F;O!KubX>Jhp~hK(%%dQpYhvn!r+*)t1;pI6?OKz6_S~RL8Th6qmD0sQdoMo_4e7 z0&5fz18NKJ>>*nf^41siWq)KR2ko8?w=MZqGSs0qB>=5q?k|k4ecJ4GHl(32SJBp+ zxWEK5Hra@__nF5Nt6G$@iV`99DR{M_B_T7ZIO^V@j5L3Kb63tOC3UFP51d99*RLFB zOeHgBs5?KmH~QACoreLAvQ><>peX)lOCB?|Hf^p3(X_ZTOn7+s;1gtUhc&S8xk!&A z(;b=-H|O)gX?vb1BDqX#2uCxpRh|<(H5(g7ZutuTc@#e2QAoRqQG!(S`Hc1v_y(6) zv0q>*SvUTKe6imB?w-CZ6TZB(P}5khBSvtA_3|~#y;$|Bil>%O`W)6ylDe^yX?qB@E`zr7= zT(h<1RYE1nBezbnJqE-jeN=KgXQ#<~HkQdC+kzrQD*3zpGo8)!xnnDIke6@q8PPXu z_Oq!vKDoHTf3@oOjMtR}uSE`(&A{-31@n25F6W^o{udB6@X`!meH|JJ&OQIm zSP<_ujm{fbAwRirJB%Q(qw}!M5W8KmQSU~8{M#sOsbH3v`WD=>N1NYvMLCbGFT1R< zNM;jnfezJ&=oD$8g6GB8H{gP-0Ywt}d;sm+b2w&Lt@))uk}i*j) z$=D`WOxZH6gaz8UBHy!_WK>N;VI6-R*&r2-?V-P8TDz$6*)L}cLR{XjQ3WdKH+KC~$i@+5-J%8Mlfu^p253y2k zwdnGlDoCE$2!=*x&L+l^!vJG$xbcln&Z`Mu6_K~((lQG%*7t<2Px>>vll|PPyk5;r zsHkmh@4snvpnfgeB{>pzChynbH`Aw~8nQ=#2yAqFCag8bTyi`h@n?gp%g3lCx5sr2 z&U)GUSx&Bp?{=4GZERz0-;;HzXtt_zxC5FQ;x7-KsYVT# z7nf)iwArg&h)d-V?jTEQJGN8N6PL?Q2pEE<^4X9Y`79xUR^b2b0p1@%;~OB5$)X)8 zu3;)9IvTNv*X|e||1OlIJ&D(8F$P)?c}Ouu(gN*gjkAiJ@-kG? z4MQh1bEY`C&h&?C&h_hx7!RPosqfGN^w@z-NZ|r}R-QuCQPt)G%HN%T zx9Y+UpG*arE!qhFtF*D|zXvwZ^blB%k@w5o;nl`O-fZ3AiOex=;tQBU;+YtQs z4dy^#u2dqGisv4qzE~e&gF#q%g*B9m4S(;AVDWH}F^`7XQcKki__vgT{7&E21VztB z_B>l{b(i@AMwMThnf5`|Euk%m4Feb-H8SA!)zIT#8=~ZEkqt+yz_*4!HlD2)?sxEA znOw)9OJCA#)wbd11#Xw%LwpYi0aIZk^#4}!8=N{@C7g^N`LZ4UO-Q3n^PdSy3}vK*hC9n^2IdDb)x?l)hjP!q)+5?DHY*vAmQm6yzjea($@m*@>(E7CtCDc;mLu(U0I=7yBng^W*|oz>-GNkPxREH zbL;l;RXMsyZHttbRr2!q(=$vR9MO=k60Mp`LF>&C-s_1}@Dl&qPSP&)E0tiNUx%hD z6Ka}x!7taWrZW)~UA}Gmy)gf0KFL?Mui8>I4r%{>T?7I+DKVE3qVNBi+CCulyQ(Kg zn6|8Gx!bNas+}*o&aXo7e?A*HDt^nahJ_AI{dn#mn+sKi%FE_ML-NRW;!AHPR@_&I!N*VTU0mjhw>lhmi(W&IV64u!WtlFS7|BDBk9Yg;8T^lKj@#c z`F94GSm-RV&SOn*C@Jq+L(&f1Vou#Sjd%Xv&Ig z-lp7n60Aex1nAQ;+PNDWztWe3>r0@6=-zxu8da(0gGvq!uDlh>iBa|Lj-)sFWy0zP zwD#1beamPnL6U-=5`B8}6=HrH=bxQ+*K(^xHxL6)r!U9dV*Fl`F zkpG$`$DvX8x;FkhXUnLkGx+W6q~#1`+FApO>Uah!meN#3ufcR3paA+z?7E`{0DPcS z5oDE0s1cpsY8NWZg$UTyH@3R|bElLafC>J4Go%-l%ZRSGgtf-RKvAI|x}aXq+b|>& zpo!93^p!m6#_$)y>_qQi8sfKnxp7w>a3U?1y)#}Rx|jJrC7_|{^QHaIkNYrzGy7Bx z)ZDeuR=^_p@yeDw@2}qZo0F~FN7AqTF%vIuF1v%5jSVh-O#{V0`Lv}57!Fg)2_+Tc zy#xMN*RhI@mvcDJImoHEx<>cfb zGCQCUg2xKkSr}$-v%=V-Ps^7O(syZCd(E<~9h3ggmI_#b4`SW@=Ro>Jz00OEX1EL0 zGN87cCZyqRCn1~-*+c1gU9O=-d|W8T2xr2;O(G(dshwH*?=6F!)(70RYw|aG_STR% zy(F{s^I;XmWMW**ROrB#)YR!@VHo192BAqtvDoN10 z;K=CAW_VROl$pu;X^V2zQ3FhbTc_cFP8e7P#E`U@;<0j$8i{}zG{E6@AsqG6!yL^k0hVDv^F}^}(hI z=$OwBad1l_|2qJ%JRWYBE0PZiAiMps5W>EJ0Wt2khiy!5XY1V4@wH!nAa;%IJJW!& z7JJZbjY+DYu6v|HzKWD+(5;WEn3!LE%ztbAjEYP-U^E0EZT$N2-C(gJ=v3=rAGrY% zmwxA?;zDe9uRBXWA|xaveRn`)I;>%G$B_7WdcU`Jd1LS*&E6}08I@%HuhV%)Bo*mK zh4R)W0?r>riAmByh570F5<8s5W?3g2EU`VuVTBAI?To_ z0+RoIXp+%sH8Q}#+fF@%#FMumjVU18*m*;BJj&Nd|n}~zGt9T!gv9f-1WSkF%QN|z4 z{>Q&w7Vu6Fut+ANA$JRzm=rDMuma=R-#tBf6%-Yb(9w=F`? z);T(&H;0`-@NcNMZD4>#YwWtJ{pVYt6ohys_d+)lZ$2~FO~GHdhzd^r3orY z6>ApoTXxe5)SI@FX3rfwwVFhFAAP`$2jg@%1pj>V zaA%<}!`O_$g|EhCp7ckhz1-c`9p%X9N9s!B>15ffj(ezoix z){12;HPTvl;`*V5!j8Bz{M~H9HW&Q=Vh_NqH+qEM#s6Iw6a7;FVE=7M62P1Prldkl z=JG#3c@LP*zqPqIShO7@@{mJf=6jiy#)Ahzj1tI95J#vQK2AJ*iam6T9ZHC!m64Tw zYt|kmys--2(DL%~cb`2IK6UC;f0d)L2NLNbA|wD->6e1qE?>EF;mQ^J(&MD~;&JD^y35&FgOR2n^gjMh5VIuHn zzsRo2EF|<`IF(M&`aD=@n9y*nu?&e)QBy1K{dCXB(dObD$?>6;K$a&YCXUZx%vRLk z$q0f}6g^f7Uk;NS5pgzrFMGf`=Z&$C-pco~{DB5TL&Kc>3?F?iUfu;>yHk{ulsOvN zvLj0uYl$`*aXtm_-n}!zI-6Ik7eJ#u=)B-P-)VbnP=pz%8&W;+N1{5eqmxQf9-RnzMg zGFT7)qHDu8=yFi5;KUw*8W$%ww=%auskuV@T6BSAhSyf8=mwgdfrchOLv0A1t=7fu z9mnR3TS#$Kq~S9X)Zc~vVl_<;DpOzZu$cz{Xymmi<2=>e+?yJC-E^DQw2gGt20 zlwEK1k_OA6kuTP*0PZXNvILRQgguos?}(w7jRafCC0tBrtt}_fxy$|G2VgHb=E3pt z{rfnd;{&T@I9tzfj1kW}5~Z+G6b9x;F$Z9YfxcRN=m?568-7Qg$`cPm4j=gV;r1GF zJ&y)EaP3g~`3lOETbk9q37?yr=Zv=#P>3_H8zg~Uxw>O!XK!!u!xX4*qBK{?R`p!{ zP!;joe(JLxHyH^@$@q-T23BB4t}+S+KPKDNaInBo|7w!4!%+?&TI+sC7ckVmnIkH+ zpQxnc@tutnL9Z)}-lKSo4nRR?8OykCj0icD)5WEvq%5^V?Jhl!?^Y*!YpERL^A_7p zxneTqq4v?a@p$ z6PUxpWY;g9{*x|bl3R}@1yC})$qwl!UES}$+e;})e+s$E zBfi;k@g6I`3aUMUu9IrZi_$2;K8sg1U7XX0MqFH+-(^l78diV*tfPT~_qy}Agp>X? zLYVPEoa*6~1qR@^#xivIx#&FQ+0hGx7B$3V868C=2DtHpKbw~;nJh85dFX6IBr z$BAS7uFAiQeKkzwa+~M*&$dY>E#J9qLj!$5ec6IHD?{wyo&p>3)2BAP7#1-(dBt<% z2yaZ{{o1OkDtR@9)o+dhwgD|a;O+{8+iV(IrhoC)v|@DA)l5@0CpCP>f{z}{`i`4R z)a^}qwx%f2A^mXE0TfPs5_{vD2K>8AW_CXL8x0r-*VR#S0L871$2)vjFXY|)V{b5$ zZJE%@dme{6F&4oId-YW5_107m>ML1;Zb8?nFM4jk@sAu}XPv^^K^^T*IL=T^GvW3_ zB0pF-yh0prb@8>{K$9~FJ2@O!_O}aW1YY_X{u(Wy*Fu8K76X6tUr~Z4O{bmHV!|0h zx53mZylhtrmB{}4sQT##a?Xnt@^RbEbG& z&&^gE8Q(ZvGY2e!txz=7SqY|IPulGz!bHCT27b%)`^p;y9A;kplAfBX6d z#J+6N_*yNio}jh35_*yA?{0jT%T$ZJNrsJgQ+?hCsYQ;aq(Rr>i=8x*cdDUr&D23QBhmzLu2m;5z7@7 zN3$yZOpU7}ICLLAyTkw6Y{g*3Y{jnGPPQ_c*SKMF$pc*(exD^& z*p6|SZh~cCxaO(V=9e85?zpP$CQ^%KFW!V{$Wd^o#N;)%Vyr@uaifvpwcwaK#hsl* zgEt=xs=^R#YY(m*Udi)JI82|?ei&jZ1$s*c(gfqm^`(iosI_!;IY@WMUc@|t#gFFK=k6zr`p~Rif3f^%;CUwuB2<99Mr7lS05^| z>SOetZv_k-{?-T>>x*G9#TJMT6M3MclU%f};BDUl8h5N9aaW@D12vedBy%#Bkc?U$ zUUZq(!2_$f(8#h4h$C!U)q4vrC`G8Nrjgn8_xEps^3V2UoI!kbMvXr8ZNfPf))Iq8!5x>x3c^<5 zjPZD5)~aS2c6N8E0ZV2WZwQw$JKlfSQqTRikYas(eUPv5$h6$Y-hc8nY?;3GpLN$I zB_*>FS)|5bU?r%JN@b3Yj1>A2QeXnCWb#8~_xlHh=#k5X6i_G{ubP3dFGrAVEwqD@ zN}lhZ_XyXGsYhYa5HuTwrBt7cf0Dwr+l%6Tb%Ye1Z$%P>a^`B<3^Qpvi`}|4WGbZ! zCFN>-3*FqdgMwS|oe_?r|iVt6~ z3DU-TpBFO6-fAlc4qi|C#{RRx+jFxvxovlxY3NVHl{-57G+NXNFy&ROe1Y020)oiV z{3g}m_HtFX4vQBHV}uWV?YuZ{kP@24)qREBLlA{iqYpOlqx(*#1l<;Ae#2#Pk58@b z%wvg5OV+1PpB}H-iMXyF%l6)bn`k(06&1rGeIFIi5XLfGMdhX9ZBOdLl=GGaZE)q{ z#cTebJMB$D^6za%Qd@{&9EYwCxPZ$mDpKz8S~O+7|CVQZcd06ET~4cY-}(;e?`va( zif;0p6wAq3ZdSk{;TebwR+*Y}_!Jl8)WvkQEJ6SHt)(Alk4(w+4DVe!P+s9UO!D6^ zpgvnu*1u69vDJM;{?(W#-(QY)wt4)t+{?eYO#c5~qx@rf$>*u({{6{NAiRh8o90er z3nUc#M@9u%{;#+oyrKC&`RZ<-?k>K@vqVHh(2mz>{T+4?o(Wc0SKrN2e+&&t*1U#w z&s55vgMOOIPpSoKs600`WcXR@Rw&^6t10LY8hu|@RyMvO(uAJ$zbRC4aB_0uRjdO* zNi7S62!&p+eJ_*imVCN!mWzukxhV4NknZQLZ#Ot@+-P3?*x1XSX5;+djB8|`11?q9 z*AD{|E?OQ8p!Syi+&4z|oSg+&6`FvkMHvLaVI;e;2p0)RP(bsK^)@gBhI)JD?Lh7Y zNGpAX0~{0>7?|g9u6g{@-vKcb5)$qM`tJW|ED3|9=V^4A52NMNFDHgWrqX{e*v4xe zNTBwE2ftt~=7##pIsPNnlm0QZ?5q1=!fEFsR$t2wh>O0);_Q&CaQ%cFuOX&L23*!ls%^42M@#Oe_n?w+DJa zF4E|VBcS(cF69*Ap5G;SW%H6=ygRJfF{IR-K`Jc^>Un2qUAYel@`UMJW?D$Rnau#c zru!V)46uPE)&*kDr@O!P#GwlKWmHr!z_3qjaxyX`lG>D?^D;LY!GIUN{*$4&xY&9q zj~s#?BP$zQ*lfSEN==CF&~Ytb(RkYED|af}jsTq8y)4PbChNuKd@sW$6Jg)6ZMy8X z{ig9qW@8)gi|UQB&fGPi9JX2Bk=4faqAg7}XIkG<0G1uVhmHCVmmEjanG7aAxvlCb zr`>)~q_h}6nuhSc4Qj=ik{JaB1*WKgGvqga_$>8jnbJM73RY)&r_D9DxG1vkEsKg= zPcmr^4&9bP)dBAD*!^x5o1X5lb(#J8^-w!b4i5QAD+uWOOWD!N$HyK0dvu_Rar4$K z=c9e+vQ+`)GHFxVyb-({t0JhZ9q%ZgD=8@v8pTbZjxC=(a&&ZD2F>C_HzE2+DtdY@ zF;P)@;0UtqdGl)%Rc{g$Eh6VZ+t}@c{UJUUn@QsN%Qlcc9`FhN&{nJTntR}*m4mO~ zS$aWvE!V7 zNsuFrxa%E`H*m*&zi$uwTon+|K&eU>)l9uNCWHO(3RI4j9jXB1h4(e`EVjQE?msA6 z6CKXD85*5~Wht-*xt005B@&@Qq=z(;#_uer*0ElES6MeJ$137-zTBM3P-@~!{ zj6HH!0Js^ywY{BNUM^?@GarOZSWBv6Xn(FGzz0w{Y)yIh=gf%>CAIWdyT+dRPEX2qd>*(l2 z9+0~k71cXfHk1M;0EZH8(ZoKypV3vK}*&UO;tcCWo9LM@H7$V2^(Q z@X+M=o-8K=#*A^8tY(RlrtpEwY~&-RiHz^18+gqYWK>!9&h=oKAW&uV3Ym z=7Ycv#AmKd)|BOp*xEgjHjNt1mwc9lKh53O1fGO}NKEKFU&H=eSYXMGm8MqBVxxMl z%8Z*9ijfEIlg=TDIN*k~_R^?u?$}_!8M|X=^!#kFkQ@8X-LdA4U%Drk<|+*xx7_m@ z3Xav8Z+Y%MDLI0~TI@=A&9<2mgM5sZBypd8 z=t3VPyZ@Q~dZxZlzI*6qE#q=eR80s2qs>?q7c{# z;$Rz2C}nLfzX=Io*0#A>Tar0xMe(hUYqDJ(4->o!LsyPd=gbtB*E<%Iy?*JOzg z7WwNI67baOB9pj>dIU5_aeCb^IDd{@F<0{}4Chjp$^k_@it_-O65#c!bK}*KO(%u! zuEjYrvLQ@~&LZUJOh3Z`%=RAGqle_;z|mnjifS%ES^WZ)LCN0plHJ;pO9sUvbl>EM z08eHgY(39h?b9rhkgP0!aa0XAOSJ=l-y2@7Jp#7F9KXx}vp&OqhBQ&8FYdUhD(%p~ zC~w5J)q8P%KIm%zk;sxa*vwtnyaHH}<7i+W`wf%xRMuPO%-utYcRtY?5qImJtwrn} z{As4h7GCna5ZthNb|_Ut$c+HXTMQv{*di(@ILzfNfAVo&%p$|FnC1ya61>Q!envPF zh@5)i$kmwe<7x}@*n7UeE2JipDLWTLeUHUuqb?WV&Mmw@r6XrFmY~ir(ui0S34&&0 zS{iEVCY|4ZAA0NPs>(ShAwoi0{XxkcG5jR`jROJ#HX)NUyI!|Jy?n&sm$*)Yq|NkG zp(91+2*(mPI{orL-^l7T&n-@spdgd;E|H4Flk`i@-R|IqzB%u>=}Ol|7gSPR`$}!w zoKNh=#b6tfPeG|B`g9Eq4W?eRZRFvj?1ikcsi{v~YQ#~0`eKWmSXDc=-lU`$WG6;u z5h_o;cCkab=H3!Z{PyIYj_6Z$YMR6CJkzW2wkzcPD^{enaYAlZ<%UvI8&%%ISUxt7 zSwpsCv4{G9O{yn-+{Ncf(CM8Q1Vk;(GudZa#bMWt&ocD#SU6|R_DkII zJ@zW$t!R#!o_-vAF21+$R|$vT=-AjpbKv>7UZu_KEY*0`L>(eHG1F~mGYy9uV3kvS%%JsoV#*!f-Ts_0ibN6P6v&#R~Bxl zjFqA$)um_nZc`?bjYwKHf;LO-n!H#};i6)3FXYv4-2YTF+y4m&YhmCii+E-YYg^-J z$>ZwplHN~i8aOG$a`&>AArU`Ph;_rCyl{KO{xm@ynTtVRc*R&xwe#$gZ_*BHJC=@8 z4qXQ&x#YdQy_Svax$H&hZT;#;gMZ)!?TD`IjFFKgFN3P|q!na9=w8@PSiw_(a51I` zis(hX(yA2CT$u(dUz_KH)Xu58B7Wr?b0zJJ&A@i~VW}?Y%A7P6e0_sgGpK&W3y}2w z&X}=)&ymOGVfy3+j_P$VlO4p=B?tK5A#tf^_SHiQIg*m1b4dM%A>UCO8bOIxOfS6? zv)Tn4DX=bFNWDUx)IWRdV2(iL`^_b{0|+Oar>485K6XQD%nTqpsfZsytJoL*N1^jy zgx))h$E@WyZ2tN41r~?c(`*%g8)q?30l^3@x?OQl&Y+pG`hl}AxJfw%DK%@A6A_gKh`#>oxs?{=&oc(xYCZQHG(e!d8-+4P(>-Dseuy_1d3n9y z+?AB|eEja6Giy>e1}KCZx}52EkCGSk zA+KtkCV3Y;*U6sIQfkq4f0P0#`Z#g)oW%4%Gd*EmvLat6f6t{#{EIr;yVUh?A22~X zD+dpcjJ|#{@xnDS$Y5N6q<7RAn%UD9lTJ^cHUmcM@op(cLTleAKpC;w>8+wegOAC8 z4Mv=VjA`okCsz}$C52@Px*{Bs9vVHDA$CEEW?mJ7vasi%V;|_28YosH1;h2nGsuXO z33kbvQ}*9WWvGm^h9mKZx1p7<9vz2J6U7tEaf1Jye?sv6`>OH(59$1m=>EU>>VUji a9D!G8`Blp=NlPc-M?psQe(}A>{{I7?z{_<2 diff --git a/analysis/figures/F3_coverage_validity.png b/analysis/figures/F3_coverage_validity.png index 5884f20e9269ff0f95f38c7ba9db67b232c7b74a..d7c6f43241139e15b655a7076f0758c0c78d361c 100644 GIT binary patch literal 42631 zcmbrmbyQVt)ICargoJ=d3xY^@x0Hf}pmc|HcZVn?A&r!TfJk?PAR#CXA}JvyDV=v6 z-|xHk{>HffU4{+@hr>S4eqyaT=bCGWt18RmVUuGcAtB*CQjmFygoIKG|CO*X;1v;< zdol1$*i}}`Ro%hD)x+4?9O;R%tD~)htF4s@t-HCiii!cX= z-T(a#b_eGd9L_&cPv9yy92K-(kdST@A^s!3$Qpsq%K!UFMoPmoeQOpyMZ@hFbN3;~ z?CdN9{G+>z6&4&wn`1(YvBSdgzVIa$M%c*6EZgktgV|Zm>s6n*>kJQf4-1dQ@5+7> zgH^a9POINDnmoQ=Y@cS_>Nly-dKm^^y@Owl4zURc;Y%nahz0Q-9Sx&Y2EK?1$m)sU zXEjNqg~x~&kc0kzzu|6anPyRmVzt_+ zG#!b*uVS{;{pKx`8r=ik+D~W^OFc3D4KAzq+&0FZR$C5}`E5&j2Z)ido_on8j0GbDESy_ZRK=9vBs+6b!<47|#PSk+QJp=5g(^QlDUXUSjzjeR6_w3ebr0rCt)vi{d z>ia;Q+Ehp1dY9GcgUv~m7`&^EtIP9B>v3jyYc#LzkDA>TA8fzxZ*pvMRML z9Ble=17@#=$CNQr6^PlJV>+>(UXK|fDgt$?^mD6}1 z=~P?Z{q|1acroqPX>zUI^oB|{+cayP{p?$btCNQ@I43BZgGp?Q}Yx$nHbw*|M?-}xG!!L&Q-t(cE{GUM9?uU&6K$%V__ zw+xDTTR}arIP}d5%IpdXvkBDlYX4QM7VzBcYf4&xH;+8 zvL{-~Z6i_FGugv!xK(FCU3T%Pka`mN#RtcU7cj{EevY=@KxEV|;w0L?=R0jS(wzRB zYW+%6mFA->M}MY97g;tZN~*?+)O(0H3`93KyHp~GAF%E;|BYVxGgTEw{de&TPXF3) zE}M|cZ`|V)L8nEnjfoPQsRox9c|8fN+D~^zi#5xbCSRQF{_<|)wHndN6!kLk@Us~! zd_QP@K6ARiQl*qBIZq-^B&Qzwto z1uODhR=+2~C582lU*ps0h+mzUD#8qPiq*=56D5(bJeNuJ-0s_qj3}6%jkjBTL?po| ze6%?^2r;ZuVcHo!k6lGa$aPIM?=a>hq|J@Vu(kPWuEpn=scx!n-oN;`KZPrIWqq`O zRFjy)K#xYm_D4C>&g%Dhoi0D>fKE2uTHLz9o(Piur%mp5{YB~pIlfn`gIT(;t_3k2 z{MB_Au(NbGCd+8-v%a{rT%B#IUA5FVwaM4*)K`6Lbld!JYy}Y_Z@T<4$~;dcYi`@+ zH*9r>>!lb8)>2L%34xO{t>49$6S_St0kj3H z7pH$a#kOD5_zz8we0V~!^4fQOB%cc7FTH$pbt%ULet@gie8K~DoypbWvQ}qLPS&S+ zVd4^3Sr611FP93Rm!z4)q4I8n$S}h zz|%5=7RMNQq?hw~h(^XPUy(af>XLXh4 z{iLstU5*{Txo7o=zx1QJ9(>CXX)llVvzyyxxz`vHQFN!9Bm?mb{`B1Nd_o0>QNs_z zy&x0THlvPU_hUSjqvOIODv!E5(a809Qo>2b{Xb};vM$e$UpG;mCpo=q;W&%-G=&w} z*ZST6|EvfqGw+@L@=NZm@a&spe+9Crvji=s2ScPn5(>o;s+hdp$( z{%$EW`cg=0cpm2^x^=9%a&JbErYBS+F3o1C$>JIVu3HalQ8907LBuzkeK43M;n1?R zztW#n@3bU$r;tX}bExr87|q%b_wLtv)3tVPUKqad7iXJghy2Zqwl+SP59!HLMLbGp zSLe%(Lz3=S9{&o-?#A-oy^`gHZNBM(OU6k%(}7m@<*0LJjwb*9|9c%w6ObgxNXn@m zw4WJ`v3#_B)2?2v48?w&-ciM~BW|}+( z*DF(x{C8J+NZxzu)tRTGe{QvS1q+aSmMZ2`+iAY=B+4_u_dJS>xDf>t)wDpuV>`V+ zrsuDz*w=&yrpF1*VkeB5d^XSh^dOD>nZqn_Bw|!5JU%@bRq33aC^vrC((1N3q2{mf z8yDtf6~ zHM>f9@8@W%PvhqH9E}J>nUiOs zbF^JJ*H06sEP1CJxFyvcyBJdSUzoSK9lH56iXCrFuWF<(tYHpreN+`Go&KD9X9oX) zxJ4M%pOtdejgC2$G+rx8@tcpn)~i4$bX!EWAfgd;lzh?qejuvp(GZ{Y*bvP?xyzT1 z5X{5VUl0b~dhRU>9SSDwO8e=YtmVeQ%-y2Usj()pfBO3U!`C;xg&gKYpVZfda)=d{ z8~I2~zx1Q;xPY`WRw;UPuraR4p;=~75`IDWON@JhO;~}vhyN$+Q{}~um2RtPU!$g*M@Llh8F2&Ne$n+!h}63MS^QNv%I?+q&U6d@Rnp@`W`?=38XL`)E}4>C z`IuKtc+*2nUrc5ON#GO4`wJf%d7af+NS1WM=!9nOcMPAu3V-g|#$}xAyziM<)kvOjs@ZUktik-e*sMjjna;E+_XJ6QmrRSMq2IHZM4QP{Jk4Z^O-w=3 z^Cf@k_#W`f>Y#Yj%nkCd2Sn{O6m`C}i8 zxW5(mJF^t}d^a#!nY<* zi~#neerNiEYohV2-0n}HU7!)=<^de@*Sd26m`kpEFuUgiu1f+^jC{hizXuUeKx#axZC=$HNZGlfQjeh8zYaI~I!JaZj9j9G3Na#`5bbq=_ zzjk_L?vEYnl|j@(4+Sf1{V>HRF|&-aS4cjpZp(dJj>^WDpPUjtXVTpz}(ikB9g zU2|Mcmg?W)L33NMCZ+syn^h|tQ;HHvP3g8s8jGqo=DceA)Kelt3BkRfrN=jD5}LTE z0mjKkIWwH7ex^yRi2ub!RwGHvY5e)6V}RvaFX~aN-_Pyz69PAkOo5!WGYXYdZrSVJ zH^an~v=w$0Zyag)zs9x$94v|Htb5wN8O6-$&dcmDvFXw^yS3fN@045-BU{Io%ZzMF z8uejuAsS!HwVpc|aKiZ$j(P7*w%56#Y8@TZWb3;V^J#ZKR`Sc)SCw4?^ z`*GR}7{mx1d#Kdixxy$P{IzqZ$NWG>K+f>ZuohYYmfoImO#65}RsNVG|2M-U#8LKb zneRT>`0R_WV!h1^!wUBJiNj8=cV#`G$KV4m#FG(GoC6Tv5|e|oEP&&t$FTn0164*d!Z z+gdM9mSW;s4|5r*H%5gvQb=drom_pCn#Db8|Lv3S+oq*?kLz)0ur2$2Pxs9)b4~#+ zZa(0M!dh@3Dfm2V>}yG3NZWE)_xnb=x6DwqAJ4~yk*VW8PqDcf#IYhn?y(&Do*}|V zE4vWUcDs3UyjZgiqx^&@WZG$BAcGg#VP~~fH0jj2?eGh1gR@o3t!?%(XLkX8TH=_^ z$?NMM>$Xqk*Ek*Bf}A4o=sA8&>EX%Em=hm&-;APUf8ldQPT!<8S=jhI+lVh)%cF}> zKb=lG`j+zLUCh-{6^TU1{~MSLtzvhz!g@WaeLLlo&{$@1dsHp$Px_*7`$+@NQ}8p~lOqv5)q9NwzM6K}3*dsp#| zAdozWWkix)zcCN}=*=ZT5dXyOT(uwK-3vj*LB#?P#Hssz8|KdzJ*nMaodr6OuJW`TQ2m)K#K9P_z5Ss zeX&N-)4XXr7kfwV&ZBA!6jYy6=aZ$bGjX3IrN2=v9fFBLt4_RD-D(cY5Ej?nJ8XKS zI;odU-@APfDhMP z6I>FHBxS3#$IL})7GuheI?tZa?Lt`~e9sb=YtI^OJKs9nsH~X88t!`j``fz`2EE-X ztI-dRb-4#PrSg{|(`gBWlk;jt_$mfjReH1AnfA9lHZSkUepKV&Hp!4*8_AEHHV=*> zdg7Ut4#AVnFKW01V{@hOWjv3n0Y-p?)`Vfe<;t#q=0%B6)nINz*q*oTubMlXuTAEU zek8Gdc8~s1hus4x1RZDN3EqTlR8UvZ0@9q^Q$nMUj2~05YLi(!m#t} zItUCj{k}Ex-hVw;2BlZ(sVls08LZAWd9Ya+eQdw(j8aMumcM!9RaIPlj6=>^Vk@W> zm;8|qnZ1wJ=iA!4c+=CgDojE%_od83`0ajL;v3zY&85VaQ{?m^uPU;8!Qg*9>)}m~ zS3thYhMmvG*@r76enKkdwxO}9fM+50GsK`bB{72V#k~a6W%RyZAUuRS`^CA`dF&Wu z5D(Yd=~CV3SaQ@^Ui=by#)jJv*1asCx!`Jab#Z#Af&RFLwvLd&kdwR4dF8Q~>>o=a z9N+V;x<-A+2=tXf@skBqeSF*d>CsX`w7#9P-mkbCXc-N&M5bp)?H1W>75}^=`_@@| zvfGJ&xN~EgC|B>~xxKBT=C4E4xdI^fn-2o$lW}g@{&|MiagVLpAeS7|S$8%y!My!^ zzEF*x>TZYlXr;Mg(=pGI!#66X$dFZ;8%)b{vgkXVAWFElsEXgHigAjf;ImFhdX6$M z79%*-t@|&J(QEYDeVNCS>#Ue7WH;^?WU!4&YVwKrp?ls?7GvJ})4|WJdA|lgR#|Ny zON{;Jno(A%?pZ?3YSqZnHBLEay{$%5#whGumHOPnP62WseIXD>EVe!w?_gf$+co9? z2r1R8jX8xI%hflGx+0hD%RTE55YTM`w`#GS;%4)=+wmGEUZ@%rck#OSb*SFSw5k8ge507< zkg=uwvl$c}YQ;h^oBa!q)a1nB(t%{sJ-bqf5%>+Ngu0H}=caE;(jcosy;8@R!zWKb&dmT#4v8 zRA8e2O*+{PnLZZF*~XgLOXS?`?l9*rLT%0eW8xbxzh;GgcB0z#2sgzwNhpj~DUAqS z8?VkMM8C1rNf}VR{hgL!zTFH7O+_$2H2ZtT2h?a3OUh4N9Y_V=E|G)r(&f5dlQB{^ zJk5!bMvXCFW+{9d_|y;(b&~yTV`{A)^$L^O6sdL$R>S9RrCx}SJqtQ<1+it}*(B!6 zv6%Lxe=vJTrL7oDF`CNmjH7W@c4zas?rF(gVZ0Dj4MoeX&~J3*a+{J$W|mxLP2X-` zXF;PcEO|Z3mQolz`d%(ICYnlQBaHM^AiewZy}P^=oQx{|`+fI6V7f?L?1m^ux96cN zy*=PdSo>Vm=)U#y)caZQYfo1CfRIN)3jG~>KjQl;jyRtMNBzMCe(tO=dItw!_; z1yzS(QLM}Cv2VZpBUT4jUiy7`(Ihn{9m$Cs>#;--#Ka&2Rz%TwBBt#+iKPkj4&y}uK zjw)7)vsm=%%MGXn9or|1r9tv%M5uqD6ZQPsA-Jh$n<3)ywd%!S{bfcMC1PATWX^kwvf7QVMWSB&G6)3}kDNEQ z-sw)XP^4}I;$|??#r<@im+BQYd;%RP85qtZEiJOT7M%a85MEP23YS6GobWCIaXJ~e zm{Z<3-hb27{0C?yd3Da%vz4!j4DPndJkFT^O(cg8lot?H1L2?KAgJ3r2yo5Zhw*UV zVMfmm-2{Yw=z&%&~*;;aX&3aMr?%a3|fmusX+_5ivup9&|zHzLd>tRN&aNelBPQ#17njh>yElp>G(@J zi`{4IXAQSeJJP7`DMIDiV_ZD-5zDrV#fg`yHus0~VsA<>Qrs}o%>wisRD@ISH7Y?{;&fITx zU4M%~R@QpbSO4ad@N&-9mK}<8-Z?J?#{o(IFD{eC2?q#OV)A84LW53M&ve{cukG!^OChw``$aqr(kKE=kSZn+<-!|_}Rcrq1b10wV z<>_jcIlybK`th*di9{AnUNIyql)m`;lct}_FzABe00k4j?V!4p5k0h?@S7ETb2%(e z^Qg+=+w$j7oHIKTDa|hRESrf>alne!i)_=qQ-$3=UfKI>5cf-$zYIjBxe|Ob3gE*F zU{1bLD)$8b#qoTAp|JZOvK1@m7n3-c_t98SI%>&9J+yepG&%K4}(Kp@x%3JOLS* zOfEk5{!4tDPrC1=Lk~h`Qbb>dpvgTg({K9Vm-)s#L)h)HJTezI2Iz;eH*PWZMv$-^ zbW(A0Pvyws;4#9Lb0>|pPp%;-W*BhzaGGd;(&F;^`%R!ENVCg)IwXYB;FjEG4w2KT zC$1b4VLhEASMZ%?{tF2xttq#WP9JtTSL_&zQSt}4!D(y9B32(+6c{Ou2{(S(j3=ZaUC16R{31Do5W5*;x}Ex1 zHt^N(-kOo?9f`T+0iP(pLHB^q$qZ;SKjiVSqJ|d*F?jP6TE#xE!DI-Tm2GlXX@47%I-6n;G%j0?*R1)167thKj^M|ZI+ zyz??Tj)4;e4Ncl8-J74Rjox7#OOKu%rz4?d(>8f$+xyP67a(l!4JxS{${h3@gau3b zL6S3beKBiE4hd2bDI6Q$ZNkYEC<=bXMHoqz7E1D`A55j2)mb(hfo*5_mGJC zVKU#1<=`qb+TGfzEGol=8sKum9m%8Jt6!JlAT{>|?y12!?X&JNAC9sZq7r{tU60&` z+8BF+qI|wsU1>*-Lfh-~?1NIOYxThR%94{DlBLBRHq!>IW;%>mGGyLbp{MAMI=JhM zkZecY9{PV^>E{Y&Uta#%np`W?sMdky<0{4e6g|h3F<~KleDE+PN56rWhDL; z7(sl)^xn^BTI`Ipcjd=Z%pKZ4X`9NB#F7SHs4W*QjtR;!>TBP$@9g>ns-gnans|{q z*9E4L1R0;A$Kq2u5s$;r>)N2oFO2=BnpDtvIc#E#O}jGh(7(5?+me&{1~2H-GcHRF zoK^$FXLmT$-dxfy_6ifG;shk0P;^fbP!>DRyb&x)YK{zIITcI4lA&fmK~~a|qEbY= zc|y>mkivm#jYgH?8KS5Z)$z$uZ-;r{q(_EENlqGR@z1`~g3P6whKgh~D~iv6wfsWF z=<)VvXSFo!&CtUKP+FueSNnoI=s;0zJnMb*rzpNc*kjw4?bq@j=iRI5J3MmtU*8Sv z>0(Gf^RP%5Q%ZRmb|_B}y~_X9xZ#|7NeQ|32BZF61^H;MBTe7lsmzg{FcxNEBf7h- zQNKQ+$T<>3(Jg%OvHnEUI7ULrIz^(+65b)uF*YeFqZn2LIszb6nm zcsE`9^IqCIa%%@5babDE$JZf&H*#&Tl<1zLXScs3bVY0H#$sXCUN|mmyVBWgNJFOY z=;f0`N1@xl>Fd;WTXF1pclM1|-nR}wL3&lPmD!Y2b0bhtCl>;p&DTa=4XdYHbcJy{ zFLI99VMLvoRFmYE$t;VV?1$e-Q*f&k8-$|M(i|hcNaH#u zUq;*h@spg`*Eh5Zk%CLz0ypj#KK#aueEs>~yQRsUht!U4Qr#ge%qa>_oy9_cM&K<> zdGFUTt(kHycsIDtKPC#w#{6>*h4S+56sqOE;y0_XyO6qi{ojj}ytmr1SblP$vS6aU zs_z)GvjmV=iI9NswDJ@P*|xWoOS+)IX+5pzTJ-q=D)Ak;*)Nj@K2LD)d%A>AzK)B; zn0@^aUwL{$(r~<%dhbRg3(0<7JU@x^qn~ERi-WBvi#S`iL1A=|_iDSoSe5u`Uk!zw zMSM;dx8B&hKfU?DmD6fKMVAvBbcT=x$Da>aKKu3oC5S$of*^`QKsh84^hfy(M=p~t zto7mC*Q5PsP>K9}7#4#_NBa(Mjg|EyxfLYjXLzJ^L0KHks6nP(>wyCPWz7EZe&w+1SQwH$$iVY$c>Ju zlCYUwyHduc{?~L0SHy7-O+YV{zFQh^a^XqYoDXlB?RUq!hdggTjOR~Jm<-Sk4mmi^w!y3Hl%a8P;8QWz?weT$As#M^k#If^n^^^r2Ft^XmH#nDvWg-wJqd zPbl@G+K+yCbG1MloVmTG5j=7aBPx*I90fDodi>)f@V6ZFqns&F#J-y;GnBxyP~(Ie z{nZH2X~)4D|4Z+0ru^9!G)a-0obQOwCBH4@74g3c;2bqv)T@x+PW0F$ zA#&*04vTwyuisSTFA(#$Y`XklPKmw%vNLfm=6#?b-=x-vhP(q#Dp!g-hLZm0TLNCI z86La=#8qh@siJ<5qnF><0+YsY<%1Iw1P?7p z3OQc}yOq2#CLXz&A0F1{iU6*j#IW9`-PE8CM#|zhgU_&)kcAc_vIz+Uh0nTQ8ey{Q z)h~=x651{hO9SAlk--SYLCY3IxATdZ`s3D4^R|U%T1}Scy?3Apf5WMuJVA;HFVSf` zncEkIjT)OtUi=%JKf|+$3oh-9mx-i`kvodF3WZvkjHZ$~3~SGudyTt&!ql{VEkVCftT#B%x@P;P<=q1KTXpA2vsjow%VP3{}A(u zlw^B^vWyJcNhUJj$p`G1iqHPXJZ9q7zZjuWr$fYBtY9rcEAy4$?)w<#5hffUBz2PvLETx6Rz^jY+oA=&#Lh-> zD_|EDOPI6qd9$aJ7)exmH+;OOAC-ypspz`dtZ~ogHdY!@$``>{GOOFWP&uP^LFR}lr@qI4R&WO5g5TzCLf@(O!jD`z*bc3yAHY&7 zje$#+z2g(l_(ad(<^&Z9Bj;VD2Y3oJ;icQGgoF(kf7Lj9Um~OCof|I%N!d8w=%v(* zvb`(WZgt7?;=4>ERmVLP)O2<^F9wynmdFG5j>m;lN}hzK|I}+Ish1G8N<}~Iu)G)o zj>jhQ3L6y;{R9Vg?FZ$Ij~x=hGOnBmooX2*_Nosi^*E;k3bO6q*+_gA=cY1ay=65| zUMRfdFTqG7lYQe!;A$bnRPgGlfhS%)z5%L5EfHay(DzKS;=gJH)r%>!)mEblPC zOr5x(pk zgu(ddjIHY59Hv`|2-Vi4*;z&gI8({T0|g0LJkf0ztd;V(8=t<%Ayn?C@XGIf*ww{6 zO(ZKg9;x^2&lp)p)stm@3s2KKDX}3vzg#7(9Z{6`a(;h^GhN>nK2cw=Mk`wj3;m69 z^s|*t42izv+Io*D6q=`lEl&v$KCp1eR zA}MuAD;H0#BZD-XIbr#QiDjfR);@GlXZj->aWjf{*La{#2GeeIa&W0!v@)TQ8A>8%bi=I+8^xDwx00VrOu}^tqN(x5vkFJGBzC;{&o+o zLY*ug_YU_i z)!nB5hBip6^#t;&sNb0e1fFVv1RMCdRKCAy`a+9DC=2+eR&yV+Q zxns{@vp1~tG5pW(-k@`S^E)f#1b;|_8HVY!mJpkV`m1fNUjhG+pqtDTwI6(UwJ6`+CaR}U$F6dPM ztk3no3*?fP$_?p&dzRH) zvyeI}lbmBC4QvRZEdbFv+t$}tmlK;QgO8=S7QI)$Q?W@$p>#1tO2P1haI!WF2O|{6 zEM>hV!7T6BqC+gLgr_%IqI(~Qly%6NtjS|12^s{n96F0$LUnpmsS_qt5&XB`e63+- z_HzKE_rWjRy7<5oHUcg{j1uA{6K&O=&!~vX0pM+I|FiY`PF8hiSpsh^5}o2N`B~FHYG%d69KMp91q4sCO}Jl|4x>glJw(tXz9oUB%ySJw$1cRvj7uHo`QiaiTSj)N6`pa8`dng81Y*Yyh|QM)FNu_7^Sea z%CFi$8TSrv=gooYjMaDC$F`h3<3j(Y$`(o*i2(eZJD*y9z~;4|`EeG&doRqB3K8U} zyFa=EiJ@YNBMl@imANY>gri>$c(nJL90w6sZc+*7wHdbhrUtQ~r`OLh7Y8+K4k+_x zU0EarF<6qFIe124;|UXu zB76d2x|q*%ge(3s{mrfWH(^lZFes=CkM;u(+C=m(@!7RAx%xu8Ov`c?D{LtjDlu=3 zBzApf;D{ATEp*t7@E1@Do#=-~07VGmHNle5cvMoC`-(Dq0|{~CH*<>%Cqk^nPXeHS z;T|0PnQCHJC44>L7aTNx2XXfvCn%KaHPC@6XOgB9s(c5PAe&i8@Ilh(mC9~}GX|$8 z^2z{WViDw=4_E}~@3crc7~h!;_d3*sKGLoi2Sd^aQTSXDv;R?6^zGy5Wb>de=xHgm zJSDkYU-^~z2dc9|L`#4ch{xdN&4Ys3K4>P2h?1O`1(Y%Of$@pxFL zFwY)vby^Tb#w9J{Duf!N0=V{jW0WPN*8NU=iC$=Fx{K(A@R)`L=Z)zltO|C4nf0-s zYc85CkJ>y@HmJrFt0UVs@_DUG); zceTTk9q_{cY81H1as-KCpi#iRJ_Y}njWz7CxU%e`L11;}U`zffp20MhLS1iikGNwI zVs@Tkn}3{rkJ)`Kh*{HvL9VD5dpU{vpu^D5Mt8ZUpY^Q*D7 zi)t{E_BfKGb%;<0mZoZ<3O|HDB_^nY^f_L(s+sr4K(&Gzpmy#K0v4+j9oZsL?iIc9 zvh2lTi0&W45fD?f-qF>gCSq`!-jYHz-+}kL2*37ui4FsBvz80ckVZZ}V@Y<_rk7rL z<6RH)-ngqB*!NlRU3k(40))Z3o_p7Xp-4FLDHoU~i5q>$-MZkQA96h+ZWWW;X0`qz z92~BO2=H}TFCfn&YHN0h0ua;{=I!@>5)9Be)IqDpnXp=|^^bC6=iLPVtZV+rLQMb*3)nY-5xeNTtF#xmUsHm`8+%=q72()|to0jfF7L|j@{ z)`5~$NlYRQvUV3*&05v>(sA3PK+C;=&C`wB`+>ppV4Aa)iard?`nJ`BtA)zNxp7(m z!u<=gP0>nDC_lf5%H??@^S=F{g{JBZxnOc@+*!7Px$tRz#PndCu$#3_ADcYF=b7C7 zNl|lwHpttvmtF^8E4VR1Cvx$;#H8`fzLtN)4r}6J^pYgY3vM?KYN}k8`LQp&7p%v4 zpomrGuU;+I0#!uEJ3`KR7*$6bnZh_iS!8aqzgGSuvdC=Cx_c#K;|AY)tx7=YDP002 zmj|9s`x5$(gg>G$0Iv#l@+aQRasy3nbU*4S?TjFl9A*j44@-wob+H&xcD zcUqQ{r5Ufw62iiykr{6-{l$HO2WhrMik0w-YQH9Gqt6UrGL;OGuQ%K@L2CV!QGsa4 zDngdCw*jf3TA^yID3Qw)3PO?T&*(94gDd%kUE=HZaWw9HX*AjA zwAN0ACgEU1K8S)kymjGlQCj`aYy*YEJy$RAtJEtZJ7?6H#j?-b6P-<$v__hklAuzQ;_YcZO(JX{&j_oo=V^}?@ z@&i4VGxc2WunA60x=5r7kz=DfL^~BKhL*RC$<9dq1$<(2& zB_oclw^uS8_+6c_j2=$uv%+r?u(1>L7UUP(4|mScc6Y^vq*i9)r9x)B0)KQ(l-ZpGoFY92g=zP~a)%>sbs+)#gP z@H=;}E_j0AU?n;=BR%ggS&|+?xWomC#0Iv)94*3vBD0LSQeJC6%M595oQS5eU8@3t zG=7Fd+pNZ0UaO~whBn508t~6VQ;S*K9qJS)Q~hf>Mxfwd`Dpi(I$;iQPi&{5Ef`GB z`p|~R0Xb&xkQv6QzES@fVa`HWm_l)Gf2}a>p`61VmLgV#NN@%9%iREoj0T6D~=K`l2C7w_}Rwy|);81#Ks`7$q`XVgKxO%97(q)sC`5X%MO(=u|V+ zQK;emy9_bj*bA`61S}c{8p;UN&^hft8haLLKq1A`i<2ZA2q_ISkyStOD$Ku!K%)*Q)Ti6+iJSD zWJ=|UjEq?t;5sV@URt{p0F{2YWd`htk9fw3TpMk1_RSNwh&==s7E?fk^9oeq@JeBm8{>K~o~3=e`%9 z1YcbZgl|OuzhM~>;TM@_LBPrTVdsGRO&sBWdITjJUQ5|uzoBh;qYxfU5Z6}kxtH^y zUJT(M%v0#Np(XSF?UHqSFEh#nAWC_as%(t8FhcreXnQ~G9z`z!Pu1p15Gtl`3$u)j z)H;l)5{%+cOckMNdaPJe<@4vyh+Sm%dJ=n>0iHeZS8J~)Ci@Ut(@ch7_5IOZIFdLF zOQmdTZKv{}P=&ZrzGv;dhHoANdN!MH^-J5O-n)WQ+2Qg!P`?>>aHDBMSK}3p3pd!S zJ`*G77%7wNR%d1;R}u$fBkYsSOF$jGEjR&@|I`cQgDd)Wxj`G@?-H1uxMYZIC>j^p z6P}3FQYuIoda}&Qjfp7Q4(y%mnIbHB(d9sD-@Z#^Lh`ah$XhTi(C7)g(y-rPYpP0~ zTGVr4^BEnTfal5u7Nf~8UL`0mmP~^AP_a9js<^<)`(VwO>d`#}&V*gi2NCO_>H+jM z=_rbVvu$Qg2D$g3-3_YYPfK+8E@6@0)XXSAF$$5A&qUG)7APDb0usjnb&5nN^z{RcDLICKy8jBGwe>g> zS_zmQ6T#WVZOMS`2y6+k?^ypo+Sgg9kA`4fok1FqFf6INU#vNJ2B-O9r9*+6JDujE9axI%WjokV(&PJ+mAmqccNp z=dcA(YrGr4CYw?C47`n(RCb4f&+6f%DV^vcoU?A?=*zuvA#ltHxWpX4C39pMfKH*$ z0R-uj6)G_F@p+XDVR&#wimZ%`MMriouwyge6loE!u=!<(eSyHijeJT5nLi1jw+(s_ zurD14L?f+14ubf_6#(7K>DYkXeSk~zm5rM(kVz1jHt-_@@v}AhlDOC0_x?XPUuX(6 zr(3Te4ZM@woB-)T5x4`WQ!>P1Y=#wn=_I?922c?oUeO3y!O-GygAj7!!4{Cx^8XUGGD=%Zmape6D;u1J;&Ata*rYj7A5 z=3G0okikUQItZ@&Z!&FH%At?FIJvL{v4w8BcEDv4?^BQL)!njG) z$rNiz$CF)SNN$DSg|~oWyZ@0c5BrD#CeErWnssbQ=uz(4S{UAXAo8f<=pqoAOY!cv z0BAg}beI>%Gl-!SqJxNX_*GaCp+g5#Mb`veU)K3E8fBxRQul(W9MO9~-wF2XdyKx| zsU+t#AtwGmtp?<-%J2BuewzP63Lis4E2t%K1Y@CAh@Fs2@OiO+v(d)%YW@(vU5wiC8dBDPiX~)xsbh9zAv8gGZ_wN- z<%+&7? zMR^~c)&AtctySs#AW+UMX@6!*p_Y||a8;5M2G2nNX==HGw2Ni<2b7K6HyE(hRyN1) z<{%Nq2C?kR--W<^b09tJ8?PT87R>MX z1}ek`i|-og5k+`Jh^Wgng*2{X+-@lX&w069R@w>aQ;$5H0ZK)J=TIZ1%NY6zVSp`J zL57$p$Yx`#=x`tK-OU}B^$~^^8EqC#yP0}5UEG|^1#lj?CfO*Jp%SW<*MrI6gT5G4 zLr|apt|U@Zc=CUd_TF(lwtxS)riRg*N<&I%P^l!67Mi4q($=7%snF6ANkycR(4wLt z+B-Cew4<~r8VV^E-sz8}Bu_n+_e$LI0*gm>rrJdg7@Ua#kRHSTQ`R+D$> zE!zy?>m`rfL7Ia%_-D4GCB}Vt7c0)ZN7=yE7KS%{~ zbh9Q>Z*x`cIh;p(9&1p#Wjy_E*cJ=BPKUvpa@LcBFT9zUuN;MxGIfoiRPHMDD3AR8$l?erqd&KBNMhc;zU_zj!+##NaanOT;xjOWJIr!zW@6=fcf~ z10F=|Z`+ox7yoB9gB-L$TP*yhZ?rDopHRCkaa#1;7q}IYJ3DZ9IZBQV#9ibK&*qm| zW`7u!4}dMs-^9Ef+yU?`Tgg&BLTdEgLqiEpMc*R(d@dy0W2Y5n<_n)rI9Di2dnwDdQT-Z`=M^;`AMHgM(W;f};M> zUYD2ccM3zjZvWlnTI>jac;cB7`I);ctOkaBMAXlec z-g7j1sUDE}f`+@xXj!aXfqvyVKb%jjvVo%7b==nzmI-7R1ee5>(TYsy5JpEJ6ZB-L zBLv*{VCz$M}0Ye@Z)oYk{lju53OwLZ@q$iwh=-b6=!6cbtqv zXP^lowUgPKVQ1mR=YHbnX*zMgMr~26lH$B$%xw$(+5)x6^%>QfZc6ByGoZr8r#;wM{CC zDDQYbainpNnKA7b4n0+CI$8zjKTeIFo|JgsNquD`&=C~hlyY#e3K82+U5?t@RC>l` zJ|>)I<-#L;Ye~LeUHCzfZxiB%;*QL|=|EAEud-2&#j5Qwy48+;t^C|QsGmJmh>ch7sW&S8au*cCo^7oGwgYlc;;SvsuWDa=QV@q|&Y4fACwcL; zuNzVurra4L#85!c&9q-T!h<)gQbQ@l(R7sM`XjX152|?^Q#sp@l~Dgp=xTwx!$F||#Jh4oi`U&ND%+B|l6zN0bC32y@SL&bw)x@U zvQ6DCF&u6+;RYTpXiuvGsV<`8wc`t2M%k4`i0;u-R-4$A&T5-*IMqyZl;<<|!#|KFlUSI65V2>kwM1j*o z%@K`+6Win{dY14Q-FbCur%nU6(BG=a#)RHqXt0yvU)M4(FWoVqu6;$FbpG?_CwC}IfC zaBG}67lgFp620;{E}g~={=2TkE#R!EY!^`=V!)C2#ZMr5QPcT+)$#4*jZj}I=nmdJ zPiN&~etYN6na=;KKy={8-Uv!7=v?ANl6k$NIEBNDxU)K*KYv$_UDnH2X)oW*Qgzf$~m7Fk(9Q|4p${2#CL+#>n8a69iIHnbp{151|x)T zSNNCiFSgORwe7xwQplnS=bwC)frx2I{!!9)i^Ms8Z8?}=Z3<5$t>UD#<>)Ag-+iaQ!m#PE{Jps?5b zKrgmx9(&L)eao{*Lezx*aL{3xS;R$ccSM(ky{OY0{!qa#hc`hzoFgg!r|uiJ8#KCN zU6~u2(IV_h`2N(FChudmVa}Z$5#Q?8jBUnt3~fa|KO(8XncF0=vLc4W&-d*37lmr=jgLcqFTckj|Q+(xzJ}d-QTP2 zw6&kM9M< zppHjpPwMV_-&ubkkVrtL)`(LJn15c;T^Hp?7W!e?Oq6{`kvx-DX&s%9H%u|-TTrBt z-PQoPED?+IH~Zs4LlTbMU(PZkGwI#Gzn8d6O+C@a-rRNSJ`viyjFsU1NF-2H2CE#g zU3haIK45ra0UONpwd?zWCg3!?J%*(9;ge-{IMFLCiaNd|q2rh55XCk|; zkFjU!$=(VL`rFwxu5Du6$wmp0Z#AGp8E7pY>2$?SVF@R>sCR@m z3ur)R*IU4Rp=GQ^N78loNB~znQCfHfb{dIyI9Tc{U|zNp78}gF5E9}hjg=tkIu1X8 zj4S$tbosO2l>O@j&n*JMpPD(-=hyiqO^BaKW4J6LjEciv27uDO1Lg$g<7V;w`R>Fv zyCy!zTZXB0@1X~VnCC&~^mw-+{i^kLKU?VqckuBsE!0wrtUe2F@-|Hq3X=lpGs0fb z(=zN|oE?*Z04jmV6%G{l_lHjfa7rJSv=ii?U96=RJ(Ag!$c6x>dOWK7SaA!fmtjHg z;3xZRFR|10`3(_S(u$QHOA1o{@;3J#cF${AHv}SyviRXm&8AO>#X3HowvnHS1FFP` zBW(^f(AAfF6G0gHYkI@<*$_y_VBt34WAihag^qJhn@H8)g52CiK?a`#Xz@0E7W|WK zwbXyybyNnQ9STH`W`M%MIZI?&ZEK=w2ND1vO%h*K>e##p(~F3#NpbekCu$FlWW9uf zMWrGkA>>^yQRiyca!Fb$r*^}_njQ;ibQ@U+f8>m05|flcjdSLqwUBe22Sc8Bt?D0E zdlffprfG2tc?6ROi0TYdDTh&3K+t;qCC8r8m!E7-4(S1(@g_SKE&Jb(WYSLsijrba zCCfYw>g|^|R65l0Bi9LIcrO+uV0!~v{i8ojsBA5kr94JQ$yBE>_#;Jtoml7{P1aJP zcCCZ+MUSLb(!*i@8p@*0ZwjGS;#Ba?`_fZts3Za@9X~+UQ)4tA2qxkC z+dV6*2!N#csz~`BWQDa)%r|QdfB-yGREBNzCA_Vs=BxRcPX4lr)axJ$t*>?!LD(BZ z|A@KPP)`e=>hW5yB14B_MSC@cq+G5QS zZ9*FFG?EJU3W=hrv(p%ZcFMu*qZfs({ICH=$tCoEnk*4iNlCJ4J8_Ya=JbqQ*?KXxn6$Z4%02xmh+C0e0ICyKekgLIJKLAT;U zMXu5qx}YY2xTt=>4-(b&8LyMyxynI)N^iXNA(RpEZ+!Zzs7O<}UjD2u0YAzsP_met zEa#t$z~1Y4*|8plZd>M21LB7*3=XZgpy@=zR`c3VwC^ZVU>>v)@a2udp@5dSJqPoKu+3-klzCJw|8yd6^5cX6!lhEzQ zD>hK{meFOS_+_`KIxpdoQmtj~@nZJFWr?RfkzCobw=cGyQOkDJ+9$lV9#&nQJ9k50 z{eY5~aH)j9^siG9^AXQNHC#%)WAj7Mn;r4FYF8QT^50haFS^c1C2~@!X4Sxk;^W#@>X#E+<>a}$o+Xai=KGOFw@$>VB*-9Ck++2}1 z*1V}75;U&2D|zJFj9)V8T+%;s#{u@Ed}(&Iy;r#`2Mc!Cx2S*`=H5DU@>QG^*Q+BS zGqJ4KX6RQRkSMpA=gFWal0qn&jhHXRZaKVP=wQ4JPvD+b>Z((gQ*<(pIi|%{W@_l@ zhalklCbIZCE@8-SlAYqUnz|j}siV1QV=n@qb-Rlpu4lRN?CFRPZK&G2Et1rR(UE(5 zT8UZ1G@W}hmB>y6K#Dk4g}yL?dhG^)s+acyD4rYugx8FC(>_&`4DZOM->#(0a@mt$ zZ1g1o{hZfjkiTFy@-{a`J?BcO?`I(2U8b}aWt7b1DV|9azz3mfrh1)^AQF~)vVg6! zlUMRiEywhi_J^@3x!*^BR%yDlnp1wcnj%lTzREso+NAVFuH1)g8yqEE^KHMn(TlB1 zdBpKM;z32`LtwqEKLC3>ISs_ab9Xy@=yi)XR$fXKo0GX)*oHlugAA z*`_xmzy}8Ha;+ITOS~QCR?pi{>vj%r4K{EE)Vpj%P4>vZ@(eYIjyezM2pz!9=QMVV3#y$DMI)VmRL_x% zO;eYSPR7^(r0dS$P?2bNQ-VV88B5#~oC={+hk=X6#;Rf1XC#$T#+~?Su$Hpx_T?+u zM#EttYWxhwfW%*eNMQ{Vw`jiLi33(YLtoI5^CrL~hevXb9b!(G_Nxps+Z_qI1_4ns zTMj4!QiV~U?+ja;d`I-0|9KxUR7RI9_R=Tkl)f3|{=Sb^Skq0gSZX)Fplk)v5 z|53N9X=W1RBi@Zo&^5R8q11SgWdRL~{U?Lvf}o;`S6|!-G9O;MRGd@sJ6#%oqD!NC zHAU$p7_o%U^r&sQGX!PEhb@6y>6d$Z7W>R(PoINcHDJP*i18EICO^M=su$B zmpE`^i*5NTx(Z4?G#bViR`EROgjn%8DP8mlG;jS1EMcCr2b4dlC_moI;a4_-`63KN z18R)XA#75EP?YDMeo11%iPrVz2>A3_Th6)-(I59YjUfBlT>$rms#V=oZykk*j$%>J zw^FWv|25X>qq@w=616kaqS8FL14IB4I;F^=#%JZ-A1`5{SdMI-5_4$OkG9fFZe)f- z^Dp(;8&E5Z{PSxBlP~cYDT+dU9tzM8QK-{!@b;3-l$)|{$sS+cHha&JhGEU&<*HoG?H%ceea04Ay=~r8c1~GH89FmBrG+D1qC4M_EGj*Qr?g_x zz0YC?P=#xx>+DBBaK?HFZr2FzPd&wqg6hM!1%OG>P=Smk1k*CE8TJE*C6;Tw=%V!z zDre(Q-|(&y1xAKhX#t2I;R`6o^;Vp2e3~59=fJ#N!}RV_0qdlD3|f!6Z-QQ^E@%B| zv%tx$7#ZBsdMEKPpnaS=N}32nw`ps zEnuZYT`>k-?A1U7rZ{;z64?TIglyj3>HWEmVG=b}=5>8aH26tVb;<8sEF7!W%j$AJ z=8IEhtmZxUYY!}?$yaK=;%2WXEW-^t0EXg=VkcSy&o!Cj{mbIF@>$mYP(6QNvVQpv zp;sRiLHaaD9Tresma*LE6CB|-=#3tO@2Py>sKP2&CI{ZlMu^rB&B1jP=%buGgRWfp zN%PHU2Dkt6fw?r(tQ2|lN!7gC&jbwfu))mkRx^w`6-F_lO~B*tG??emRDBe-#sVOF z=Un$G2tmPyX%>yKc<>AM!znD+<;Xnrl$pIk-AWmVQ#B(kO=O&z@>x(x9qhxKGL33+`u%P+{J;#uvQQ*D*RR!WmqYTE|+X7louS*+lep!GTYt>y*;5l3(0^exg^EQY=pERxb?83^{0JcXfDI(nkeQM^SLYI!Q zavZb*rS6zj9>M5FdV?^@vm5yNEr?X_xA>qz^487TUe=o)1M2>;I=k0vlfjPL(XZ=r z>dV#NY|=C`&A2LjJ1^tvffga=C%voJ-jnkfPGqCLY902%m_bWVw7p)K2favS7=Ie; z&{wF_T>RvK>ECgP`xhGX?Nus#3FvlT>&RFp$58(8i`1UaMb&G6 zF&AO*!_#WdIB0WFj5n__0pif9XA%9>8x2(XO(g+HOxUgNXF4@8VV~(7 zbMpaH|LU9r4)WpA^G|2!wVqF=C9>%%RPo$;@4FrlmTrFubj0P)ZJ2y;BB#Wvonhx^ zGK-VCVqFb^*!0PW%Dpch#*x4JZe1sL%~Psrrg14_mf0T{St(w9z~o7qx3LTBy<2}a z# zLRxf%#?a$ZARL|=r8OJ^IYsN0&Xt+F_$cu+iKDM>kLDjNNWN!>=Jgrd^BHxgErV^J zsI#l)n>8<^$1gg?t{DDpPag6(=Xq40HGiiQ5L3(fPN=(3F}Zx*wzmNRL5y;e8|`Y2 zEw`P#`eVakZUg_x@6k@%ygvFLnBA3PcF#2PBkHWL0@(~|QH49)2Q|iDL>#%eZmVh2 zfj^OBZo7kzBsT?1hH~ zcHI$WbKt6r9Dg`^a$G+NV!F;b1y0ci1^?J$M@#q>67d>vVOtA*OmSi2avr(z0D~9^XP5$F4JY6#(I-X?KF}U=3QA|qE%H`0|LjeqvtCM zbd_1Pc_Jj)dmgw`)^8_P1^R5(^lxV>ft}3XQm^VlKo@ZqL75BGCeDz$^B=$|+plhA zZR;Tbde{U(;GT?v1SEdGZ4Y#}bji?Dy+mh`Z!RrlXSL=C7A27Ysj&8I-A%~Hqy#VM zP=?U#AiFO>qwvg^w|Wx7Xp>NoMDZ$X?{>)Jw_C}~d>yc)Gzv=o#TBNNzdv!mcI4h> z6N)%KiI#B=w-6{>@e5T3tr?C`^7zkUMn%RBc4a{-Xy~Nldat-*$MnAByTd0u^5{Qc zRDE3&;}~H2F1m-rohjJj%qTb>Pg<+aG01m@dyox8wH{8p;u$ng;WZWyR8G+=g z;0GIFd%OWT#n@L6%gpJdI%bX2dgG>Au11>$h2S;p?|{#u_I`t$abD@A4Q@;$3f*Lm zfS4@cmRHGfAFG?7uvs^8S)l^HZ*Qn5lF)RCT~U04p6_}Blva#@(58NCzTK8@NohB~}lWTy1GWA_(X1)0u5>-_&*n*7tESi&<2d z&bZJ%GS|L{+{~~$H4X3~9~0j2@x?UqKy1q0J(1c2uDanuIHoW-Aqa zUgZx02GGQ8g9@7GwaS3Ie3r1zZICvvHC4B6ZlRnU2aXDn_*T^M9bxn$-5(#25euC= z9~`1`$)UjwvRc!~+W(45%_QlsHF`vY=Np!B(EZ4V3d3fRj6A>Gc7{qPfTE#Miyo}L z=%9k~`#S4p{o&L#(G%b4{f(CMKM6iE_yoShQAz8^8<*BZSL0C%2?-RIoQF?|>&qS} zxdtP4_PYAqENH0SB;te3MFdRk2N*fISk~>HNN_@5kVa6h!Dgjo-MBzO`jl_>RYs*wa-c zo;`i4<)!VC^IyuYZ8VJxglH$}9~N7q*a?OZ`sc<-pr}JEzwSN-LoJ7=%ah;jau+e7 z1LS}D>|JOXM{FT)*79@qhsI3@44x`cE}qc%wiC`*TW4KdOjeaA)P*fr1UZpi{fC5> zv|bs%&q6A*`pjRo*9ovIfA*BFQkjtR`I$*z3`kY(w@Y2+CyUwrFrBVPUw0ap+y1oz z#or)9QY-O8CQ7{F<%&_YJNA;uq}W38niAQrGlS<61KluKyNN%!N&I36MO~|L z_REtun1rM_aQq`?zchXvLQ85U0V8$#xbaFzu%j@y>|*JaV@QW4BGHkiq~N+YUnk^P zelt~QkSSQCLRnN_*d?^T_VT(bE=<@nXmhsnRWWR$-eLVofkP*lQjZa3-QrzeACUZW zNhVNGt*w)4=Jvo2Y|z~gOgh$ubEb?PCSvMu1tFe($oa^u{J4D)yQJsdqfn_$%Oz+N zZGie%{}!-5(zgP%AT`FjpXzy`;f81(7uiI+1hVlBqY=PlaQWtH4BlxOhnlplA(Sd{ zx2xO+fm`*>&s5gZhQ~3q1D~G{I4x-Te0Bp+{n_9|wnthY)FsT>@Jxp;c=-UzJ>y>(s!`eD(EMxF!E?ZlT^9+&80*PhJ4_)lX1#%K+l{tum;XT&CTJZB6?^r z7xUJ3>Gy~AQ?3`X&?fV2QAK0pb&V^)-4J1mv|_1`;q0N?*4l7vU|$P^B~{=ObEqABE)$N^(RGx7-sWMS{C)2#NNxHC!aua39_AW+dvmS@G^kc@h z%qGO~++nGSDU9ubbhP3N7k|{>lBH#XPP711&fw zGO&+TlIi>sQPP4uD;Ifjz542P`=YIAXfFYf>S*ojWiSRMB#5CVSzEcO`oXW-Vu$ta z1`w22y*%hMB%KZqnn1K=7ABR{gOeWuO4DI`Ti#1wT@R;$!%Q(tA{U#?vKtzY)|>M3 zL@(~mPzZm}RRVqV9|#G%_~;JmhL2+d-6;RT!Q>5_((uJ|(279@z-e+|%}s{DP87)7 zEo+8_)TARM!g`53$-XNugT5}H@r+0@UA?0}!iV!jc`R88xYK!`r(TCxY<37ssTqD> zBq9_8TA!<(Vb&MKwYEPy|76%QkL|jWl$E0hM9`o9@HeS1+kK*O{Yq!3{l8SR!@`!x z;g%n=THn|55wBd7*SC;=1YZ%-SD%d>WYGP-p5tJCNY&cHjhJ^ymHK(Q8@ zZ?^2$Yp)87F_^DbQ(MWgS+#$?nX|myk7N8I4{oW@2|VUVs(r^7^GSf8DXOYn_B9Tr zO4X7^&Vnlf$-~&Tbhw!<&r9zKq6L3b&cRs5l4tdJ2*#r6Hl@bI#2-D_!{x$>Q#lx=1|VM&>C79|;%X zk8^zcC@IpamQ~C9>7Y>bjlDNFXX@a!@)%@Id#{(xKPqx|a*H6fWF(SCCrq*U<_maZqG2#w6B}*<@ilZDfHtH*AyDt2H$29u3 zvR}LmZK&fnjMgh~9k+?H2unB;{p_j6{qjqO^)G5tw?!Aa(^DI2we7tr(fz_$gIhV# z;nu0k%3N65;Un5;vM!ZCQham)+NwNxX!Yqr}c(%L(-g0dr7kT@g zc66pn_RloE=F{@M{jminW^D;*4FP1*X5cLwVyA;Pp?rEug=(enUdfG&Yp=S7LDlg| zjTi$op+ku?1Cd6D(x@^Z6@yQri|NLRr6qHBN4%npoM?(DsC>wr*G(w zJXn`Qa%GbHnM`$V&wkxZ_NH6qcFBnah%sMjtaqO=f)1-_@W$so|9}Tkee~JDJ3D8? zM6PP9H16AN&?cs8n1Q*JXPsJ;v>mk+Be9>mG#ul+K@9<*6I>%ft4b_6bsts2$z;CL zIO#t~L1@qxR~}i{M0O2z$GYJMOe`VZ!kLgS7(ZNI)V1xM+5dzF>SN{QW;)Y!cuPAC zk#>Rji==Um<_@;`>hV1<#FBN-cch^yYrp22&KJO*E(NTz#D?*nZE-@EZ`IM^#2S4g zlow6n77Fu;wmhA~xkHYw)rysf-y~hl!W~$$Ga%wTLJ8IUABkEKPH+v1JW@AJ)4FNZ z8_yD=+_aQM0d8a|i{kcGHUojIf;F{QMAw!WwiAJuWlpwe3yoyUEI=_S^vcOqyl#SD zKZZKGR^5FEa~L^#W_{2&%D7H1H~cE!@`*aVHSC&U`i6i@M)mf#-tbNRUfWT)5ptTt z2f?MA^ULr7d_4E02X(Okd#M$ca3$g7_F<0jW+JUZd*^;@E#j-Hr#LqvSiVV5@DTf> zG}PVRbLiM#139V;a(5-2C)lXbuW!E+F$vL#%n}xMh>TjCIsI`Kit(H1-_LzSaR9-- z=`OFyO_iy{(CGM}3j4+MW03dgwX@M8mZD}ET-SJ4NdlcvYg7l6j>ulL^(_Z}qOzxY zs%jd0rT55iv>mGGOx_8V>QCSzj1~H^w_lU8Rzuw9o#$70`-lfig_RY#CX%z6A8Foi z_~`vdFhg@gJ9fqG%Kr@vo zF-A~N5~7(4V&-K=6MdBuq)mKDm3bBM0<6-qwEG8BL3X^vqZq<_%av)h?Fb&Iy_uhA^%XZBRsDO@F?7&}=JjBL97}YGiPpyva&dD^mwf z`22YIaLdDqU?bU=h3M2IX1=_$#tQKI01Ydk*g zJ@5{%ftQ~bP@}6ThTji7Mm;D3>hYALLcK4BhQ<<4*s2}T5*FroezgY=;+F7H%ekuR z{z7f$sBHBXyAFBtxBU}PY#4iS&3`)a~UA_F_5i zGI*(wU~`_k?UTw^FKhnDow0=;Bl<}4x!;?E@{){swv+Oipw9BenVxIsFd-(q>k zdH%PcYHOL?B(0heFq`znH>j2)&o80KOsl9ki)yd%T1a5p443V10fEA1lZG(>6lK#L zCQ%NB75^oln%>8PhVrdlClytWs6a03hKlVc5h<|SY7cSTl)G*Zf~;O8P?%4t9W-fH z?EeID$_hYSe!!HB>_sG6XL?$`nA+@3}I>3#x5tbW_Icy=nX!!OzX}wTV&u3-HSE1BB!+;3sZd+o9bZmZ-%D! zE&Oe!H6o_)^=B4m;JL=_bh%^x>BaTBjJfM4beoePX3k7Q@D}THxcWlK2(}?azGUs; z#}ZbV{hRC(t-hBeyLgw#6Va5cs?%T5)JSK}-9H$3HtPeuhnCOlDW$kgWGdVkW-vaK zw%>?5z*=N=R5F?)5m=A)xJ`(&hU^WO!Q}r^NKHh4kCU?A{wou+sDUllrVNo}V=~8g zvCb14MO&i@I_li5IgrXsB2fFRM_1R97V3yd$vW8zIo zBAZA6Wsi_r4+a>ELP36Jm!*fTm)#Lb`<6}PK3DeB)CbmBvwXn-`~7&_IrhuNYHYG` zFpY-F26do%*^g7>;5mA1Lz9M73LhMlle}oRpWV#yI~)gj&s$7Sz9$-iDQ4$lKb{% zJp>8`9W|+PWVhq_Gbmh%ymxf8fGw_%Z86M#MIb1ZjiF?YLEBqZN3&EWOkls}^^E*9 z3_le-WLI>0Jl>}M>K^~PO}k=B$p<-6nH>%YYK1>L|9v~SHqUDy*(em(=QsW^y)@lu zr<&D4BJRv_(orZVw<`nPZU2rSvYTjvY(n$4T>et268eXh#&D9^t~Vg+NGjtdL^TLo z!)yF%sX$UWcODn8zJ(ekL*m-UP5CQ&>cdZ$;L1oU=-?^4o-pm>>769e4L&C289GEk z6nXjPb(&~XR0f^!`GTm_W3XA^yXs6WrXvIbZ}(b!BkkRw!T$i)GDg%AQ3VR-TK_0r zjn!4N0#u$WTxRie)7k#a)=T`3hE*a6!&l8Ye%)j-f%1=k*ssXxo1|Qx4YJ}lp^_6 z%5|j_sUh&<)sq+<$Y#mR1pJpKb(rtGyzejF1~X_o4;s*3?IaJ8RfMDHLg@cq+rR%_ zK>{Z0&|3Zt2qt7G5s`AQz@<~@*3f5fYBu;K{+1J_9_@fcMucw5OLWCjDkz`=BBQyT$nweI6r0J1%He!3UK%7(Zl8A^-%)e z{@r|(j1|1T-vYmM7rvP|wNV!St4xIX8lhd`7dOsp(i?>4xzy7?JdfYj3gUXMWmc)} zMe}&0$XJFEh+%!d*Q_fMk6>ES=Qnd(+xmShfP*|FI!zLqJ&_&UNqUT~PvIQLmt18Os<0so`i$ttL!) zUk)^xON?K`3G!*Uaav2xF-(P9vV2=aK{z)2NP(z9ec_;$QoToO__Y$X?>x^~z!v7X zybfgiA)z>z{*lpV)T3i$JdDa(D!D}Qt<)B@MFGn=1YRISE4OsR*^1Dyz`ooF81=qP z{_7%XSyGeMb^tS#5nOit5fiIOfBWY$MEWSYKwdQ7dtv*sZn<2J?^?h+=yTA;)2&=U z_EPo;9kweESqDajZ~^)Z*?eB9&{Lll&T5|&_9^_xto(cyP}Ti*M!niG3C112US~nen@BFvPECbYto zW_^%w^r^j%FoVuuWtkOI3pJ>t&0sZ~nmf^r4}UowzHm&%6LgY0#Excl#MILS9j^J< z#|XGD_>^{R!B*F*d4l6+P3YD;>rxU3A>VC?JQ&g?8q7$*HFd|h&Tccc&S1pCXQjpF zX07cfl%2l;KAe)-m`}BtQZN<^(2GZ50Vm0{lwO3h0H6A!U4FA3es`}L2Mlez@r-!b zUmraEX`xZ<3J!mxn#Z(fpLvJ-Mr=?p7M|(9$0tdGUFe0S^5vg&R>w#k9YGMuLNKq_db2g7DivpI4ntkIZ934H6_S$jaQjk&tds>LPnf{Sl0(O z`Sk3}8oNE&tEuV!Bh=`Qd|SYqp9_Q^A`${Q{fLRWnVQ&}f{88EomeAd(~a+yO1g302OU;ko{hHb zIucWNx14m)`c}Q-Cz#8#eqP6Of~Aqf;KqGD?A6q_7t+5SK5e~OF%&KDx52A0be(Wa z-j=%t3+*2S883fUO1nqnSx(wA6&MGMw}4yGzCVJ18IOtgM}unsK{ODOXtB=TQY+zm z$K)&XIc1Y!_Zi~Y{^SF$-od+QVg5kiKQ8AAmooPBG4#wT+01Ws zToa7dszUdE;h5f@n7ZcpMI9)NFW2kb+VR@4LE48WIq2&;+40Cc+A3Nxb?MQbK+1s= zK}uEgvv0Zt_jSFMpkljD{r6JO!+Vg-@T5}j`dXfACpUJo%BV_lUE+M(_^ROa34Hy1 z&|{ELkLDv)3|pB7Z{27;JJh&u^URQMY1%uU>XDhOMQ4cPWbV_AdZQ-OV>rxKmFYqo zF0y!j+16wPERWa-z7nbnKjR|y=8I=2)A1qhJ)*}8Q{I-}T7Syn3kd?&igueT)-2s7RvwRgDz{HF1yv!Mqd(n*!f`!D(s(3d>0?M` zDB!gj4U%-N$Xm9G&rXBczhryMC9^jPS#0j1vnkUX}C zmHtMfFy-8q2fl~Xb*nR|=O@TrB&BFF-wM+t)Rf-4T>GnBpG-i$ku3eq>8rJ}Kmauz zlLj~=t)q8#UqPrKh@JQZI{4%s=Rues+)QvK$lRy$%AySMdaSjFgubY*KW_n}Y&Ky` z>V2^{x9gU!C6q3^cH|f8q?b@P6LV|FCEforz6|mIy_nEY>9XS2R!lQ8OSkUw(yi+k zw-D%%D~EISQ#7Fb=3_pGx6@SFi8$mW?+Qq3_%S7gC#}M21>G(wmMQ~oCjcczVrs}g zKcOXoiTE3sqDP8MJ>{jOL7+Sv6UT~XpODfHHOCP8G2>2Ba9)!<_LRfy&O0)NM*k$3 z|4BrvlLWs(2piO5&Y+tinRd^Hh~okGR(PTQ%M53?1F-E$yp(oF3n9v`{Ut~Hq2fYk z1=6UY3cAD!rTEzH?q`1Em5*+dgPaJyXZTPHKz*NN=V)KML}gsm8=RZ$U&7+Jz2e+*9air8t0=xZ`^;y>GpNv-&-#|?D^IUYo+gf&@__%WqJ5me=bEb%Dh7_Ii(zQTOLSV3>d9<`b1qFsRR%&M}oZbjEKN#_e5l3HZkeOfHofutobEb z>Nyr;7<*!UzgvJKETpM-brgR;|KRfK++83xHa*8B%BQXw1Uljxz$g}0bO2j7MtbI` zM?cqtxvtepe-f{d;XwCLOpxIL!*iCjj}fWaBB9r7dQtg(=QUxS>?;lMtlq;eG%NVT zb>L}7rv*j~XzVepD@5<&Lt^|GliIxYaH8&V56LDnwwVwAHE2Wirv|J_8D17 zW7DMMoHoG}&(5+EavFBu*SVWL7P2}U8qy2guVLoD+uH%jO3v0dvQ;zSBi4xEi{OrA zbPr$!#Nfutu@zmwdVwFY0e(?&yu0mvDDVIIJ~+6=Ahv0lf5oo00>&ikmF#83s{)I1 zR<4(oeDA&jkGZ9aqk&uLoNrP0UOx*Ts7i*wXtz6UXa+x!890pEtGr;9jUs)vP#1#W zC)%LY%R`(Uq?=@PLN^3FAM?rXCzXV{s6=}a(Wv}AWQ=^dI9d$vt~!$j<~*9R!oCwH zs0}9^qkxWcUH|$6i#V(L{J7nwnf&yPhnIC{qBj}H{#?OM-n)_0-H&Y#D1UWKx^S3I z3;T;)d8SEWHbxA{wv3~!XE>wucjITQO-(m6_Y!PzA}t;cWe0b8p-pFfyzm@uzzhEp zb*&BL#_otn_^!V2O>yp-+%Ip7AG9<1^#E7XVi69_ywmN%oN=4vJBIZKTM%QxDRkGo z#6xpNS}Tru&`-Qg(p9Dzl9cJkaHtX%Yv;u0rR~1=0?qQEeZIP2y>THKond{I$2Z)A9Vf<`2D=-`E97;(`MX^gK!N=K=J?J>-E67cxy8J zf+%WaF5vH;vN&alS8y9%N6wwj31qZ{@y2C`Y1G0wHsbEBe(NXBn!7&c(vtlewO)Zw zsav_*-y8F7hqZ!khafovv>+V9kymGf}y%fuw`{CgQxvIoo1 z`jPJb90#Ccb(-5i4__b{KzQLPZ|@1n3oRxbC&=Td;Dh7YVuPY@dWu#0h{NT%cs~-> zg%?+iz!<1<6lU^8w%au+V~}^&bP@Wcj120xpD}=eqiN1q1P0+`=yZmMP{ku&Veo+5 zZiml}OPk_lXFoHu6K7&CMup*{R7+g&U2Fj5&n{xwU#|RiapoOwM|JF6d}&dR>IVCo z^mSoVAES3tha+G_^+NNA|KeO9S030*i#Ld0IrW9SITETGoK=a>ae0H_EL))Z#Pk}= zDn!AA20pxmnIhS-DA^zRlu8W<;7v9+S-LskKwC@;fJQkAbLe!w`!Yo3Za#oc?1RML z1>+jB^XJMZAzc+0SQ7v$)kA({3&6PrFDZw`Mdj{JN&dIRr%uf8?EyG=^pRztNl(IF zCyUDrKR7S}J6#~HAb1V-@DP5y$6FxcG5>Jri2@dGCy5jFav~LJ;XuFhpe=HC`CG<+ z35B=L$V#U0*l0<-`B@eDez*U?>aS0fGJ{3Sa$VxYqkqbXtzOH=tKhu}sASmVd)sF6 z7Y{z)%lru`wGOr`XiYI9=R>4!S7-7=t3U~D1pc`P=Zv}0UB}`w23(gU(~DfiE3SnI zsd{F-$r~=Q0cSGH0{!6T7(`3deH}D(A9-bjLg>5TdX;9DWn)JWl9vK=Q1KncHos<0K$0TNX^*9klB=&fd5?no0e=m@mZ7W?y2JB&)z2{l!`|M{Db zMbZudIVD*r?3W65?qlw+d+Ca}k4R;?W6yiTdo1xH?Wz@7_O;&UYq>DnmP6m{ zc!HcP2)6`<@biIFwnVJCIbXN86L#>_C;chG#@V$hoM)E4&0v8Qzn%vcumFsI$BM@G zdRfqv9oMqlr@3!%n!l50j&1Ux)E@CT#eMtemgYwBGcjZ$w-oME!sJ z3pUAst2H4cP6fw!uF!1WtH;W!hYX-=5W&FYj4?G8slOv<|6anL2RP8=O31u0#vMek z-FXC*x<$@@Wv5tB{JBPA?npaj+#Xmqp<=V)|L2-nZOAx=w99bNq*oH$*nj6apkenr z`-)GglQEbXIM4~X_@VBK1Yw(rJ{=ZchcoYpdC3+AAX7})=EpO*O5iu3yU%)D@V{h? z7kPJJ_cBI}C`1r^GX3ZyDv`DJ@35st0k%7U=Ol>Ibt&`}mQbxb^1I&8!#=zP^-J3# zudFNId&h?sl@$iv!yqzaP(-Mby*VxM`z-1t0XU7W!+2C&?%e2l;wXto!7;0e9oN$& z4fzFR&d1o?pWmm_3NM#4Y7;tW3qZqRo@Zv}^G?7UI#0mBi!+x{SlgoJkBTvagu1-^)aS;i3726Iem%B0{@%90vM|4h zT+e6XXK+JeFd;|I^e#vZldtgY^~sf4fSgb6QN+K?e;&rxFt80`^)Ek5s$qDao&}1n z?(J+CONae2d{PR<`x)<2S>r;~1*6mW%6X-)k)VE0qpz*LLTAmm3!|F zPe1EWN$KP{t~UmNPF))~V4avC6V(d|Ry#6{V3Qf1i7^Pyo5tg%PiUvbs;e{;eNrlN zgj0EwQN%v>nZwrQMC>JE5P0XAR`?#J2_|NI#ovQ#SJI-Au=PUNd`}fy;m2G&`CLR` zwRz8BVAoUuHgOXY7To}ZL>$Fsd;bGCQaV>AKs4JsmApDsZcSXySr@_Zl{&1w5*ls-}Pb0~Bkmq9EY1JdYho^vF%*qInlq!J%6 zGK!P=Wr|>*A#f41aahA9$Sbu48hgF9yQaBheyUZ++Yjr4Ux9vZwe!Ajqc0pUg<&9w zFcc^;*!hNiKD;9LOKpWIru{nq#_2d+0#)d6U*HpvCK*>S06uc)BJa7T&d71AKRxXv zV;Lr{Q)#i9e{WoJpyQ!<)RI?f(xb|&reV8yG$$MWc;F}AMC8;1){lYQN}b7631E6T z8RYY%^c8xKQKFXLF5_B`wCB7T&l5fC>5L}C9-@Ty?t%LKOPk}Vov!8RtLT$A>!&u& zy&Z+oNs+!fOE{DsKZM*g4Ao zzuK<+FX=Om=e5~Zrj{(lh1P6t<&mWp%jf`WRyG@iA!B4_OIMUdC{E2|4Le*^qYLf8 zmZqh77D}mUQ}cMa%JPV^P)m!n(wL_1{qzUyxA7NzUp&X>^SrN*r(n|c3D)wc<4^sj zoI6BKw&&{nV7yrerX{>SSqZvZ+l^(}O#^HYRC50nvMb#{Sy?{!(&2S=O!J=k4ECJM z%MeoimXMP}-;*6bf{>5kz1So_xIu{UAxIsYb$)5iiQcY;Jn&X$$Y3pN(b8>SN1O`q z4~GN2xmV7mgJ7=DDvW8=45T?qilKxML^CE5H5Sl_-dOF8xVV^u)lK56>oLPGMqSbY z>FyL5McsrbOS0t}a|)T);P?scDhZ~+O7|wYY&`tCi>d3N(E5N6a<7|-@TX1lP&;rd zWX?iHESW9G7>o0*mfk=}y3E${nS*f2nb)_dq7|@*?^?SWO;iZzQ9+P5wiAL_GAmD42r0*-#}DLS<>%PrVuSW5MiAG! zMKX1Gu`wOZIVSE@yBYipi#>9VX{K-A22tf*itoJ_a}rL$O!VOsv8K5VJzmDNXa)kv zDfvWPgpBaB@q@UpLqOor*ZyYP0TL#Q+gzWz$(Cd6-m7hMH%_)44rH(o=1vSgcw!{2 zl(Kn#Yfhb%F)=a~OK`W$ZnJ$~XFbPU+5wDQzHOPd)fzG(GBO|jZ zCIe>+)@N$XNzbg5K*!bNHs!VdGXllP_qpIzSAz9nt{!VE>^lAxKcK?#Z{ykP<*5`4Ucb5O&Yv3XL(<%mv^2bmqwn`!>p`x~Yaqqftj?t>e<=K&|63F|? zUv`Ddd5nYCoac9Zc4D&6@v|e3%S|rv+<}^>DwFcQD?%vA!SyWm@c(2edt`Jt5LIT?A34RR_|@g!`Whf zYKv==*=1qra`WoGry0zn-{e$LA0*;f7&+Ah!)n``m|Qf`pjNEMDHLSJc-CmoM*~1Lo($6X9FwT ziDEhVs-3Fe5CtE_xpsxVl#*}4gp%MA{Hrxk{?|4^su4jYgEl}yIOHN}o1m-zQ=|GHy|COgVa%kzxUfyp S{b*zye&4eEnU|Le6aEEO%ugx+ literal 43788 zcmbrmWmJ`W^e#%5grIZ^s3;u*QqmwG2+|-S(v5V7NS9I~C5V7@ExHj9kd&5?kWOg? z&Rp*QJ?DNnAMO};kKuN2g|%LOG3Qfrg*{c4Bfz7=LqS0ycq}iihJu1t48N3cu;5RG zUopqPZ$i#8+Ro4H&79qgolH@l7&|-I*gM-;zF=@Qb$VrKZ^z5Q&CS8h#$e&>?C?s6 zlhgKpe+7rVlR2l;6#5Cg3a*2^&MOoYauVb(RP&5sc&{(+kEJEl-BY(_Fl*G8$FO&0 z=m};+WF)U&$3#`V`E_2`{VxF!-H?nakleBO_y0 zY)eZ5@-HOO(6!+I8S!xc|9=54dLx?jO&aBNk$ZA6bP4MDYL9E3mi3#yrMIQ@1q<@DC>ulBQZYWs$d8#9GEyv-#M)#u7Ad(MyU4pcv=Hg}*KZ$AA^-MFnU-;~uSXV+zZUhnF%)?_}Ap1`J4zIt_axmO9bq?w z8)uFXroD;Z{87;lt{%U;laTuGl|r5K+ClsJM43tXs}*IBwVb2_-^H$&d%YhI#c}U@rdh;=H%gazshbRU5mQ*Q~EJI*lUBd zs@ZRTSE@PS1jG9(V$<)uOA~S#o_G1(_dbOvLa-O#n6{=a+is@rY*D_+`*bPh(d8gf zfYf!uX(MZH`v!ss7uGZIgD^rt?|o_n+Z`+LFT(WBTqWoF;0 zmL6GO&)_zZ&F!uC`19FjrcUF^u0Rtk2y%65gaPE6U(Jg14TGrWJP1=ze!M=5nHaZU|1m?I$UGYM+(fQ&jKF zHtKO0HWh9}tvDv>oYyi1h$UiLH9z>{PO8LsSVJht8(bk)VfOuPhVR*9v8%tYxPx!^ z{e`DVxgJTc^n>SFv3{*{ZAs%%%I&9`B?h*ZgP8%hRkNhW#=|oBEQhrXd=52wzrDL_ zBJvylofyX?rwbcEy~T}7!d%K{Gcj0hJB^@-xu=IHoXu6q5GyMlC^c?B@TKH=St{bb z+w$vdEdQBYM(Z>D9SxnAN)PWsSHrdmR@q!`-R+r1YL=%A_~`o{3JKvt?n}%4|c^gnRu= z@|CliiPBX>$~l#<~b;h4{wKJ@%{3ngXVa_n!Xoyv#abPpyZLZ?@i`5DWu;gy|3;-)#rCS<2H8q_3T#+;l>AkTiuN_@BMFg zm5JWmR)`;5>AI>g?fvXWQ)M+;4ht=_BVaqlx8Ve3(EOgpm-rlb`ZGCu{TQOW-XzX0 z_h@JtI#ds=Zf@=2`0t|H6xx4Jd0=vpeQ;&aN9MoUf<5QTOuBS;;2zNCP8Z6@F?u!=J^DMi5UrTR zp>w6lZ})Tcz&HJ}Xmb0{L`Z4L;?+Z^rNOOG@o81y5@Gnck$5a%gWm7&ugaU6Hgg85XxENg1vDZVjhJuv^=X^U zHTkdy_!~_Mh8lGo4t|sxvTJtTN-^ySC8jc(KKm6yUpS4u7J+XAV&_Ws?bdAwS7$E1 zz_`f@b7AkxD(i784@Nn9F&wMNOzfSEM(EIb8H;}g9>utFwOW^5aJe1Gt$p8!h^n5R z`Cz-ShT2bYyM;&=(?$-jRE**<+*Yp>VdFjBWs9SpCQ)jIq0%oqI98~WUA-tO?EAOo zD=jpH!LB>XsS}pCqV|_#o5vm)CK2SE(ZrFIepe@G8%kG~7l&f@jfaUN%O=wJ)8F{# zKZtmgW8JN3e3q;Hd3&dqt{jFgliP=CzR)|lG_gx~ za-KOzD@Z{he3Abw_ouH&I{k(!)c%m%CE8YjJ!qfniz$4T3%g>1!SDDi)fxw&=Fo2; zK0o=eez-MN6>%t4?(nOP+pfJ@!X%{AQ*gM>#k!G5Sx-oqLOF%++rqun9xlQf1pTl! zy@~UDPLe^m%TmuIeU3m0C&#Rfn@V+9F1BEUYh>v>f8>;4xzn;j239fR9A6fNW(}@P zp&xeH>2dQFt$F6+)?CV#SPa8ZoO!y3tbJGhbWQY)G9Svn`@NiL(Roy@jThJz0_R>f z=g|Vw_}N|2!vc|}R2i5mk1m`mr-_pnJyQ)_5=vm1XnO-qn_Z4wyy`^(W^AdY&TnE5 zY!%B37f-jnr~FBDPs}WoetWs}>3Zvs${T#jR58NGJvAy&5Sb1yzH`UXh%s!dgHTDV^f|WNd2~g;yji1|xya1+tj!G-WBPV!)Ya+glJ@g? z`+TS~3yPwi2ZM@AKZKlDpH@zn5=5!7bzb~*+*^=oj4Z-PtoF0U`XN-PRgym9Lz3EP zpKK9)Onrkt-E;k3$noR1%wc0%&Y$2Yz2z}gNK+~|YDLS~*_pNDHQ_1R;-=gPH+_*c z`)WWmx2p%LsZ2p{)FC=eM3Cd7+m)czHwi6Bcb}0{+7(QYT8O>8=61}cjp!wu4X3+==YNastw5Zq7>g_KC z6xc*^T+gU-e2;eKN_ZWkh(2o;>8?ep^)bk5U`}@uPGI45mML>K>`GQH+fJ5$J@fN6 z`V0N(ECd&kluT;br`ae=mjKGaO~wRyL=GkYv;- zo#Uq2xyBhYK{B|?`g5-qJ(*Z!o<{J$OMV|k)+qR9LF#Cy`3m97;N`@2xa><8YNErn zS$!#DKv_!TC%_lBwCGuPbv5j_Hhh~~3b3bHYpd^UgNNt!4%F*koy)Ro(R18(2{859 zZyR5=h7r-YDV${fjM@|QW_f55&)OP7a+w#Ih9A?Khlv*8xY#LkE+r*L=Wwt!;=0li zlE=E+aI~;)FJd{&e8D6j`^>?=1M!u=EntF03XOo*fU zQxsJ>QAR^_vQ7*4gGk5EjegoS&>3zc&)8oAqm1@B-ss-dFuTYms_ZfxW#XEJcDjH9 zM`4BhUpBg>n=1iAiI|?vhsS&OV;N6te2=$hEV|U}xgA`H#dg#EE-1nmW#b-aB+k4&X>OaSv9|@aA%o9h^>2byPwByEA6~@#@aXY3 zchRPw$df>U+GR0_u$|^(keLr}CT|#r{)E6PJs}C2b{y+_$A@aTvl1YZi0RZ!(6QI! zf3{+=wSDi0(-Gfu2B`2!S635jR!`xR3e8M( zdE?SAAQ>M%ort!a4sp%p-#-JM)PWCG>Jbzr#9J9Yr?3p4xQ#zy2I$&gE^httOPjHq zo^RJWucw^GN6jOyVI)}kS^G-YXl{at*uunzn<{wA1fQ;ky5`L35z$f^TkN?lL$ns+ zbbfmq9no~Ma82}1J*!&G2KnXHgyo!I_cLSBhrKq`{5GY8G3TrnRiYXvH|z+x8PE+fSsAi5tzc7!vxWYR>^GvrI4Jq{!vj73sh8U2r1` zsG{i%hx?YGgTJyv)e!b#LeVDw6+WxSLaf4pBxG7822rlG3au0}5nI;HXjU7P)7Ai12$LeZ zD35hYjRIVRDKO`0B~VR&Vpq*>Uh|)p(zHgzqVq{y!pl1 z;gFr&8o7SVnH%E4y{+*%oqQsoc?CbA71`ZT*|?yT`J&1>+|Y}V^P@|Aa1zR#two{OYJP?U6$8GI=7WjWdkwVhaZE{$*ZEd>*&(c%Ue>_YjLNIWrNS}(mQ0iFLGK{w)aNU)Bw5< z+qECQ8@-DfToLh$bv}RYH1vNEu-ENtQ4hEZp?oih8AC&YZ_eaMKPNrJ>I1;}%$96U zB<>xLQ47kRKx&G1opXNrzJzHI2qo4;M$9wWGvqPDUmM!FB;YphYy?knsE(;n#ZXOr+mMT=+GdhRRA$qLc`pX-mw`2w-s!`~*Md?aCJo6X!V}4<7Sf4^7#8#n%I>#Hn`n@&3~f{u-Wpw$_$WW1j266}nk;FeTmMKxQWRzd8IUT$yk zx8smH>Fo&w4mlJGd{La>QO2dtUQSPEJR0?-TQ$+iJen4#EkR=+j^1Fk+3#6QaoFNf zOx|fHiI~g~-ZS}@BK!T`FRjugn9|ZJ-tp(2{{4A&K?nP&^CT*F_ z`krqIj+19M%c|~A`0jrHK4ZVvDWE!?=!SSiA)3r<9zA0J(cji|E@<*3|9L%Px$$^* z7kYopmhjFoin}h?Be(WJV%a)439x$z7@@iA5P24P@TwQF zgW)m~$*#25qzrU?)jIdcj4=(Euy6F4`6#qNb+5}MXo!)F7Uw&OD^u~w?-%iDJZAcJ zXcaN4_19}1lTm{0Ddm65F*=8l+;;wiJ?D*f@73?Vv+oo+HC$N%-h6%QQh}mP#@Xuu z*~Bc7^7Bm~jSO{}IAhGkKQ%9N*%&Ju+u=2OP42aY+l?&=Oy9h}y=N2vljCGwWr+*q>x#UJ@|GGdFr#Uh}LJx(fYW7*Xhpnq<{Tvi?yJ@2qvIcD79~bd4(&s*G*?(9Gr`)7S=$NB3 zc*g{^w6=U;afkl6j8nqohdqF}oRO|64*Kj|K~1CPwlczVGK8qCUuq-v?B+9Abqh1We-S?s zt+E{+vN~FsLmZ5|*iPeDncmNk)6()11w%zJCpfeJ-DmWEGz*$S?pBoC??I@61gWwe zAryD$YM*7rFrvqp-ek#p>#t@Aifeqjs85Xy>t3V+gjYSNdN=k`dgo0;kx6eQS|>D3GdDRG-0aB ztRJJ#R77v@Ext;LyZb-irVaK2v(MRq&=eJ?fry}HI}m*ep{bAJEG&sRLD(Q5SmDImKi z_J<)x&fnhNMxDV*7vE^NZ#wO?57+?e%?wp4&ggK~Yv&uRL)P;o=Eo%LMSnoe!Ln%# zWvQF4w(kX)cWBQGdZ}8LG_D=v4eqW4dFeX09S%_N304s&^Fc(P0KL4uukj^x1}lKi zWnKMS98*Z8^|*2%ao|aoOoT!(9%(3qDD>p7oTe)o6aW?N%vPW{CYs-&r!=!(2bp(1u^L`{us7`uf=`4b_MzH z+=&{cMiR7*9cv?2qhB6FA9Sf5`TS(WbiTbAw5B05XM;vhjR=YQdIO4&QiL}^Tvfv( z=U@ZHsgN0zpQ5c&upM;jU5k){F&+CAv>G0hPJG%N5HoY_sLz9zVNYK8updpM*!`^e zIQm8 z{nK5PG2-cO3CArHhFrR0_Ol{0ofM*7K8c)$#kqK@1!W$8emgC6;9htDo-nZyuo(f} zNb&OT=`y$tNw5}cwm1qIA|6>A<>mu)8!i)xUn9uJTMoPE{o(=0x5em1h1xu?agZ26 zkG&HMnxkkZmK6hIDz^rM(k;rLlQ9?hS{2(HYnGdCvmSD~RaUpn`%{PP3vlW-ReV4n zSHmN{|0%cJaou!(>DyWTR{k@#Ypo<`gx)8Md{01IUrNw2hzloW)qS*keze63WiRZI zueM+#IOql%*B>}wg^)-B^g38r)QS!!KLlS6g1)4vnx;yRF z*?99o`!&O?-DR|D6Z$kRW68>uD{!q`sEvhACN{_muzDWtn8!;b( zf;P1>sXYd=u=CC9OeTZ1ugpR4PGx&ur~c?CD$_eCdt*k7Qg84*X%mC5V(1^uE+0B! zFeTO_@wlHN2* zZXWTHAj$N;V{FF}!XI=j}p)AWTl?}SrPjXZ$*{;RFA=AGnZ&`P^k3^?0a19f$JDAWv zVBnICkV_@t>PvZ`h^{dmf6aTu2 zJwq;t-ec#+$So0&E=3QcGcj_L;*vr~_O`_twpW+r^m_$*7$#RDJoj}s(JdGGMAV4! zaa>WErg8Mkd~tM%OJ9(nIXMH_yoZT*)4}6mr+)NF|m^%^BUa zAFiZBI1$fz^R3MSL;x#-7;V(P`GSAyg6lePRV zcEWRk-VyL4Ij!&omuy>;rFy=3cUNWI(h=Pk+xZqQM(WX5OVepS4Dr~-Vo>SUEZ*1m zb}dGt4EFKpb~+KePh!VXpz5WM z$%nNyL?F$1?CTUYXGG{)c%BrK?fJG#SHGOu#Fn`6g5UoXl%dH}W9r2Bo{m#na?BF) z&;Ei_CUF=9W_Qola<&))g1||MSe)ucxP!P%_-k&GEeTa?Whb<`#F>6k3fKXi*n0<-Q?av$Ty z(2x#oN*&ams#&7bF$x0N+X0}g-@h`!m4NkO% z;P#Ztm)mnM<`pD}av#OXg^8ow(0tlO5`Zq&!A(Ne?N23veP?!_A z;PvFL36Prax^3S^m*+2NA6UgJ6H1}PgeeLh?)0$azvsdw-iTqK8L*9S-#-xTe2->x zL#a+OHk!s{v*aYULNq6aSWDfxMOxCEvRwx=a-M-grX}L|hxQ}y+E%X!qcN5cOzq+= zax6w3dbQNu>d42ff)RnJ06VbuYk}b1v8@tBjh(pDNdJl|d zPX?(3bR|I^*@4ZZEqpYANL@loo(|76)?RF-?N%0)K&n)`u8YVaEiJ={?86QV7ORC9 zog&KLvI{B3hUo547zdJ~smo;(6|-UIj;d;)YOuz-`j1B}zX(j+>fkkOyA1RvW@<{7 z4vsU@nnjnJp76)D0uy}9sKU7YjnIDNxJ8DY|Gus}Q5<6&!~tl15Tb`MwxOAn)08oC zu~CPuLQL~=E@4vak5>xzr^WEoO|6|WZ`jG&_h1z5*kJ8nt1#74a#JAOTRLJKo< z@HMeBMsp_)>wTU1W2C#bQJaFw)Y|<(788wepTOI(gIw{4W@jdD6Q7z9mxe!!e5NwH z^84&?H1zRZe<#z`;WtCiQq4L-d7KuwhHbGT&t6oLW|v4W37_nT;ikyDl#7nYU3L+V zGjtJaWevX3txA3wypDF_B};`tjQQx3!L0}1llc@{1DF7(i@grhPgj~?%4(5_aNwgu zT}qkORQuQhhZG?E!PSiy+#ks;#r&<<|1MDK|HII#9qkxw>L?Q9LgI%veta}b5qB{D z_tnG%4_8^dLZTya)Q5g_^x3_BN%xNl+#DJ_Kz|Ug%1d>aj6SN6y3qMAN5n{#;sy~O zIqLEW{>#ky_W%BnFa_7>HR)DzKmxj?RGY&pKcCZ*{3JApek?ZCtk-16N1rqo0EJg> z-L6yr6^|N@pS`^RxTPdGC=KKfiQRy&!U>C6%xwwB2#Pw!XQolaF1cgP$ha?BX<&e`dGBinun_5YcbGWfoZ6#rU3Uime{Co`WZNC$ zLK*sRfjq{_ZQj=P?ae>^+BfJB$x#d_zFJz6^}! zldxA;G~6G;%Qtaj1RYg*;d?7*D2t0HRt|A8n4}b{n-#^wBRG?}D$9uWz0HSs6b- zr`fdmRhK60rW-U?lZZ}H)D_jqySJH9DQ=h5;;RJ7XY$cC zj20Le(fv)_wp0BC0!JrllehSQ!JuYA#4V6V%Vmkc=zZdOp$kJ+;tNM^q9^HJ{ zA)Q!1CB|opC}}Y++2`DEJ;4n8E`=89%E;))YA3XtW|7ByE;i-`-UzPO7$>ak8ESdOqsk?r(%qHFLe;%$oKu>g|%h zr-J4w0`|qNGax?{8sfjb^W;mwm$L8InbBkwOiUiJ&YkcCJa+0|KW>GT3In9$N)2L+ z$RVfoxz&E9KkWe~Dk>}5lO-OLC4N1G4F~64$AykFzjL4NOE(ZZy0)gPA3VRtRCt2A zxJez1h{e~B5^upFq-nkP+@ZSfy6lD3cyUp0r~DMs*sA1JMqiA{gs_~L4z_tdN|R7d zfKbt;cPbOkI(SnmMIHXQwupDL_<~;vfq#yS9GP+NC&pp<>uc%(On&?Jw&MELkVFnh z5<+`;*>%a8$XHKxqDkb~i-kl-zeVIDk*ggb^#*i?yv-oqyut)pF-abW2WF|l!UXsf zRY6F$RAr;|x0Q%twPjTol!`-jE+;BrrK^8jc0=xC@_buI?;oF`rz!V6t-F5*(eYwl~}$C-$Cf$@>C;khAgN zZKd6eE;6a+GpAQ+@$1N`bUTkL02fu>Qv6I_-^DyEv|BVeCQ#xIJ_W-J5jqjRZo8uV z@0u4cuc4wTZ)6BMGInm^+##lzcjoS4^|&T+&FGFt=2)xER>AMl-)P${7ekMG*}92j z7lGdiSxj=PQ@DdGau^X<1OYee-53U9+CC_6caa3SFrOk_QjAxPxRj1XJZkcig80F{ zwe1Kck6B+{z#UoDEuug`HOoJq_FOALJ#Mvb>C;4RFT9Y87+#;^%Tx0=1k`OohsTX{c<^aKb}BhzS43qbiFQkC$2*N*>cS{6OMP>QR{`+<;o?eObN zg@8z&%TZm4=_Ala8Ofb=;>_NB`;mv$5vKi%^(DHN31h-g0^ajb2i@6jRv9|RJDQV; zEN%G)OJ82(cH*SlGFEf9Q}?#KrV$~`Jy#97;Sf&#Kp=`nQDyvAGMW~XyNnzgAGyPi zBhocg3d@U1mxowI($Y8MulF$65}7YGoc86+lfTVJL*y zr~&mF>-YJO)=Q*sYT#NIotg04X|SviIH60c?X+l=nT#PCL0kGVUI; zusQsV>dO|Iw@_BCM^WP6zZ|5R{11E%O1xq~m(?AjN=?HG`|l+P;uC&WTE0_C6B?Zw z4C($ar5<=EgEYz?RPT>Zup&Ffq_LFyp&V!VOGO+)*0Kj~NEN?z6xD-I5cLZ%V8`S6 z??E%SXX@`F@e;A7b9%?iLeN+LJpV}X-+Q-Y_Km_HRIz)#Wq$lgHCq-c8(9<6()E9N zcsvS>0YQ9Ax2ec27nI~(Y_)T4wHWvTP#G}Wg1~y?y{kIL7P*I zN5c>D zHOOdZC()SRt|*O0ZshylfE{?ej%<*2_|GPv4|<78Ux|@(0z8;bI!PSQzuYJXr8vi< zNZRZ_a$z9syc$~wm~Er#cXyl=6`utI*mnzwx1VwUn>UEL^>n88Rj$ocB83F09p0Zy)OD?khK0YP~?Vw zL7Ci;-#FYDKkPcnM+)M|`2NA_kW`SaQz2}fXDwtn?cm|@LIKtFX82v+gJ4lSBr_D> z-FdM;}mxr7W9T#3><;(XMKg*3qj!&r2Pq_BE?=OX6Az0o+6cq3M z!&jfvt-8~MtMO?BpF-lzXv`Xq6?r3AKooeGl`(xnGmyys0rJ+dOhq?S0<0U{_mENO zeNfgF3p9(E@89oP?wYN4Q!)FVavQ|5;Yu`4Y~OY=b-haEPy?vSRyTb`8wVo5Q`$1B=1SY zCAtnul;LNty>$0N+Ftu55I9+5E(ATc2QFlCr`)ec3s#>9IuP1g+frBm4&gp_pQuiy zMI1OX5`VNe#!E_Widiv{4a!1R3wE*m z?iB&_b|Han_e zgrOW9xPXwdFA)5;>{9pyrWUKJv`W7uInrRnYwAR^9xqP#S0h#iGR)@{Zr%M@(qqN2 zLt1fzV6~0g^T6VweQl;*;Ex{sxPK?KjZXKM$9lvd>$9@L70f(>J`Vq>2n^=H9B!Q_ zZPN@mDNhRUTA-Mj0`e4k90AMey#fe+ieLj;hfwn;K#p%Y4aMx7@E(7z_P&5Kb65^h z{jrTVgtYeo%DsbXyiq8_98CT1O3tu}viOQ&`NP6=P~l_WK?~{Qxu;Fu9vd!@ z=lvm$fCI$?%%zdgb{XES+YRc`p$?q=FB*P=lo66?Yx!s}3^0b2#n+di-Y>y`2%n{!Y>XB0{Ex80Q4KwOxWZf|A35*T zTkCU<(0tT#FA?UQdceoJ4GxD|m6e9aa)K5!44rV(@q&kW3gOcBav*)Ca+!%m`l+STNg%r( zq8oym96%(;37;3YEJEfxuYzN%NV`lDEP6)doB~|EJ5#pf#m`4X7|r|e`e8F@YCR9X zxVH0r!M@3+9ZMbFmNFo#KZH&x6T|I!3U%1}LbF0vpNK{q1RRwMXyY+-t@UwAW}NJNv-6moXqucJVQX5Sa>0vSY@edw)Rz*XBu&h*zwO- zWBREu72)xh#?a?50V|aPvr3dUW`x<@U}>0fH5GEfsl9FVB5LT@Q>P>cr=evK#Ht#+ zgdrnBImb9PqaV58)bHm|3)ro$&|_))Yk0D}O%{7-4UWqGup%cMNR=gI-rliZ)!3-* zHLiguk?gCFhRwd*cD%(Oz+ok*2@yhB}y!^JX z_ljbUhhEHk^*}7|B?^qV(TQRU#b&mlq@&lCe+?&%aLj;2LNAbcnrzn%FxK+s-t28r zNNlbSWiNV*nY-wwMRkzC)se!ag~GRTUCbmk&5}7B~u@y0L6!_t+fk~xxcy1`Q)}1X4mBwOZMNDeXi<$3PXY1c5>no zKQ+BsU8=Qsoph=BEHXE0oY&Mm*7I{kbmi2UN@-8wkU9n1-lrAY0OLAmnUeOw;jM5! zC|WpY*DDz_J1hy9-%>MWK&;{OrG)R%q`6)7%c_$4aAMVLvqI4y4l+&m?o&|I78fVO zY!)5EE`?I*{^x!+t0u_+G)xf#EMj7rMC4o^IHG*dzaR0&gkJXE4dX$_CX9z%$hsxL zA`xM27LKNQU-GX~rNsKd`4)E$CV7n1DS9p=@yc_G!AO)SwAc?;J1~AR{!Kku9RbOU zdk{}39EgciXEO&pJS3|THytq^!e2HB^_0`_9=s`<(w)u~%6_Wti884mZN~j)y;|sX zUi!_x&sAV2aSzqum=vXz@tF2}%14#8vxXB?DPOUbp8&kTk}D!rg9)2lIZepaHX)>C zRL6j=&e^z3#I7coHz>5UP5WL=;0J^x-XB5ZT2Abs-^%>Jx=tGq|i6}}b{ z>HM}+;w5CNgWMN?gNE2QK$ZQwbAwsQB28&|Q!HIl0j1{&>^lIuvto))2w`tex@f&(z+Knj0 z4Fhe@>j-m)#|w3C6se*DMf@agMSUNOfdK#2Z3dW_MFO_&9!D3qF>e6orrSSjZOqrv zenZA87+VQLHp(H@qr*EALL?RWD5?yy`oyoPp1$e16q?poTMC3&%iCOOXrwQGHv1VO zb*(h{$aoZdh0+^js{UwN-qw7Xf8sA(poYfbRq3c^lUC_+`v>))LdN&nffX@eqr2uX zf7ZJx14Sd4Io?lTW85CyW@c*TfkG#_rz zli_pp{`WdE%>#q2OdwCk{Hgo~peqO9-g-&ZHt=(bh}gce=NjZlBWW_aQs8kk8l z1}yv&M16D@q{t~a({F;>F02YTNf}}y_drw|{B!^ptQ!zaVUYn`dxopi0l!Z#zJ}jv zh~MpS(w3I~h?-$5?6w^bJzO4wJ+~o*7GQ8YtJFk+R%@c!ulX+sZafe!Mh;r0%@U!q z-zS7L^K&4jFFT)wJV0XDRwDWwfZ|EWF%=|ZUCzW;<-#sTK&8A3V)R>pdKT*_$_X!f z;vXv{aTFyj3H|%FE#PveXch5)xy2&;ql`k`L6^RO;{qDs@Ts$k!X>9^a0Cq?bI? z3Pl%`=_eUtjqz}@rlYzC{&DB8sE0b^MS8hc14*12H(52oPhdg$_csp%)K&m{CejMK zuI}Jy7ibiGz6a0b_xWo1V@?6<<;fz^Kcv^qCy%RLWtH&2qUh+-yfIpT7tY5Hy6zV(!W|ez&H;c>aG@P;T1l_Q~y=<(eq8cDp zK7`9>Ig3$}kg-9QSirL%*U%sa_lsKMW;9c=R0!dyP7=^itH^1XBI1P6)X}E?ReniW zNu@!hd1%&SbD;Tu&rD!Ry=&fhG-d4s-V`PGi`U4r1u~BL#n=?&D3LSyed(fle_q4E z4Ezdg>LsMNQ>LfBXs{y3W;3KU@c5qzMiOlO#Xqi@>I2Dbg>C@94}Q&DUP1=)==U0cR1w#c zU1PBEqAuxQ@mT%HYrTn@<)BQui%3t4DZkl<_Y4zQ(jq*erx_RAohg0)2fATx-W@KW ztNbp)g#cvC#}_DQ>VGrG+>x!NSddueWQmOjTMk%xpCeK>0PyGX8fo(UER??0ZT_y} z-;x4RMpi!EBP(EIPVh^-goN0mk&JB4+csowu zR8No+e8v+H>PK}KfNd|z-o=xERt|1>B6RmPr0rI;EDM`)2OOr6xuMr6bM8P?t&c`- z5EJVlBX0G^x1OMwQhngFY>U>GzSVc29sm&Efb;xd&7{$8b@1a`j^{$}VNml&@pC=9 z*#^r-dgPm@ig>6MU@=&HPv*|9w3&P>!KxYxP-_NZ`@sNWc(Xt>w-{rA2c%WGZUD~d z^R5RMNdnbr7*WEqiT}SzpCOBGWgb?+ zYAER=u@%M}!RRAbZ{wsekOEjQyvBxf5z?&Ku_(YUY?G0eHfvky0Z*my&pl*&!G{eB zT??>=)a-%NRHdaS=ql1rsO8`DE^psdB`@FM74usOEfaf6WJOaQD7j=u9k;mX4mOag?II zHqUeazCzF#$TyC!h=aK^ocnozCd5Lt!BFfrbYv8KfA~}f$soobK0Xe=d9UVi zC}X@f%{Y|o8fRKK5V>64 z0cQLb94J7-K;69g(OXx;zX9?LjNvaWm`g!5eF*B|r2)*_0+fr?Gd~Vx@lH0-FoLD= z%tU@tMV|p5{4BAc4Sj#riI9T%e%KwFhzqmlyj|mC3siQK6-$Z_b2*^3sCIKgVlk_2> zf3MdZZH%j+HvD^+GekmcC)5K1vQP@;!>fJcfE>0gZc&$k4i ztChceoaQoBWv#U}5o|=PLE!s$PwwK;Fe!X?3-wBQDgHr?9R!HU5HLqn@>}!QhpQw* z-+LgtjLjd;w+6lCHCN6k4B~~(%kuM=_W8J>-bMq=LW`Bl6=f;J{lBJ~Zu#7fcF9Xa0{!w@f05w6bcnDG=!)*{9$2v7t=| z{Zq|T>@Znef*UkH<6c*tc)%cj7zp~yx6UZ)!^*epODTyQ&vl_ZasZ4PSGzwe{$vWt zJ&{h}DLC|h>j)t`pXKnnJPcau=#ZvKW+r42Y}mx5%kaK!9)l8V-4e=kPjGOO737!n zv#!^yN@RbK`OyQC0E^7AfKZmy%$Eyuepf|TaFy$n*G&c37{e9i?-1Xzcun)rNl-K= z9-^MLFu8wJ$)TPX+cglZfX=E_{6oAFN;2$*$;(l?z}dg2f4mwl0YBjwZXzjJ6c)5! z^3~B>S=Wi*2C(kS-2w$zc_--Efy2fadrSr#5z%27o_Y)s%-`En>D$NSLf46ducZb^ z9+JgVVN8DmFlJ0$0|85I&b=oy@1Yj0nyRbE%MFx4{=VNN-A)&;8r*L!4emixyKLPwo?rjqMXO!<25S3GOXir=vK;`*PeONLSz04AqT##bx;qfN_mnJ7>Tj+ ze)p#plwtnv_C8z7EprWS$BtbDZ-z?{QtPbr~wxeLR2@ zG>6vFy>~R9tp8eLY)%Im^+ zVV-=WpY)Ut8k=IVara?SkutJXApV&NwHA)mbuMY`T*?{~jz92t73W5oLC{HuruYNslGUh?}{ zzitxbfZ1kCA>#!o^oJ2a($T7aS~3P-bm|`8()4 z6@Yw?hrP+0$5c@tx6JWym`u=D0f|)Ph9)Y7RL`<#Q#uM&{32F!%WB_Qwf@3p5xXJX zE^J!KkAA|TtT^-nqOt6;PEVcF(T5Kd^2dh&4n9V4!=Ngya~KTtBBLanWy8VoSv4D4BC@W zR)335bY?ZI^HTlpr^?qJX^g9neDqhzXVSsD3^jqtsokePhbInrefsTY=i5Jr zQ+xEzBJPu~Yo7S`hc)fXhdHw?k?mv98o%fCyKWTpvz6s78sm(13vS(Fc)V7^Iqk}T zgkfpqoZea0(Ps2-W65kPPoobqH8$m1R#!!22x_WEg?;7HzVPUkwT4p-WX=zQk7)CS zvuJutKU`V6jqAt?w^+_2VV!xFc@LQYdnhfT`}TZ%7FF@|DG6w)J3WlZkuhj8;d+ZH z=FY?kfm7_eF1TlYILm$Gj_RaLRnx_DO$^()SeO#zM6pSQ6Lx6WQkx4_v-$)+52ffQ zn^(9W63N*u#=~kEW%g%-PFTi8gCbQCM@zY+1&ppEADs3P0Z37xMhMvdgtUIcPeKUb)SB{>xcSS|CP51 z#yi9F30!5WP*&=?S_>87*@EbGK|9xf6p&hbC0#0!fpO<>xWN1Mvzyp=p*d#HmeRhU zzahMI@P;#-Ht#7{X}*3B)|zbM@3HHCZaX>Q{&GS@Kj`e+;(Zpg&?j8k_v}WM-f>?A ziBw(oOHcXKCS{&=N(Q0d;Soto_E;5k)N}1|9wX*U$wh6A9Ww5{`j^$j`{!0TN4{9O zJNxbY;Sp)zq0;863o7|ZM!oV4EO)7Y=^k1%zJHp=d1J94X=>l_DK5uew|DMG z1kROy3|!yiUv-e3=E9hBUgd*0lrM2W8;tX`9J2vMI<8};D|o&DBxZ_T%ut(q$4OC$ zF@@%0aFr$c6YUple12&4jLyZzW`|*;T zH`kg3Bo?%4ttNdN6mlNbE^N-*{k6a=WepnPPDhBw+J8GdU$ra2U{Agx{6*t`05adB zrd{I@a*g|ggQLqT?KSi(4OdS*^Igrabz!I=R>a|Q7{4J`DZ1HMd#fZT!IYLDTlp8` zq>X8-a1&waun7%XfoJ^7on8I`zduTtwEOw1xZLZ*4--wuJ^D>6^*#C2#0_Yo+P=CJ za0*(g-aQ(i5<$~2<(-Xg`}Ui;x|Q3sfmsu;)rsBoFxqi-o27uEe4hM$hvJG5=kG-){{0m_ zk$Vr*giSEeF_}kybdp_{VfgSmLq1x@E}D+v4Lbw`QdDcS#pe=Jy~Md!CJ$N5*R9`U zSfRGh4Dga%pvHl2Knx1w=Vfi4F&zkaYR~OkKXm$yD7^xJ9&)WuFB)&{4oz85=nlF5 zBa_*1@ocUcTVYGMRa>>$^N}%)3)Q_tI9f9+yP7#mJ7rQBoNdvrTYnI%TJ5%Uo z^3Xuwx9n=#@0W}7*nRy}$_hf>V! zXUmv8w&U}TEBg7JCjBdG-F{sXs6LRs*PbK%*&t)VVSO8TN@-SJ=zw5y>Kj*rqSrp! zO4V9>T8}*+eZKHHzn9$4Dh}mG-H>Ty@>{MX zx%Zr8EBJOtw13q`0}kF=;UaQ%j0S}7TSAmg_xH}$(5(iNQS&ce+bptJFc#&BF5Hf7 z|4ft77IxhzV?*E-W9aj2eN{t$`SgbcuP~s4klRSPo)#cu)H+MQ)-HmkJyLDC)b@?b zg?`wJ!Pj$5?Uv4F{rt;^e=w~XZx$c)rn{~waJ7t+lyp?=LXRF*ii9V38&RuTNNRfV z)j9?YzrYoA_^>#d#k8E4(%sH0lf0Q8yYHIi_b{?NWw<{;v|>isO`}dpY$egnKK7gS zP;Tj`3c7V0tyvuUl!5~Su<6pO^Cn$ur}9*k&5Wm8t9gjRC1}qc%*#L**Ql%!7)p5$ z0NCtMhmgG7d94~VbjW;;^-jwuNjqlxbiY2RVKI%7@XS7u{Y0b>nC zn)6ht9Za`%Z;Elal#R$9g?<*c(>tR*p`zDsZ78k!`JCr=60!6y@TGfOoS(Jhvu5ey z4#Zi;S-2Icm9(-65qV}Mxt&5RoZ9p=qv=Ugr|GD-&>ZoLwNh@4rX)#cl!aZfWvY@c zdvT^y`a`*gQ0Lev6mq|qhU9plEk4I{pa%H4G$k1P3$C;AXZL@$OOZxiNWUiRU zQDfx@TpWd`tx)GASH=5M_f(5R>S`7Z910pUXXd{Zym1;8-~YVMYj}MpWWogl`H}3@ z?E)IFA5?47Z|f`F9Mj3)Q8nrQiO(kOG}aAwa*A0TMV@<$LxyQE6@_{yjnVv@_50uL zc|q-z{d{`;Mg610nS*6fAFsCB3;pmy>!T<*4~wLhvrM29-Awk=(>)QZXf+=JJbHLR z{E6mTdY^tpzZthzqbs^>Iu?n^!H-&vk4CO%E~CohPpuGACdXx5hNLgk4@$T3$jRv^ zhHkIC(wKKa{MJ<{mu^v~7rA%y*w==?pIZ!EzHlE8OlZV8Ij1C0PF!N6C=6Dou7hNb zF6?t^2qZFt>`!)Er*+&+=5VW~-v0N=d3AQviaFD(yhl~1spiMbt>@(dNAZiWSzp?9 zFn8m+w)4NAop7>VM;(9}*NPv-SKaIdISbeap955wq>IRlH@x_M zJkVrz@gN&l)!8IJxi{`(?e8kODul;Qt&N@)%7s40eN+(yH9uXBgID>vKLNp-;g65z zn!o?SVWFp7!!~l}S=~?d<4{QeN{P#b)@n(uC=mDj$Wp6Ed(tR%vEJm1q)4HZ3H!PB zS8sMNnQucWyEpIqUUl_EwaX8o7kU8zrtsY1Xu&|Bi}o z^BTU=-HMJBJkdWRl$?pcyG0ORpoHGUVoIq%nq*$M*rZwhjXIp1BL&0Fc# z_UW~W!~8z=;Wl0)Z{6(SS9gri(>lNUDEuXUXrDR+RvzlGh;IbM7!5Rc^teQLDT7d< z|FW9&?997CDP=05946D$wVRk471!$10qfu7;Lw^9Vj(D16ep@L2&ncdSO%GlgZA{> zQ3eiYgbCV(ShYqQyE#fYmF@NEf`@?gD|Kf|o}BU8&TlT4L?ePHa3|FPc^T#2RaI?VBz( z3cUi|!}yCuKEFQU}>Vz(Sc|DBo~=Sv|~ zi!rp@Mc+6;=*cL%@yQ$ywE+Ir1;xNLN+CI-ysBeXw0?DKyB1Cw0i>%%LuTf<3#6mA zo$uZeKvko5u-)?ZxtdqoHFJ^zRP-dSMGela=AtlR!~*@(CG~J=NT^`q=!G#@Rcyw? zLrn5D2TlqKhK=IH{J$)M4Z8nIwG7qSe?q>{7_~9DNWNI;`81(uhq*LOD9f@;XWQ)# zFeB*qqZl|+`4??EU!RIoyJ+R{X5dy0%8Xz@cWEbpI$l(tKMK;&Q8h5x6L5e5 z@|S%6&OD0ssI6T;N)4#WzGWZWUDxP=d1b1JvrC%ZP!p%<3z${e+~A-uKuPqf&0<4; zsJQ;w@_?nPGf(%>1794{#TT~;B0dnUf+@+Lf6gzhRSDjrZsyU21C!-_G z_BavHp@Cffu|^j2H%YsbZd`MHhwD@G|Ye%!GK2>wz%a1lA_y{dGv+Sw91oAPHX}dsW?GshFFet7z5d3R9^_Gn^C}pCxxCNMj&Rf{z~#PE)b2z| ziZZwJ=!XiQ0!~J)#P3kWf0kC76x!%we_uIDp8 zCdOZ|QWd+)yCVd41@*9QP#T#!6Z=;%Oi6wNU>R-QrGXhERGnLl^Q{PAE-De9I5bDX z-|V0<`2>)-b4;JxnQ(dlutsY?lqRL;f7_t!J91ea4@<7(4D9N@d>&16C_<5H;qBvN zyu01tB-oyC*t^j?f0?wU2>c*Px^Dpbd>zOVO3@4P%z!w|DZK{kf^0?R&y+3(7H)@@ zIrxRNNy4d6y!J-bt9xrky-%I4)T^XQpX@1gWE=NX()6QTjQAg?nloTb$TTypk#D;- zPZ#PlA`IV9!7RwUQ~px2FSCT(qVKh5d*=3INJ=sFL9OFu2n}OiYy~Xj1V|;;i^JJrML`)dhL;;-~ZxkU*!|XWSi69 z=tAr{L=o{EZ1#Z|^gSgiRFs|8XpE>teED7d^tfV@!Y56CuaOmWQwA{ehSumt#iJq= zmb0+t$u$%~C*WFirs#6bE(j7IWN)V2S{aapMx~{~*H9nK%Br2x6n#;~+^Rs{qo6Ww zwar8Ks-$^GqhP?(3R||&{3v8ca!EMbDrJzA`M%^-QpH5wN?(86Nih1UvvVPorj@fR zYUe47hbRw)wW9)0F;evN8FF!${QBAi$=$lh4RP}BHkHpFD%fvOI-4A)udhESj>|t5 zRDnXbmp0p3AffjmtWjJWEWmk$c`zuj>u$8MAoD6eHI2JpmKpQ-i(%F0?Ck@98)(xQ z8PUjC(&cJ?t%Y;zZr>NE_B)*zss<(O*L78+FXApu$(5jlL}Iaxt>;jU@CO&H@F)fk z4?FS}Gu^PCO*_c+ERy;`Xdr~3wg{GR2nhh5W^Giok-0tIxQ!BWnSG1F$a;?bigxRj z1XI>9n?ug9?Ly!R%Ms1sTNNYMr(3o7HJRK?5aN{G@Zp`20WE- z)A1tmg7yi?>D4tE*rNa9g&Eq~k~7ciw3!TO!#{K)B%=}cAi~F>GTRsY9`{-k_+DE> zF;S!3glX?IZX7v>QQ6o?3R^&ndrQhX$cv8Qz)Gq&7h2b~XIHun1bMChkEH4EPFs)I zf+jPlDwqr*TFvmukw4c~=y=%W3(g^HL)Mk)KPUp+c3D7F^%a@D4iPVx*FD!Xy5tY- zz*DoSg+y$pA$nJt9cC+`(QLOyoVQi@7=?+aR?74VcgtTYHHl7ZKn2_*G(qQ`Bu=U(P{Er58|hmfNqnu{rmS^h@iN}c-sPR*@DtBkj5T1!vznE9J{yFa_x zxD>&^+tp>@HuM8`!#_2@^@S87xrW2V!aZz*+LoZhw(V_aZ)52=d9SBCF4}+K5mY3@ zmKbXAc8ZF(K<`l+-~`2uH_!mTnM(Z${%jhij!xE7JNFr3J-_^xE+RO2TVI-L(>QSm z@8ARu`K5*HjQV%ID|iqTh8@^@wUoa#yYr(^jMom z3_~94MQ{n_#OEARXLrb8cntpIuB?+^lcp0@FbV`FCi5(W2gGTXZJ7^VLT5Go7&%L% zG>bO-{;@i5Hak8%hzP_+Mf|7#-^KG_C0!z!*jGjYV8*^-AGcCvqCjAP7brmsjsr7 zZ-+hD8sTL15uTCYqp~N|o;a^s<2OHEV*06GzeJU;I74j#aZNgyC@`fcyC1(^RWbYd z&fGjnbApplVGlJKwDA(xq)M|`c)Yd2zG8s$Lgna?Fm-j&5`WYLwn5jNoXRhO9NCZyn8${SY zx`?!E)3}fZoAnViMo1`Az2MG6^9rJk7br@2bK|5l0>8-gFFmL00dVTTtO2CVOL+~G zPz^|uJoNjajYm|(4~F{Y&ZxD7RY?0u^?UN7HnhIMp(ZYtlPTYh(4n985cooAZNoQVPpyT zu;o~J;hW<16#bKa7Dsp(^(_;FgakbpHCduBN2y=;J|j z5T#jjwUE|xFI}2ZIREZM^ywjoj!@U^^qYVG5VuDhN!uVZ3iuoa1Sc77J=q#I(N6AI zaA#mrOD_0Q>++kgY~l$;o+e0he<_XmV0(3e%rP#PO7uvDF)y4D`R&(46Iq5eEYU$EW&Df6=vc|?}q3QnS z*Cy)2&P;<*_K^|i8&}tlx(4hC43F+Wd8)orCZQirJ^ow5@Ne4w?uMVXp?d2 z21iEu2*u!N^ZhM#3#)dW`PM?Q6na-*EDs!a@gtY@0o?Y}i%dr!<#OGV1PI6Q7CdrM zU&NBonqvP|$gE?`6}?LXTKAE~PF(!TpX3nZNs=QYn7Lr%n50kmid&-WpajkSh{yZ` zQ$=7=o$Ug*RsS`h{epMYNEF9ltM{QV%i6vq?cnRqM#^LU3)(>>T(#BY;mEu zMArF*IbTbqE{qtFbtB=1(<<@|wrkg+MQm@C;5&Ml1Ex;0dEVk`n0 zR3(S*>^OjHtbRwSI=hOO`AQ&MU$Qsh2JeBZe%TWH;FA?f>L z&CVuwU*Cy2_1a9`$>G@B1B2Jj$v7_PzjvDn`cVEy+dzHOjZAtOo%zgJVcX1+^iS@g z_3o=1)ainmR$UWg;G%vmD%#1WCMHJ1)yv4jrB?6ncBkO)QM0pKl^9}PyPXjy_EefK z9D|N)PKKAKm{5yJS_yMB{DA<~|w0>YA+=4!r3d+e?Hf`Zcshyb{shBGb5_i(=Nj(4Oc`sw+aLn%F=E7?Y z!vmllU#qVe)aeoP`PA0=Nm#5kV6!IFvugW93fT|yOY%#fX+U=eUO z7E|esD?TN@TxGiJnz35YPhDmbZt3ho#yp);Of@ib$@uIY@ay30f?%?Y_5NlzC~uu_)7g> zdAP7Ud!l8i(6-2+(HGtGKKA3@4O=u#+tPGwyC=7OD9swB{vKaG5dXOj0{JzJxg(AX z;Kv-w*oFUMmWFP?s`yc6+?cZDw`r_pL`C@SeY=sj-{OaSy4;S^rQvBU|KfuX?KYw{ zE6#VAw|O70e0A*faN|H$I*WKhg2piKiHcMgF)K|!Y(M|UWs2d(Te4Osc+b1bT|FLq zrP6sGIXbD}J4@>6wcPZDgC7ym}=m<1n*m-a0e)^QZ0&-y{%~dc(0e|exrHz5%jw%Y0>|9vXN;V63{&XNTWX=B{nDnGJFML4 zeqWo2zxf>f0W!eBLPAP}_4RArzw+@tYB9H37+7gaASWmgmk}pH=C?H*_0^TgEnJbe zfqw;(1%?#uTR4?v*P=6#6g50l8Vdc!v=qhu-Dq>@)a>N+Hr6G#EHWRj<)pfVx zV9axq;iOQ?UVV90P04=zj79%=-7QYp4YliI(aSmKhfR@4jzhvzc7N^w%W=A66kZN6 z1H|+AQK9N

yr$4Xq0+AbcHS!d3a=`HUg&{Y#rlqy>P@0lr>X6-6vlQj@=b$lt1VL!hyxF zv}kxP)vnc}dsp0gtKF^2ZW^4Aebk42d8eX0f*T+_Z?LU2tEqOMz^XpEd_t4*Yps{1 zKetq(aJ7H{C=e!>&QOic>}zxB0$T2F%gC)tM80k1s7W@-)N`zQn0x2uYvb`sNb{U# z3?Op;U|6XhzF8GcYMO1>bo6Rcm$8rBZqK%?JLCONn8hTQ##ob8R^& ziq{+xJ#wsJ{}t0j8US1Qu@>$$m^Aak@!*9$#&#JH?L19St>N)tkA1t*&NF#&kp+%X z-@1y+cg~&Q3awy13izX(3O3pISWC=z&!Fz7bpV=Pw^~1~`m5YC>Wh1sNf6lRS0$sE z7`cQvnBGyPxBIYs=Wpgd4E1;?eA+e9?++~GsgDmHGEVay@5wQ=z0iN$*KNi!M3X;k z3tr$CLg~C5Q+?Dv*Ff!7;FLXFQ@uIH)e@Pm{CIs}VlxFAr2U>JH2JeO7l4DK%Eg-K z5Oa=(M?!t+ z2k)YrG}ULrtH-A`?!0pO0jJT0>X^=j*K?Nd7D?LuAJH_v7LCP%*=b8VL{6TGJwvt^ z`vo%M8$}0egzH=^!SXzU)jRuA8BeS)n&(a)6K=?z&z^fO#;FW`)D-b1K~Ns59Qg;t zWM^WBp=b0q+ul>U{ZWZ_;=xk(_T)rA+riVvL;;&-T6}Sb%4s_24WQza!i96SBtH@l zbC0cl`8Kn5+5XGo5zp5dHLBh+Rl+Isw^gE89PsXGv;s?GHmf(%`kk zs4&~@yBAtsy808K_3TmeH2ZMvbjy0P8uYh=^Faj;r83-FXR+TrmT`t&ZGXk+uCOg4 z7^O^BTI!=B&=b}sNIjnZM>q*GMTTY0U1B2N8ZP_VX%g5RzLT`SLLr|Rqjy5IB&xBH zl^u^Gfl2$i%Qoj#=&vCMtES!CHTbXz*HK(eH4ong({LO5PlMCFAX5v@esrVt+SCqGQy1BeElNR&N1|-nb&FSjaE{|^H3ZnudjcUiQ+AduFfExp&a}g3um#eFH8v53GfQE1@g)0J(W2 zSVPD{b>6TgTj_i82i1WH@+*WH?doWpLqF!?6;HAb!7))cZwYre6U{)H@`I-8JAt)_ zLbHVmeG*$4Q4U_?#}w7)U=Dfw`gR-AGC}RK>8k=#{0n_y)aA5NSuI5#!wCDSEUvhI z?XZbmOZ(1_{C z^^2@|1c^T)`G7N@-!h>FgKChA3**#vMxpHe<@V0S*_hH^hNSEps94J5)VC?S4feJ)! zLG6A5jmn0ORimBfvt#qi>utxjrf*!S-fBFinsq6dghIa2smj4Mpa5m{oZw%yNn?)y z=amH)2uf-=1;-)gc?odabFen!5m}3Bb@=-?G=CP^KC&Q_F{7V z@w=tPdWJUlKgbzr+KXpDzvUwJ44KE|V24)+0b-47n?VCAG`ed}CEbt=lLmAE_;Ucu$b3e=ZbAgsViDrX9by>;(`mJ@& z;o-%iU^T-dk5Woce0dW^Kd7y|UPC#{Q2uWHGTSVWZ5u|$ z89}kM|NKt-OJmnpkXryASql`QI5*MV2+ZI(8vz?40hMIiLBQrpeJMWX)*$5Fr;+^A zNMF+O#=xen?K$5c1}DcW?0$c<;g`1l}|RvWD5 z`!H)N%Wn>D<-VaT_KVB7;lJ%y1a=Gc^tOlUtoDTQX75dK=6hg=P}4f4+{bAH74GKA z&fzG}y*C*3y9GMgKO^{pI<}k_D}v<3Fssa9K$;qUyE0kT|E=X2kdUv6G0@ zhKz&vCA*wz(J+qd(k_O70u;cR;@Sq4Y`>|nX-vd)3XnR{`{*VIU%}@^uDSJ*d(fJ2 z`}vrd5=8c#>Mg~HNzu`?j*;k=O!q(PfPs#Q-(je}?BKuzCPrpBfByHs?_#|4zwe?G z&IjzJIlc)SP5x8qSuv2r^=A6XuQeT_y>{X5_Z4VjA3}Gz8~=RUn6u?j56Q5Ldn@s< z`xFiJgWrsEVqTTJT(uVweK@o4#lsJ7*n90EPSm^?Q72{kri#=8{1Zmoea>N0dR{$= z%zX6WDNXcOV2h$;NWP1eiKwxgIgdsUUpLUiLQ|rsS4H&Ftr`5nPIZMAbEU3{G|yB& zJKCI`<9E9$H!d1x^f-tpoLXHS^#5@(3 zaWFnuGug8(V=@N>WJ=%Zr=rhMjqLSfB1-h4qnD8Z12Urz4*3IcSIKt3q#a`O9^KN^ zuKI`Oo?DPNTSUu-?6wwh3RK6v2H%&6n02@M?^K zwk4W12&Sl2m|dbB1xo+xv^`%I9N{=&TH%}?x}aU~y;6Pm9UtPCG}}byn|~Omc(&B4 zzgRzlNVuK@;qo%`L%LIYj?7NdpnLA$hVPj)etle@`Z1KH^6mxGPU=>fjmIj_{h zOQl_a|IGi|xk*7`8PU;37MAMAZLColgoR9Pg(P0kF52p@bLpFU6NcfF;= zvr$p+D0c@-q33JZa==bNuOLCZ@kin4Ov0{0+Qa3rb`E!py#j#LiXwA~v=}lib@_q5 z2d!~OE?j+eKSNKM?AtwfS+S+pPmROSt5QA^WnEcx{leKF*|+tf#7Rp^St}t@4C`q4 z>0!yiVD3*xdp+Vyv8-~_)S008w$4()g%Kvye#fp>cmGHTP#JIqz5dp_Eor9w-%Fkz zDaQ6}yo$7ln5gAZG!aba#`CM>lwG6jyVC%N9dD*Lf=0sy(m<_8k_k+6l|opG)+nxZ zco}YME6~nw{qp;}JWud?SRo7OwDAE*>7i7CBz(Xf^mGe?0WdzeHgEwTlxuadl`wpk zY>#Fx<$-E>l6Lnh_&o*#;FrY%0@vd56%Ds)*aPsli#odO>I2Jp(&yl=r`mf?bW6N~ zH;UpBYgy2K_j15ajNGHP{;40CfQ=n>?t(_^fuAlIQ z8+K|H=qtO}Pj7+0nq)RX&Xf!EJk}#9~$2IL>G!TT08`JXdxW6e&H*CGJU0R0qgZkSyz!pfv51n&u4QI;`v=iz>ut2&2=2+ z$_g#Q$)TxaebWZDWbgo!lt%~P`hDY}4IIzZZ>Y(RcwJ|I2`%s7H?U6Fy+OHCKh|)N zuJz63D}W@EtANaTf)*0~3HR@06{xa4-sRqM^uoPq3vAI7ut*H!tzOMVO-G+9gd)ix zz-aA&>fQFWwfO@dDEv15!=ylESl5&)beKiB05(%c-*Q^$NX2cMu-ht-f}@K{Z$PkP ziLjgz&`_%6 z({jr=Py|fBk*Cv~(0;o%^*~UBICb4+J>q#=I1B}YOy&A@yhLMGWc=V`uw{=!tap;8 z{X80y77V1Q?`%npP-hA#eq2>7S0VWjCtU=#j^=roTO&n14&jWTux8OuSq!zS znJ8D~v^bQ%&5zec>K8vj*4`viow1b8)K2fQ_P2;&=q)k0RuMoKltI?%g7SxbUB!mw zYsN?PR?Hf-@mq;3;L}%f0RX$OC?!s-patS4%C%idPU7HzW`(1n?1ISzKlBSw@h%AXkVEwj-0tyhvvwvqlr<$Ipw-=VfbIUJVjm*XpLS=J)aySnbHTAr^4q`c%rfRW4U1T^Yixz z;8No?c=)7{j(;kY;`2_2l3#PX-IoGbpsQZ=xdtMN#6n&&-Swmz5HPJ&#^)IZN-Oh> z`$TpbN6A{r-i+Ik6zQq{N;E!^^`o7ok2HzUgHB=#U+A%T9mwiX0&6=U7pTQIa@F+i0DSKLm^ZYydwaKYgnQBT1 zp7^HIw)xbR#o$I6qnOy7giA0^CZ~o`g!Ch3qweuiuKp}Wy=@K-%w6SWV1cQ-t z#dTG!_lGjpK=c$+f`cY;3jLV$D@E}KTTiVy*MIc#VJ$CsK#wEYIPVJ=Rf~ho!D)o+ zAyhx(mIcnE=fz?^Z88#oscypMvnq%-(db!sRJDw+ey#FTU>2sQ0G_y@a_YghM4_!rMQ?iYt&7> zkQ?_&5gg0|<(KuG6KfUc*nrXkEUwR5MU?D>>5NmD0(l3Z2t+;FS$ClIo|822l;%2_?*O(qrhAnWc4UZ%qgUetE6p@9b?Xc z&K*Gh8@>|TJ{iIBx9aNHq4vbd2id>8zz((e#9&x;eu6t2jd&qtAyoQ+VL)E}lMpz7Lz* zTV7F{!@T4ln;kEShOiY37I|T!k7W$pB?_ep{vnzNR~h}h?<}PY(#PHhO8AMWg3vU< zeMHR2sk+u zWZWg2)p|ZDzBK(>J30!*aEdWgn2`y?Kd%a<`6 zgz?jiNgJBrPvY#)IEj&`D4W=P04Ai#;db zA1-V0Fy&U}8!Cp~4{0uqwx{*MY<@Yrk@9{Vedfpc_D`({N#+Scm9wYQY*StC*jMIglH`RwX$4n|7T9t| zD&)D(Ve}No)z0_tXuR9AbS7@U@289v%_O}tEP+CsQ^P2eD4Ss-+8ivINAkhOdy{6O zI?`kIcgKdxdoJ91ADJ9DrZQyix)7d6ChSC}E-Q7C=HwzwF-( zUn_G#QiBKGZ!a%yn_f*I#ec2vV^pGhN;U>~ksRe_TV8fLI+>HG=AJWqyX3Jc#p8lk`h9h)7qIOM6#*lX zCl_4}{uAxL&g*OJyHH@gWBsr2N6@qXL}0D|`~SuU@&79y*tIU@SpUBNr|Q35VP_tj z0i*-sk$?HfLeHwmdf9p@3Y=T{P5@lGeW_%Z55nG{zYF)23wj2!fnuaG2cd*s@NE;I z#8E654Da{{%If%`6rQ0#H2xH7d}makJ>B9fq;%(jPSj@<*YV4>7@eS8swTc~HHkknKCnN8R}f{0M#bZ-ciLh30VRQb25?G6+)v zICN1Mqlyk-GW8G1=vLluLXQS&$q3|?kdo>M-tXX4-7#>-6?eIiz!2THe+Y`!I5Zv< zWa693Nf0%23JJv}z@bLaPXt03zE}XwH<%l6DT)KoBksdkc1nOel;Ih`%x6n`0My)! zFfWLA_zzbt*GT&150aseM_(bI^8py(5umkU!@co+SV-|Ig(ztxuYa8zwAtW98GJ`( z&8a?y9wbgZhCTmT6*?G{?DiA-mFF@$f;OH#!L|7vltRrAoB35>D}cGWucy3F;v@v@ zqIsA5$9#alWY_hBDfjzN8FHW3DO!U5;Gf$;$l`M&T4%F8i`F5R7gdL~3ybG<;LLz?PfiLVcBRGF@N&qaK3nDaW4*ofQ| zuiE%mZS~i=m7QLFudiQpUdNV^YL3o}donoTNfkB+ILnGgHl z^!n`7c-Q4_3&dd-C~cU6hI{NhEww*i(HFb?G(H~!(SD*NlNnTXBDEh@3;FW5Fqe>T z8LOrC`}2`4%(xApkU5ib6o12ic|3dS?|Z8*CQ#cIDcP~?hqk^FS3KONIv2)W^QZHD z7VP~mT;Wl>(vR=6FS4w1d#m;XE=D5ZS<#y-083BFvp55kJ;!XpDeH=6Tp-Q2N)r42 z8bwGytwp?3^ETN!Tai1*M6a1eo0X*?aS?~(fH$Z(O5!frHAI9j+?fGnbP?V=3rZEy zw3|X@Ecn(5wkcaST#R!)?gy#P_b-4;{M@T29Iv!>_m&4|pdyuJ#SWv-lf)J#H5ZqB z*Vqr*A3VvmIX#rqJ_NK4@3{`SRtKoTq6y*w>vvg0MjIl1xWA4)n4x;cmbgYOrEg=G9zo*tl!B~&Pj`Jt z1hROqe^1yx@u;YX|F3&GQ&W(uhwGGrmaf9596YL2tJ?GZ$1aAmtHYgNJtzi`?Xjo( zhJ#=9HD4jq1`MTsc6dDA#1jhws`O+rB`tOd5NjUfg3w%06ON}vunVCGK(k9uXALx7 zfs=`O+n)GG?X?-NRl2A-45yj$7TCI;`^II;cg66pWOCC!P^Ymh-o`*L9slZsc`W?$ z0Nx?LJdpaKPNhF31+Ix_j6B_1L z-9|{0s_-Uz#uN&=7HFKlSC%rUuC*<_4kjK)5Ep^KJ;LvseMoh_^GaYWIG|9hbUdE{`=dx*iPTQr ze5T^pTpE}HK-z`c@E*^Lpzy%GmeV)#f?-%mR;Cc`{PjJG5Cg}blgo;;inCRBaY|=% zxzLSs{KKS{3h_Lhl-z|LwE);hgKWL1x3qgsKCaXc*h$N4Bd!Dz&(_b9&+<#)wpkws z0TF>{5Mr&(eBqm)00*RBVVKJlWJf3g#Z@_aM|sVJnVQw1*z>2;I|qt6FBK`FZ)r16 z+Pl>jPMyg3) zu^v;fWP6YUuI2<16Q-zVSs=wLO!-@qs8ZMJR6a$4u2HZ>Oz950dOl_}?Igf;~G zEt)`MJ$d^kyoUhF`#BH3sP+oOrFp;braSisD|EXw)*{Y07iu@yCv_zTv=|%Mf�f z@P7Q9(dcG-A7=|E5}-dl6{S;8V~S|yk^+h}?hV7)49+$OJ=L}*G=R@fnI(e{g9L@F zm8x93w1~{LWk+ML%4Ig=$U0qeh%XuCT7U?nKC1^`sMePI=9S$&dilLB1#+PyJ9hI5 zfQ74G&P6UexXpepp;5O|GEM@@-pk1r_`y{9i7Mh;U7=&}I7g=(Vy4Y?-TG4=3HyH5 z%^0aVN575fvH-gf5|>%1;T$oef(cpsv2FCVyVggRw|-pa(759|ZPuOsh&&gG*wZXH zH<)X0sPT+%&02t#pXx!R*GE3gtzNRdX)*ro7=Tf1DRX0_ zb14OWW^`+t#rBCJtu{;|V+*Dr-*2U_A}GCtaqXQ*(x!6hvNr#JZ?K*}s1D*Q)P8=r z5MJ@71GHj0Eums7pifKw0`poY1#n~`@bl0DQqmpE0^c03~_F z0R(2yP#ThF4@`^E9hm_=npn{|d}t?%MEm?TGPDABTe`sarvfBufI*fz@ShC$O_?*m z3|E7LZwj*pS4K%w6}SOVL&K(i8fpLoW(#0?RuDI^Rq<)>djt4OP*=E+io&OC{<=nnIIBYFnSFFlf)zxi z1xkX+8cs<^s6WzIlTY9Nc3^7JtULa_SbN6CqK??7qBNB`6a6dbO zBKgR3V|UT4>>?l3W(B1@2bCxVo9u zDxkjdlcvz{A#{waKZC-BM7!FV8fCf%eN}L5Umjsv z!kEFc;fn!4w|1)PSQtsmmfe9JfD*AjW?Z#x%lrfc0aE6{B)dg%3MDx#_+O&42PQID z0CKlsC*dHdct76^I7fGNQTIID)K+iWir{c3wC)N+8HufPAe&ebD)QS#o0Y@H{s9eX z#_?=G0StOgJh{_cwgf`nmIEC^@1Ib9=@+WM5S6zJx|vhpWB)1VXYk?P^no={ZC4}_ zL0x}xlB#Fm-s7)dpxUEtcAgA?lXaJ3d9X>BIcCW2yC*vtbgSg1&(Tt`7cU`^x{FfI z&qvOGmofx-(ph`=6O%|T6qy8_p-GR07G#!<+?g1kOca=IL(-7&Z9NLco0tRH%dO&F z=XEH#UbWe z0-p>r#+oc*fYa%@X@=vmZ^+?#7Lyhbf?}TknpQk<)cgXIm~p008W$ipwjXb(W8UN1 zQDRK(Dy<+H9Wid{_F>Ls46x)nyxe;`Vk_F<^q9Jx0jP@COZn}_yjmE7v5dXi*h-pO zboVK1aq6mxLsa}h<0%XK@!Y-PC70(3lwSZ8*JPVWTiI9TR1`Z)=6;oob^e>vON4)z zD9u~xE5ymC8~d5`Yc{I2GmWK>7i7rqL)xEAOCnq5WC$D!-45zEEmj;9#51(}m7JOS z_MN9d7g;$&49&!YC^4mgkiTHaeAt+K?GlQRC=mFekCQ1#Xtb=jWnK%Vz^SKM$dirp zc#%F*uJ-=RZetw@`pS?w_4qTm1$$jf5fY8dZ_{5YBTYm$&iTQXn|(E4cP)!ViPu}X zQr=%iOPt9(SN_v-d3ue!!$&(RltM+@?74RPspC_>Jn`rAd4K|4ZPuV4N6`UD%=OJkFAcx zxlj36_m)47G)NnHF3*OaH1JAmRfn#kH13Wv7tC7_vto>0Epp+uS{YYNoy6M9O$eF` z98M+^2+1BT`6CbM}gUTzrs(^=9(ykCO%e3~&TId|TZmP_Bz_`Rc4hW_nh z%LF69f->c+xpVtq-b7utpFPb~CQ)V+Oa|3^Y?e%V#moHSc?v&lQn)s{mrAx zM?xQap4uEzHYP8CZ1Vf#)*#Ta{$UZYD?4sOFxypmJN9~jlD`APWEV9`0B5^bo$~}t zGRraWT821=1|NaqG~BL1-zC3~SqG2=I7%>-1j^Ewn=bf2@enS;fUGJ!3$ANqB<*iL z3d(@+-PvJlO!u6VcE(T6reHu*AU*;Vfujq*q>C2q%gY?DhcT#^O}IY9K*PXB;B7K0 zd+ne=V98Lvgmt)pct0T9K-$0OwV7zRI>_8A=T!RT8egGF*g^_mec4B8DP3J2{L~Aw zsi{RVkkS~GG+FzOAFKTe9kAw?qxz=T#LveHKsqJ_P#^M2{3e8yAMZgZy*slvPrX$x z8(MyYPZA@Vai<0ULPNYH9eX?cI4!$(H8$R1rpXHul*+%EM7GkD>mX6X;O7#wdXtLw#mjT>9Q$X1MA+e*#P;A5$ZS`W zW1|m&mQ!I)D~{gC{Dqa()S_dhJ;3)N*E_Jlm(si*A3flrTac|@Qxwv&@HmofW@exY z9YCz9UPioTL0vFc+@GBVsj#=;*Nfm{YR$e~X_@`TbVe(TK1U~MYA_S|0_NiOI|6J27=hZwO&Ync5uGN)^ao@}C z6{?xOYw0&zlMgWVGp4vf!B9f`V(UDl4Rs36;ht>IJ3YY03x>z3vd83jPk57=Mx3ijNZR-Na&M z*lEI}PZ2_lWId`G#>Paq$I@L9SDE)PeEGUzHOIo~423%cbu-M};7nQ;vshN_q;|Q_ z0bS}J=SUms>Asl<F?|kR^yD&OA^vl4WI?sAvl) z=~9`kHMx87sa`(kP$UwwuPE$z&-dQNqx!vCmc1V%<#SioIlB^oIv2-cjDmCqRuLTw#wR}szg@^WUX!p*-o#@ zy}}*Xpp#@{Of_Sf`&h?HUN*nG6GyV1XGSL4vEH+O27zE!1Eo>(8r zDACWHM#3eU%KFdX@p@UNnC@)i%))~bqj%K>?~2#7eE}n36HhYa$&4<^d89<*F~aXO z8OVqanEi1|??=&y!K3)=Au=b1;zWu24G0gFk9EDb#r~Sh%xX$ZDA5q|_3QjH=%MB= z+{$=c9bdTh80Nxo98<7aNh7_(nab}YB9wpkE_TP#&l$2D<~u)TT9eqe)34)iQKb12 zpPXvTY6S0qV<7LuPk%gRybCXcw5~9 zX`48=-`M(R<8rGaZ3|(ODw(%+M!l1^WlE3p5MQExm$mQCte_m>;}Y5oftR68HFvj` zskOQ}QeCQ-ze()Ne-jn4ki7Ek^sfVxq>!0Ac1P4`l^SHuKr%;Qioo;+;h>?9qgA=X z$_>Pld+{5&>BW*)??vR*c=33)j8q4#wQebP((Q42ae>iaNA4!S zE!9cz4mLb999zw@bGS8~l(cmR685$D=iSI4kNDqE)cUme?TZL!0-v#g$5<=jLHrJs zB7qxig=mxN>>fuLp!H1wM|AYvLcs2dSkUSOb^PWxr-{NLx2aLeeNAxFaI}tud1Qr0 zIP>PD?8r_U18cy+OTEFue9E|RqH2e`?xpYMR1!kjuGi>pB(Y_E(8c8UBs-@H_H6y< z7H9<5tYDaEWR&9DU4ds5#75qnR*~8%#w)4M?VU84#B60@tBiHR$2JPi9KXzS&2H#BX_#2=B^ z5n=>>Ys6h(1P0EkVeCM%OGz_`vz-%`oD-^}|D> z@x$#LhXSeM2cXB0_}8O8-|;{)<;R*eWFr#&)E4fg+tpt)>R{z3-y1ta&f3;3VyD>3 z_V(K*6m}tJ&9Tu>Z>T5`Ggv#mTIli0k@M_AbcUKTBJEeK=IR$`w}0Jut|_90xus$G z$T*M!f6@~6W^yaHG;LH?B*v6E3*Alr8Z{w~#5FHXa~z(YFE6~0np}$``p-SvHr1!^ z7sgvl!qkTsC>E4vKXx=N^11V8iBXgCG;HkfDRO3lVrG@-vDTVDF3MMUNcii9;04_8 z_E?twQ|@Kf+wYKWHNke@Z^F`&U&&Wvcw<63a_CXL@n5s^Wdo@f`END8{C(@43@J*q zAgg9R)r80Ie%eqxGg>C1+njJBF+cH7_l?;r?xroe(PlhD3)Cmm9(%eCJd6O&FC7%q z-Slxw%XrS!ma2`7e7_cxe(@jpy)*%9xB0D(Z%PQI=O@JfyRG6&uD-umnR>KD zhUlNDxSZCfczB=7hf0XppoI$>GW`IhoX~h@Etc zFA19}z7(D8Ntl$?-Zi=Kh`i9HWK93AZ*WjkXpML_M)^YT!o$M6`S#C?1HJvG%l=j? z6Y;G+JNHm0t+GP}Oq6Fc+q~mTwl4zMb17}4K7P2Tyo9{+K%-kVQ?dy8Gb+*xN|uv? zv{UT@PtZSW&^KU`a<{s;*K_G-R@z>}hq`d(N=h&UVsA&bw<;OxIDEU)ByGK@<}7$c z=WWBois#8ipo5amVT8C9sK0or?R|~od^~vv1p&*{+!*D|uvS1gcV~_`;Clh+zOJ$g zp#*NyFhSbOI6tBa)I}l`?BZ9uhD;c+;uSMqbiYhPC?_|=wx(SoB9lg z9jU|1bY^CSkG8P2`)!S*)^8{_@krfX5lMziz8Wm|L%wjyW_a%Sh9DOLwAZ*tUe`|J zUl-_lB~8u4@a7MFN_p8R$8Z(`c{f)HOFqdm*yN2igLFy&{{$aOsI5pb2d^kODF09R5e^pa6o3Tb&1l&C_7a$7%N&5Zi7GUJBOEX+;im-86Fqj zJIE5uj}#a%FzGm^SPBu57So_kxr>+2^#RG6109iP41n%SSIlWggf6G}PDoJTht{h# zjGw=nv~h)TLTx9Zuz~>w$6G#Sf%q#+lcjBwvqnv-kSm@@`B-Y?E7X+&8kt&IBy3l? z-h1biu8^V1d=lj~Aj<H===h1EhEkY`OP^vWI3c-L9}FF9fWA5e&9 zcR*IzF$x5SSwj6>(x6k%P|RWQw=u2@Szrj@HJ@4tDcl4Ai$@w*EGRxjtvJC#7oNa| zXD~distrwfEN>btbBbiZAdWk-U6n?D^p{0*dFqn)JydH^N>=$i!2FYP)^u~cYd%9c z$;A#D*flqrP0k3&#!%DM1!rkPM3;MylE}DX-j0+c7yH4JcRPK$X*CmqS`GRIZ4h@e zNZ~^i$4YCzW-=vY_*x(&*?K0*)`aP^W;gJi)_8Fyf)F7yAqJ3g%|c5TkBGy zPpDm>z>vElZ=@t2T547*g2nRo+mkWUr^x9pj2g6Lg>rVE`r0Ty8jIJVP4kwelp6U7 zjbN@Uo?aUheB%K^T4A*Alv@F1nM`dX)3T>No)~s7Gx?pWJ&^Xqv^e*?uh+p`=Le!ET)jy$WJS$scmowut1|tU^ z&-Tsug_-?(5=;J-WF382$X9LXcAPKQ@K@dj(S0MTk?l%HYw$+L`udguoUfZ6!E+B? zj{Xb^r_sfe_@rH9b$iA4Wq8CHRd>u~UfR-%f8cmRG?DT26^iFjr9Gk}4b3Jp63ktz zlXKVAeN0;U_771r-gz(MOwEpV-=WOU@VJ^-GEy&X!<>Vq?w>iWh%Wec3(Ovro#&!8 zcK+3du#y|bQthS8^Pa{ht3fx*eDLZw0}pYS{gKZ~T=}N-!i|~ssoKPzAd><+$$v|_-@FHKe{%ylK>ZC|d%%0o)t>rk!!NVL2~Czx)`NZ)&kL0Y5GUbQjbNc=Mx})5%n#Wd$4%iXy%cUR360%x9);x3u zxU%K2iKeIJeYsW+K_SL`yH-j8zPtLe`EN-_OZmG=10-LG{ikSSxc4} zCz&;P3oYWCdkF5H^po_fgB3W7mDQ%|K<0M%jSA@3B=5Jf)!4L`ZshhZT(sLC<{o@_ z`9M;N>P(U+c>!82KNGJVF)`!Dm%5k5sUK267v1miyim+Dpi_DLn>j`3UaQN!V7b=% zvg9-4%4cPrL(LB>kkSR8QlAy{Hf=^1M8L8Hta@|isRqj-$dO} z51DELt2ivgu(LqQ0c4lL@akZwOJ7asXL#7LZpz zLE3uVXgxM;_rGP&Wgthy0q)F`kazQs5u3(#@Ce$23;d(kdYac!$@K)S-TPSo$s>DK z^6E~m^#4RFZg{{+__Qt>LVrB?zY-DsuLLv52K%23uK!)o$^Y9AIWau268s?nb0RLP{FxMsm|4Y`P?+rMtV{YyY0- z|DJI^ob!Hp8H2GMxbM1Rt-0o$>ke1?Ao~o16axVP;n`a`hzbG%QVRGrL4N}NQetjaN()P6jy7nYT=a-}MaH@n+2E9ZVji9mmG5rIVcK_k7k zVr#zX=R}G6z{<(Qgqn7lUB)Edp9c zuu%rPBZ-n#^54Dfjv$y!_O`LHk!vrM&iRlbSmS^RSo^Fx}?seSLe_ zz8=lD_*It5S4}doVN?9}F{ZDyvW-Pi^x8CV-T&db@J1N+hShedO&>KSl8EDZ7*)MZ zCl;P;40+s~UGq>rY&b(`+Ag5UVuaH3UpV6g(R(+y`bnZ^ZC`t$NN6?n97A$$-#^sZ zEeN_ULZqcp%>v$L`pV4sJ!w3-#(nXEHd)Up{6>Y8+cYeZS@lb37E_0*x%s=LHsAHF zu{@PFAJ6TX>byi&Z56hz^DocH2K{%jIQwiZ2^^;+!8qDBmotlaUnDwj4i`_Ay|MjU zWT?G4l9hbEJ3V0Ee$RkO$Z8bu7GEZdYRg>e6^);d-2zlyqg*GSy|U#bLLrq~x}yET z%m0vFd8X;&uQfe0^V_sHwmL>6ZKb$+|Il8N;*@As+SQRGnr9nfn&WtF<%It5e|%sU zqco%x@j6q|U%hg=zFLjNjKaKLjTKwloBb&|qf3MFWh6F%QL(@F(;w5lnd%)(tktHr zo1IrpD__X&YO)|JR)Z;A&y$S@=_ksxe~LShSP~n32}GmvQ9@1b>+7>Ra?sP$3xs8- zE-Z0=umMx)baVEJm!Zsdt{#ty>hQJp$3F${t%R8={fBs`XY@IsXj}~WRgPiABSB&Y z!jyB}$4+m==L~9og}-2&cvv1F^}qAN%8Z-qIPO@syQA)y(I#i~HreEyvR&{0w!syU z<#$s&>2xRAqeUeP!(yAWTwu|xNdNn!XQ%1c_f6=p$!U8`7{YA`mEQIB^$s#=@D`U( z2if-a_D+k4>rUdN=Pm}8`t{|KzuDu%y_J~5@I<+;rmHUjugz&Sj55KX)$48Hhl~kF z8BMa7dHl4mSyW4vxjd#bV`xy2W%IqJZpYKm?>e6uXMKG=)xcu~ zRaH+f{hIe*(43;kdE)l8p+E1E#qOl0q55Tm_4M${4cK%R=P5-+MMwkOJUs7XMQ@6j znVI{n=5EHzw0Zvg>yF&7AUZ^*RE4^H`V)wZrr>*X#w9Hyla@8tU~eeZ;COfC_-BZ3 zsmvgB{q&Vukx`n6w_D(55{Kd9!NuA7LGvN|QTn%6nrTp%ZI$I}E-tPD?}H|o%fY-r zC~Wrorr^J!-GtT6X(3uWF?=fEcc}Ddo#uc7l-YSY3$8=+|B-_V9{_# zlkoJEuSPj*h*53U(6Z^&D#czeJw9xn)eWW#{87t)_YLXxjZp_eq5qtRez|S~`;jnt zsO#=jYAlS*zQy?RIi-_I+gAl`&-yfM6^6j(J zyKu%#PW1}C*=-ZM_Io!~<&Oa*{}34-EYFh-KtzPvgB@FNeY*L*?RtYu_1(LOz9ldd zqNmnfR38r)Ti=6q&{_-v(oUN^d(&?V#Eq4%(4m3&goO7CP*)zXDMlu8q|j7tY|1=m zbEPB6{l->i-dK)`-&0L{Jg_y)e|+e8V?}nw+5JeyYh}8#)ZUJ<;o=a?F=IB6q%Cp@ zW{ywabN!o@KwC$(Vp1}*YJrvr<>u!H$!WQ$33o&-=g5 zBO(@}{th2j=6?xj6sypm-lGRBJmaMD+Io5FRED_=f?Xkok3-hOzk2sYodea$5_NU) z@afOalK!|c6hc$>hzx#b+Z=uMwsUDwrq`KMm4-{?HY@f>{$w}yNSg_*R0b6f3o8R; zw-ttbe#uJ%)21xTvJ!|Lg~ zU|M-AsT%fYtBivucXvQy5ga+rBfSKhVfe0V#2{dM$h+uL-`}5(jxMDFy2J8kxifh3 z*86O0Ofoj7@G!CgR-Tc8@FI}q;&xn~Lwx|pLDw&HRrUfITR8jS%h`>s*eRo8rVYzG zEY9H1>qIDRGY&Pg%^<}0?08<8^f(`dwmPoLoc4Y4{PefB7~|L73XIDC^&u<_msSC! zt%e(i&d~7nHc|?sOabSf*M4^%1So%h`Cf63QQ2IL*Ih~%>TZkPpQ$#_@mZ0f7r;|W z>AEH2n4#Ad!p}N~70J-b1vil^gdUUQl~32}XqNA#OPuL+?3fGOPAS1GmH0OPxCNm3x`&?X8w>0fSt~Q&%w^L1$;@@}yPr ztL?(9`G@bXxjR0;l=_k~LLpo3npx$e=;#UXQyEfStM_GE8lQbd(qxfxS7K)7!Prv9 zb38}8znOemg$6Oyn-j&V8Uv2(cNcT^s7P`xd{;Ent!$?oL(a?{kXOWha?1|~3(Z89 zO_G=tNt4B@vaNo)mfPJHVeLo?_m{!Wb0=EDAee8g?@m{89QiCbX_xEpx?-^orSiZ^ zLT3J(JIkg+%V<&0lOKKhnQa`4lnrifND8*UK0mc0@hjuMZTorYt37eoj<$w3z=3W- z9w1S*=V z_+#||ZDJd$cG2TI^->LiyIk*$REv3ohUPhs%?$TF(Z^tHQKf<7d<(L&MpC2PX7X66 zD)Vb&%TG6F+dEw7F|6%pt)6zqExz}Hqf$q99CE(|3dIIN+}w5wD4vc?Vw$5u{hF3$ z^_O{){12m>xf|KfMEV? zz3(}qx?df;IQ-)j7MAx?JZUqV_sytBu5a9f5?0S|uhI)4A+{^)*w>zb%nXSAz|PU@v)9?Cp`$C`#EIk)lfH*!YRKbH7fl0h8* z*}MqCx{yLV*NzCJ8$~eztJT-=iI>B?&Yxg5`)>s-__rEo|rbB^uL-6 zr4~h?@>oV>#_0Xfjjw4$My%^pKZss*dkJu7DK@SJPX#|l+c9F$prN51GGgR&Nn366 z(4E!o=cckWk@z?W!%5|Sar7}Z-4*)HUgvu59w#@Be%l?f=nch8&-c~45+Gr= z7+AotB{2L@ns41h2Yre(%F}#6o_{{__C6S8^9OxPJ4ee=P+GWj*3t$4t%RuahJw03 z@$A1CbJOa(sn{rO&Kpv5K_sb)IYA0>TS_7pCAfyKCeHC`L>W znz+%wS1C@I?G?weF&x5oB?<}(YCJSsC%l+M>}rPGmg{vegIUJBh7~Jj3dI6C$S0!x zx@iOF-PY5Mrh}ukb6q>wtbnODUhg-Zs)~x~*&48*G+U=FyJaHl?fPUL!sEG9xJT+I zV#EqaCViCOJe4SIouZgq-P{s!yl^iZk+0=k^f?`}Yc1ptRl0qlTDCJCgU5-E?s63C zWP(mfao_sO(EmO^NItQ)`rg>SR63Gy2S!4Q{hUIO?gfK_5>ba*+tte;vx=m4?zC6_ z{`FZRHz|Y{Ahlf}DcCPc$uE-g*e@kBVaMn)83VjADw5OQf-}J%WQNWuvi&jR^{05* zTXSvs=XEROj`f2C!w&&tC1D-d~lB%fh8J zfN$iScL3q{)Lj_US&x6tbA$17CPpUS!fM1xgl6RL{E;%IZPZ(5Jk z1M-%Fe~GqJuw8D5ypC^QOjhbZ;Q76M{3&2oK#6IFk`-AK2_-Lkl(kLo2r)T6o+g@< zx8|0g)fr_Wx_-g4o*Z>3h2e$a4)lD;U^$c!!zJ9SgsFao&1Ka{!N|_8f~DVH_i9#i zge~F}GZ0ZHPU-8CvS!9Qzo3P_PEzdHt~6hE-szO`bOk*=Ds09=&xMbwhyLO2bYv&A zFT;ItJ6YqEf<$QUL^e#4^w}m>L<*nfXvcxmLU#nan)`Nzv)&xxcXli!iH@eIRr}by z11)*WCWundxsi{3GQ0i+l%mz3V`&`J1Lf?3=(3!VRHPn?oavt?ij>(~pVww?k7z22 zQnltU@!g9G-uSqvl*=58CNv#4DX<^s$S+|y|OEpce z3!%fTb+=Sd+QZ71DJ9JnXG3h;nQD{J1(?wV-Ci=s*5$LViapoBt`^VJ39Pp=?V)WO zL%08|le#{T`!W&|JV|-^9QlW9(gFxwYRlLJjMO!jxB=O`qu;98WyWU%uW}_Cve~ez zrv6a;&1?Q}lpkOicTs@pe6~xpD(c#R@nmgIayMd_^FmG8ri;Xs)jFol6hU7^5T}dF z9VPHZ0V~f=Q`?u{Z!(N;I?<^3Ma0fb^c_b;E>rx8b7%s}0d!wjvF@I!0mXrU9v>ct zOMsHruedQB2WrmI)CWuHZY>_&Exg~3QUS})$Qc(@N^;DI%U62!V9EJ;Ha-?!7kkBi zEi@hyl5}4l`GXe81{4ey1_GvJ6kb_+Kc`*pSt=LGLgats3}L!JHd!3YwR>7_)T`Bi z*5?c*f}I@B65N~~w_Yy3`idI8TB-O963&z|FZ=KZ7_Nd^}$= zc|G1=s$1i}i^!J8GxV~I>k6gODHFCeW{t4r4_-!~I`OFFvgdgxy7J~?z=_?7E@`=o z((vl>(*KbWR7$Gh&-tm25F?$@*cMV~iEgbugC&dWQ$jW^x#dpsF8=Ab=m-wfmPdYn z*N%vGG(#W9dw4<1;eA*2TlC+3C%@7FmGhN1ysX`;K&J0=*sJm<&f9o;mo0r1Na&DA z_RE1Xo$RW*cvF><$2>|_dxF1RIby$NfL#?)s3-2D7}s-BwuF>Tl1+L8j-Ed&u;S&?6|=%Ybtm5 zFSUL@(B|N;62;}jUe1kKWbys)jN`?;cwl#I`*(|fTOf8SKqNpSsQ$xjOlP2Ymz9G_ z>HcQ7@^-&BZ{~FYersG4XUMCi&7si`8%z~j`W#lmt(RoT71=v3?^%M{< zW-YQ(E%K_^P;3MkYk1lgpA=;O+3k9?egFN#(Auz^0p%p&S;rvlxD$Elud zl*j$2PoFk|`Yza+F8u%fe%eqN@w;U_L(sLR`2)T!EQ;d-gVKJsMw-CLtyb*9oPxV% z6#5f$k#;cbb819-C*+xjFoT3c^1Am;7q(amV?fnHY+=7L`Q~tjmE=>4N|HZ+ zKFV;(|1=-sCg3obKUlSITzR5?MfM1rA2}G5j=5CJ;Fhnxt=$Q1DG)!1?;Rg>`sbHQkoa7oc~)` z7~t!KlJJ-p9SZr6NDmVZQU@A(UYR`uq=eh;wPiIVnF4Ax>J)6q$Mhn|aMY5Sd8dM3 z{;Doy>P4;AMWT>`l{w<;i*I6Wf&G`ZtnxVYoTWVFWP zTE2x0kh1f?*=mDWBkR<1`Vp9Y?79s&C)A)Fou2Ec^z^lxA>;c3L9JE$ zSjGr7A5;ILA>?hAKbjk*@nP?k&Xdu10ft+|Sz&~4@8aFhME`eIJlJ)*Gd(v*Cewm7 z&|3WH5LfP;67PcBpSBu!cTYIRuG=Jczwe2%PI8>;jB5o|uwnLC;d zJVo8)ghIzYe(|kJZkjJtq;fr&8z6>N_iWC;P!!J55g+?2W;f5T9*!)6z#r>iK+f+V z9v#wcL3@D>>#Ok!6zYkgD2-u^;ey5;dww4At$p(ECz=Q|Q0tt^PKmoD!pY0_E)JEb?i#f6Wa3a{LGAFcJ6R%XS#826$&fYl=H9-;uL!|OeSl{!lIYzs zhg>2-rOxhZEdRQ@dMVJHu(@=;%!v%^#dV8ISbF*AC{C8rk$;~QNip&2cs(%=uB(tg zNx{cavVY!(x5X~o$}`d4rGP`MRvD@D0vA`=0euJI#cZe8{aFGu`valXsPV%hoY7x( zr9md#6_=rX{;a0)3pQ0n%`#lFVfqX+qB-s|7)N`D?t!sRlj(Evw)9ph34zTTWWsbf zj-#yn4*#MaiY=jl4v}&L<{!75dWxQHPh?)2^F;SXnSCag>5pz?6VsAxT!R4;aP8;i z-=N|`i$6+LQ)Kq7=SdhkT$73GXO<}?SRvJ1ynnyjw-pz7(9+P*<&iF68LQcr`xQ%hVw?c8qaYo8c1pE;s zt0L#+m?Dt@Z|gq89pPy?rXC{@Ofi&NB9JCa9LT!S^8m<+XEWCD%XR!FRKV@)?NvhS zTKsXY12|jg2SvrpE2BJbiasg$)pDZPQV0Xh0J+zN{0&f&1eSyvS${H;-op=9`UMUz z)~e1GBo2K};pCBh#yyzu9)4I1R|kF^)OEKcsSjvDv-P&Rfx(|Y?h<>GP*AV|ImpF{ zKOBCyhzz-Vy>i=&t<8dD1s^hWmdmaH**W0~Kw^5#A3FF&IITPHKoOV^XhWy_^JyZj zi%}`;;$2Xt>WbbSNqzeo+%#0En8jY%dXc6sq5uw|!RQ2&K95d#E}i@!kzI7y3<~YC z%wWN!BKOp!Bn=~9%TAx=Kr}Ta&n4f>;xoRtzkJBp;)*;!dAEyrz3WX$Ar8K{_?Su~YZmq2Y)OpP8zDhE7nIAi zxLh4W;sWUbnXb(L^e@+cY9W3Ku@N5lOI*(X=ToIaVqxrZ(R*aY)?1^}fGzmRGj2u; zf(^I`u|AL-83!~elLn=BHcbOXWv9!T5ZXk3sp1+?{oiZNdmgI9!O1GwSlj{E;`5fHg}r+ewaR$x^19H}Eoptgx*Xibn=0cH2E0Fn4 z26;w`rc^8{(Ijx|Lprc4(VVHMukPSPt6+UdIc&}ZjW{?={BozXG6{o#r57xY5~MmJ z0}lzdu4C6Hb7dgB{I&3Hadp`1_YSxnMT+98MHjbkIhG-4fwOE@X<8^zDWeDjRK91* z=b74GAV#Qv4b$#mZ@+Vb&$t?b-_f!}-Q5>5yRSeCFZF#ZC&wxF81brt>5LZ-SNZG6 z+6x=ujJ{0(x?CEu$_0{~D-^%KDHo?MRxj64_YJhN3?Jz4uesIoF>c00BtXJmfCi@~ z06m)lOmY=Q?Vf}+ZrdQB=3fDYyakF~ZiH?yN@M0twhP%MKq9N1NBpn?^yAVW6bdhP z%66M&V<}7Oc0tg0ywS0_#FXGe5blsVF3z(4~P$e#NY|)!7T3(n4 zWuYz?n;~{)WI%&N@DmndNe5(jIvV`Wr&xCZL@`1y8+~nT%GIH2@AFB`buhLXx6Xgw z^A(!}Ppg!uF+0ivc1b-z%=gBnVG+55OJevC&ZW0#e%w6jF0_3T*)sE&Jp-d6ABFMjiA8ZirC5QM~a% z%jt(}Kn|!MAzM;(xn%aXP2#LA%in-(tsFQE_*`|)<(syRp^)2M;_?OJ32bDG)~}@Z7<`E_J3K%pQ}Ub#Dy-5^%nuOG|?pI4$e~ct#Q(F!-dfun9Map5*uc6uGWk*e2Yoq>qMzIA=un`5tAHA2&cm@gS z-2kCk%L5qjrjD;qWskZL$Xl}@5En$4(Gx@zq8P0wFC6A<8=CI4e5y&Gu1=4peVLM| zb;#i*Fg!_G7~2|$bObnt1DQ}2k+kPoK|ukuHhL8nTe=J6S?z_x9ezm#0xNrTapT(% zakSCa-6&g3YMt8-{=}z+FL{Vor@vYTS9|<@e7F=# zZV0qf6pti_mhrlzLXo5}a1^wh%-YY~R8>^e+)I?tti=Zyp^a?Mh&jc9KVWCP+SN=% zkpJXt<`c14IQ26k6o1PhNJsdI6!9yanHwAbEe~(I<+>S695z_b4$L?ODadrrS6ri= z@|ThQV=b+0OcDDT&^hs;hTuk{^w&(8YDHgONmK>ey&K@ly81X}I8NGwR-xB4zJ5+S zDi!|(j90+Z-9M|F7{OdV)>CE5U}5eN^}M0+=(-YFh`=CT3MJ%J%)z8s6^z4OtTYsx z)Q`q|$Vy6LfTNx58E6fqeQbuvR@wnj2K)73P=|SpU=5=Ildb)xz|{ak=>fdAa=AJ+ zrs0#TqbFokRFyh;=f|C>Y@O&V=h*Dtw}m;wpMVKQeU}9nSD=SbfqMd2Wh#4$TOS{8 zSrEHxUg4I9_MS;k-+j|GDCYJY>z`=xbZ)2*v|F~WXp*6K$mYdO?O@ZbcVQltn&P{_ z49u}(c6M!7`imxk5%fNhN-{fWT2t$nlJO{BI9KQlHVOy&!>D>x(i_d#^3vVojTx}Y zN=K(7l3I^fVM4xGJsLDj-G0ff8?>L!IsCJVg7 zHolchl2nQj$6Hl#9mmd|d(2NZP$@4aRxBOBN7S)ao&s1S-GtB+$21O&Eb&<4*HfLl zuiG}5s=hX!}pIg|uS5 zT~H)wkjUot?tWd!cinH9%Xm^GHjd3nMUn)O4UhzJisIzE;;e0JaClduDgDH~4bYIF z`!hpErac2+OPQ2HEp}zoBsM=SMlw~1QEH21BiP+gcA>iMRhJ!$TvJNyPsPX`Y{pa~ z9ZNk#m?+Gin#hbrM6ZS=>7!KFv&Hv9Wi`cOaFK5D=-(AxKC67TD!3ngO-GX(No+)z ziK(yWxVRzQ zZgHwJ5=4C5*<M@ zHomXb%gtm6dOtUt78X{tRfU=mCW+KzP9p(E`7YM%KN#MtgvAL*xc|RD2WhY+U?XOI zW$kW#5fm^wwpW_(@MDvnzx`&a6be+)2t6y_!Su#jIrTs60!-8Fd&fvm}6aY~`+ zuxFqqCml6%olf^RN+gp)u|dpC2KZNM`qSAPuE$0po|dZ{UV*<<{^ariIp=ZqF`|$Q zB^N`x7q~3C?RYi*$N_gzQYw!{nth&-=X|5%2_R+qfVQ_*Qd$vjk?AQjK0dCeLzOxG z(?W6kZ(XVI{n;P4EQ{tLfWt~OI)TzE51cw?e``N|-mFy`A z^G=3h4_78|yV4q*-g4HNt1$d7=vsmW0^t4}CPxc32?>eDz_C}MLh47?rONh)lKRx- zWX-LInYjZXk`&g@HaT0sU2LodEi+b44KPp|_A&wIn33;81LS#D^>e*Ff#g&Z_rZJ~ z51-^@zppBeF}i&vN#V(rlm{Cpg8QzFv~p13$ua|eH5Q|;neN0y{8T>s%s;NJw+GE9 zzh5Z!Msl<<0Y}GIaCx1B)=PuJISBP+wTbNZRQdP6A2c+QFcxE|#44sj5cU89uiA<4 z@p#u-ndNg5?sWhcCZ|Zq$gCSmy#Q?8;f*5YeFv~eVbBt=1C-c)0&%U9j*(GUdOKXt$bHV}!s-Va7z9I>-R+l(WZ*T%!Xy&yUa7{=3&`{k1jK{l9fQ zXhKTBNH;0%vp2)#N~Zq{>P<`(B{C0nEd`!K8P{0dnpWFVfM!*Gn!ktZp;mo!t+&6b z0vj9olP?Gs{#BqD>qA|*(2_?B-jfl~<+_d&;qb3|WbZEG*-?r5XkF~jZJr!6fC>-; zKL1;}Zm2yfnIo+!+WRFlTXKc2T>1-bJoSj*hdZVgep3j#C`sgdepfF|LzWt#e*P|r zb=yVPLl{~BuVg`ZOv0s=DMCrZ7QGS|1j6@LTN0=w^QTRl4R3%6F_XYhoUf&Iy3$_P zuAyobRNQL##70jzHDO|Hx2xu!zMW4RrXT&M-^Afgzb5pT9H3%P0)CSrlBune$v1ZM zT3a{$-(D3$qaFK(hm(NQq3Afw5*4NPz?00zxZMU>++~T8o?aU@r#@T^6!t7t?3x#a ze)l&U(@Twxt4Q<-tPS#@JVu|9rbK$;DyJqru@DF|G0zfLYxQ!Sq=?zGiQ<`%pup2l z*RK8;vO_wp{nv6RGB&osYElYw`{Dk!^=~__*B+)f0I z=LY4ZHg0Q5=I9I8w-m&ueLVFhIJWc*3?<_tmfesI0i;Q>$GblzdN|w-W6?3c$p>o0 z4;q2?ybe;lCR-@_!`Ga9P%o6Gy1%All9rZ!Tdq_0L;I&WOY{B700##cgG;ndSFk{T z0H5iO$7^6q0w{?E7%=PSZUHm1W}VK$4Jn9lF=L=A_f?U;cJUzUDvX5iz1G2WEvBdy zr0IMTn3a^4R$|}}BGKR<5y9ZkREcDg2w$Qs=VfRjDRnKI-*>EnVzdHD6vd;S=g{aH zsMr&2%br`g4Gonctp&ko*!{rmsOA*cIvs>Svif!mHKxLkZpUV>zFfSgvGukL0>=6aCBE_{Om7YR2?F%Q&qxXbWhb-#`|<5=sLkr&)e^?jY=4{!;u${x& zuQ$GL%%v=~y!*yt)h}aMr7yOj3EKw@!L_M(Og&hFYs`TB`}zcnm>QtX*EYUjYkUJZ zU;F0rkmHqRMdH&J42^g5Ao0NM&ev{$#Uck`=j7x>&+7qXf(j(QG8jSqE=+H)F}i6e zdfC!{-d7(yWK30LI?{NtY9XDd=wj9R{(O@va4F_1W{Ii*Lr&$C>#SLFu|^xXucosr zQ0z7nG&VK+xr1b8XrCMw_6tpyvv2fs9+;|kz_-y zz&^aPT}!jBP`>rw&QI*9Cp zENXisff;GCe(b&d?D;+LA-)BL;bfbMq3v{~u|Dt^tIb&gkKACE*pih%j3$sSF!ng= z`TXv#Brjo$ZVfiGG=Oc22OP>f@VN%OeMw+RrX}xA2dplYYH}1CwzvIl__fFZI5=B) z92BDTRd)K2&honGRIT4vk$;FLwpi&y{G!j|`gt^kkOpcjIyY`D^_B;mj?3oz_#3@> zVJJzzte2+n{y6 zo@(_xHTT@BhQZ0K1YoZ21({G?DRfo2`lEDsOiaDk!sUW{)=rJAL+HI^wnb&gF)(x_ z!Xxw0Co%T}y}DswXeeIBAC>e5{urdnfs+)WIjXtokjGV7$p*WHHQ+Q?xfQaSVA=-$ zc)@z$UFU260vs}>)PxdDNK3l*t?GbuDn#-EgC>T>iA)?ckZ!{0Q~xl(%J@u|2v5nt z_gvi*7*5Q3zrNfNtu+VvOg_{5K!5}odj!5RD*zrcAGrVVh=>YxU>K)+KSxTacR_`s z+>nrvunuSqHMTh5ELCE&0!C=n0URX(9mb$lK(-W#bFr|zTet+csyIp10$3L->tj1G z@m{>h2Wflf(SyQ6WVk*$`wHl19A~a=0I0w*LLDg0;^d~Gw6td6iQT3{G{gs7jrC|K=ZaMsogJQ9 zUPDH=2Z1Y;1~fZV_bX(9=7GJX_NDThTuXHqX~11z%=cJd5|GD&ia0G;Uc8hW$zyKS zitMu{aImib0LdCd7&$Nn0}`Nz2Xwn#r)drlvZ}qio-6ua!a9ecmzDs<^9h7|^Gs?Dj6RCJH~y z95&E(p6VMnQ#D3U$4mzr{oOLgkBJEt)ktCucCv?O(au-^r<~31O#C#GU?EcpjIjxF z!5HOy!1#$;N-P$pA@;1nVfAL7IAODxJyo_guyL>()sBE6 zk+%8PnnRV8&3VhM9<+A6Ox}_FDAXu?y_R5w#RRlkHW~>T>}N;ygE$w@c!k+s#B**a zqb2j#qu-6jM;*rnBx$!ik*_$wHo)#_pi!C-gN*&Yc#Bp}++gRJNlc>Yq`98}Zs-@of_)q^>5 z1Xw;sha>1E9n{Fj2;{(d(yTD(cVflx>XaFF=mqd*E=ubDt11U>-r@j4x4oI1AU=^c z+)C81#H-ItduWMa4?aL~FA;iJ38kvvv64Lsb={UJ?So(oESgpYct7l9`kxQI+7xbarqU8e^NYP9Q;sG#F1W|L%Ne63~Db zlO(1Yv6Lb@t&}!I`WyU&#Aq0RDLL}$dd63jq^eKYtFOh;1Nq1MuaPG8iA3v{>H9HR zs$!Es7&QsnoKzn@eAIO~acf7?c;(Bn4^fdEP#UfzfONDb_cyCjm_ zIgV^n+U{+`L(W=EYqe)UgrAxk4rIaBeP8z%~a?)bz>dHZ5UxA zsG>?bddEaRS^5e)Z$x1bj?s7kb$Svhl3M;y2&0ph|0kAF^)uG43yN~0+kgXey64-| z(VvM2s;laE8+`SwtV;e;VTm}Pz_4zyXZ*>9=I(12sAMbUE3siiZWJ*F$Wy`hQrRB> z$rFlbj*F3ukN{T9FR}!mAU%jp83w+HFFRbq7y}Shz(roBiXa~RKN5O^5BT{1Qy~98 z70~~0|Nj`?8X<=P>kGyYiNJ{-*=HJG@cw)M(Ms1kbhpCUTOVZN6R__3gHSO|fnZ_f z3s2KPxa~6=W$@vCQrxiUB~J-h z>gEA>hlJ2t!Nk64Ln<((_jc+Yirk3 zA1cLW0$hb3&T24!(Oa71F0L&QxJM0G#%2Yt^YFdkz{{dv4{}4;V zPq%#?c$S-ZbQE#fQVNfXO5FZaIPj{ng(gonHVJeLy@kb4=<6&2!KE~40I=KrAgr^) z5m`9~fKqC^vIsY835xx|58t}zPDA4q#r6w-%Oi$RIUqGHjU~(XB6Y_=W?3@@V2*yc zPp4ttDH60ajpx~dW`OTI&)VP;n|M*{*|J4#aRW3P(DCtU?rRzY!jcYjqPQU&_!@$AslO--AZo6|x_ZEDwgmE|>pbWvYuc}$ zGhOZo1f5&gl3DIxGr=R9&q*I6OD{Y?QN;MxS^&sVvMa*Q{z0eKQrXC}>&r8BhsdPC z{kaDDJeg<={Y{V|i=Y={6fp;PAY>M^JGq?G8oA;R{-^f^0xNT@#bC-yV7GGJ^(0_X zQ%YTIyaS1tnry#b|} z^W4AVxH4f`$V(g}c75o?E`_+mj$J?+P(dp)TuWgvSn2NBnM*@I{-@d5S&ace#LWMd z%@yyimWo;5TzhAr13jnbeSb2}U3M^_ewqLM5N4~m@9i2y%s}(+V+YB#Ap$9dnX4`< zlDaqIUdZ~`R0Hj!E~W3v1JZ8Sz1y@s3G43qqX?g1|IP8EtyFH_@5Jw$$1i&adhu2L z+8kkty)k#JBmMCqFS>sAx>9QZ7k?$?ssiX6S)d-p)1gFM*sAaV4u%O2rNozM02kC+ zPk%7Wuhgk~_f1HWVz4U|djRlc82ZV1+>jWZH{ref1@n4+@*`hc)F`kDn(<%%!#gu1 zErNwFv+^kh)jy;s*Ut<7T51-Dw4rmlN=E%WNvjYpsA6{$@7V0kkqRxN+{Mxn7~_p$ zKZbW+Mzc+h=KI2n1Zz5;`=!VGES6>nq|Ky%Ai2e5AYfDwXRyvWt1TldwY%Xj~ZnXmbt4LUc?ve)me8wB1vZ-Cn&0^0~M8F)Gvc zgsN^8(vik#-H)!xr~a4u7A5R+;-Skp61i=VhFZFr zO?C(8oztT8`1UlC5&Yy{8-syavQZ@c$;X z14&#y-vUF5)udUpf;~zylTwUeyrJ*;HyB!-adPO z?xl+J2`QDG@r~zMpdwIPLo8rdsIXmZ(O#G?QGeyCn~Wye=}dF-IFmEDKlP33-D~^b z3XWik%h*rg4WYZhP*b!E`Y)yF?~GVv_&tuYp%s8u9sABe!v_z?8!;jqbKNECaa-d( zlYYQkab)Td;0bRuvsPFL6KDnYeAh{~3e2iqfqu|wYI9Z)-uoUzjRAnc2+RxeDdrA_FH5Mh904 z9MnXU@s+4R0;b@wA}yJbfuX-hIS(r}GCiy%A+x4y$IZvdmr@0Mzon_Sk84=^L!kM< zJ}v#`RmkV&-hb8OrU#*TxCHBm3m_9yTX+yP(JNqj0wPxc^v!bcNkJ%k1w8+f>q+Vb zdttGMhcP)ID$wFIDD8*n5EAQF?dexh-b@`|tx!iE_Fr$E4yl+~VS)%>up^FKj+EpC++pnPT<8$(T)ILD7?GJ!z+SgY)~f)%T+zR^Qn znA<48;z&U($c)spP?-BwChD*;NRr3*9>U)L*H+c2^`$Ncc!~hz!J@}^>1!JsCDtOr zv2JU5S#&bQ9|InF_=1@CyPDy z0u|n31o&8Hkb4p~hQ5~yrb5)toIYeErJE4#j`r{t4Iu4;r>Ud7@~YGH?-IX)HsjLBpD>)btQS67Mj2o^igZk@hq_ER z7RJ8JrD3WvsEdarGY?4tGCWY^NkSNs=G1}j$NqYzs?5P_7#L0zq(NXsa~-0i%0-Yn zGQQ-;#YMFu@&rw)@|R|2W_>^hDBGZ8%*{=x?)JK)w>Wx9lN<1i6EvExjr%4YG;LDS zzIhL^lWVZ-j0%hX*7#dT*zdn-GzX(U_$d2KVebDc%)%o%{0R1b0aqEP7)rjS82hii z{|#Cr;szL)p~q?GSlvZgNFe`1{8K=#;?tQ$xE;1>A%%n?p}~iX$r24cQhvw zL5h-jbd;L#UBWF*D--oqD)hN$9kPe$loT9U41J5DAVi%it?Cd24`vx4VAhYXZnp)z z-^p;FSNSF=H?np-|L5xoP1ac{!xrD|e+q-j4Aq86z)D~Ad_wiV`UYekdf-Wdy6v-o zR{U>kYo^SgPd<(C1nd(yazDQU+EG7{qgX&NQ+_G|Dv%^i1JCY8O$Y?l)W$5(6{sMB zxD3b)C;-oAv;czc1Y8?XusS)h901+J^Bgxo zscpcgS?gN_o@#xq;U$oLC8>}Va&eNHyxUy^&rRzMz`=*`gN;d;>laqZ{eeo_217_- z`fT>YrU5#H0=$D5{RliDd*#D2ggwW%-oJ=TK%v6diY?WRgn?WW1cEQ9-t^LrL%V0f zeO_H+VU$tvEX63kGldT~#xTMuP)3qK8-4G}$=;p~sO@6QfUyF8q7auPMz~q?DW{O@ z4js1eQ2= zPKzy`;0+@Po}j{sUVM!!Ab(^W0G_0i)JUU9NrI47iyC+c;U3Kaz?@Iq17{t)7L|tg z7^eaZ4w2yw%;?or@brwYwr#j}Q& zAc=&}E**~D>l_GQX$%s5r_X&fJy>KcINcnA(FT?$gH5xC$e`yu#LEqObV#^9erDB& zs-b3x0rxEEZ@#ihfSgeciveyzg3O4L-=Sms1di^&LQ?xR0$yA}M@uWE|2k@tj%M|L z;^mtT(8ydc5DVS~;v0CYz>vWAosotJ&xamP z_|7zHM5YS|Jje26!!QuXTRa7Euf?J7?*~4w-3SkZoc=@WRT$IS|HdxK*B!q-1}KEw z@R`~^KzjOT`bprs{WsDeh@HKoIZ|*7*PA8xhdWRBBAuu-G$egv!vx;+@@;-z4P-Li z|3VZV;>N)O(S}%sW@RRH7^sIJzsmuQZo*c9>M42TTho@P#0-sw*&whmUg8?jWY^XL ze%;j&uzeHzfWa@_1#}chthrU1RUKaYRbmWa3Xwo6d~gI6@($@Ah(?LpTdV1|BWZVF z+jG?idJFu4r<_JI6ZWKJhZQ{WV;|u!&{_Whg#l6Ash=Q7!$8640?2d@ z(J>?^Uvr1b-q*X>my}gXaQ=!Zv_5TDPin8~70EkKK#zllGry3WBz>pSW~6XoJ~T8` z0KxKMgZ=w>Qm<&yz>6d#`?Al@YOJYZ9t>rNVyCJflD)i1Wo+vG>He;Q&uGxqM1G21 z#Vo4LEjBU|jLBmj$IMGnHH&_tbcK%?c;4w@!t}w{z+`M7;wYtT-wA_o9gi0fo zko2<<4M{?NIUIy7f6o8Nz?)S71HQz9z}QfhE01!q2bX z$aAik>|`TDILzZ^Nrvr|iK>EV@1>Car-te@q2c_ExA6dz>rI@yOYVq2`UUwN^p{QJ zp`rq-h5ls|+J$pc3%Upr>+9>I0pfwpEmQU?WX&`H{@Nffzb^neyMU_LK1=65N?6H# zKe~Q>=DZ|ZIju%B*Gr^qYx{|M?LBt+1E|^6CL4xIOpPlPM3>RGAlW%(>Q=Z6(nDqt zQq(r?j2aXcd2M}bRR@+nG@R>7?D&;(yK8~EXMT!hpby+4w>2me8!NJcV~O53=?IuzRUN_un{W&5&hT{ujj5LGKLbHt^+Y5er$I!zWi)7 z0P_tOpd!@F`((~0p+Hx>pm2OIUsj_x%na%BD-S)x~^sB-Ve1gQ?&22k$$%6 za^KVSukvcTaIPUz3j72_0*Yv|YND0XB#l7g%Q_(bGu(OotCpDa(zk6f0?+|NN;nwlnVV z)~@*&yt*-h*8K=Ab&_HP@fr=)lHwb|bUvzgRuGvol0W25h1~-Mync;D=535;B>pIL zJbd_e87>kQ0MAQkntnM39HL+q;i@X-OZUD7UL4+msn?dmP`^?nGM3Ghs8|r`7;#am zGD;w3)hLMtIpz>lO59$ck#Y>{9ix+y*g;#(>+DS^eY`h&ap%~;zP_FHP|5Pf$x=xt z>N)%3BAsM$DG-qxdK2QNC~1L|77PF#1g$=r`Mu91oSgUI*4cp zbTmT!RRDw8GynYjwcaUjxJYLJv}6>|1(5>@KKYaeLEwW{XS`6csMJ*Br2|C8Qi}N) zxaJhATWQPv2T}T-5$$*Ytihn4&h0sG|DABVhT?O@I&7Wq0Exw+XFJ^oU02a69FM^- zMScG^j##9GMnnh^lpY^WAI||jr9(BJsN)UJJOQh|D-@QQhu4fYAUQ^PEO@c*Q>FY| zX|Xo}layeu*R5uHgAQcSz(f0m+CSnYu+BVP?4Bl6e=LA4_0`sB4+Fx zptg!C=RhChc-j~rvncRJ_a%3vgJxPVz~s0xsx150m+tm5Vc+nV4jMlWtw;c+`4AK} z>9OE=h44kZ;r{BNXBkKfeoO}kSk%~U%LivfPsXe3cc@Gm9DN=Ibjc{8(r; ztDd}yFKyTm3!WTb3Y!LtX)kvssyW_%Fo24LhWp8m&ChhG4Q_aeY)U@RNu^aENWgkn zx!OJ~wLeGJ-P@}%{#wHGjl!@$0j*ohim01m9ETRneQ#vKuVmnkRakTux_3b|*YKY% zrnj(jwTjpRdO;LH^;bV+f~VYP3n?;xP@!CwhJ0XV_(+flQgclW5kQ7drZI&x4*`Hn zfOp>u40+?%GghOm!j}lmh0}mjDLhA%!`MnN25nxI2XVApB&|eY59JqW%9{a-mMM^_ zq+8>~P;@OPbfG_mb8v9b1}mpyfC}|uwkX;uzx3adb%ioiThdu$RsBwUl)r?-lsHAT z`?(qpNJcB`u*}RC!Ii{RFDB>IR$|_`8Tcdt+=1#4XEfqsk7tzIrAEcbptI(2pe28}J0lu<)O}06SlXRd z9Lmh3z&%7kKL*IW_GHhSUrNF-atk^)^y4aCD8ZCIebPYsjC4Pz1$&QDG^|Ll4$5iN zCc|`-XYHrV8F*LfV~L8%Armn*o9WPh&N0zS!-xJkrQyn&^+5D=wF_k`tg8BT)1J@_6)c8zf}YP^Wpj0jGi*1}_nm7@8%CP_Z=Uy`6vGBudIU8bC5UKFAN3Gck#{gGpX~fXsUADCxWkhh2?}XU zsFSTy|VeUE)8OwAR#5bOmYb}leeXndKc7PjT{a6*~2 zR_>4{eghko8HFj7Hmy{T74|R9-VF-B5CVZoP*4ML5ox97!?)N5bVweUHl(LW!(IIc zCR{cfa(U*p1e6OhK=nenjFy!sw6$`eW_d;?A$14@)lm~Cx#gRw-MZvyU%5Y03}One zy!J7u%bW`cT*AQd>avn2;?9BYDh3ebqWC_50_9ayc*Ef3$@uRa2PlRV=>)eTe~}4| zk^x(}B&1@E1ca74<*Q~$tJ%I{TSD8{u$T9rOetfh$aeVnV+LUKngdB_0`#LW??i#t zvgjeh;jiV{BKX)?=NJ2(^v!0`)4zjY=;y^@Sux4Xfq1uN$EA-!_Dvvn=%sO;phu2= zCKi^fP4byr3R@4~BgA1FR}1Qo;@pLMva-Dp+Ly1-3CToesguDWhSoi@{yQ&><_70` z70#dZH+A@li1rZtU5bbYL8~$(MBJ4Ye^gN5LoHc!%&;mM3_j&Ve3$6k?!l-R9;0@>rFh^SKyDe60riksY3|3qo55#%B<6c9 z^eDHKoAvb2Zha;p3kd^}etp5RHvZ^~v6G3HDDV^_Yv!4hMQ!v>*i1^KwrnF9ATiDY z5Qc`>mmw3KCr?oOIqQCQPENs#+L!c1t|yDy6z!Ij3AOLCSo2U^=cS-3T>R1AI836# z{2wgh`SW?y%EYiqU(YHf@vcyuM9~j+S~=VHDM2A2Eh9=^5;=iW}vp!s%vmsr|{yXF^a97caib^>|;z~%*jI#hx{TyUcU>Y068+swg zX=ED=YE&>{q`UrYFpYtXES3sVTV%(1PzlfSp7g-2 zU0bJ|IwQ}`c1B09Q9d8VA9dZR?vG{848R{vLiRH94+I6a+c;EG#UHxXH%r0L4s{T} z$@yeBykzR;mz11bT$;dbU*F17OkWwBkZNXR+GNjt@{iI1el6(KmHgQ!==f_% ztq5KB2NCTYLulN|XZiZO#B_5Y-5A>yDr175kjD@U+Fo5Kqkxv&z(5|J0t9}{wfLVF z@ceBH6$ETS0wA73CZ*w%l9SQ4b*c&#N@^py4L|D!Qi>h~HO(X2a87qa+njCL9gWO( z)iA!Sv-Ti5b0fdA4=N1dP=8~qhxpkbAK+7??cUaRj=&>00h(cr$uaFQ-GI6c$CvSQf<$o7 zX@rVv!+VW*nH@Ay9{ZdF$)z zKMxUtYGcVSl9$Z=cent)46R|RbmO`R=N3NqyVHG8--}kE=qoxqSHG+88Y?ldhG|e5 zpvwjXKdhsoHQktMGBC>O!b+SRvex~{{P3Iagr1y{Hs>$$Wo2d#yFk1O_6>K9BZAj{ zkX}Vm@*!_hI^sJHORoPjFhPC2zz3@%w-h{+b421b$@yl>+1idA-j4bm9!|50|xfc=Zq*G#~|*H4xicoPga5BDyW}173Ht z_aP+EyDNh8_Vo>iDZ`O&CL&jWo=(b*Dy-4}+nJl~_G1QF8%zJbI7VfGp-e67IfAT= z7V9zgCh+Q1{YJ1iv8QWY_1&S@YD{Hh6ui7o*7mo>m{S^$Cb2aDt=(~_5{41y-#LFK zt?!;PDya7}winyt0f4_Syw$VDHKs7B4;DpKX_5p9p0scI+Z-*nz)tB~Z!-x4ah}>?+H>8!deU0|{V3>xRwbRZo?f9)mHz6?jpw3tqMJ z9#e}rB&@w6sst5*dc7kW5fr+A0WC-l0RM%1Vrq4jHm(1lH~xPMGX~jMa6&UnN;bD_ z!x@Xn`1>F>7AX!vjp0h&Up*_d-nyCJ51kt>fXXEe>*YU7l;fQjP!ss^yUU+sjTpO_ zh-9h8pXyWSs+gXDd&qD2Z$B8nv=kt}0=grOF{CGi<=2|@?bE+!qz|9yx;a%+`WiBO zUEM*kG`s`Cgz%{wYO4Q+b}?Ay`9SoiynbC&3c!DHJe{Z`$_S$I5Tp=a-hU?rUKsDu z5_Aqk6!2>=T)m+>W>9NiLv4zrBp5l-Iugl#A$TcgYy)~)FGq=!cOGTo=Jy5cSS}L( zUo9ejqx0*7gnug!3Li{^awB1)5R!#pTPvi`K%P^cjz1Zb zHDUH1)~~fM;XyG)*V`q?HN$~^f>C^&i>dEgzoWAc0>a@dN|}#i?Y~Hr3UsX=+L<7r zEJVw>-XwKP^*MO>(;IjnowHJ)97+In9Fpz(!9^gn?Ow;_)$@-1$jOn>14qkHw!8oc z`jNSFS_IP1^r|0VuKH5Z(eeJ5U3_l43D|lxA`3<5iU8(g-@I?-RC3>;3B>B^dCD{{ z`=Gl$ug4nn1Ta8K1K0?0vF2KwhbJ2$faM=4=>iI+54#^A6%AaGX0fF)MM7yg&lDaASN-cd!l2k%b~10z^!glu*OxKI91pwOcx1@gM?f?(Scu%T@r zGm3Uh9J`qju(y7-v_Ug!A7DSvYWw1gB_&{4@rUNvttJf>)f{wEH6$6l)lQRrL|=G& zwEaT!1XwIBv8NMvQ#H~l(*97i&ZS#^+X%qoclhxV01mAjoS-WCw|;LOfjYSWu$FSVbhDs8R!DJ{f)?2_Crw)Gq5mE(9UH=^kI5JmiW$g z0^Yosi07OS0sB$`&qlss5;M9hACK-SfWK;PTR=TStb-1FR`G4 zO2E?f{+Zyn z9YC;%W`?tl)fajqnfQ`~mXmd*e4bb1B%yH?cK+v$ZLokP<)N8~UgM?7$oq^_LMxem zkq;cw!$LM1Up2C1F1UQ)aYTL$wZNg}ydxZxbWjW`Dk_v8Ng%Dv$`e!uixj6~<0EqHuIAs-)iQW2-T1hgyx-qorcfL+3ZDEVRM$R!D&vtg13f`&3Q{6?d7 zDuM^+8C=%o)sr?a(%Cs_^@I8h=Li|pR!J!{A7i}d66aDRtc8wm{+q;qJk&7v@o7dP z|B0{N{MP&75`K6%_vp`;olXdqN^}?ceHfY8fwhhUP+0-Al{tV9FQ5Sm)CrYhRZiY& z&ka4|U_|GQ*jYuv3c4vE-XiY0k?ahK2q4AbKtj0VJKw+Wf`B|t8Bw+T?sBejS`zvn z7gi6PToFMIW6L$*3l+8y1`#B&6`!j|frNL7xE{kcOjK;W@zuK7sPXBA(8`~TFW`L) zLka9plQJ`@sZ(e&*$cGY>U_;9VjjM&~AS9z;(E+=l7$L z&Pqsuk|c-}X;?}2*tBc_0F0I%S@b~N%L6r{m7F9Pm1_Q#DBJ`wBw1&s0-Ex8tAY}xST=g zkqWY4pnDzyHLQtkbXS>xIv1|eNi%poCvShM)*;Oj+JkN67a_nAfxhsxDJg7kOmCVqX3s41Z%q407>7BM5|tY98st^?icK z_?jwI#=kas39MWreiq25nBYliAWNJXnN|sh+ljk5X~VHdxA{VK`N_2`n%UFewJ@=K~)tI(M$=tY{p-oW^hdI8_iS;2oKzmK|^lIKM8ut!kS8$s&S}{ z$lMv7f&oy+L-Z63lu1e~P6GWv1f%~R?Q8e)N2;?fY8I%sd%g8h{*0fP*t7^FO!TeI zMvX%45}RKup}1!dDxoXIZyV;s1^Z1T#h}!IM(Mw#E3G2cTF$i#7Xhv$;~Esgh@2of z{m4eDZ*oz6sCS1z@S1s#4AbUyd+AzNwssXbc;X2N2#9x*9{cTl#>rIOXMTlo!>Z)DOJP<4yeZ2)e@L4)_e@SP)#s`*lS>OQx7{6y-`cYrO%&&kQ^x$TYuX z?=XBI^RnbQ&kv9|5pCD3}o9H7XwXezb@QA?vXa;DiA7Y~RVE-R&2NqPKw(D0#UXT^vO z)dY950DPipDZpjGhX5yRbNcK9sS=M;XB(j98Uk&<6OBT64$8H*f3G9<^W5wyCm`-t zu3{Pyut#7ZPM$n`d%#+8fd>Fcp`r@lF4UrWySvr6qp_>EgoH+H+|r2oNR*~vubTq^n7R^#c=wgfx>9XF)#5gmMU1jJ_H^F3DwgE#8>=8ut08{t-K-CE%*K=2q06EB%|&`@+wK6|au=B4+erPP`--4t1Pj$?#!5esQ_43-9Fb z&;&dv0e)gInGh9aYM)cRw$jljM&9z@(cQjjM_aG%amM~%t)m+dF$^-6OSNcd2mdJN%%nF#S`Wh}`Bjq6Pp2D$7SA zotZ4AE{h%eAI=B!?PwN-`WtM&0jv0RFJT7tgut0W94qj?a=Z*Fb>PsK{7@l_?`aq! zjgRcXqgCj0Z6Y0d!yT&S%zxjWR4Cw=@ATDuAKC?--E&G7V$cvzSF2|RsSm`SMqHCs zi|;!-U!PW})~}2cc!t{tbMs|3(d(OX?!QYP;Gh5%sc%Ao$AjY_#wxKi=E!0`t%}E?;;64wvWaWi0UekD`3o;4>xX& zm_K9(<4BDv!|FJx{tfnLr~80z#5&Wdu_2QPhQB2CL5BF#wM4ki%nW#WS18O+KD5s z`7x*gnqa5eU4Y=2Zrw7Ne7WgGV`_dT1x`*yxR6COUZ4+Zy8Y+!J4l#UQQ-k9BmP@V zRPr2BN4E3&`jqxTS~3U92<_dh25N5PvKo z&(5*Wk$TNs6=-3~+?Bm zJyO#ZCH`td(jWzY7md<;{w~nEs1bsHBhD7I-OzbDxJX$I-Z{r?H8SYrX))!H$#8)W zJR~vXLS^8=jWCwL952!>o|i^>+nT|0!LJ3sR8hFY#F_Bd#YPQp4tyOQD~UKTvF|Ek z=YY>+Ehy>bX;7kt_TlYbF_g?K;r)Pg%yZK z$_56wIXnRJ?~1^qi)UkC@U2!)6?0L5tY=f=tI%DP1Sp2*yzdjmeik)gXIJHGVRBSA zm2hH(nMQ4}EqMyQ$YljX3t<|(xUS!$@}huZsO@IZGUr+_5$&HMds}Jj_2b?M>UhFw zfI?Rig8FtItpQc5b2>6K02xj_d?x9@3J+Jg#3M+X*igKDIfxbzPo7qmIzeJ_o3t6F zsd6IUohkQ48Y&5o>dARFqi67+UwwiDSDM@z-XVhC{mQs)nEMJBm@1objHVz4Oqg}? z`Z)FEl-x%%PG~&x$qCRiGt<cKV0OR@%jn01;7DxS)Y}}JZaBChCEC|8Y2}| zrN@p9g>G(Mo&on=&Kz_vlfHmUxjgwAwB0V>wZp5bMSR&e#+))p@e9{xS(!U4kr6#K z1pC;YjO0t;XeUgY`4!P}&4aV7aO4-p(yRWMphE{P@YLB^8=cAt_AnK7_+tDA~$ z5RA@ee&bL10{+jBrtNYcUH<0m&Af4UuSojwh8F;ZBzBn{NgQM zTdf51kf0O~!$Cm>6QtH?&XcS5KKi}(Wx>1x)Yr&FCaJXbog9PuPBfwt6&o0^kS}re zLCt%7KtmpG{Fb4-X}})pb;;EHvOkjRfaDedxD^G0t!W!~h=#O;lWgT;G}|R`Xyji_ z)VGZ&BsGe5+9M5$0V4U;;+_mA+SoXC!uivE8x9$XYv-AtU((DxG$2StIs2=e2~2Ny zM4;``ZHP6|h_kw_wu9{AuQiVrU)(39GBlMK^Lh;e+3z?Kwo3;xs_q*T^|%qNUASX7 zEX|9$Ysp>wX_3PAkoJJx8v@&;7U%N8&Z<^CZ3lJ&(38S9iMP}>H0BYGAIL+sh$2y+ zqdMs6?_B6_2Io>XI7THsFOcg3w@WTr5_?6kU~yCCRQqXA(3?3DEXUv7x5tPb(@6g!R#MgXd=B$ktDjvs=e^puB7o)n!IANQ zxXHQM^JSj%L+$F@-LUt}quT&NK;n{}!RP+!icSm+TT%=+h38j~0`&hUK zr`M}ZxHKzbW#M+ZD1rMTzNB&?}h7zqt<;O zD!!1%F?=lJV|+z>uDIG=n16_zuiPzbXc|2Cxn~Q8R-s4B$`K6OKY4q)}5Ve5` z?>Qk4$y^f>;vzK%!L8EhrHs7;B!6{fxBwt#AtC;GCTz*Ni$CKa!RbnVH06SfEML8Oy(AHJL z^;LIMegC>Af#_LwyoFb@<+iPaMgVts-1}hSGtQo5raXZ4*EG(YCYqez2(PK8%pWc4 zWG6UH1(Y8yS*~gLix+~sa_hkGYF)&+VQD&g^Q0=F=&zV&qDhdz*;0~I487DPua@3C zfV>?n-Q-wm8?Q!7a0KHO>ny2Q@t}*m&Q-A-S_%}wuH=2{gwPc>kApE_hlA88j12eE z{^{o;W*%lnYu9#SF;@!7xL1b@vO2V8ggey5t}XzzLTYMmby_Jc==vp`>)1l*1hCO; zhv|XR6Tf5sUcd8kMGm=RO-2OEiX>)sM9ti_|K1j4GkQixl;iGPJ~{^X?m~Z)HvTBt z)sbCLl?GAAhTW|d!By~jsq52`6-U~nF4ls2!OR2Zqw@~YpC$)ACNhZ&5u>$}nR5%T zvP#A&vFm1CB+e11doI?m!op8L;yW&iFtM9?C=z*78N4x=vS~kuGUZ^@qUn2=3003d1)}!{Lbm%!! zVSank?leDI(4c0K;Ox%ovxd=1B{0+deWRznc9Z3yUcGvzTeY)XAyc5(a9mi9_-X-z z0BQfAp6~jM=t|KkG8c9c`w2FH5?6jBCUpoARrjY3i@ys3Vl*rvSe7oFWpy(|6sP)= zMKzeu>mjE1VWHd?!-wCA#5aL?(?0zkBsaD^c6Yvm@vz0?$5$alzMX+9b+DtWAfasX z&Tber${fUZqg&nTX=!+Bt@gs-CIiRq;zz8!ilQ72yc2ByE4QyrafpN}Vlh80FAT*Z zOjKrJKL$~f9jh~M*gU2nQD-E2AMwrrojCG`&~nfwtn~7P@6Qh+uhRlYKP{$uD%TYI zz-0y6tJG3HjfDlFrJi^RQ}M!;!GimEJmfl^=Oy(+FGu8T6saKmV1PSU4tnT=j)^x~ zXoJ>b$ED&YxD~{T6+?M9##y*yC~gtaRZUA4Wc~{Q+_(IDiJ`$QQcT;@H#9M%+-ciQbSx? z5v#xYR3c02%}+Ef;fq`zmfErEOU%HIvnnN~_voGSxqx8($Qh3nHwc`cs9q|qSZ^EL z!cj1b{ua{*GfMxr&N^Z?%#R3d#qj~fKNeFlbIBHq0$R0ox2%q94D^RNSaW6?SLqoY z+rNa{?wbdBE-ocgG|UguerPkFc#Ltu9FiL2O7=aDD5aF$0gEX&ZLrBeyxtnXy&R7! z`1bH4PmOAVQ$@F%y65sUNZkpQu~0_S32EU$fyF7F z?=96niu(V2sgWpsjyd{?b4Ts^vny=Jxu=zmbjP~mWJEoLHJafnQLiSew4;XI*o;*c zBJziNl&rr9(e#4>@^cqh_EhU!y;tL9`}}@`UOfM{#$DW;)LH71SeaWf7;dSf29svJ zDwV$zkr~QUC1u7r(T=j#EX$b9`zCLUp3*A19J(`y1;^t^gwtR%Z_WQMw?ftX{!zo&;UuT22dC9ME#D&GY{>r?4 z%>}xaAffS%C-88}r8%?Fobwk%CejyvpK*49Ors9*cv3A9=X!b*yxP~q$(XXywK#{d zwU~F-0x6Qsw$l9d&FK5flh`+g`n3Tl8RKG;Bk8yCm8-giy@LCuSh>v1_fEt7$R7o> zKr4~|HI|@{ILs`s%a!_Dy~SGI57$CUx8x)pUciHbBafXaX)?N*7x#K=xC<4_;LAHX zQiO(9noJRXk~BY%?fV#rJMtULFgV`@R$UXf$mebuJEpDDtNzxVS-c7z3w8_R5!Ow- zgE{9$6_|Fpt&b)1zlZgBaE7OesZ!N{Wp!D|Z6vv*!}*m}&55(9SbeSx~Z{tq!|c3G_wg;^$hDnN|F8`qs;$^zFkMf?=*({hGYgWuz{bbcG9zJH#bpk=qx7NZkvQky6)`Gm@f`)jaQj!ccT#GYUdby7L}OHvG5I8sPO`C$~}|A9nOjawPBNiQk-W_ zFwE>w4&b4BwqZ{$CabE4xQ%_!5vR!LWq5Oxql(aG=%7;D@Ji{L9D|rb`veN`xOb?W zCsJHyegiH$nk3gSNxj!ENeqv?2bP_s#c34YualjY)KRZG+} zXB%Dl8cpNz#^Z!CvT<)07Iw!&vnr3$U7%x zZkA?$i+cU~k$nzI#1)v?kw1vv8F8xgJ(TDa5ol3M8=_OyUaIAmom(9_oXd*{N}#aO zr9cPS6v5xi8Ov7r2f(ZrSQ^Ts8%^dtR{k!I=K(vd6$wzBRDkqXPK4nq!4Pu6r#~VV z_(RVtWdhw&sXoWs<6XaG?(AaK2uC(fO@sSmpYy-AzXx(73wfU!>7I=wt0qJDkn2+b zi|<^CLhg)>apG!kZ{KH5dhqA(r|q4UrtNj9!}%pW*hf8C#i~WGfG}|}cONZe%4BGT zZ$(VzMF^|N+CHZd6ii4`-NIZl_{Ls!s?P)*9vPv$Ix*+aL8^G(13zx$z2t@v%F;*1 zchP>XwyN@(dfFi&VOQVsDk#cE4(rd50KKv54RmeGT3~o{V^A_~^Lz(x^ zIkhaSuyOw4?jb6f;-tQq{N&&>gs(1-R=-jhz2i&$Bp9$--33e~E}1=3zp9hz6a@6~f9?Tqn2JaS?-dcIJ*{eLDi$IjX%E8OCpK0Y+j z%b@8@u|JIt;&IT`izsN!b$4N>SUgTo3?W3eyYfYl92Wah*_6a4J|`)OC8Q-IBOCdX z$WG9T_VQ}1AEy@Lz*JjQ_pKRHNjQ}B{qd}wo6}1m(2sZ2oHDLcJ5#L^OL>KUGZ)6@2#nU%A1=ijhK#?lv z5cBarQU-CMw`4qN>%3h|d_kYR>?LQP-<>={x!aNbJLvD&(}h-O?Z}^%(%mBpMJO6) z|1qw#Z-Gs$3<%6{BU_Q zg>|?+>D4@`=B`LUk0m_p%UlZ2rB_0?I)K-Q{9Idd`-MLfEqvFxzuv~4--|7klXxFjK>C;{Uj>mtH>1l$)cJX>MJ zw)NKXT<);}JIDnJ_1O?p9MF9$KGn($b*Lgs5M*oCijVGxKivF-cwi2M{ zQt9+l!k7+M5pFypw?S5gTrHa#If;SN%!;ncp=@&k24lAITN1J?EUirvZ&A+&Lar?2 z;?7-TnnM7Wh^k+C6OCT-qR!gdMW)-*s`ZNS+sMghJs|JISO0W?y}PSfH+hm$ z)eHEr;VtU;Hm4(IGA}RJVnG4TRZC6HSg_%#obWoz!^k~9^`AxmZkB2H9C^j8iN_z9 ztj+~p?pc@ZuWw4l#wirGE$$S>#y^=Z;jNa#<|O753RPVsQOgIHxS9W1$z~cwceu5E zWIZBa5gSKpG$Ga!9vX^4lJny(NxRlbnZ8N(-A%@Ro@`poPX18$@!lp!cwA%lQ_+Vh zx_9v7*cdgN=h{a{Cp$Ife@eO5H%j&`y-RWnVIzlPwWYLW=I?SBiIaG`F~JT>f?6zH zYfk^(sLQn{4f&=Ghbp-S5h)X1Tg7t*mgxTRWkBVKoDO{i?^Yw^QDbuA=UvJk~*0AD1l4x;Nc+y+# zh2nA(O0vnBH%=J1P2(!^+XP%UDL?JMFv|GRjTnPl88|r=BXUr%ww_R{3d^Z9dE(6R{YKahH|89^op&MLC$!%$_mgc=`JQ zOV#mLX5aObF`2{Q+F6M!PTlrT7>Hb8XUw5L_ep`pXYrC!sd1A&aDrAYQXOiUaYL0K zMX_x!GBrQO-?g^f;R$~&PXFj>O6;G!2~LVtj-|Iu$s-NQU0<76 znFVhRl(@ul(K42h@>VM`P%TD(ozJl}3?KNN6cSR5oJckZ5kdB@wna>z4#Y$S*e1+Q z#BI{ujG_jBzE}9Q3N_W7ys~O4!TBJ`OHn<$B>G2uTE<|`c}BA;xi!jiL;b1u`+l;qm~3aT!?w(q*h zP9ax1<*okri??*YmHvR}hE2isUdv=hW^?b#w;)eQ+B(e50Xvzol&wDdr=g+Y>K7hI zOxQLOLk7QMP?yxN!sLX+{x?Y3$5WhP$u^bZE6L>`l^!mr7+_8k$cf_X??-P9D2I$9tx$@f(+-l@Phi&2SCAPF$?3u07?w)viE`C_= z-^npJA1%g4QEeur&+6j3D=(V5;!}{GE5sX1B<(6~1;@QDvz=qG@Fkv00WfxCLSeaU$^DVht0!l+>VhTp=o3cDY*QN!TMEwC{+1G8 zkd7{jYmN#JbiCu@Wx@fXucJUX9#s5yp0|!jL|*QOfZx}Pdq7Jjege89DmFHo)cKNT zJcm}#ybImG+XX5fPu_8PqE%3LCC)F)MrG6t9(U$rIH{zy%USl4B|NpEBhJI^{pGwG zFas*`_{PQ;LiBd4I#`?ECcpFef5}M15*_dsz0L^$4Hh*Vk$SCB-vTeoJ3Gzl6NK#{ z@dF1^*!m)^BCCx98NY{kmM=URORp&4$BB%kS|TsE60B}*46Ur0y5a9fN|^nx1GNE{ z{#f!njdYZTgaOA9EuDgSLIzMl7FhM)c^vwTCM(CPUdL0tNR7u>vax|t5g*+98nAJh z=2P#0E3U-($<7}GQyNHH6}F3p?y4|5`RM4ym(d3~)k@B-$GlCy)YH}o1UeT>CbCx4 zJa7dUZj{aYBtZU45YjtD583sQF_Zr|ngv+H5fTlxoYV;-ZarP&8~9-a5n?0r5$21{ zHaXQP)1C|4DFW-Yc8&}?jh(A^en9?6W*cmEKKU*AZFhhfoSZY9CKW_2yDo)BH1S24 zfuULp#~Ckf=?rh!wm5-GPOt1+`9Q_JAc^UahxxAmQHX#ln{D z>88AE{h3a8dtV6@dsf~k(Hr%W%Jsb_=v&gKXZDr{QXGv^AQ&q|&caWgUh41ABp#ho zcOA$%DOAmEWO~~c9$>1ud?M|)prw8xWjR0Boxr)|Y$D+za{ z+|XC9*+Xg|kM&0l{uxAA{I*(MUktDwL0`Cn(Px~eQ={ryWLaaDlXrcUlQ+B0OxqPz z>r1G}y4(sZ$RXV!6(`6!;3B{?Dvf*&waiS@OQOwczZyXSkAx!CoX24jXr7&!(Hs+r zrHc+x$PC{t>JaI&mIK7l zhvBDGb_&m#sIq=Z#KkX>#T4mD;#ozZ*VYrtqj?1FVo$k zv3fSfMp>rMNl;VPi+md*4tS8?uD8wukGGHxNro}ny@NIdW=bdU)n8xp)&cF8isf=! zXUPpNrfE4n^POj6Y<#jK!107YaFdS_mRD4AL6k&B6b?(_X*0uSZGxTf5?zR!vY=QF))5K(su5<{yubbFc_2r*VbwU6VpM>!agkYQ z?(atTP^NbMzcpfF>L-8i%pIlA@}9CxDo3KvsEzJEjV+OKO$G$T)lOc zkZI5{^pPRN*jN%vHpMm1oA2IaA~*AM-&1<7q=&sPd1#TBjs7w~SY|A2yis47F)=}} zMrzI*DTKt`-GdKY#k*|jk~L0^lY@RPay@+*hed&CXMKE(Rd#~ztmHUxnpV)@V%b+t zQFluHM|MJm+6U4yZbkZ}??l_c9H#_m?8^X!lzA^?#sK?GNqhnf&ElOXROBwBtDFS@SZ}F z0UwolV?W6O-nuGm85CcDU0~HDZi*C2=9E#em2X%D46GQj&i!dOn9I8mDcE*vxQyu? zmRWJEqYwF;;H2xj6nMEC47W)J>=41j3rO*TSS$!wR<2G?5prNiX#2$Fx9-U7&D`%I zRKaL9D1Y`o|2*#Z1OM5-Op?$pxpF;>A zsxG!Az9(%D>$W>hARl|Lb<^v`+GB18;nB^|Tk$wm-B$gMip#Ix=uJi{7C-d4|EBCT z)pwJ!FGUN4vI%ur;aji28g&tZy0|XT0Un)iD4!m2+c{~()&PfaH==0Z8v<7Za!W)G z*nMy+Bo)Cytu_W-5@{AhD#mZU%f87(3J$PH3ofc&b=nrw-#K^ z7f)-b|6FI~6e{+hXKd7Y)K3CdKg;TLyH+pcK25%muMrR2hmni3^nY6QC$l4$?IKn) ze|P<)qfRkoY7QzI+A?q#A;P zMA=9REIIh`-zD*$0f*PLvhg(McHQ8r6HJQy>kRsAgX&-4nk%7N5zS&Jy-e2C{}f7>#lSch zAL%mR%o&q}JPgfe0m3bK)#6Vt^+MWvU* zq!{Mq9l69r)0axt71@~(|B6CE?)DagDslfWp=0TDO@D9G(|)@B-YK^8#>XGmF3g-F z$Db$zEZQDY&-a?Z1zezp zBBl$hlvD-afWd?g-1}-b&-I)ln{<8PA`WK%waJFs7Dv3WE1aP}!XZ?aYa4FKgO;ZX zjBl7rPH>S*{;@?ojR70`Gt%1-4bS&7g{$X({Sz|?=UO>{aGZ)Bdk8(pz}D(m;lCP| z&C1GJr1v;n7x|4OL2}U-kqc>%TSNmo5c}wU-5^{9t1$KW-30-+EmMd;KIaAIX7wBe zio2u8Ma(FAa?_R4RXE1GVTZ9oi!I3deFCQ+&w0muUUwKXg_KD(IK^J|kOU14GK6|0uqAJ-AHW0|!iGZ9l;Ci@ zcIWjR%ps5<8j1t4YH-(C*#FOebL6j@UXB9}K(EQ8=G9A(+<|+U69TjDVI{M6Tc=g6 zL9tj9osz;0>u$><(p6_H*nJ-Ifpvmc+4U;3k%0Sv3t@M@?QD+MssKB}oM7xx;K7|L zUBHvZK`UDas(WPY#MPtjnVvy@j*Mk6@UK0)bQhEPQ7d4GGp5uH(F3d`X0Dkn59Q>z|6G8(2-+DY3EoO4F*vS)oJ<6+Dmk5# z+?N2S{wq*DrZ55916*eYt8Mj9;lkS_5VBdXUL1!YdYgq zHUWCCTqIt9bO&zc0pHMyCAhuH_30!}L$N}Cee3&nV3+<9UyC!Y@BgUnOv9mW-#2b2 z5m_UeC`;Kr$xfn8vP8))TTvc+3S(cYN3_V25-R%=MV9Q#2xXl?vOPj(Fl3owir;m2 z{C{u%?|vOehgZz<{eJKJy3YGLKW845t~uQnT;b+0fpGyAJk^F3SX@hbhOVu_Hvlzj zci1uNa?SP@=>ljOmnmX*VN5mEWGI)>NTdsT0D-D_5%qpPBNYC_u;-&>$o#)8g>Y&* zyE+Z{rCT?!0OI;9jN7^dXBF-ae{vMiWP_p8U2v)P7_=QCG z4kit6^_ba01!&a7P?`&uny3N-mLWEAl4gKm6yXh=to}JM;SE!*c-jDfF;;_I7d>8`2Taw@`u6_qPh>H|RerfgxE#xI^;MnC{^bKr3N;5g4gVen}7u#mR5TWh(IadwpxH)ctj^kPd6AOUr0 zpg4hE_@MU=90)8CCbV<}`+U82$)zA#6-i&=Q61eon5E1y4Uc6WJAZXVI)V&*mD?5F zKUW){I@g5$4yt{rK#ho6$HI!EGZuHEk!~OY7Kya5x~%jrftrtRe?gSR^2k;^e72TS z>jg68mR)j$RI0A-Qapa?s0l&PkbV|0ix!H}3j6&~?n z`<)CaUkqJPuol?M1n~i*V5V=YMXu36hBMZ1a($TO0rOhH_;Of5CP= zJ(HZd#lsc7dWd^N%)+4u?&ZU1@I9gxMTOTdKpw%H^0la#LL6^erJ!)PcH9Xb(m6xJ ziz7AL!*QmXXLM-`e}PffqX~vR2g};kgJ2w|;N~_8nE`E-X*c%yGWMzhBSzQITleVG z%;qRfP_<>?#ztwco#5ljDQG~m1HKT;oj-kv4|azpOa08dKLm<@&UB61m5>7~X1boN z3)OLLzozVcqbKsz(syqCcCAi`m8!?!S#80vZ_+l1nO0bS?D~tNHkCz8H7C?`8LndD zA&sigaGa@=)Kt0L7WvpjTCl>kOJEb=p=e&@CtOJQoro{^RYo~Y*phAqd)SVBfk}Z9 zYbA)A*Dh3;&;A;hebFM!kGfX}XvOP?=3pqWlD02tyh;5&=7+RU451tG;83w_=sQAN z!Q900N~C)mXPPl4w%xdehq4Iqs>-Z~5V;E2PO|0LWvNO>jc|F3fN=3ejJVyzZU@79 z>p$CKOr1cWLt-weB}Lgfws-4Z$qFWvs)rxYAJc=Zr(rZ(DY1{Kd2=rt&2jC<3WcO+ zciLY)+fAmX6^P84HNCwvX&^0q8OzNGGK`EUoQqS%9RqGyNyBy!-eUt%bHz@bs^dD9 zFTFkM{Y5OV=u%;_@M#wt2R?c(2Gcd_FZddXaXqM&B8gZZ>+3uY0GC}Xr^8gOVBq8_ z2Lx%YYh@afpe%3&>JTtdil?%NS>c?0{=_{ViEfO2!u|fp!kE|ER)0RwR)ve3n>F@P z@K%tP9j5(iR&8J9eugoAs_QPQFBeOMcPLY{&OQkke*Pu4)H7rbR*8A2OASt);@@Wp zRjU8`j#vGrq;!p+owlwlhV=K!7NXA)$^pCa2>=%NLooNBmF~>;dfbyILvPgya>8O| zk~2o}b;t2#12b!(1z615Rf+&ARhc5JR_ERjVS*Oam zwp(GPgNgdW8~!i8MqO6OGtW)vA)t+w*V=VCF2a%9?#OukfmcJUu=!aG0qgalBRAXJ znS%@ZYIO&&3B;0eQ}qlsr<;t#W)XmnXap`5wcA5lul<3rUn}Ahf1Z_J%lg$YBaJ`a z8lC;N^xC2lC46e>|GhVk0D_3J)8cRXtYBfev&oEW|9D`NBP7~CGBaXBGf8~rA2z4a z61Ch$Dho8ua%F3Z*L6tLIn(!4fTP%R>$dp*1N@@#8oCc3Mmg@Ez6ilsMVbUl8u#5f z>Gnt2_`WAYtDQ2X9azd(E`tG6%8vsSChIAGaYMzB*fKn>@c6g& z-t7E7v#PKC55@Es2)Fb-Pl+I20mOD^C1+`;#**&Kl$iE3Er1|uuIO}{ZDAE6GK-Q? zP$&Nst{e;(e5stwxWT9S>bnc0=I3>o3ZOWwR$Ct}>K${JF&yEeS`nY%g(xvlHjnRbzRL`-aNk~pTls0d;=xT^RxN9JAr)!L1m z@V4XYMpvk|Yyzn&th9JC=Xt7q6J>Yui1Fw6?0L0}JWEg79Y>oV6XoL)ZtG232>#!z5Z+ z${O*wj=*+}UNTbu4DO7>a@HLBRNcmdD1Et<8MA#|to+iI@O)if5xZsmj>oQwH-hyN zS?ZUgIKQZLhr7^%xF0&!-?B+yFk+I6dGT^1XKK|2!!YvciPQyXOJ$a=x&K%rnGcX< z8(X=aH77{9rIm4(Ec~vUj&$c#dU5S-Q+14coTmd<j+GhE69!V)L6sB*okzZvl&KB(G*MBKg^Wj&u$^8^ zhmzRRGy_YHHb3w&IUD)|;Om!m32?fmnwBns$^NKm{|pdfo`B@1J|;lqw1-y0eo;SD zUw!XKfh0#gPlLmnncq-eO$5%Vdyv;eOAPF#Akl5MPF;&UbXK;Hxmbl`12jT1UZtHg zTjxOgVAi*Dims2Wn^Uj2&rkmDj8(<|GJy$F1^#rZN z2u2eB#p^I>999$86mphdQI9?R1`x-EzG{_k9FTmnD-wM{Ua-E-;OO;WUY z(|n3UsF}*sGUNCMZARCq8BN!6->rUrfo1dmZzN*3ILNo+;`+!e|G)bA^9f+Qft?x+ z69b1Q$D>Y+f5;xDyncPSA^~Qx^(haeZ9$a(@5{6M!KR(#kuL-`h3IS)Qe0OLLLD{O zGLl$@fRjL_0QvA#3b=f|)M`%K1)yOItB0DJyHoq0`6!fQ_KB;mfeoyoMDLan(LSJe2mhs^#r+c`a64Jv>Mk z3oMQVxC12-1*igCpGfd2yO0EF?3WWCJa`Z=TwD*;{W@CKsS8y*T58XN|I3fM&lVtd^9&?WaG+Wrhm(7A5IykIe<<&%Xm@B-##Tds zBYquRJjDcgLN?oVSV%UwX{&;A{39$$Tro!Z}Oq~%&DqU8-QVI+wnQr-utJj5wIn|7Eo+P zuoC&)d0`Nd01+*z!N9E@6dc>xrs&po3CdZ!5TOlp2lEv1yoQWSmp|l855QV=9h$eK zBVYi3?>A>$OD`!8MwyB&Cu0*6RfV384n-I*Iwwa4YN0!P_M;uc@77 zxv7BSXDkeyh0j1|a@K(9BQ#@Em90@#KxQ>M=_GCccDqZH0fZlXPM9{L9#Qq}P2 z2T*(pY0r+iUz}WlB*q+cC&GU>``lbyn~*3~%rST>MsTn?ZjrB&y!7$q2J8q)Y_b3e zbJ~KpWj8*724)byix7Yl4NQe}gEvc%nmgA`7L2^SAZR8Hhje_i?7J0T=$+@!pp6HU zM<5dk?9&>cYemdq;S9jz4)_L*;1HKNIS{^=6XEwx2vWr8Dp>Vy00!Md?l0xWU(n6% z#unw$+!VNfy4V4rz_ceH`V97Gwo~WAccs|fN~_rc>S1{eMo(@D(JCxU-zaJ<2ZOaq zRHxl4e7=+BzTYu>T~IDw%?%pzzp9-foWw&L=-kA(@HZgVrH2_|!L<8#af)=W4~OJl zehk?|~RoWG& zD{PuI=)(@Fj`=^aAxW^K)x0Ag8d8UvK(aglKvU2i;KF5-z@q|oL{CcS3sWXftJjsK=< zLWnsm3{6ukRhU8IFc-RP~Oe~Fo2<1E6|t#YNFrD2q6 zo-Tru{Z`6oZ38UH#H|=Xdqr011o;xTdx9)_9e^gb$b1jWOPUV0CJEHSr8Q zY`IrP(Qd2X8ojG3?={9lRzDN9PCVQN-+7sa6a>M@!D30xI5iA(QuXhCkvrn{utF#+TJZd*%v|yIp!_(kgn$*jXe%tX=tb3)8 zeBYIHojzn_l;-MiB|9P;2S8Zu} ztLDX^wx$3tfS&t+-yn!8b-aG~=l#vkuc59O)JKU3itYm41 zg9}nb4e{&Ee!ZPl5GTyPuMU!GcAW+%1KkyDkbUc}`p+-8g?+2W(3i(2D(V?H2D?XXO+9e+i!$Ch z1*{s?{d0!tUc-s`ExJn6|V;bNAAe_KYsKscLX&2rY0)k&!e?o%F5+MXao5@$U*o zh`j}Hy%9thCgRmP3!797q$fJ33qZL|J^~nW_*4R|)wS|!5Oz7Ab}d7x!4gUS^xf%R_4$OkFFr@0 z+TSSS)XMRl>Acl-;m)dc{2s+ItgfghtLq2T`@6cJn#u79s&N8(Qvtyh)EkS%D}(i| z6a0~bC<+7yj?uvAMe*gt1{M-Gn0H(f-XZ|*M- zdh|SUE49P)wD-LQo9;27wKv7dr5bz}Jj<z&upoKDDXGVfcatawxtFb~OKn zmE+Pubhr&s9hPl5W?xx^8fw;nNY9KxQ5A3~Eu8=`{Qp%^`~Uo1rT!hp9*2`A+Q$tQ P>EPcvpjnrkbd3BDuo!-( literal 47197 zcmc$GWmJ@J^eqh{64Kp>h;%nn(jq7wN{4iJh;&O!OP6$)fJ1jmOAp;J)P3f+{_EaP z_uF0T&RVQ-#+iBFdEWD!v(Mi9OxQaWxo4PUm5_$^$^eozOOL(iAJLJmY%&XzZ^ z<@x4^&0eHmK6robQ{v)IHAzTtIuhmY`CO&-H5A2p%jmr~!bUQVie?ay)EJlq~lcT{#fcs_aZ#NqaY zf6+cjE_$5W>NEl{@c@Nv84Ju?1w;9t9kY zW+u2E{FNU}dQG|ff}LIU;r`0-;_7>9>Pw;s8Q9ek`{qC*z9HV{F15JP?sP&s9nJHd zaVh9@>9(_vfDwHgi)~&FB6r8#&r6*FHD*zs zD_?_yZ%qB=4}E0k{LmUtZ$J#tZfCejl5C}3&F<#K($*qe1rQQn=c%wl-)qSR*QW;?6W*A(3P z|JGS=&Kmk920ta{>SVyo9mV0Pp^XTo&eE;7iR*ZL(A76^UqiiF4Z_(NcvX3Rx%M12 zL(B*E_Vef5TC3@jF(GeF`omYhaKx!jM9l8<88W7>Ug zO9~vko4l{rLV{tzcOa}Bw}wb@C|={;bZXpYv_c?t4B~ z=)Ejd$$0PcTgY{vxvHv420Ll)*QX$;YWOXP*G=q(h%L!;`Ii;cbg-J$7~(64uS0t{ zH2c8f*YA^c$n?qTAerA?GfrmeZRd@|w(T8_#ESNh7jiz^W3vZfA8^4CeD6*me0wKP zNqu~NWN@3*$1RoUHC6vbhe4Zid173GOGj6zMi8T)iZ`B~olT?j2#x36N?88P&o90@ z=udd@Wx1h3LNoF4;a2tXV4*M?`6v}cS#d#S^*0}}>(wB~VcyxdSBFc}9Bguv7 zgE$ld*<9?`SQ3MnXx;+k8k-|4l(4aXqCxkoA}<&jNA0&NZEoWw!J~P(Xr(mkZC2AP7rwnWa>ilz zi;WqrCZEY19oKzM65*8N5 znMu#qc)ybuqXRA>v2kh}v7xux>!YdSgvo=)>-89@*r!jQf@e%kOth%nR}mFKXb!hu z9EVDe6VkY$q-Kb+dlKwgzk?VbRfik23Cj zkB1qvksp@Dx(&8D#-$(2(>90sSDdd?R(x-V8x2)dRB|qEZf=r@4cxn)lauT2_}uo< z#}M4KPsfyN6u-|g{(aVRcRKj`m1YIc0e9}tEb8mBu`9o=RC89dYG386g9Y>5g@(!# z(tP$Pa{l_eyR*@JSWD3Px$~fLjHnsNVBR@ri`Axzg|}Lb_Br{gSz<|jeSL3Yy7xqF z7X=i}%nG_8;KCa&9o=88u-9TAee(XeyWQw>>uS8@_jqr<`?s#NyJOn(0;YQh;3ytiAYnO( z!E<9sjN27&t=MI;dSOY2%O2!#v1u~@Y-eQe{ok{h-F?BXt8~|+wkx6=kB8%K9NiMl zN`nm_+vZ&+MeWc2q^pQb4;B~62ClC4I`yBr(FOxipRyYITqK&FZ4DFIwX8=AM{a); z?|Hbt)vahfRZ#2=C!DT68xbBgM%)7D66pe}LyeweKVy<)5Sn(g^X})c@Nlk~*9NU~ zyTp_uUlN#AclVT7k|*}wQoB^UOp_fdN>WPb_IRlC=4R^?qbueAKI#NTlOg2 z1X<|g-U({7ado@*aThw@6~ya}U(UbAUq0R#u7`{B7zoEQ0XW`X`5rcE3TJn9FS5KR z+#gp(gRh8|jX}ZuAMIX`9Mev_AOD<=(6+s+d+mMD<5c8}9< z@TquWxBE2)PdQHVRQ=>#_vgkxGwsdtynNY@QGd)I(DCe^Kj11MkJ_N(VQFoU98!_R z@4HHS-3o)2Mo}BCt*t?w#;EFpF5F)q&j&*EUFS_%vd0!${)NAq!a>zU9@1}d)lTF4 zQZfTF!?NS`@h9UYQ17Fzo{2p@7VA9g5-CL5gKLsJ={Qr#Ka8RG*1ldri$XK zd=_0zmYQAoMoU3_Lw%yDr?oydIhwAAWaS~$@KolAx!&L75fhs?olhvf2Bl83@+gj8 zE?hL^u$?Vm)GmVpB?UvnVUbwRK0qhLe*tzzAzppdyvKII|FgmFG4=7V2?J_aF74&ij&*HZlWF=;sD zshCn(%=A9O+4=5dl6(SFG0hlN(ptB)->)^r*>U2GkqIA=ud%P}X2fund_cA4+lxaS zH}t$lsA;emrz;#?@)f>D_~Uc3W_iDf-d8}Y9(vLDC+%YH-}lQKcKzll4EvJ?qYO?7 zv;i1o*gVT;%hA8dr2o0`(m}(DZNl*Hy(o#v`CFo+6UWVg;>|wit*x!UWe~5`5NcoM zla-&Th)j_roWI?V1O{iCx?#M%n{~4WHBY84_vei*CkvuBHR(8)`~l;93Yu5-1Wa@3t^Wh03hQ zvI9IHE@O9l%|jcIl)6#0~y^Np_~TYd!ha4e0>;*Gjbgd_Ra#6IqETdEFOdrc4~G zd^m+uY6-n7?jqUOl1lf-gqR=Y4z0Tz0mGVTq4>~3o%Ob(e{n;E()arqq_^_&oJg3r zSFhW@pF7?fwek#MbBv!CQai17B2Gq6!g$-(zL8Tbn}h1b!X138l+IV)r=Olue9}Cw6J*R086LzxA}30 zyxW+E6N|?y%O#l}ZDN|es}vu9?Z30{*yV^|L#f?&0smEjr?D#7kqD= zyqw5aobIsexOWZ7$mi5T;=LI_Lc?5j=MX^4SRa@4>^&)A6v7vrd@ z${14Ox7UvzZrrvceAp=UjH)K&GlYN1IiVY3Y!t%wx842Z^T~&q%KSpi+!!j+ziW=( zOcx;f=A1V6kFfpJ-R(Zd&X+eVLyyI-ZAgWji(n0GN(Vc2f41Ss@-?Wz#W1g48O=pz z=+TE+`Zjo8+CFZ(S}v^*=s@-&;mB1STlLwuw$G%X;9U(Y2$gw z0wuRubP*wF({SzF}1Hc^nis14PkNnh%!c1z zgmCIC&PTG&h-WUluRpsG4A}lUYN_NKaP<}2v9-9Uc@TCq`cQ-Jr`UTRwOUm~_RD&f zZPHL`*w@5c{`;j%@%A5&rwmve$UiRF`OO(Uk4+8pWS?DZ#_YeBM^?HU&%t^xKS;f^ z&m!=Y3Lzk@&}iWGUGpDm!S>_h(&|s$->hU&^xMtGb7i24XxavH6p21uUepyy6AtDI zZ@#;Z8+h*KKW}pd#Sp$kgmirR7PL zQq2vmAmo#gR+-G{U$dvxTp0)8j{aL~scu4)Ctz9?Gx5upvZ5VkW&F1) zcc)Ckl4{R#1~Ihsh+VTia>#wiuT2VR+fSfW;WihaiF__T$v3za@=E-}2?CeKb&o<_hO4 zgA0);?#T$QZVqgGEDfZ%J@#ClpunTyFbgdu135U5vP;OGQG(+s=J3BkRBHQle!HZt z;dJ3v>KVbHE!o>G5wDRzGo^$uc9#qvCxUBQsQ@N5x5QpCg{Q*;JNPN->80VP3&g&c zQ^te;?t?oKv>6CC+-5Y)mtrc4m|pMQ?i}Qjj?LfBXirw~5}?YhnCm%nQ^Ob^-~c|P zA9S#7@pZomoS7A#HM9SmKhBiIJ7ZE#6VR(n*Bbw-^S*hPV)nAPAR@!S4T97UezzQ2Kv zY#Uqd&MJ3AkBCucWDHKUelz1fb9ur!v&&pF+n z;J!7s4=)rDmb>Y#QoaSX#4kP$0ild zn%&b}8t~0m!c2n^&HSa%Eky?oo3=Awori?;)5wm~AxQF#CAPDfP2?e;$OM^5dhHmM z7YaK|^JlCoB6iSGDid$*&$7dOpk{>5eGMV|;Z&dY|62!+&)QCxjK~(y@c@#MCT@%$ z9Hyz-F&8{oHk0&9D<0b!bSu=b_KL=X`h`ep*&+<1BM`T)Fum~xo9&Sdm4+xpR3Qmw z+s?RoTzeXhG5Y&B85u)vFN^0r9#a1A$OKe{)Jtf&BW(m^E7k`m<}L1W&jF^scb7R1 zM2y>cSjW<_E0a5genc}N=Pb{@_aZ^$f3Q_NCH23eF2L<|$Q~8ug3iREc89viak#5A z{L^|AZF%D-)_yUg8!Riieg_?DdSS%o#ei_^`Oj(Dm*Z)B#PP#vkF^DnY8zXoi4QBI zN0uj2dS6iItIF0~S@yh1G0>U6ldPdilf@-ijluX6Lv3gxuB_Ub3X9O`zASV0Md0EjS4O(46yw+Ah==j}nm zhk6V~nK6bic2Sd!d)W?VsSJ+%jDfGk;YacPiGcDC_V)ANb45G2~I{;lv&SJKtTFqlAAzrAd zx4G^66sRd6SwV0{UitQ{`Gj7*Uasj&6ZTr63kpUQNiilla*34fO-3^CYxQ&K4DUc! z^>(wFDZ)=C`p7 z^rL5Yb5mz8RZ}>+%GIXa{qhKOXDjqSbVmFB6i_)2k~MpNCbI4mwjJ@{eC<#Y^!oeM z?-+;kus@t@b+X*A14%^8y$y&iQa3OTV{CEy26F6P1uA)q)!8oTl*|1gIeN;|MHdlW zfiEC#k=k`jOe%rJ@jE*{e(cD@hWtz?&l52o`?@L&ipo3~d=_BTkMEPsner@1@#HGR z2oDm2n7`8ziU)IV%ka5O+~VblC^R5e{q6C4g>kv^TCBHfchq}CMx{W7&{Ds4#uF;E z_Y_m312pai!e`8UQ7=&`Dai}|{#K&&WFnU1=nl*b9DWG2tl@YSQuBG8c`9PJn?jUG z-QNe$Eq`nXE`$Bcx24&_3$FrR?wns9+j&hEsy3vSHWfAXahg}MA?)u29vCq_ms2>? zG9Pk(C@wCR?J|)gCb$*%y>kyM4DV^3SZZ-|Yg8uq0*Pdg#-y_QTPsg!s9e_b9%bYstJ*m4+QXF>-&NE4x~hNN%BZ*uG;z>Bx?I z`Esv7!1}r_c>f7k(hfn4Vcm6a@IJQh$=-v?%mwxVx_&keR@$pL#5WyscM5-2QI;rz zowkciP`JYS%GLs%g0>V*$YaOfJbpO;^oARQE5Dhe;@d3Z!s`mzY;Xk(&tPq9MJps| zEHvt2?U@#7GU}w&uF(Suw0V)NFQN%k5lx;PKc&3KkBjSJ_b{)N0!2QObd)+q=Hvbe zgj#c+!i4hpjuQjph}}HM@0dTQVXr7fXgLp%-s1k{arEf)M})Q=S@Z~u@QQ6y`CflYM3kxb`_~gz@*G_U;aKeV z?dQjb`x16mF>(iQ0o9r5rwae(&9gpi<8PA;_g#zp-WID4xZamk#2nxdBrE$k-g9}Dn0~b>O97#+=#rSeQ)_b1 z(`H0JP<;!%BS*i9on0e-3HvM3X%i~6lnwlMcGy9TV|vk^=0$!XsWfB72E=ttK!$l-oke7lMSlb=7Bo;uyGSzR z#tZbbvJDMkc)#Y*WO@Eg84=8;a2iiB3&3R?%vG@zlIt~#F?0ot#;2QfR3006>88wz z{zHEgRK;G7ikw9*EWke)lgL*%oMK*P>fm>+J^+lqp7{Y}ah?OhdU>t{Y1W)A&!E;h zNhD^SG~mU#2*XCu-FY5@ve!{S`%p$LPq7;C^yE?c@a^|p3faN5rr`(mR=9kmhG52V zP%vw`<#Y;tNF8Bne*r&7aU71}uf=~kz7Q9T{dTszbv(eV@uSm#X4S{f z8T68FcXt-%Ppg(+maz zUAdLP<70bEfXeX-6X2p5i4olZ_OnDla+SJu*IG_uw!K>Rd-QQxsxcp>yS01{KZ96P z4)MEVsba0FTs@D?#NZjYWC4W%syF#+*-Q0VU)SOmx7ko^vVmuR0o8vh9IBCj*!Us_ z(S=O%gJgD5vc5wvAzzFnrQ%eJlCtvaT2T`YIJHHnvFre$EMa0w8d( z6Z_xi&*;HQpK8J{a?KHqR&8=zTn?MRSRx~*$Ek_sQgdA^=o+g5d0vO*!v={YUAO;6 z4Zk=ntUZBmhgE^{(%g4J@?E_`PjH=&E3U}HvSSfXl+gsx((-a%6-se6T(a(${T(xyKZFpJ;Z^Y zk2I(z{0vAx1E-sVy-Gl(veznF$$T|h$eMV*h)VxmK&rS})a%my_FZV`9TKwAsi0yq zTO#Pp1lkTT3eN6;qQk(>UO9K*ce9zyWj;b7_;`2Lt0aDZo^S`4yx!%}#BKE?1;@jo zV?b*UP8F;7YU%@a_H?hZ<1cGOcz8S}AxobWj_5|+I#{U)qhq!A5qi8^ay9rF8Ub%$ULgL zuIImY#_4QHT{5QdeC4(NgSi%ext!k%jm36VF zM6B->7;_EJ!a*NQljU#^dF4swS(N)3g5Wxo*Z^rmdN0>Rn}O9f&Gz)UXuP{oha7zXRDs6Ll3Go8sLb947&?ze+Hvu~A_C zgjnBMt!gw5YzJ>9YQObEB{3j)WKO#DZE1}*0p0ZcO~B8OHD<~es3o-c$*HNk$L-g@ z*jMMuJXn#p= zVILD6j^~At)dpFYvGOa7yK4y@MB!h&YD1Uyqq>FkDk|guEHG%-7&zDZB6j>Z)xskfM7^KZ& z9areo%L_e+<=A`FR<>($X^zOV+$_J*Ownl^bd-Jn)OH|wA}40l^8-c{P8<%*5GA_2 zpu=<>GK#m9jiz`B=MFJ_5#~9Sw!_97$O0+6qa=8@2G_+|L&!n1s;NF-|-*TubG%+QIp96c`5afs%IYF2T)80zuShtNRxWbh$|TU~k~+2^doOT$#$?lS7i)s208-Yp#FheXz?3L z7ewa6M%*2QG`EP)S%}j0I6tDsWU1d|!_GluKHt z&T4Z49#cXBj#rD8JK|w9&xld{`poE7rEh4@318cKU5Yy^+%9)IIS|n3Tt!V^N%5ni z^j|{!)$W_02IZ|!VpEHirHxrn6{+nzfK<-S7ne{R>+|L~$Ow7(`;3p+NpD%lPu_E! zJ5WIMmc@`nu_h#_-%xUV@@jt6a~KP;1N*C_xRwDD+#ZFoJm#lru7-Gj8;j^foHjOI ziBJC<{TbQMv{Hwz4s~_)eBfMR%K*!({?}<6Bdq`(hpJ4r!dO?pu2+(14n~o5*%ol- zADQbC#(?xg6&3Yi-onS%UXqDKGXJ4#bLX08O7LJTCupGJIbNI-i0opx zdO59n0Mg7yL7QA$ zg!2X60P118Y@(~-x19J%cNEl~h@s#>f$`lXmCkL290S8a1?2bqXdKjWV6iBX&DBi~ zd6OWhfb;1QY=dly00vk-R8rjE>~TBOOT|=uJnLifOC%F(wHbi4!vCmk_;Q3Vu4G5z zp#7j(4~ZMOj5Y1{$3M@PYRam!o-Of>+kv1L1&6Y`>c5zH;4|vsMsv&7CZG=;pT*ns z7^5h=6%`7i!3&7G-?{r2t=R3z!=2ZPt+2RrPiy>b^^fN_nsXd!OG*?Ep}ozW8w4x9 zer~@L@+kIefAwWlb4UNkUqwH$7onb%ZwReu@{jT^M z2mObxiPg9r;V623URlm4bM>4+Oy53TXx-@B&r_mlFo$T|x#U-xDs8szwmY-CMWZk9 z@G`m)YOqih{df0bm)?BC>B6e^*?7dgqurUx8^n{K>yyqJTd-{z{_bP+T}bwq_(>0b zflW*(b8J=Ek3j6SMouKMR1r)p=ESdD3Ko};y>KNaSlPeN-^$8Zr%~)yz66AZQW&Gv zU~vY?ZwT%*gbgu5E$QZJv&zFlLkFf z4a6NS$ATe)D)o=H;w;6fP4C%+^qLu%`hKxR{nWX!u%o7eBSC?gF05>!6w>)7F5d^M z)T?c_{{9*Q9cnxxIw>Lk?}g}f$vdG=PCV(TXU2INLtmfZ_uzB1tpO|@_ z{*u$qb^TA&w1azmVHPnSUfc`LAW|A!y=4KH)Urk2peaDlh9|wCr5vO=!0Zl(XG!#RCrt$x6)T`?h$DJG^9Tc6wsHz~ zvI0Pd)2L4Iy3~7c?N?LZCDHld?}uT_FC9K$KU;E{^x?YLsEXg${wyq9D1)5c16*yB zbbmU?Zk6TX;n7PKlXvfXHq398CF*$|yVe~t2wrSb(_Xus4kCaRILUq?KI+nis4Sqd zMKhRh2<$s)6D)RYJFUpykZfyLrx6-&M%=PFc%(`uzTw>1*qx&C6SN6k={ zct_utHqc(2g06LtdsKuTIMou_JMP}w0!&MH6(&&v6n@~ING`fJF*7S!JplE{N*ANk z_wn8>oyqdNa0&t&m!mF*ECVLjDqT%Y4LiNWeB?)QW_P2*I)k8~HfjgJ?#X|XFF}o# z2bz@)mD|6bC(}KZ$jiCW5Z?|A0f1j^T{^(7QBGpjrURI&2~R2M(iqvOwYTLubyL-)mF+qQ__)W3?1l!Y&V59GOSNkXK$0cy)l$#2igeM^*Nd*F^LrNNt*SCv?eJPRoBvTpPN zWqH2Et#S~4`HNUNf_A&dqz^F9^M?jR`K$kaP+QB9ndx^AE3^tTJ+^26hW{>mGOMq0_=N| zZb+Ky)w`dIqH)8nPix~WW@L}Qi6y2WJ_$14zoq$WJ@+ezWhsc2Nj8)I-92bnYz9z2 zZSbXyb8MU)ffzdl8@qIfHc&ir>;ZHg(?h~!Xw1RE!SZdM7e9eD&C2;UnAkwe zLbrknCvflUR}@R&;t-++NIVYsU-|)lG5EI*h?!F!6+%)Pb(vh$9THR8B@> z8!0SAB{!!zmcPp0nUXmVV}~%84h*M{~rI}?;Y42 z#Y5Q&;i)Ic{pV$U0+JAiGiwsctJ{ub=FDyq>3Nz5tOQB(r(9{8| z<#c_s>p@AiH)to>mc(DZDwPhG;!i))oXlbI_2ES?R{;ZXbRk9qf?H0a1#ODRU7atHSJdq{`9?vO*Fe$djLV2 zQB+jFJYB#cR}aYc&=sI}zh97YYN7{;u2K)UETPxzjjKpEK%0HK1gzB$aln_NifSp; zeGQ&c1Xz@TlT!^)36<49o99XAwJd;_=`VlUZh3mU-}M^$?ovyGkSj&${UxtWsbPn& z6*b&Kjrx>mt_*TxHI^-zv)=mv)Lo7J6M?@+;47E_m+nN78a)80VT+!7<&MC*W8CJ8 zE$}$2=M=@j$XIOr>hJSi3qf|M`+gSjLlpIs zWIe|)cpZbfL!!uE0KD*Sn-8cq?%2kEBGn{W1K**v!@c@|bkagS2ZX9O=;%05k_F{r zylifpASx;KudyWlGq0zbXU4&EwP~N(!D5qoruV5loRz`rBPaxpa`-p^C{`@OAAh{a zbbo`Y&E9$(*t}Gx&Fukb4GkvHtR}ML!To@B*ocWWU+4rXbm~el3A%s(o~l+AyOB19 zg;%>C=+yb47>ofz3T|YD@n@oEs;b?}zVqHp)FiqUG4$MJ1Q;RcFFbn#pc#oUNmxM< za|utNvk2QXEG!DIrHmrBuRyu)0}h`ri^KtTfM1&Gz|&j518l$*7DrfUAOYa=34m+i z<9r#QE9T4uZH@hhEA(3w+HiPUpiEABnVuyyEieF02>-Hk#pUGWRC``ng)F=w8uQfM zNe;brJz8GEU_(q*Maht90^5AE8%JD&e$sq6jS=uzuhe|hObGSIK`-^D;bc)1bvqBh zC0Y9?eJu3;4dRuTlJ3A~txq)tidaAhsS@=!Cg{Ix?fE4o6ZblnUcR^%%;8cYDKlmC z7oI+rX*Y3gHP{U6|Nc!GJ2^UH9uUvJM+qOfyd7lkn2Q|v3#J$NrM{*)WTG zh;Ig{-)hRs7S(cDZ}(P*NKKyTfAXVfjM|Rg4lw$=Z%Z&T5+bHKQxMeyw$lZ`q!_#tpFSp#UX&M%7z;no=CG#k0a?i6E9UA)X{>0f(;np)cx;h;L_h9gp z2G@xo^3%#(OdYTFOwnqo$JrKk%seYjF%f_c8(P4gYhOXvVPzcA?@j3&SlT%YLGqA`jH3C}5bqSVaf`u3<}6Oi-l zwW3%_U{WPwV{413|iU0Dsy<~)Dg*K<}sD`zt%(16k0?PIB zGq=OsY_;j2hQw1~kI?K&KU@jjMj|0mCnF(^F`|H+0R&&Nbq!jB(Sl;iE_#?V)Cf1I z6xUwI>B9Z6{GXDZqYtpEPUzQ@TiDY;duF3@iJBPT)zWrG#>6B8k)^bZDY}lC#6=9~y%*lU40%%5^|Gm%Y0|@1FTiZX`&4y( zqN1O_^v}b3oO(QZjmS_g$TF|MB2|u{9tL#@%uy?_X@irG?UEi7^SZRD6Ky5J=)_z6 zhmHy$EH^EY>vtAz60Og&OoL@ijzcoGAMu-bGw!AODdl|55R{wT_j1Ik(;U>E_l<{| znGBEpj9x8hY!9CUKI7Gy_9J>Od(b?+3y7;Ro{04k@qyK162;+(_}(=Rbkg-rK%(Cq zV~ms`YwaWN{qi$k=;h3AB1H~4Kl|fPve*9Wa-pHa5RCT-Q8|m+qithBa0GmR2;Z(( zi&L48D%#;}5BrjjWFprGtxILANK1ExAexb4CjHm@1_Jf(x+%7ZnUJ~pRtt_CYTyXK zA_?mg5$FU%y)n>l;0jY*F^Lge^Z5UNe<#*?w_BL?)9>-&zY+()qn^w3hiX83H`FI) zX0pM1SRhp?zJDLjz{E7bn5;)v*>;%*W=2wd_dz<54q%2)hV`h@9(#f5BZ>+y*lyuw zabc#Thc3uW2~XKu-vLAMy1frDYQv|vCIBJd5Yu2fm<09g{`Sn)E(k0P^A6GwGCFpV z#6EE1CV;Vr{@uyK{Lj7N@5^;~1wm8r(|k1ZJ(yp3e`F3a*+h;Mde@pb7)jx>nqq$P z>{%iyw^<)JRCW@(VY?paXtpLU;okilFqrWZc-i5A7pOn_3}A=@4wCBb?{BXLLDypn z;FB#i_kV+Hk0B)8@reMbN#*8hc;oH-9-ey(;=mj5&Zo;>N5!3$0tbKV5*F+_^ee#F zSy)Vrn{`%YJFLVr04}8=y*U6zrG$hRsslhTnD)_o#)0?tH@RIaEqwN?ibDBlb1*p> z;GBMdS$1cto~s$HCqEl~fl7>$yoVm`xK~w9AUZ4~&4ZcptC`cBCnoOgjam8+~~(Vk|%+tdve(HCgYz zYdLCHe{2T~24*?r64!_g$R>_XPSXyAFKB5c`RtbWIaa3!!K}mKDd%&w<5l?i;H8on z@mWqtU+C{IHq|TwXj<~535ZF(Vuzry|~d z&V_P1--oW(T&dte02w+w?*(B(bI6QmfvIap}Wd}+gTL2u+jH1yvF2+Qjm_XhHs z_+U8UE0duVPBzr!cR&(xc}#qzSpqRWcax_O^VZUz1qzI%I2zVKRlA+T7;E%HA{bk7 zgm4dN5w|83@loR`Umap{8c$PO=Dy7B!F1wnBT0bmYbVPffg`RJraHRhP(L{u$?ka zQxlkmf(T@U!`QqDfMM-E9PW$=AAOe!!G@TH_0E@@8!&@d4GdO%{wK#)GIaat4; zotIO6Yhm(LQS!}NXEzJ8Ai2F(q(r_^8(wJWbba{W5WGgMCCHrZnhTK&*5!gw3<U zI|aq2@EoK`V$idtezaxv`T+Tt)93>9rs=M=ck5<;83OenU#J#pz>cX(*)FV?K{PCQ z%dbsbr>%LYPH~=8K#s=eg>}cxJ2b1R_x31 z=g~9EIJlc|mw)OJs!o~l5e{Q^N9tcxH)n7<^6mfYB*AsH&(C(Vi+bM`7WmO3^zB;< z$-$p6x3^s#(+4$!qXuu@lqZ!!zGdpNeT)%rS)5C%(r1O}OO@9X(p`@VV7D!9RiVbH z$3X7Oql;JTryb@fDr?p;@DvR)e`sANG>rPLqNnsho2s#t$Cy=7EF9UQ;Is$=e(zLV z7X^@BJ;y;FEH^tH&krdZ1|Fq9QRIc;m3}~2`{QULxHnr_slfQj;0(<2m1>6|@CsM4 zQwXOm&Bw$u+ix-zzTL`1j<0j;FgaYV3{*gg_vrWK=u`E<%e+7Ru(G_}j(c(x$bJ&Y z?vj3iy3)Emql=C)V^|NZa9yY|@N->2ef)(;?#M9@$76qucM}n1cH<$$`y27!eN0zS^&oEqYEd-xNJ7;gQ3QRz3sVOp+k}f#^ zqNt1@mj}*vS?ThPw%8+>wJkol;C!362!|5E$!#PhITxmgQc;hEFAa%SS;_jYsf)n6 za0(R20buAa!E9I`u0X$9wC{cfpPNgEWRu?Ygy9FnzWj42>h%DYhCU#E=M-c|Mn=|> zzWrME@-WDn?$pj>cb6qr`Zy)OI_~;PEU}o?dbps2Er-*>vICs~5p{T>+9o4}CMQ&hi0iOyMg?+OUg507s% zNM)jw^(mG<0JQhPMGz!=4FOt#iU|%+lkoBxyUyg>qmR_su}=m5!LT>`;SrnN&cGTK549cu zr)k*gp>a^auJQ&XhB{e}%0v43^XHM{ix%yFcJ_NCh)WlC{KCdJ)<|}Vr2*5~j^KBi zR1>tecw@xlB=N-TLc*`lDc^5|&*4azIqCh-77~d7{!(^DTo@|`(_WEM{(R79@7*70 zNSUg1kLI3V`CXtIseE73hQEiC>_)9VM;%bR+y)(5ESQ+qf^)hxZmQC~5;PiV4}r9{ z=foOy45~lkm8SU}fqCpczV66>!`vhZ$}8pnO)O6Kj1MNySCj`+N&2w*?Y_~;Aapx{ z`CTfN5g;v+p=6|rUjW)DQDKC>8vzqUKte8l7)D57KHkU_Gyaai1yeX&V)rQC&Hu}0 zMJ`JlDG2#P+lr0p+8L5Afi@Cyf}Sgzh? zo(}MD32ge!eW0T#L%TC#krl9C^>3~K<0A7dOmsL-ZdSFS(CHB~hS8aLY1wL(*H5BY zQaB?*Ly;nO5w{36^>UDul$JtYQh?!S(`nWU;e|cH0o_`QPnQ*8eS>r*8eFp0# zPX`dD6N^0S7)8XQE5Q~@r8WXRoF;tASOE{QIvMnkJTv?@bMHV71(M~4v_?Id$eDi+ z6L0)6=B5WUEVO7Jvpm6xwN|Cw-)RTPqlF%|?i2;VBA(|(@L}DA7C1MUD*+-u&;p3Q z@NuVNgSKWXzS%#Kr1U}PI3Lv&;2=kBl~c4+Nkt_Ij3t%s(is~YKP`_54UN4y+a3gD z%Fr?|co81zXMaJ5f6^bqF@n5+)kf2zjcPH7=08OZDF!SU7&x7}004|pr1hA_MKqXU zfCT`1LIQmk7)0tPUbP#KPc&Uzz8HFB zIOurx=iq09;rAl4X$mnu29|;AbYSmesS!#6qsmYretu2;^}#-&Z=&AUznY_CV)(sg zr8ki7l)p*KmHO$Po-rElsTk|K=VMx3QlbGr;)c5N2Hd1Eo5J>dzoQ5!3-uY-^e* z7jm;v13}g)L{d60*8_t0`To2*(Q6$Oloa&>B{fnd*2lFF>M03oy3-6Sv$xV;dw432 zrq^i!ZD$0dp)OJ{jAqqu)&TrA<6ttoX3Zra)6_sG!35aB%N{oHRUB0^;`XKFq82!n z$-s_48KnmyY2crlmR3?RnE2`irI7QxGz>G!UW@MDncZ$V{yV@`}SNwCS|L36ZFE2wY(hG55 zrKGl^TLBuG;vldhnu4iiKBstmwfVuhdPUYIk3Rs9hV}_x zup7@mg1*@MW#=6b95tG9gpXR)KZ147g**UH7}UXCq+Uwelzv#K6YFWh+)r7N3vdWw z<&7d2cxN}@b=39(sCE@#j^P~`5od%O*8tX*2h87eZ7e{Kj;9;)E|0+csR*CWp9eo~ zL59l%SZI&6VB#S!S?LW8TZIyn1L0XH?qp;8S6R7}NO)KNhLuSPqup$n%mdi1Ho>HS z;?&gC=OZx9?g95|nSf?%qU6J?Cq%Dx>osP{`D~6abKb=&JXqK5XC9|Hv{*fgz2Cbe zZ}n6HaqO-{x-nV*MK`g_`=D;t1hl|VIPEA1(g>L2lV$l>#tcYxs4ws;eusCl4b11eA19@dZVmtU<9$CMV1rp3CN*3M zUaq~DbW+B-jLnkzl~h3YL}f!uk(Q*{-g(k5>M+nqxxl0L=J)lss>pMC8!+8E2)LmmXCE+Kk43^E0q@Xj+?1uUH}cKjZXLG$`GR%$WL={iTzgeS zGq_s}K0XP6)%>Tk8$dvB0)j3cOow*w>*L2E@9MgSZs z`}er-f3HW*^*NvOe!t$Y*K@t^0T`_hR%S-IyDwa(oG-l~Qk>h}Ublzyf&ey=^rqV= zuV)9`C(7t)vD2g}NSq6(CDm`-$V9Za$P#+J3$!biM&ZvDgosata`fPl7ttj>Gi&6B z;J^JfbIcJWgm$ny2cBD9W|VM@vjA8d=I;&`Np*16zQAD$Xr^K9`Is!rOi!W;GQn-i&u3 z7U1hH9z84L#I#CLCcQwZ055cgm}ke34!fp4?IRSg{aruW<4x46tzV3dlzb;t0?*-OY5(k zW*NG9k@_c`lJ}`ZV%aVW`iC3@f>JS7>{ z=#T27f#<#zBJ*!|hQ4-^`^N#7>FZZ%TnVD$l|^wAeDEW zH09>6Ij$0ItxP7^(9=ZgPXsD-M}e`&!SUDyczape`>RSyLj**zwgRVF3M^Ko;YUXv za|z1~^8^+j>;rnXE7sBoz(PJqDk`C90SXle5g;lEUt+sAS?4{x8OL!IX)|Q_Ua*da zt;f>S>Wzy9f38KSe-H^6okv~Sg1_|c@up&uN(Ak0M+or|f3jDbB({bCHUiS0m<4)xq#xOV6W>q5A=g*G?Pdz zRR05Kl#0Q{!$=-66VuBv5$bT(`=w6^LM1`pJ7W#Uxw+-_n#vXX3hNa)Y@Mm~Q}|X2 zEkjiVhMK5|&PTp`8WD5Hyx<+x5P5%n`%vftC#tWjuOn%r{8H2eR40~`aSCeT}~DH_;K4aw~=$b4lfMVq7DpvNU(x|)*Qf7$?&=%;pMAvgRHN5)=dZk zDb`FCyEal`Sz39K@3TU{h9g3UPJW(;&d?5k(@LP~d>gJsp7t8llI{T4#03rxj(i0N zR8{`CWNf>n&Nnq4sJ^L@ja=apm*wY!k3*710tj<0Z zJSyy1L3ESVIf!Qw1uSqTGF?EY+~W!c0s|%x*$zqLn6w{n#Gps{Y9yi)9wqGdpE^YY!SN zidr5*r3Xl3d;y$tEewBx=<9iPG85rF36ndX*&{vh-he~V)ZAyOH7zF6>BIgaO zd_d8?OhC2;`mI?50qBkU5Irs(BX`|eK8ADz9;^%Y_)sgmSW2&THZ_3nry81OXYG8W ziRI4*$Ixkvj1M5P?`U=vwjQ^s`qZD45#2XCx{^0~CZZ?48u6>4l}=^efbGNohEgOq zH}|l>4gQ6hVSKWN57bi1Uv}Oc?CG!uEOJQZFt3-mkG)WYz@6+cO^K}B4tbb!0BI2A z?r@>2I}sj85EpaZv8pr{pr!Rsw<1e%tgd$BzccdPHF!D)DCC~SNKAbipLd?5^N_0j zh5lDfxB6Hm)%QW2F7eguWYUdKJX=UpI6adb9&~bpLmX4!x^dfw!6IkP-qbWj0y_J# zRSpBpen!Zr)GZO~A^??_r{Tbpo|8G5$`PC!e4x?$d<>k!JMn` z@3bs!J+$@qB+3#0i*|ZVKf@ac6G|4(L4StgM|bMHX!P>>EEGU%@)SPJd!WdUwcA4? zyL_W!S~2dq3P67@&X#MN zN_sN8GUu33!<{v^9d8aVm0pZNy!UL1=cQ^j&f~{_cij+%MQ<%fXPtS`Q%P&e^xr9w zpD{0*QCu03>jBasJggv6MZFbzggAkN#nJM{}v03bL@5lR{_61b-~pt9xw{fjd@ z)s8NrD=Jtu8EuG)jPwp*JNgX7cvw7MRnnu7Z}ASyO#BdfrK`2z@X`KBMv^FsALDu; zeM77|lWBJ3Th$#dxn^b^6fCc&ODxr=1x>Rn;4M8*OIlq@b%CN}yL>N_LQv9<)6(^u zANMhC0&cfk5c;QyS{9?4)8Mm1QB3j>5QB}Gp&T$BmbOBkC@-Cz>ZefcKe$REC*T$G zlzJJ)YhA7PjA@4ok5b|}P~OW{UdV|@>Ju`ar$+DF1p58Q#Ph(-4*3H41Sg{2N7We! zk3mrs5E&$srOEkwJ@duiQ20IRy*v%sgrYDiE4IdG2`{Cepg6c{x_uW@X8EqP9gex# zexh|?k&Mi`&8i%V2~g=TQl#eY_5mni)T1`^f-5J!>(hNzg1)|M!Q)r%=V9`rC?cW{ z7Dxs$A~GAm8Nf(9pLpKDkc_?y5wb`oU!+a42zW-aLOP|7u!|2>=95ZcjxD)U2Z|k| zZ07$l#D>)#aNmNCEQg3jMD$1utJ-H^(tJ5keRd*nj&t!8bLa@e*Y~5Bng8*@Md8w& z9}K3gpeNGL?HVn=erxMTL|8zGj)gZ;4|3R# zi2$&39yvh}dlJ<~-LDT!+#o_i0y>N`VMNy^fE!W+RwzJB6!pXk8Lf}ud+SA98>J;= z15t*DNq~Q_K+Ax$i$h3ATi>8>m~etD;NBMWHDCn8)66Z*c0OvY!Ep1cOvRL8*#Y|f zvHA}Ba6%B-AQmxnE%xWm!Gr3hcEh9$dH{h)OM22Mq`~|`H-=2BVI|ZB@(iX~GTsH9 z);UYeB>)+mhQ~R(REEq_t*q8LtUL~c!&E}#(3+gIkbG(hn#N%X>=81`pFwGBc*--F)+Dfn zses&xhu7WP3-Fx+m9$cq#m&8oC@(?b3zi5i;-(^9#(Pt=U(OO9FooCw;KAY7n|48Q4!rtimVeE zM}puu$CGV?kgY_+Y3(W+g=>^ko34C!@f59*fQB7Op&2O8*aZw2*x0OG4%AHmdqJf( z{8mH8CH|BqIAZR%6kmx5GEEB(^*x+KF6+({JHYy>kBM1Y9 z%>snt=VS(Y&N{={^VLS)+LGpWb^UC-M$~;_v&y`WVi!;-CFh|c4il^3*D&U` zM8U$+TEA=IjSTrY>=YvRfjU?6#$FfZ%!${i5qQDcMU0GSyqfOr5Gs{Cp3X&~zQu{R zP3I0+HH8h|c+OhN@ng$(Er`7YEe+JacBp@EVji@&`y;JTM5A%R;%r9(0r40j36`=J zv2&)ee`&F{OV{r_nO$J2nK8|Te-|dx0+RR2M%ZqMP|zdT&8@=xGE84Es4$2Od|$Du zKAe#l+Wy5MA=i_of(^#p`b-G#KsMvqrS$I(KffY9I>kJZH=G~KvZDRJLXKKm2e-$t zz;MLJZc12pRgQaX#3zh>T@v`u6P~kv^yst;p$+3rnp3_sw{5)F6|tC6wg&@_>+H1dQP4a5;X@Ch&^Ca1mLRiAj``GiQ2&H+a?G`!0V^RaMBd1 zFC?KVB%0;k3%$!qXt%95c1UYUxLQC%c=<>ltrE*>1f6QpT_@vBVGf)D8h{)&f@%oH zJQi3E;1i_zz_AJwjYXU5gz4tTLos<_?~u*oy&y%q_XE`g!-vNYznPFYs~|h ztD$mGGZV(f#*|jT13V;w1&^aR@?|dA{&jmfJ3GdJz)%F00*Dso)lP(ARO8LPJ;cHO zaD)Z`8%u8?ubiB}Rv#+&q2zC$A3V|xddi{|h%MCs>B3+geiY7Zz=KIj{>53|h{?B1 zzAnwCVSG|{DLyVPy8_JjuJA8*aqn$jc33}3icBg3$1#GSp%v(@5%PBhDyc6E9i2_p z{?D&rr~yFceYCf}=EeugjsrF&sHUVOZ$1mo6}+wH0ogl=xEKHP`2Mo-W)?ZW3AXTE z0I1X-1-*vQE9&ts7@D>B0{`I<_^cg%-aN!kZ^1u9XbW8@e)Upf~H9sJ% zff8mnMP(lK{QreTf2ixge9v?L9CLH_!|K>+!509p?e}mJ7mWOKj4up)C;QVTDSVbs z1?t9gQ}T!St^vS$+N zdRVzNUVFM^{`(*zC~08gws6BalMo3HP7#N9M;Y6#Pe?-p6Izm zm&2Y|-Qh;Y7Q+tZjHpFX0j&_NN%rD&{MvY@!xNv zTw8aL%6%3_%7lc3d}JrOev-gMM_+iQ0L6B`FH878D|zZbtU_i5K?AHsq1B9!N_i-D2^=I<%hk^1a3fkI%5?eXjMi z^=$%>1uEmLG(k*^+jfC<#?Pne6P>Y8?>~)R)xSUzk}F-IqpGSqTBwtSl%ji_z??*d ziMv%Dc1CotOa$5&M~qGtyN#nKgUcOQQEfYhhWe|A3wJFG&jG=``LQ=lz; zN92uQ$Z6|w1|k~MZ3uABPz!GWaR*kwxM|_YQkmQ*|2t#n_dl=?*6PLVg2v9gH?hOr z_XIk#-0YQOfDyC9duW8UZFJfjI_d#<4=+sxowxtz;%11d++vMu9P&xuPu@zs2mQo- zhky`FbCL#55DUq(ojKPcqlmmg`OP1WnS^I z5sV#72$j&Dnn471+Ur8lcnyGma|xtIPy%PA2UcMPSGLGP=Xf^uXo@+6X%9V&V7?^+ z{3jRBOJW_QJ!gpRzV_o-S|?zQ6+70D!<+#UjlW-LXZ$qvuO5$%_=$kvwA(?z*n3cZ z@&Ep^J<8t9%*?q(2V`t8UK)~7Gdm7-a8b$3=Q2FxY;ic{9{wo2e6ODB+-Z4@!44c|jDbvwGc5)d;O9E9c7_dy@t z0kh>S?z)?v;yi$t-$$xJIC>qzwOBBbkpjb>bWV&Dev$7>)k-A5syxB z47LcrpL|#ii90v`1yLQVKLpLHflDrOfk@%JJiJJ<_URKCodFx8lVGB5u}5PyNhct_; z!R+G4LGG@-mit$`UUrA6m5=P0C9qmvh;<~`1s#)J8}K>R4k236$2|!%P`&We{&ppJ zexs5iy|Ea%%^`#Q&&?Gtffb(FP~iJP?WlZ*H0*O?B7@|M!dFs2q2+Y*kXjvR98SmV z>x&k%nK|jtY1kCLZz#Yv6bqaBX3ai;WQRH5$|;h?7Q%rq^$+r8ac^At@x3U#O{PMy zIphzUIm(mh98Fow8ZR@evZ;{XYYw$Mcq=#jPCfimaT0P(kqh5B^kNfA5 z!}CF~Ckz1|5G~Jm`ZNz%k-TI%PgY_>P)pY~2Vb#R<3M*> zWrkR7N*E)UEdw?dpZ!>$Pj@ew0GEXaqx@A0hNX0~!^Sw$YOD1D1ghyrZz->D0mwNA zjFHJ}seTt~q0TGHz>6u)lE7%nbFExZJq>+}~UAGJC>ihs|8Zw1db^BtRZQ z&ecFZItMAN{IzAkz{-IE25B6@^opIKorUN$SZ+dXITN@&=6R`hgVrKX`P*?2+v0gS zQAmk>`3a{=jH*D&WAIeqrcEkwH$y@NfAK`;Y*GLgbK$#9HCb2Oy4^O~9cSJodoT8Rf)?2`l!O!}S1p z!@vaxptSqvvI@QQ0j#X=a?x!mticPapO}9de>ylo0p+|l5?}$D%{}mhC?LB~FSVEg z56}5s>6S+D%nWUY(6kTj0v7)aqy^UcvO%_ ztlpS6QGm8 z!4aem`w@{u)17k;=#s@D5bV%X_k3wd%-@ZfeKmEzS@Ez6DK zF!V^#{ID&Y-~1Y^THNys3tj!WD)vAjW=>LJbvh=!uMh^sT{`I>K6r2lr6Uv!(uPHy zlT?Q)!UX{=;1=!Eij{^+Dfx2&)F_!_lGxGwi8H1Php3nsyl0@@2bu!!AhN2q52uB? zKmx;ofRCJ=7!tj62)LjZC;m?hcJ9j{(52)5tpy{I|8F!T@(lm21-m*_>)GnU13(c1 zw%thfCcLs^Q*fS)(Q@W`M$5{i0j)0pB!^ty>nhn~gj+wC9~nUsy#LD-(IVuq+XWyM zrl108yk74uj{n!T%lB}koDC(S7vF3FDGM`MN{W(x-=C`tKRNN~o@5C2rvTsb|2aG0 zOAk-dK*ck(5hVr&Zv{jQfaPj!hjZzF!+&ZpIl-7yGzUueBnWco01z-A!0Qri9zFJs z|0p}gPG_(pcvHd>aJ0rkW|in3nPw2?bGBqapiQT>%{G)P_Wu5U7_RZ91K-oyXAX(8 zutb&RLXe=V;Djvs0k(_>-Fc~5AR;sX4SquJ(@TLi$!LX(VHt@2gdiM{wgw_@b)Y=c zjy5V*MTQ2{@#*6PGhe1~LFAI2IR@ynvqyIEkbtrLh1JcA==8GOzDl0382p>P{ zP9GOp87}2T;)D>79AYM@lGd3`sH9=WHO?}mZ?|7n(pn2ug7h+V9Nlq6&1ot`Mtq^q z+|V%5IU1nfWTa^;9!#uwrm>RSPU5GwS2-M^_CbwEe()9A{HBy1e=o{_G z@G^VbvIex`1!+yR>0hT~c$$gr!~h9!U*5ke7?4s!z6;D0(i3ijMBQkXG6F~INX;ZA zxahFnKX}V+>_zLuV52bCAxD>IpLjW#JL?#oqO%n2LeH$nz>&w+^+0|L>vmGUw| z+VO_+olI|^G4}{ig`b(yj;sH%{_p&u3vHhea|H}m%hNw%h<>LN(XZ|Tc52Aj?K(lx z90e-|!87^Si4_h3O$Uebp8Y-JAiT>evuqeCIN%~%$R2Mft4h3aTtsx2MSL5%fdGja zD7i9sSceq)Y0XYi@*LvL`d3)gpXRrb8u{x=9XpA(N&TT=24h|BeyuRH zn7~mNb{g7nZMTM#hO?eDTltg%fUN}bGa&+0AbF!FXopMH&`Dth$Dw|iuwV5D^Zu!e zIhVN|241qF!mSN}&N=S`0-Gn#p}1&jY8vvA=BE4~bW}%q6rG>+(mU3)vNC$&a4}U} zDa#|D$p9i22nOpa);+$>FB#_~dwVDg+30Ua_u3%qpM{$`#Iklk6n{3z?nuKBfYrE->W=0CyyhTrnD*XLsvX zXWJvoc+ai7A0n@B%q^;u8{v#iq=BVN@o&L4iQLmWHxJmlLsEL<7blhCs-c=|V$JZ3;sKV1kFOrWU#VkLZs6(=Bl$g7o zdO6jPu#@%@f@PiQnAKVT%ATXb-T z9InXnM?AmCsLmDLzXR7S34urVxNtU1`ChKw3Vlh0;^FJ=GTSh5}D78wfn4kouGGoYzrgW>y~ga@L_*n}<|VPR4$ zk)DwQR778zv;YIi%9rz=j}mRagUokM18qfD!C&wAV}!F=lv~RGB0x>Ilk1ha)>M_0 z!bDAYpv%zci|d{WjGllVMcve;>=68zOa@XJ@F%!3NjY-TEivGO`9LyPTGvQnS=9{oiO?y|a| zLQ)13Djf?;Ie1OA;WDO{;atj;e-0rh2Mb+0%Ui+#%$6!!a^PJ6#Y)S~z_S4+#688| zLit=@4MkUSIF%H=fa9YJ#!y$>`}gk|&D2*4+`7xIDll8;xX(e0h)lzA2shPLsF|8m z0oPNVN2tVKk)il$I}kK)Gp2r6t({lfeXoI@y;f9LUX^HB?<1=0n>)_LsJ4+m^uAYg zO}Z*Ea{S15U8&43KMiQzjxoz9sX8a7N%;{LIVisZ}@W(Zo_5hVG zs%tJcdVa&w{(O04dae#PDK!WHg*p#~Qlth^ih^CB3WPO~NTc+q5@ft3sn`3`!oSd& zsjbjXck7cW0AzDF{N{Wn;@(271bo&B1WskE?FXi_;1K!MJE@_gBc0atdu>{He0zGDm914LyKO*7o!gXb9kcC+BJ(-|zFe6m4s=_!kx4|DKN&x3~G>ruk&Z(aQ^AQ(-zz37Y@3Xol1 zBv_e<*Um}3KKcy3E)}ME#X(;>cxDh=Lghlbkg zrTZNNjA(|4;PbI8mmDPEFJy~c`g0!cOuoy9DXQc&xB3a%B(;!Yl_wnDY79kl7=+~_ z=>bdk1-Q*45rUP{=TfebGpcu&#VE>r*HF7bCGv~s^(Zh8bt@GO zvmg{QS=xYKvsn5C=O26Y?21={0oUzmMQA{508V?dfxu~XB!i+kA7v`aIc{UttXGs4 z77aRNnmc$wZr#C05vFO8p(vY)AI*3F*UEP@p^r~ zqEmfR^0~68L}b)6!iZrQzzQqHlfCBaPo~qFKF0Nua)!+(#O#?8>C$j%a=m2*b73|2 zWoR=DX!No*YM8#fqvWCWrw_w}eYChme>8UpnMFU#wbL)@KfR_In9?^!m zLfe3qq~pr*x|S)9;vW>8165rKNoQBf$>%cQ4kBm9Wjz~N<(v>2kzWEU5K^Fbeh2PX zs{FWvNJI9{1uhpC8a}o~S>Zka^l7LJ4L&zmvrK&R%E?O(2nH#^h0io+tv?^+{8g8>0dEgtc+PFf{UdkoZX~yBq;bS z#=*Ir+tmL(npn15KmKT7NS$>niLboJmn)A~^v=*T_JX@WyoD?45oOrsF~Y~OvzUtZ zPvz*@&kM(Qv}|5Ur#1WHD&c?_01V8RO(N8mF-(0Vm+L;rcK-MywDQi0!5ZE!3O?$)8wP{g=zBhbO>EC)x8--H&$CbS-fN7PkKbVCvYe@EJ!lFiTg&eUQM_habvBOl$O#z3igzc4qJn?#fH3ljB_7NdR%e<))lO@Fn2dEt_Zb)vOCx#r0H@>EfY29nNp+fF%O8#F z+D6yQxb*e)XAxDy+M9zLo24=kGhE$Z322#%k4wV135(`&jcwwf# zkvg<~hcjdnE;@u;BXn$ZV=`wNpjYnJr%SpL@jgI-|MEI$I!dhvnrCU#`1gOmIp~03 zo0S2-j~k&owtN@99+P#+wK|`+r*U;udXUX}HeJwvyYeP}H9iYl?51K^%+;e1kYELe z{>Nmxo~imVnaV`S6OuF^dBJ+Y9j^&{TseduOXDDdJ>JIV>EO_M=5-SY2y2^G;mW#B zkb4+;AQr=pBE8%p;a)`D2-rL~pw3y#Yvb@s+tfTf%e3BLv-J$(yfn{AxfG>s^cx}* zYA0y!cmb6&)Y@HizZyyv9_Zt7{N;r-(7X~1RM`=NDfkSF++E$NYHejx-qRu%$)}mn zv$@3cKvwgCW+-+iS4iC{zVLY?;mJ<+72{v<&Bbw`KG`lwJrX((M;&7Lc<*+CjUOR? z6$@U&`s+fhQx(6zk8>gcFz2{+LN6Q|Eo$d8?q}ewz627ds!f^DVICujz@Bv~$MS)m z&u35d3UXt!j;0iajsxUk6m)vLh3(!S#S3fzVDneT=+l@!za`@_3iEqz1^l-AN-K!B zl%yL8^$i{~)ph^_JwMze0+REI*?O?5pW(P*rikV zTNoEHdl7ntK_!TO`(_BHj3Z3rs6XPA|Mhqm65ESu1NfQ|dj*0KP5iy0rKKr@X~xE+ zwTONsXQ>pmKQpfkR|Ys0XC`Dy?=`OX3Jm+<4?sD9zNV(h=@-dPkaYANn0oQk%$>?i z?Umh9>!scAJQN{B^m7=ZVASf@3!!5;GQ1a5QjH07Jv# zvnZq$Qs?niDarpmIYU~PhFD6wUiPP2(Ocoa>}A?@)O-79fMYFy4$q*0)}N)oItng( zG|z)#o?r5qh>kHuQv1xMTe$efum8VD!^P}CJ!nm^W)Gl;EWLd+dSgDnSEu1caB0U0P6VNESiQRO$j6OO zhsIgBpZ5Ko@e>SClT*ft25)g3LlZ?5I7>w5R&aGHEM}jqnaaPcdh(TQ!qcba75z~^ zVgTzaRml4j2ZlUs+Ur)M-p-RUT`uu!sx>D>z_e5J-5L#T9IfuRz_uN6Rx|Zi>PtlQ zj2adZjV%9$EdG!yfY?Sn@0+zm78QNFkwsQI`lWsn1%-DTi5YG&BYUTPbvNPcIQOuh z_|3(sha`(3#8-Q>otOb9(f_u306_adsfqCB*GG!#g4b!=#Q#WWYq?iC0=oLmQs{+1v$s;O-1L-wu3l)y^Ha);Cq_Tr=M*!|pGX z+@4fc9%$~f_w*tNirspz%^2u^;^p zKR_7xTY|tyWoU-#etYDY9poW9ZMt`F*iwxTLO-N$Z+v-smOkAVB1H5hCHsA1zf``; zkLDvf6j@>?g3p(ePk}4rZNaqs&hm=Q0`&jh=QTQL>){U3${H7VhKQ!c4jMv8(L`usj^(Yp=(nBnI;eC zsa2l&Y*1bE=e+&LD7u)NcrP?Vayoa3nfL;!%7F%qOBQPynEFIGHk5+^kAJJml`VJu zA;l@fE116N9R858nPZk~24)I1;k6bGe;JzO1Qsj=^t06=InMga=+#uXU9~uE0%-TL zPBWr=RLDg_1h{vkcJfYpfSb9*=$Q0Z$3H~*Ep{D!HH?a^#?KnjBj1i0p>F-OwIEMC z`^#0nux7t;=<@s&OU66edToX0f~Pq}yd@28*%6nEg(@dJ4=?xA*+oR%4Wl(6D z#Vf1FH-y$#Zb5<&qIr_r8naz~M|t1JF~HOKIUS>p4^Pd(W)w}U>x4_X3QMc72Zg5D zUmCR=&WgcCmIv|SCn8*`68}UEKkU`hUIbcIdL&nkIV)dQ!z_M(%^~LF(~I9po;SMG z83%9olHRHF(`u>?{^OD!T#h?KFOqY7{B25~{WEZI!CflOy5DU6u6c|mt31-^9=Gqu z-Xy%A4z&B=KB2LDV<|p`y|v}}5lyYWZ@SD3n(goPveQ!VUJ|Wky8bBW<@C+eqc%oz zn~^lPVdtX(i8TNU4#ft#M}52|lJPWs%_l@zE!{&rvhH#kw!;_u^_f(wxlh}8h>%#f zgFhuMi|XeP0UuBPi1y|)R<*N3<#fDRyDxLmM}kprm4><%TG`o2G*MM2%0O;G%6ZiO zY*emF(1cDzD8XH}r7hN1zIK3qzKl;!K;6^hph8F3@Id)vyI|JJ$@?-?^P0y9Wyvbn z9;llVbw?%ZDkxX14(TfF*%o%iG7i+;)J^0gfzjBHQL|f~HDM@Z;bB<8()C-cqDMY~ zyf2?bi|Pp9Ci662AD48w`t(Mr@Oe9iBcx*PG3kDNq(NIOYng zSM3)18ItS0e6-673L+^_Q;M8loXb?hem}C{@bC$uB+D1fy*9gIh5$U--_mYZi0 zgO_s-gzFtIcGjyrc<2GLEo%_%R2H>{YJ zQ1B$X%cJhQ$nzHtR%tX%jp0Sq%5F4!5IIr~U1aL6J3Biru`1T3Lj7egwAFs`AQH7) z@>!P-^mrI=DyusTO+j zL+x>&9e(Y4ES}IqsyO#-aABT9x@)Y<{OC2>+XVSzs9&FF{yrxQ;I zKDn9dir}1)iDkZn@ev8%EQR1tiyD{`ruxY^-luAPxKk4`CDw%B0vgtJ1$oyrZ6axC#JZ(4e+4G!Rcl& z4VQX%&$K^Ci$3@+C)8DQdL!4OW-W2H;VcI(7856=c6^{o&3EbZt9mZu*yP6xyRf0r z_uv)_($}CLejlG;JL%w$BG?~e_pOp9!GkGfqU<^74@55(f_3YImrgXbvgfbYf3-W- zZvE9m)pmFKmb2O!qEmca42|lGL&Ox#N6QTzbx)`)(rpWqr7#Q)m{~s!lFOF!C*)!C zfod-}oKIQyqni_sarx>pvazlGno$)0?S6KyX)~g14D>N#eS22~+@zAKHw8}F@ueFk zQ%9GSgo7kUAY^o|rOmojeKdMv+?a3BiQQN`JnaTU@np(d&y{ zTBmkS&`D=!z+2UUx9Z5jl;CER!zUqRoU=T< zwJfUV@h;E{TH5%`OgF)I*340VM$FW2XC4#MDRAguJDcKNJs$Y`r6K%A4f2gV98s5e`1o<)k6c-K0h9en&ob-0-#E_7uPsi=DsCoEA_eifsB9N=%X5Yo+Q+RU zRBwi;DVFLYpsWr$vr98a^JH2htsyHU5NY)nao`bbZ|4qNGb(pZcYb!RYk7*7!l7$C zz;J&mSm}Z{>66$y&K6qs=Y`=T(?(19o;}NCggYtrqhSe@OPn?>6eB=y-3eL{7hICC z)n~l^vX_#zp9kjrS!p&o*F@^K868XGbwct|auVoI(nUFy31?FVD-`W9l)r=aG*N{6 z$V(juvHn%=0w-H{R%+5KI0B+5xQfx7gQKT#N1H@BAXk_>twC>*S<(I>MP_MD{lZ=6 zR@JpC?|@vZ{8ICw3P%ZlTWf6#!w2m#2@xb`U@xirJ1xc6-FoABsX$Nuc;xZsI-6o2 zFE1}cD|PvAhiEuYKA_>I2pZ0Mdq)A6VJ1SFyR71Gxb;*{n$UQ0vlM6G%ZGjTJuQ*l z=-c+4E9m@oSWEgl^Q)>dj5W|Ps#CObDF-dg&5cJbD~n(#?LqA0ic6P)kLOf@IB@4g znonMflh&`l8Ait9DzxYfFRLF-@-1yk^|L^u&mj-FwT>9x2%lcYxp(#SNLD5~&cj-& zGDpnO7?{~}3!vnlC63uilr8-;|C`B(JtZwwU@qrfzJ^U$ngA2`9d#5KvW!<1Zg!`v zPiH2?C$D|db#IyuyO(%;;FSYu%mxoC))v;lc_3|Q#8d33g3rTN4buSXs#^x!<%Z*` zM{xkeNQ7ywY}dB;SzTD)pX8SDuU7n0*`qSMd2PjBb>R_VBYR8xR%J(5U-IT(oh+86 zVD$bcPYW^cUVHUOI!0E0mUG5Ntz0mtt{CBdiu9ef5r~ef?xLLmdr*H0BcZg;F*vTw&Sq)OI5PAeum z^1zQ(OOKeFY@uxIo@Q^}tp=aQ2-k_Ba!A+mR?IDNP6yu<#I;e#9xCPM>-1i~bYZ1Y z2ijg4hfK9_K|9>`*bBnGsOH34>r`Ld0Hd8yB!y7<*L$1=LJ6OeW(=!-e;s5LVhfyk z@TW?A-Y{86Z8%TgBC#u$8^wPD*61PwlzEaBMdF@GL^;%X>=~oQRq?yPGY!2#baTjn zPNHqXqSdS0^Nl-HGhg)=u~YODQRgevX-y4M#X^uuiF3RR&U9q!Jl%-)+Y?T0rM!z* zMQCuUvHmzLI#}I~uhdXx^cqPp|CH#hYhM+Tf0Uf0=dg^Y8*-7s8%n81EVMAA2mK@Ysw(p2Y3}9xaX{*6-@3@i7Fjg(HM!(-@`Y)m863}A>;F7$ zN^=}rJCFG+$IOqq1Z-J-)JH%^hrN!*oc&t*(%?2D6soJ2Ebs2$@RCLo~#a5fHt4ohUSu+mltl% z>D5Q27d_L-PZhaNU0{pW-JaPw;w3j{(PVhb+2l!5*Tj%S(MZZM>pQ_0@Mni95p|E=FBb1pGR{QW=ZmA$nv`vcg0LHqgIXk=RPwGZw7VJCJCaM-tKTEc z)Q5XX$PDKDHw(L4e$1-$)(=BJ@FZhF^YT}(zt9NP$sZqM}Dyo7{16|y0m7bmHy zO3o_U=RUX_%gXler%8v&aiNrK;wk2fcora$#mGusB5>wh4Y7Ur#;ZtRov`2y(g3Zk z{7AC?geTWHM7@$MzQ<6~B>m7TIYT+%WT`>d*?xE9PejBbh>IA0dg)|gzlT4Lo;W6S zm2;AE81oM5Lc}qMDq##?`t@%AWzQp*uRPW1sOl(YCR8(6e<_)%+j03C)A@K{?D3{K;|;)ljo0~C+DwR^ z(!KZQN47JCu=XaQ|JJR6VouBW5;ePvw>9W!{h8obZQf5-w6JCwZwJRm5j5vJ%Ip|> z$7STet=B;eu`t~j;{+CJvujiJ#l43pWgYkT{oiq@b-a3t8o;(K{~B0ux!+m6$*$d2}Gb4iVTXOS1T=Ml3#N4dc@E9V18}x zdaSCSTYtFEhC~#sqWA8twz$ z3sU4m87J;UaxYjU+_s3!3+d5Cg$%6+0so^}V+z-=8R;JbyL)|+y(3Ax;3a$7#fqsr z>RNpfEnX{MXKmQRml%ZI^b9r&zYpg;N{q#9{P6^T&Nl49DkEf;BaDr#ZmNHDIstnHmIbB_mzrNB2EO(8fqFyJy`RHUm&AP}M^wl)Pc&ixBG{`Xs5 zGT&5(!cL4bXo&GF(!TdG?5ksWU(MM#%_ggbOHM(`3k7xyioY&7tN_d*7og zsh}?8YIJ?)Rr!UfG{YJvb+FESX(LJ?G2T%wrF3vCYLjw;fE`6+FVHMkA~t$HUbuB6 z>zdv9^1qzlrq5V-gsQ~e>V5JySV$x1o)P6harus64pBe~n~yv@DyI_8XA=Dz&S`c( zyt-A=(P0w9|-y~(fZ8i_oGsnDBu_9Lk{mYlOrI!!{_=T57w+xz%=>9R5=;pZEyaH!;6we?6%}2KgNJ)` zPH%Tdf_%HdRDjEzmmYN2)_o7q=)!>mq9E10#_^0=4tO$guwL8c8n!zk(h*Vq;L4W`FX8XsU zmvYQcVam+kuO4cJfPWTgfV?jx(`kiP6PfiFH5eI}nC$&4Ap|J>{_oBWtD55NmY$Qe zr*j;(AZ3Q18h-POjF)~t;TGsxo%hu2sCAu?I;Cslti1Ae9*6 zdLJqG|5D3jersf8mgK)+gSC=k_4A)d=`XBx;sozi2<*3ep_?{o?QnsUPUx4rF z_H<+5$5po;lYmt3*MeDiN!k9MhD7Pop9@OX$#$;_$N6YdCSe3p-sL9pCwfDGB700WCwqKvmmGa(q4en zbB>w~&G;KVkStgNJRnDvPtSe~S4eg_4O=1}+5uwj z|7{f;L~^Jo#N5xWHo@JCo9=d8qNj3&{9^oDf<-l)Tzftzhy^bhZ#!%%bo9zg)5pCY z!+XG`DMuK`Io0THXWS+TQ-|SSCfDem0j)gJY7id@K2Ni|U)WRmLYZ`2!x8$TPCy3G zR_c~n1IcqR>gOClcA8a9iPzQMq37oc!_*~iCCC11Qpf?dvBg>Ic&jcvUL>ta^jpJ7 ziI)DLbFjmv_Z$Xc3x)Q-pw$cbBNWzaR=s959}WJs9D7yoO!lVMUaOZJ7NtTs+~{#i z^G!r^d;k|{JP_vN!wyAZS-};`RRoLcS z!njKqfs{$XSDLZ#9q}%34{WX(RHO^XQh=@HMTF`jxzya139=fDOe^~#84%6Nb$Qi- z=*&bT#^gkyyNr*r!me+6(Eg}fhtCKlD=YtA-}C1rT{vIos!qoHm7+<-{-6JF*M9Inf2##)fuMbV*2L#Q4>Y!?`iPWZ*RR9-y?i06$?%9bzhRNJL@->0m9%8mBjc-FN%NE`m1QvR*#qUf^W{&o+>z?3 zpWAwWDrjtJuOvAZ$0V$oxY8hnbDkTggk9ANk$$pB%5^ZNw%%5G6??KYBs+xyOSlyfN3$!)H)bI{8NU#mUS8fmCVxNxtHgR zR#!^uS|=onY(6zZiQheSq?`Q_3?Leq){BUS1Fd|z+cTP-u4-kmIT7(!ybzi0#)AhW ztF=(AIK36xA_otN$CnJCQE>a7$tUe|TX0x_!RiFYm%HXW%oo)s6h%i)3vd18tLas4nJ=_U$~0=pe=w4gnr%!MPh=qe*CYefYJm$KaabNsw>TK+DzW~nK9P774;zWpm zdfpWptE|pfuO!gAbZN=LdK?@Bd5Vf!crPs0Yv$ukItQcKOViFiwWsM$&51M09t}I) zAk$TzUq0dD&9o`J>Fu>4H&EqezdzkzWQ~!<^}Z@G&~sZF|6jeGc{r5++xI0)QBjCw zi*H%7FNGK-RLCCLMTjhAUuFmiC0dv)*^{mK+IJEnd$KmNGqx--cDm0`zvI5|=bz`# z=jk8EQInauuIsa%=llJ7E9EX~X{P>tc=Wi>oSfJ2^1BB;NpavH7*=|Gq)~xNz*)#X z&OmY+-V)o7|D9oAi3ewTEbu@kr=V-a8uW1PAH6qmy2P^Lc70mRi>`YknLk4<9`1e8 zPWR*3(Ri0-vn+Goqb2r8Q~T(hlrt{M1ljC1dYhjs(Of7_Z)DS*$`veQE&!h-P#8ce zbtD#Ik$!_it;N*y=Tz^n*o1UBupGANSgviLy(K>F>*>NNurZ#-@gsstN_ly~a66fV z)+qJoC1cvSI)4*St=|c|_-sbzESl8FGEr^&Tb2@l3DWMOiUEZt9_WczDTAQC5(OHb ztN?ywo&Z#W!D2a)V-^%d0A1}kPmth|{wx^|N7CfqfL+WkD}VBt_|3M^6{3Sdul#Y! z_$cy_W40Jzr_X-(>O51-dYv4zsS9s=?-2& z>s7?Z2UglEWYRMmG_wn65 zuomP>{3XdR_FXwUy(3K9^s4Y(MdVtYq+0O6^;5&eFNfUjr1oT)l(?oN5awy%W{FvD zgX8H06VuecEM8!O(@TDGZW@Zh*P+4P;^y0U<#JX~OH`CAW%Wa2>KHf~b%R+v?=CO1 z{MyroCOAZ#0x(cmx)=6)+?4g*&U4~3f!hvLunGzAz^4z~dppqgV&&@P&;YSqiWnOw z&B&9UJ)I1MYlW3k6fapm1lh6PJpaHc|N3-I(22H)kJ~p>bMUX36^?sWOyxUG~J&haj2&=otFKR}~XYA>^L_()VMqxjnP z!C)^9-0_h#xeSB>S2!m6ayIBKTNCMsyCG*6&&O65-;OyGpVVVEq7z$_G9+qj(r@gx zY|S-qjC`HKxwj>{>8ecbxXI?@85>s&IWvJ68S4ym<50dBFmi@XU%Jb#2|*P|7GA^= zz`d=Ss$d1MiB_yH%&JOtZ6XKNkA#*Z?_I4}YlM6hqvGLYIAOEkc;TJ@=Oru<`&8iv zS_21_i$aiL0cbJ{SI(o#-SJZQ&w-&@)cWsh8M%V6$eq5W3GNxj-0?b6eAyNG+IO3jPk&e;Df(YYmB30MI zufd6#hpsEwxJY6Yv8R7UzQuKJev?Cp-tr1jQ6Xge6%@F2_WltoPAvq`Iv@L~8$Fdu zV{6?`=puxYOa3Q0Ze*-92xvAix-Y`&2^NkY$a~DIwA0QC7 zu8=Q`tRraIx(*dW#Hr#fxZfZe~=};*FN~lqd`s7 zO37@JocOF#E1Pzy6s7ACpbig3lu&VbD(9<)5^sZGDz)p$d|3q85Iw>rXeRE0B+uR# z^l|#crk9AT1Q!3dPSkk^&FK7#fB=N*aRN;JK%F2NP9Ll4FL5401T=kOA)HfOzzEK? z!y&cGg1usIpS=r0!dGwK3PP@6G?>UMP4i9>21*>`fmx)PnwFLh{5wQv4cTZUE&%fH z>CiAFIrU1m?NhzZclcRxi<^fB(KhV6nLODQ{7e#bqvStmr34n)eOG}-n*RHy(OZ1F zN*y>D?s!mbCr7=02rQOT1hcU&L02rk`}=nxpraH#4h|laO9FYLF#sp$R(|b)rT@Z4 zVJ$?F4U9WObd1vtRn-D-4zSeW0bK&9>s7TPp{oAe>UASED9AgOcp^r?B`W9EuV3@S zPhaeSI1#&5tWl|xBX|~Wg_T5u55Rjoj8_n*g`&q#p0E-s0RNbPu>s?qa?Le})Dql> z$xaFn32vR!-tc_iMl-A$c<)*XGN9_U22~9n3OFZaMc}&?R{;u_LxIYC1uDeGmX?+d ziJI_51R=^s=$i*}EtRizs|k&bjRkkuL2|heposa_VnFTBw8hbQ7F;ajn#f>^X7F~` zPa7>lr9+m8Y4xoH1JWEdKzGA936k$?tsHpFTn@5u zJRB;FL5^otwd`E8^}eapc3sSi0EEzR1y-z=DKK*J0M5#^K9mY4-MMeSpjA$EndwRi z7DnWhM&ugPr)2Ze*z9f$=?Ag7bCm*z??V0wM`sAz`$C;U@_0%k+K!=eJ`JN>)(I;E znTU-pkZF-KO4$nf?z@9ExXM-n`1{^G5Oc015CEqN)=Ae`BSB9 z3k*nQRUeT@scFoWf$+J!ZpZA;JgDhE0+Y2hHRo9IK4>n zflCayV-GY`$s?l+{p(TEU7jW|Bg6YYKyN9S`~f_B=u#@eX{U$Q7gi@@pxi#6LitI5 zdzP*6kjeiY)bR*ZjwQE;8T z=e+i-NB3ZtfCp+3%M!lqs4+++SSwn5@Vq~ZC-Nlb*Tn&M7TP$C6LWV{Z zSr`seTo;_trt5OU&VvRW@KD}o_p2GDK31{;(rTYI7VLq8>b(1qkTcPL zI<9;ngU)arVT)}M5qy_gug{I9bH}`@;d13Ws8c}$Pa{|v`2a{*J zIQ&+|%Qa#|>we)FnJw^*Vo_)J&M3G{HY?6UWI<};c_T)RM7i0%^&p{7N_Z;Mv&&dE zsp4VSD%k%_@EkUr!<~6_P30(|(CM6Ts}&!uNX=%dtl3?sUuQ(WbuzVoOedirD903K zZmomr6n(NYJ21J?e!Ql_HV-66hn9V(@~M->@@@~8`^2J{BDg&*5vpaopjG&FSy`Fl zygRQiPC?V*@AtvIg|?w!Q7+8V**vR&gVN*0YC6*ez1Pn*$&$^s>Y}4^+l&Xt-_4^Q zGu`Qbm#p(kbmr1j5=^AT&!b{TuyCx_USij*_cjH`6v|xu4XpC70IzBou^>|g2MDWM zKTt0sYZMf(UVVOm5sp@CEn;3f@=uBWu$Z#?$EV){G=L)U`!GD#J-%yT#5V}i+G|JD z5l|oNz@#=YW{x*UZCc-V&XY8r53dtI9*#nmzeEPmXhe)?AQp>#>VA3Kh?* z-8Vwd^%i{7TZXLW#JQb=u?cHyZXrTaxr;=Y#dHkYN{4+~r`0MxM5KI(>K$S8d)$d2 zbj^q#%sxz=?zzm-Naq$6u6SV;M6-A#(tl+V;kUa(*i!m>u8Q|X&nMi)L&|ri!T_%H{pWg-IGqXPoB7$`VV>C%DD!pJczTWrP6>E6j^e%GmY^NUb zqJNayNJKSG1?*n{vTO4_?p6=Te0ZuLx>Nbp2rw_!u#RL;9dI`)EGU2$bH2bQrH^9; z(p1itk7kw1U8ch6L~NDq>UtHs7ypi`b9*M@;^Gc9%|CINMJKwwf4he6jT{M>rr9l^ zz?SbEH23n~{45r8Rsd2j=r^WUkxpoBX~g{V=g%~x^NC_sPqsHez&FL*UY=W{@L|cg zdz=L3c>Tcwpp~=$cA{mMr6n%o&-9C8rPrSD1R`MjF^MT@DmVTDIyr4dwn%FKS%TTn zmJ(NJTgoM(TSMsa?66(95PI!XEOd`J!bH`1ujJOcu3GyY30UXA!PT6sn9tSSfyRBE){U`ia&c|P8E5@|?_)9G zAxJBDG7~Y?A}FV9hsjDgdc=s4#4Q8S#bnQ~cM<5whSduB~62WzY1=|Y{)$mys?Xn*~LXY#X^?6)% zRDRnDgY-w#uM~U=Pt23~)Mvsvk#!4~_sMicEmaP&t-@^R*axZeSDD6|D$c6m8RL#p z5__VvWS->&%`S>H6^cTI@+aZQHE#{u>N(v$a^A02)pWd(iv!g(gR!^U8aj+lAOXIi5*OX^CiDBNQ zv}vysoA9k?tua{M3v^%Cl}NiM+rc|`i{Q^Oz`R|W>ltQ<`?ne#fE_wNtd>Cq`7GJC z^$M3KSYL7s84`Ww_^kv!L$8wAM;%aRz37M(TOjYrF_E);CL`!2>|+g)UB?h5sHp+b zqE{Sy2W|F`+0G-}JwMEshiQ6%2Ud?S9q}_}I^hSLfLoUXnvyNNZV=u9o|BbLipd*d zJ$K+ba0mis0gBbL|0R;s8HF9>a~cHsn?K5ua{o*b&#^D-+wGB<=T>JYIxhzuyb`jvdL?<1@3gkwqp;Dg z!L4u1*qP{c7VIe-Ui)vU$|=M^I{&+&3rx%wbDb2n%)9gjHtcOb%)3$pM6mur8>}{$ z3;Z#Mu8W#i%G(||7lJ$niKA{wzuJo-g`kAM;ReYRZbl@$Q~-?Ihf(boN``l-G& zujP7-=h?SAr%1=e?DK&?cx*>e4j5(yfE6wWd(t~7*ryMh*}-!udU*`)iY{rl%UOh< z|AxDuXAeJnFxDKqqTHKr82MYcecb~3sP=t_drm6V%C3dg-G^PZLDuin`-jdd!hSd7 z>4UTp1}X(a+~Gk*v_SLw*%0&GQ8R#xr^`~Ahq!Qi3G>wj z{s?7Rc;vzZ>%n<9^R-wX0F0g$7uSL9Do;kz;SLLnFyVV2QwsmUILOKU=1fmo>B=}Z zrLk+rm?gCq_GN@Wk%8)bvi>*#``Gr|??ef0lPo1A7Sm86?7e!d&O(h!B!>{3a|CGvI?bDZ;wUwEN_G#xV*j>P`fpK zx|j3h)YIka$G9|da;iJqJi9`0Tr{rMgxBon!X3u^|f5F<4_;v9m#m-=aAf#~9Yis*MG9SAO@HFQG5h@*v@^Ki6THqft!#N? z?#GWJrZ(6wINVNgOacH2nrxAAsiCp)hJ(Y0?G4x{I!ka(AP_a-Hs~HIOrc;oQ9dS9 zX4#)>>Vl%z$qk~_|(y|ywJ*XgXD(7=(}@G zFQJ58@HY2NO_;$!1MQa}c!HY== z364lB2-92l!23#KI-l(F)lchSM9bOWSl1oOe=|vPV}x$3l7%58WVXT+Jo|ekQP#cX z1fb|j+))(-0D)O}`K`&tieL%N%&pL#VE?#})Yh8xbQ&MnY~+}lvZ(t)?> zrPkaStOO8V?=~D5WU)g~H9;5Yj=v-e23z(xB!t;W!LlJ#v9(Kt0kgNJOJXsBQ-_|p zv`JT*}tb~Q6Ivn=!8fP>Vk>gwTgQ0n+wkp=;gp15-kf}`*%EDFmX)Xpl}ELr5h%XGOYeMfEq&rh zzOMom%^xQm-!%6_F*=pIdT*W3UFqp;2&h|Bl|8!tFVP`c1q#XR!u`5bqD3JRbN2L`+k@<%UO4?{(V?p>TY)X1J+eG`H0=MVCt<2*=wh^* zIK^aJJs@i-WS(6J?s#g2AAr@LlQj-D_G*teUZF?SMJ7Jbpx?ALkZGjHOc7<~Gm{e1GjiodkkKZz zg&Ye?mVrT;H4jA~(dKuLKp4i?JUhE~QpD^7x$3;gmAjtwp@0f8qh#15*U;27aNp^o zK4u=bczHykM^EWz14Myj#)S_oLaL9Rg5pm>yN@tO&9(Z#Qqx^%t9dJfA+`f3ClCY? zImhb*NjupXkBCH0dymxGq#P(H@mhTe*Cr@)Ub*Y~G8p~)T`1fMH22lzS2j(DULq(8 zX;WO5*649p>U!M7)mr95vS(tc-kOeo|CU)AdLg}fuQRP=p{Q}o{6v$*$PtHWZ~l+3 zsuCkA90ZL^)O-68)OpB*;3!P}_@2?V5zXCpM53EBNf~65x(pb_oW-Ep9H3QJYmuo= z_E*Wb03e?>Efn%N6Z-1dBy7U#w{ehwsUq|la-I6e_z)y_WY!PT1Mee>dA;KZxL@YB zznpe1@@|7ph7hP=U5r}k-=OHZX)(SGJ-zB{wrgV2mFQEXV=mb86#pE==eUcS%8PVF z+>8fmpSJL!e%h0Q54;pS6VJM~O}p)>m1@97cH9szXlRKTK-zhovY(Vrr2xys5b4wB$n>*xlzVfPVp)fpev&b0 zxv`rM42ZP!xzEcbV} zviq!|h@aCSPoh1xFCgW)9NT;$+~7I&NaQCN%8?v2=g_LSiVXLX*#>@bEo@);VW%S4o(ceY4KM2`_a_l!~o!L_O>q{T#3oD#QY(A}OKU!w2* zsP$!)=$>`mt&g3Tc!%+#^>Ffdu!M%ty@2AR9*d?b-u2F<*>9sw#$LBFY zCh=v`t|p_aJ)5|+H(yV>%yXS&BR6b3*cux@@Ci{l0Mwz=`k5uJ*|a3x6KYS1%fNEH53~eKTo>i-i?cW53Dg zYWr2@`juthsSpp3T2D$s0y6=VN_b8VLxX|kQ$|MV-lSLR>5FwInndK}EahfngjTIF z)@09~J(GI#CYaB7{;H+ndPqN zEX527(;MD7H)Xl@UC7D(6ZGoBIgd-ZdaHVuW5jtYm= zc)NLG>dV6?A9kjz3FovDm)&o#Pi^MhPo)o6da~Tk4>o@LP4x-g7HpEOF@Efc&pPQl z9@E*~lC#dWu;!lT6}$3?9;Q!md#WIC9Azqii-{SDhE1CDnV*jjCf6Ro654bYLU;P* zw3mRCj>mjlH@NX+XU5FF;ic_L4)gao-2J4murTXMFwll;{b`RMJxVK78Ou==ec$^1 zg<+s2eefIQpSlwzhI!%#;I=nz8yXrm21$_jF^O0^IwR=~+k#0pDh%3jOlE5A8&><0 z1vDw0mmV$NS{F;p%a1N&)7qdzjsLcn!^%y2XYrrf{M@}I(s4`MX%C|+Dn!1xQ!NFKB-7m`APG=lV z1?yg}FsbG}GiVErSReY)oh;z0vM0YJXIZ<=W1nj^lBfEgJN*0Gf8@GY?>F(++Mfwl z4e#pv%ti|yIUY}%6?HCkMygT^x}ANBcF90p*z0|@n=^WVibJ1>noR|oDW7u`D&Qk z<~Z~uQbkPx=&U3c=jQ|+`@aKZWMl$sWbjByO>G+pZMFwq-lu1*>e9u~K_ZXi!2R&? zV;);2qgIVrUKm}~0v-_&{olWTm451r*S->8uzz1;CoT8;onXU<)|r^}@ltq^b?Yy@bCLNKdEL@N( zqXmj)5;GbiU3M^C`AP{-kJt7|1{*Xx9*R?zp6~a$QIfy1n6av;V|8uR5vS-5hXojr z#02e%efW?$vMm}k7eOmIJZJowNwcCLDk^JNNc+iw|4+d!(O;)(u1Vvo1?Dyno6pFPH#cD_AIgW6(uaqE@~7W?^@)xG{S(YC%9v$BV$c$ZWr zl%vEl>v1<{i(%chyAAm_4X3lNs+iU1rx%^{DN&MC?p$-1Uw8MwFV8?cQ=gj}$ye{j zWUG5y62kFTuut%^;9e~q1A|nhfbDV@LHV2MhKmj6AG`Y)1WY*=-#Lv!^?KsD{RG`F zh_9}%L#CYe7sMCqz;frEc0Zl*3nXAB{c(AAM$BiormInA95G`5C!V{zzrsS(r>LlC z-&sR4h!BEhRk!a)BPiJLtKp}E>-F(iNVlA{G>6+^53gT>Wlbj@h|g9HU@;R@dYtL~ z&XSE~v3wHab3R#aW+Wvg#m;}Rmf_bG%}kbZ+b@bcQ4SFwFw^7?vCXZ z6k=id^9jV6T$7kTaU7yO>iKGr_Z8iPl;o4MX5e+ro}>5NKh7=l#Ah{fnjHJV3G67;;-Rc<&bu<>8>b7gE%ge zZ+}6sG4YH#CgPP|bM1T`j#L-1c-3YJPTHhsVTxEN=J3B*HRi4^`dGufVFCbDI0P3u00h z#YR=^u{&@Z((ivPnKh+ zrKPp4@|p~`fXHJshHGejw|a7PuRs>BA{hL=_IQsivAat>pLZF{Pc7wyrR*22rl z!+qychjecwv_rei(Vn;D3PRU*4Sl@Y84l@RZWsMCijm5yit5W6B2S!_A|!+RP7e>2 zFJHfEn7@T`Fq!M%g8+0)GKkE9@^su`=-TVo23S0%)v)5P?jK!V_(T#5Xvfyi6{Uw7 zj`M;v#kGTOHiueubFpcGF)2$L+}_ectP@m$`)PWB^0zbwvi=nuwQQVf<&uV$_VQn1W7+=hPQzJC3Bn2RfCqWRHgDSj|*D$g~`Vm+59uF&n(A##(` z^~EtZ2v{#HX+1W}>`gC+VT;iMO;Bx2`S7y#S+_$uQS13<3YU}}YWz^gzp{WEbO zLz2j;@Z^3rD#qDkQ@Bp+#NFIaet!icM@nf0QI6C{p$vKQQ1som?GpPr{EC}sW_41} zs{ToDDwCcVi01G^v}~F<4l8Vq88)b{=h1*<+gNyHWJZIIlNBcBg3ROK+w;|wwqt|F z6d_MR9i)$w=?08QkWMqU?cEQGglyX8T2#m{A?)|A;ug_sirp2b$1+xp9?YeI%uRNm z-z^3}0?I`@6#O%aUuLbVvQsIrAYyN)vciqy8uM$92^*H**69F*s@<1^0sz3#ado-d zh#iX6CRZst<$&@v#`s=|NxvF7G;L52UFiBK@}ka+UI>~UPicUO*=B7vS(e+1##e_1 zvQ7RU$uLT8`ODKePizp{IV-*;$ct{4I$*W0P7HQ(7OE`VRIes?^HS`p>cPijz}}DX zw#J@)Wj5!o3}a^onRdcT zp}O)gF2cQKV?lvO=%voVT;ClUM>r14l9>aYZdHExp#K zDl9nBTw+edt#!hZQ)o~YR`AT@l2!g@jtm}@BQ@2vmOd-Whc-FswZDMuwG$x8mQfd2 zFSen%R=co=?xt5Hr+D5>EbuS3g_JT9w;Qeh`OZlg>s5yP$7j^$aFH%iz;!eZ$7#3z zbS1$LpP4^&7)x2|aIm6N>9fdN}ZAJ%n%RYx_F5>op%#)h_1wHJ8m*p{}vX`ik zunt=m^O=Z<2sD8-@pn@WOA~}6q((QiD&NL-f_FdY^%Q3zQ@1%>oLNx;X$7KLnM}%uPNa zn2->R>{rYy4e;ACo3$>_V+=KAagCva1?NfTJ)nl zt%kk!zy1>dS#~qfTY`I7f`tp;vT$&mr`;1iZMeOe# zEsO?UfG_|3tbHDnnkT9^f_kJfXd#a3>xWGSOZ1z)0JvDub9b^lU!^j?>P=;-gLdouo}tZ`zox(N0hMq=NZRGlE)a57OtS8f%h6&z@)Wn^WEveu@+agnw6H|8?*Jb zt?jZ0E=LzwY5HHv`@+>MkfUD}46JZmCq^&X6G(!=H}~OjngBIqq}3z$w!i2XQDE74?VuV)L8U#tN4GitGsER-r*@as!OM*t9}SCG zmt0JmnabBjYtr>Il|a-{38A@$>x|2gfo2ftn|n=MshRIgEY+>Q<8Ww=n+<8lo=lcS zoOAg`+Yf8|oY3Z;Rv35^sKpQ3O+f@2p+z$z4UrW)m}W&FK1dez;UVLil?^)ky7!hj zzY#K`bAI2d*ZmWVR)p}`@WMBji5g6cMi@gud>*i= zzSe6N(+kSo{4-+d7uv`#=q3|o`5o4I*zD^6~f zN`A&}KlXe9o0!L_jQZb11e(Q0|^q+?0uW%{$C_GlqdI`Kuv=2?0oEurMQtZw z3(mM#^?D_stSY=`Eo)zz0!eMrrGTN^XJeBCLzK4lZED`uC4 z0Q%v54DPnun@cPE!K-#E&vCB8dmDK^w1own#B9&Cc1Da_w*^qqxCiI`Ng?|u%~=XV zO=11|N1H{BUANbBx4GSRXZL>Lm7M&dQ%yYk6()3}p)35{qWGcugO=k-wcry~jba}W zE)25}`(X!y*mVcWR}Oy`F7A1y6Heogl&HeWEF@(a&^n5F;v!CxHC#Qc?{=ZwDXL&b z4BEqcTlBIr2QFIYoSE&HnY`q$p35iZpwWrsVR%~yQ-_SoXPV4^n0{iC@`U`bKxBiE zbXcYa2PE{!*=E6@7+C+nV>oI%bmyrriaJwSpMB}ao?+D_idPOf`3x(Ni4&&nUW9QOZpJBrJL&CQ5u4H{#N};yJzZ*JfY+x)sbiZNR zIO2NV2Y2|k$5u^FbJBcbbyVVT2<9r&9Am*4s67?utJ%77MmZO=q3jUrDeH;3$0$EWv!j>FUoF0RM_fk;>A4^43>OKC zq@!~dD7+$JTEEO}|JnW+Wl@6%8F#79^n6iD&syAN3muY$oQpbYizDB>QxBziGy`)5~-N z5}3{Y_clXU*Gh`>@;$)PEN^}_bh#2P&Nwx5Y;!DOe;^!Qh`uIiu)aa<7c&cTUh8V? z(s5O@QAhl@I@bD7fl=5etA_KB21udu8RKDqA$p{qxs5J6$D#R`(_0W*3R29a^72F= z@I@4gyi_@}+=DHa`Z;KCZ;H^|hS4MX^A73u%WIIyRx#taVboV$%tbl#B=l~eOky78 zi5~sxp6`{IzYisKoE6gF4Kl47P6GkbW zpz4HF&bs8g*nZ*(Lh&Rg4mSftmYqTLL$}VJ!<@Y)&TB7rfvIMZ(Id#2KmWyaUF>iDvG-Pz#zg| za244ee|dxKT%yTkfi335p9@Gowuw-}t1jNwI#}3KH!XA8K}GqHSbr{j4&t0bP@<@) z&{8((RRPV4ECv?~=95QET&en~-XE6I7C683rH&q)TpjjJx!GrDPGz+>G~J_w?^dA% zj`?yuG{0Xclcq}I&QWZs#o8M1a~4pv=`hz~%es%QSG%6K$|95=x}Ug_FDaAKi-lHt zlk#VY>WoH_bHCM{4Bp63may%?inT3tbZd**!V|96f20uBumxVa$kb3#z@HV#hV903 z1EAi>_mNtG@Dp+o>YGBe%X^myo-*&4guttenFERzzp_FU^WtdlH*UUsepOl?4vgz3 z)jru(8<|zkqOM+wq9%1}E62U0ml5hAed%fuDC&OwRQFTF>--p%v+eXgacUoVUx!O88Or4`8{V+|RV8e}X|D-=#0cqBYUsD4J>XH> zn6*~a8}?k1T{akx&j+F{#UBfNQCls~A=|KtA9LV=U|4m(kby~-SC#QCier0*M68s# z)cVf2oNSr6;Yhq>Hub<_(1mHL+X5;S{b4b)OPht307CQ>#)4ie>5A@AcV(3%+u(0GAtxW#{X_`ZivwS3 z&WgedSvHtA1(1L63}*8*r=MdN5z%@Qs_9f#70r;$b*AegabZ3GVDdd@H#p*Ik5__b zc=I@_&1emv0ky+mQtqw|{>cY!F@DXgotq`gmxg^m|CALv0d6>?8=$77v}fe+jK~cr zG%mH?kuG(T4o9?0Rga$B{>Bw5*bvtwyDyf{8Ok^2+a#4RY5XI`5~5kbQr310N;j)S zA_GYceZF+KiN)nYznQ!52njC*QuozISdmsf=WJeh{LNXSaUVoF$@&m>na%z+QH`?! z@I8d+Tg?o4aW})t3M5BuG1izy(!c$2)tx_9nS1q87&D6f;CXhM(eMoB#U55p8d;@c zVy56?RQ4wI4K1vVCF%V8Yi2Ci_fOnNkoU&ozb1qZ*sLd>bc?{{;iKHTjj4tCC;VGN zTH?PNM7xo?gmEbO$Mw6>V1$a6m=uhA{KJCcLEh^mTrhjZspE_=L8bGswa?z$Vw+2V z^B-2lvXgJldVa{S$b!C$3G)xq@2hwZvve}Es11>0k0MyTd!wbwfBC9s>ZdB>M+;`6QHPk$ zHMqCMd{R+nge!hc!sc${D8`m}BPRWal-o4Njwo}T#usB@iPLz&nLHJZ8~y=}vHo*IRoJ^}+gxx@k|WJ3b+jbwk?PcVA0={rNLeZ7dMkSU1OtV&me7 z1A>B-Gwkfw2gPA!#(LH_0Bd&xQSPI?GqE2LZ7M>?eB`Xf?s2xKNAlvs*R7da-nm8M zS%;geb0+pmaZAfWgl1G?x2^(e>er<#t*J8=yYA<$@t>mJK0cWIjp{|(gMd>Hg8WlyDT(gb3MM>B=CigCvJjC| zK+m#(;Hm6&2B=@!qJ|68i0*^K!${}Dl}>ElX==mnm>yni6Xxs7(^wyL+^(X=8zx$b zV1_+6lm)i>qi>-_p$JX4&TGnH9`Gd<9A`Wh*EqPiv18hHIH554l%A=0%1qdbdUS3`{2TRfcAtBK+ z(M-|C?l7Z%AV+8CZ9f|>2`*pq)sn9j!IQL#RnNq-| z1CWiXrh&OY%Tr(eP8)s$5W_N{E!hStfCBROpG``VWgK55Q*rwTGie;s%QC})FcdhNKxL~?>*y(4+sT8C|=sjTRC z&sNVj@{5AG*CE$b%WPyoUGP2v=79X(8c;SQkCf5K?m0K5DiW8EC{NTXNa+AUR@s0_ zS4{V*jgWfb#}2Aj3f*;b92=Ck^fsG`)=q%CP|@|BuLhgph27d)RY$kK&AxfQWQMo^ zVDBRXezcau?Yhc(F*uAiPBI9pqynoj&)tE99##8imc2Q*Ii@-1eXh-#^+8Haz7XZ?yzA5~ zOy;tNV0pG7`IU+$Cu8Fr7R`z`s{9kRjzw;OwLIrj1NDW_Tt}j>h>-%y ze6j1PrF|(1fohrV^xT||C8$?MM4rI&R)FB2$;wQm<`E%=0L(HH6~+UvM+@IGgL?HF z2u(RpsBEh?ysksa9w(ycQPGB1pbI_$0cyjlESycdP9sUI{;>De21laUU%zAHlz@OI z^e}fSjxJ_2-x97fe8E_-{wpyu0nwAHgx!Bp_#fJDrc04?XW=Yu*{A)!XM!$dxN z8K4?w@e$Edw>`WsFIa9N5r{1q29=(!ut=zB=SdSkUT6*4sQnfnpC=L3+L>S7T=5rh z+r?E1Ao>SyiVKBclkw#2(w)%HQVmb?mK$~xPE=WE@I{PP%tTTr{8XkH2RJVT&KcNJRG!!9aH zArGx+(H4fJ2TohtP}7jn$0k;RNZuS~pM-yPewjUlO|FWs1NLz8cHHp+J6$@Ywg5BD~<;OH^;A1L$bY#0sm&V6l&sj7oZ^J%wi@|&({n83Vf0T^78)iHQN)M zDc@~`$O8wO4<+A6A~f})K$%;n4Q`W5EcNnF-aZz;sr3?gt<(;w4SQw{zwe?a+UW`z zOy9#I_B*Eb#Xg1J7$^Py$%=9&bhYPWouq*ypIJOdDt~3u`_g5M9QI@khA=uK?^ynWgUYH_y zN@4xqLI|bHg!SLjp@U-)mDsXtTln8%H*SJXYXn6gIxYB4Ae$sZ7Tb~Brp8@OQp8O`Ytdl%E&8y|P_2I3uGc@I_tte2UK~>N|sh*|O~}Cq?q2{=lbC z+xj%@+~Zg016Zfrkat@cm5%P#{Ud{4s*xm|ABm|F1l?;ePI^{bjo>-V=weH5WMNt= zFBhPesp8Y$it#+TDGn{5k2N=@qzZpq*tX?%Z$r6J=e z;w%HtBKc$MTz;EDHv%IKpf!oM`*ZpT231rV;_4XkR)MU-%tYGE#Bb4*rK+3FB!ZZC z6e*RvZF$|dS8_RGPBLOP^e^HnbS!4j4#)I*mCpMShN&dC)Y(P{I^=s6ibqn)KbN@) zD*Co&gho&93!<;An?j|36HqDxz{XT{3n~beAGTWSA|@(FlI;F{@mU!1VoNZk!O{ah zvy00W*)dMvb`F78^y=S+3VwY4wraq>leX14=P+F7GT=>Q=fSnTy5 zS6?LMd@e4Bt?{QR{`ur9Q~JGSS#twLxzDw#d zchBF_x79X=2^C3S6Y062!m4s2MskV5niw7X(Co|0a~Tb(_1lVXo4V}sthn$X;4XQa z#S%e;{lG5b_jI-#SuL7F391rWIFUK#65H-{HRqf`&QouE!m?~rMgrq)@l@YDj`$`Y z^n5qlOuN+SY_uZmZZQHv7R-AB$XpP5s$7m7vqIHPhKHIJO|$Ol-}#+OOtIT9eeg*& z^T=rAvbeM(%LO01uwq_6=|6uA->hOnmV&&XAZp?|evrgJ;D7IY{>)s{`^pw&Lu_fX zhU=@^GNvLoY7#eO*@ymzVZMR1Z$I_=r10DvhYcH6Tjz4x@*@f*MIDlRILPzdW}`X$ zHg+1rWH5UW3q?n~@L&G&jph*x_al@Wiovy;6?<=)d7FBGrbN^K9akjRWK#;14XUf} zP~4m>jHM-l__}Bo>=(rD(?5OCk2g#sC7REGY+W|qaJCRc=K-%U_bGSRPewMtf&5+j z@)ajjL(=5uG51TCcTU&PfQ`YdAU^BnPOO6t#iAJ{W|CQhrD1~$;5Wd0BZ(LG1W9wy z{cA^cS2VVgeT!lPo~m+7cr2}DLH4&8$m)~Byj)#DhuI@b@KCwoK9(L66!0%$0w-}_ ze0@o%M1!C*eI~SqPc=jW_z!y{#jy_zZah(BR!q>=;{4}HD9-Va^vYDt0DYGs;nphc zWkPvqR#cyjtdYn??Ia$QW8zDU?ahUvOKkT5lJnr?tyh-GPrQ@nDK*Z{xXub4rQ*GoGSRdctB}y*6HA{gcYTX`(mKza7@v89Wd>Pg#iGv13e5|C ze8+cERwD02mrgR`ak_}o>zI(8Wa%Aqw!EK^AD#9WO2xgFmN_C2ZY6qZ<4)X#sDIE;#RZQ%W9Q|DyxH4(Xz;x3>V^jk{UQp? zifiki2EQBxO0hLd2l2kRC8vhAOOv9Zy`Hzih}L{dhn|6vAc!80;S1c&o#bag4qzNVK^puK3G>u}|z_|>YMMn}F* zQWfYmY7n({s0Q7Cy{b7JCz;R3l$okJ!-Xf=a|!^vXh=(d$%+?L_ed>MfCG=-hK+cG z^8Q8Y&RKQ_k?IJ(-Kbb7R9RW-cZ@YrdyMs|FU@E~wzryGU04TOrP@MQWNIms1hsf5 zKib1Vw32;^4(!$d5(dNP>N!Od-S|Kv zQEI`9L(dN z_)^Y{rc(e&(g9P5lBYq(kwge-CxzoQ4JcetEAE-ddfuG>&;rI$K8G!a${B|Qxgsy{ zNLo*PtZ}`C>qbf&`OMEJUZSmEUX(mtojRU`i=%CYM@Q2kND5C}3O*IoP6Yd26GA1V zEzkfY>}uepQr$&gA{OxjLZcD z1RMac*_~h1zyzEpk$Zce7I_i{-C@^Mn|?nH+PeUwvJ9MQCF}MG_vEG?x;*a2FK5P= zGigm#aC7W7$E|T1i>ZoeAmB;^C)}_B;GIILgbHD4wzjsa%+@L@Q9#rXY9n0rr}EIa zrfOnZVKW8X2)W64aT-Qo1T9X-6OdR+<3r~s%Lpd5fNO6Vn1yZj=0zBR^VP+Cy2|$ zC;dUo6NL^yb68j|BWSX*W}Qzq@xUdSKYvm_rmZaUoEj=lU8wp>#IzmQ;SRGr5O!0h za7nXr$RW9)n+EFHgH-M~$D_3Y9xJom+4>MKC^a1_BOQ#oPyh{DU2K2J4IKJvm-oE* zEFW~7ZcQlbl9dUQ173}0eSO^qxL6Z((80p5Cd|$aA4+W{emRaP+mF_VBvDflMvgk* ziq(WPyM4_OMc&Z;%bi?;7r2uMt z%9pJ*CieWqQe(+sJP(kAFI-Q|4VZd?0-5hUh$tn0;fGtnui@vw;HgR-#DAE+GJxz* zrZJYM+6|>Jp$h%bePWQ%fA0v8W=fE`_-fl#KD#Ld9m5T_Hg%4!zHA}So3)v@4I1tp zrta3V>`47zw6QI;B^42LAI!eI&S@IqZYBP@a7)2w7oM5<%o(UXA$_9B(1X(Qz-Qt; zK8-;2fHQqd9eM`qdh+=35?Gy&B})u16%wIol-(U29kC_ekr8B;RZVGAk{}57w(CyI ziu?O_6wBKWe-^4}mP%Codx5e5d&_4!%n^6-41367Dn1QUYAWqBKCO=* zKX#pLPpN)_M;|krQZA(#WXeU!iOKTYzz6mp!$;afDV|6&nNFFG6~YHq{BWrX^WfCn zsEiPRlRm9EAmW*sE;j*gUbfQoBd7(**$>-mhVR_fu_)+E%Z!^O;j}`e6j0iyPgVS- zmcD7mqnCFKm^{8$KH52y?EE%?jJqbHG=y|!4eK08HQkxgk-_I+FM}Juk}*|c@!*#t9JPEDaEtlK4k$akvrZW!7EsTdIwt3@7VLfW zh;@GYOLscR#~MQ2-VuMuLs1!5;ysLC_EmvwQ0-Lsjy#*EcU-H=1#}G&#{b{7x%~gqEA#*5Bms@d zn6ff5-vA)&K!g!sgNmxEf=WwCg@fD?5E%Gv6s{=?SkFx0RC-KFS-6__k2VbrYtm(p zvZFynMv!f^08hjW*k_2%4X_P&Au<4>ilrzmW<0d6;=<*eufY--k5{s?tH?%mKHG};|C%bvZ-Q_!0@b`gm%O1SqYO&SK^GucJxwBJBJ`TqTT>kL*{ z+g;mIhix|k=*}3f2l9~ z3#_U~+RyH0T5?c;xCk)v!Q?NMDlBL6PSfwkhCgtp$tjvi^+Rvs%REja5Vj4QyitK` zJLA+8v2MOjC+{=?zeBs>K)Uz~-Csx(#dA} zLDh8K!#h_XkF;eEIvU>9yOg)Iv`A@bC8`u?DAw4mr}L#>oSj8=@!XN#r=K+9y9(de zsvz9}BKeMa*L49t#N6R1;5H{+6XJ8cd+*tAj*|Ec0v|~? z2<*(DbAgTB^W(>C6japQ7$cO{ir4;cQI1#pQq)Bb5KT2U1CO{64Pi_wIn?Ml@l6)I-Tux91I0=#GJ4Gcq!^wGRyrVj+xOLzJ#t90@Ve2?=Hz4Q|z~7BfE$ z{L{=QN=%)>_iV;_G=|Ddlfes191*ojXvfl+uU11sUC-ovq_Mtxo25oGEFA?|35yuFiXh&wpouR&i#l*Y@L6w+pC zg@CYaAT*KF)lId2OQRye_>kXof1#Dh{|X2&_Ww2AMuYWbdZ3fXlv`dwL4BPxG(0h} z0J!tK#t`a@qX0wg*D>#dU?+>vNi=vyb~TXPrGW!J@aLU=v}6xi1lUda6!`BROEMSW z1;@aYG~D+9BPCG?mB4a89#L1laW%2JkJzHlJTsbYz;gUfwdis=Rww5CaP`$Qyb)L; zhg2B%WEp^TwXvS(2Znc`QgRp}VkJATkoa5uZg9WcGUd3_4^Mx%vL{DHJp(rq4fq== z;D}!ieYFz~q>j@XK1cZhlYPGM8NjTuH`kZ4pyhC65x9Vu zR3#GH@l8_@R(#YBT){Kl1{2`$hy^aiTzmq%QF6|Y!J-uiRQ^?J(I>*Q1v{--8o>GC z;zNd4UwdPu$j*U|gOJ~%3B*##2?;;C;lOI?<>H@#SwSQYVyZ^F^>2uBA!{5s&9p9# zH;t^5fGf6hETR-TrOO5}nuf{0M9fDu8<>xB>PnDI=;D7cFeFqVa5p(%lcgI#w00wS z?{TqCpWvgbA}>O%J#3r`_QtulwLs-p1+!CitMroPusJ(FE>S_zQ+{f?QCKAMxnMEXCT?RR{%xrA84b5`o18+#_tR5d;hkkWY2}~1ynGcqINsJv1CVz;D6F!fm6P=v^dR^DwRWnhuJz1i! zPf&^IYjdtlZW(2b#T*!@i^a&1nruTTzq=N8HX6+m?uUZ^HfJd11yU zM>&b;y2 z4

  • jqsnJd>PBa+v$vQ)DClf;yYRdLy}71OI{REk}se?ID%JnB&x1A-Wl>YW`4hp z{6H&ONPf0EZA69rg(2v?1)Yao^G!Y!)l7(}68rhUHz(e^H{SSO<>-shb&rOI$M2r^ zGfqhtHXW?P9&S%pE5j-+A&4RrSkx5}{g#VdSnD!Yw&$1@*xhu-^}Fa4(F-L$|vz<9kh~hn|>$|DTF@<-OACt zSU^Vys7FA3J3<5D$mnb>as{c4dh zG0#C;>myH3PY@~wZAT~r5UD5Pa*X2Ey`51N!3ajs_73CQXmCdZXqHE~3#zg_-RWuY zv?K1m7z9V!+~AfYWLLsVL!r^IUMJ8lIDB1UF*RHtlbsIwv~-5uf#Uk_utyP&=!+l_ zs-A*`5}3zkX$msf0mA0ih-kn){f7?gUK#O&?%Kk}8<%%K^?gMzd#;t8Z0?Qj(0X`=t(7in5 zT8>ENVwRGpG1kgCi6!or4gq0d2$mz7o|2SwBWJ|zMkfH{|YWlp?35Ou&)A zB*_nEH$5ZcSgTB3pBor>K<3VJEwu;L>#)GTQsS%3OXBvm_+pQ1mwF5FPgC?Iyz&C7GI~Dy#c!2Isfi53hRt?AxlevcaGN90c{yGLlMMc6@aHAo! z`FF|dU{;s^`Ev=-K}H8WwQq0dfO0>ZPmMUk!2^&v7)+d2{*b8ZTp_e)0Jlmw>A`dW z<_cvIA=1{|0u?b6G@7`uB_oET!(dHYjt0o4_~ogMQhXpI zR}E;&lY0HyTTDVi22e@cU5x)m;{%QfiGqSctWmPF1e00;18Dw;1$Az|>(wnV$wmXC zgk(X?-F?>DwmmXF-5gUGEz&8v2_xJG)~P%D2Iflus+H9WG-!YX;!-$C%VG2{t*p&U;MNXk~qenTQHD;tS;?E%Q?m~Yw_f%1pg%ZP67yOgIa$XbUZI?ZZgAR)tgtHy>%0Oa702OrC`a`ni}LXPtkLTG)cGAAs(JT-Pg4&{=EBcmHm}4jjPUCNxtE zAu3J~^i&<|05Uhlqb{M)ynAm1e{3TyQRpww6gwNLg^E~{5mPqC>>hyq0%b|WpFeL% z_??Pr)dBqr7Ka1*<6v5DL%@jT{mz!?_FJE8Y$O+OQBf<>wqWs*x;rMsZugfn@EEOw zK4gp*bZk-;T^E2DM)(1?9sSREQv^3&Y~%y`I052y1uXAt%6{2Fgmpl>-2Wa-H(@>oO8{FfkAJ_v|E(mvm-atU>-?3? zUtkOwglqCj7JA=pEQ^<1#BA7|ot-7XI)tbfWdclQ>#DIxUWk>q5VZq`l;QcoGE<`f zNOU0N(a+R6Rzi0{Kh1}gKi}12M?m=#9fODnVk%5cRS=O+pxQuq;wH;XXfq_kVgUD* zrm^|jK4!|Ywg=iDzJl1;0Qk#>Q3{xAkAhPgHh*~tSkxxBu#Cmra$WL656}j$7CTX- zBlPdhBq}aIRgdgJzZUr+ucA1Bf{rc=idS+zJ1K-62A0xzeKLj2-HB(S4W73*c}x0sc4eaf>XPZmohB%< za@1u4B4hv*2okYzqTdFilUBYka?UE;-n)z-FI!kIml}3Vm{|htLhu!ZwL40T9N7SM z9eoE57iU0zn{P-3&d($h+9K}L#uLO}T0C)B2cUse0?|zE!iYIu&s2g$pY=YqG;_4` z0O$q)#XoQcSQ? zJOakBubk9I@0pM9LMmbdj_LQPhnWPpikTQ((&ZvjMwZBYYWNlEoqt2^y3)0ecg ze|}7D$-v2}3|@?pniu`W&#wzq;y=)cU;vUn19>e&B4jxH#8`X`QBgmai_hE$FM3z{ zeD2`Iix&sWaG-(70AFsJ&CE|0$TI4EB9km)MYW%RieciVbY6JSuPFbBS9nuk(2gZq zr!0EZYY2vpaGpK2(Xk{8{tY6I>gFlfQHKKmi?;KQ$FhI_e}qbukwRo=&ytl@GA<+I zvR8>@L@p$=kS%*>l@XV5Wo2X}nE-Q_xt?o=zFx;znf1+NQM-haz-s{1BycsRG<#hs&(O@U#Ae)oqJ=^mxY8h@1HZ8 zU}Ixb{Y6jrPZ?Wzs#Ww}h~h~JGg71cBJy%)c>d9-u|@7c&HX`)yy)vC>mEKA={XnM z^oM?DJ;E=9HU@k$imp(MZsj zBVQiJdmjIzoAuoHJ3_Q%J!8j9i~m)cN03LV13TC!>rK3s_n(C%s&a!69Lr1KQ?Jc( zS^4Ww5!c3_`|RIYe35P)e^^GY@!2E~V_0Qwh?HX3x!&Eh)hCOOk_)6xAY47zv3dSO z+VjR_Y+p-Tn-bx!HwUkjz5shR%3eDqHC#1XU#a{K^zEBJSy*3Q6W(N%TdDh_S@=WB z8Gfa=^En%VkB_!6pK^+>Ag46Fu6Couw0xL2;VKNDLr71bWoO0HD|q?eAf9>Tg)`Aw zmj=3Q>5AZ?T+L^(aw9h*AVZ8P(c79DkM;`P+}_R`e2|*gVxksvGQ%po?!TrDF*HMh*`BO}vFoy6cTFH%O%Bhu#8VPn`c zW{F`P1eX6LtG^%-*yS;K_hFf*6FoeO0Kl5@BU`m&y zf;g>7OPg@I5pVS5ziw8s2u7#YCV*Q1uj(JqN@8wFN~5hQO02I|NUrV&zx8SS-}|fkot&7DS1pnJg-f%1hC3Oxi^%E zRFmR=F|8pABBU=xL-v+36dgG*$482)j*iMzgDasp&#d4p}zx0eB zU)6RH*?KX5KI!8(DM!MVmQ_I;mvM8&F?dFi06)Vtq0WqY}&RKBl5p2ONv#9do7 z9xBW(fNiIdVFcuIbb`c}0}bCAWC3-+1ezb(8xLGea4-Xf*Ncaj{R9E;I3|6E`)xlT zjn_bg#mY3lQ0D1{;G@{wq(M5Ru&Ql7bQpR)TDiJZUR(I8fv%QHce@tJUf}I%mTe1cGRCmo|QW zYa6|xAfRQ<(3ogTTy(ZW-mKPzf(wmMU3j=H0@xnR1TA+cb^YnOdHxPNo0R@fsp>)c z;U754Qsi*XqxYVIHAz5a8j4L-Z*tC^7orFjvBj z6(QHSHC#f2gx#^L+*n&uZS5$~S>kq2rK&~^FW%{=$qp~Hm&l09U2AbBCt1~vx8BW0 zSq2jkZ>suteitdVGqo`hjbRcO)a7Xuza^?{j5HijV{)0lmbm}*GT2L$ZR`WTAv3O? z)`_ro8L69QG1ceq;MWG!Y|tFn`Yx4Cf0!=7#xoRs7v*`1O<27?OP%92EQgFJZA=RY z-O3U%kJ475x!MAoK@KEgU`Abs^5trwepv$m7fBwQhIfUq?psjUFxGKHO--hJQ@()F zk9eevCa%3cRH4y#U%LVG6gAr@om=z{qR4rvqIC-LGYeTTO+!V-ulCKe6c1C_(9CA_UR#9g zcJthvH`Dt)|Nb~5^5gzaWSd8ka0SN8l=eG2#^%UeLHJNcXs3P+FQw?{EI8+L{8nIm zK38pkYeDQ(3QHk0A?-QXe9hIvyg`g(x=HkL+)o9sP{~!fHZ;DAO+CoItq$oM(WEpX zS1Sf&WhAbbVlZgpFnXceNvMyPJ5N-*G1Ez^wY_5(Den$b_$@Z)D z@4<^01PiWP&d8U$o>0oeA<4747b@RBKfHA z*{iik@0&MhSEWR3MRDHD84-y%;x$(5z6SbL2ZGOf)k5F92EN98ksx4cs>*E#Q0Xu{ z5xq7GzM(?GwG-4A=jtSo4KwDbhorS*v%B z=e{M3zOY`m6%?M9JB)r!;?;O_IDv5h7) zB!EJBBs?WKraVLv&t7VuSti!?b}_sGNfQxesA!^mv;oM^vMzGPR`cPX;-nwF_$0?5 z#JFy%rWQLFOi_8CPHEIBPD&v@p6y)B7m=kO2~K1Sv9YmtDn`hMgAToJOVd9&Jn>Cd zJNMHMN*14YvWb;bxHYCaopYp9dZp$vSPK&>be^XC^50UjoRL#tSoC0U>QDEw{r{|i z29#52=Z=w=aJdEZ@M>;p7NVM7Xbaqb1h;?Q4hU5LsGEt2(P?*vb`1^T_to-}fFMC@ z$unPoA%!><=hEuIcKrJRmk)_cin#7OeyfOJ15XaNNSWnWv(;z_%177}1lG$6H(x{Jd!j(p-Y<>D;af z)kB3aH9kB%jL7GQgXA?AXR!;3sV_V8k{Wjrp?Nnaw6j%sF{HNPqdk)c#)z|0u85{k%|>=2^6b=Po*@f*gm|qr$Q7G)ScPG8 zGs`Yw4o6~6M3YI8sX~ODwV(&C0jM{7K0s&3^W77IQp_~Uu);PTnIJ9edNROd2JWhD zp(8z`=0)+>XV*z;OA>LB?XhANaUFZ-3li?S`3|o9&m!T>&fM7LJ&<|Zoj*!*f~Xq{ zZ3|73%-s*PuChR#{S^e!3xf>ef@1Y`wO@S2Gmmtg4|D4>%{VCPWUD!u{pxZtJvWRT zH8;qbO*5@_YZi-RX?XS80noGpKWJ-=*jHO23QZ*Q6$ZT+(3h!jf&u~>KD(<*@n`*A zZw*8H8w2g>IXia~aU7J=AJ5(Bc_E4_p)ecp>cG2%*kGHYc*R~zsa_F}` zARR!-FpL^{h;q9#$eYo^0d$pxMII9DI3y)&cvIm@->jtguz)R|^%|+#j{USe-CCoq1dZc( za5{b(G{YtW{;33{QXPe2piW^C842mW`&Q@aW(X;eBTyz`i9O-)=P6~%5}tnw5qbok#M1eo?*H>D-jFK^bdo#bqP(rC)}$k@v~oJyvY4yQ@sJlN%%~d6 zKUQ0RPO|>n&PZ49s37_5jmw` zCq_@Z7sXKzlH26$hWR{0oGS9s42+*V%h#^on%mku8KUzOf4kH2s|{l>IYOHfiR;p&+A zw%7qq{xm_EjokgJhfO*a4WdcQwu)v?o+a5>WGJpJX6^OAlbE?%BzsY;u1xz>+D(`C zMfTGtTBTg(m=Tr<4pqW+Srh*CD;Zou_vO~HrJJk!+Boe;^=(bp2|2{y=~sB{ z?95d7JQwk1$T`B+MKtq#tn{w?)_TVx{>>j^ox7C5vaAQY7^hyWVP+jKPL$Q2=PlIN?g#j;9Ef>%qL(mdTYLT!LrpcJs za<*fa=JTzE%-pqZ=~3N@`%5!#L^*2JGCRpnlddF|%KK2^mwA1CWERZ!WZ#*qpBAKM zt}N%LbqOLl)~<9%vcoe=5~ZAJ?whd9IlIey`ojp9ij8}KaJ%=Glx-t?3b9aIG_rIJ z0-dVormD%t5T!VI>f4j;Meve974x#sKQ3za@fZN}TV2!%Cw?WKR&>+W)hPtFLXq4r zHY$ZXpH6~`9xvEPAj3mWTs;dYxJpNRLTTf^ZT-EI@#rJ#r-Nv+N6M!va(iH~PtL-DoEmAZa`UOF(B2x9S8 zE)2}*q|~o~L~XG6Am~0aXx+O-yS-CC%xrvz^h)_}TN6VWx*@~Pc>XUpKg0kmBC@`J z0sDFOFHmBP!8o^k9MXS(ZOy2eGWd>+F>~ea!1tfoo>?f{*Dsp;5BhXUoypfa1>ahZ zzAzZDH5cal5}~gSPJe4?379=$MiTSfc(Yxng}hbRgHtk;^?)JgXm?`SyQ-N>02&Maqsb!`ho?b?VSi6p8JOYfdjv2 z4j0Hf0{nJyvBw&|p(nC$cl(~ijGphpX9=h2e$p_b4?lpoomeEod|CpDWI$l1umMGs z?e69(4~610N3t)ULHzwKj*ZMLe0e1@AD&lvD1r8-^egvMbPfq5ldj$Fw(0!p0~ASX zj+osuA8X5FboMW>A<|LTWlC{2v_{;hMCwEK^W`rwy22ek7oM{}ZJyuU0T7Tzr_TYr1 z<}~KL`Iglyud<7dE)R4DI^(qlv{Qz-W+8=P?}b6~+TKN)r_UEPD_?qvyeT6cVEipV^Z&(oe&Y*6x2ybm77)Ac(`U1FUae&3Nk zrG9_gE+itNN9^(KMsXC)S-o5ec=fr2KV$6u>kgln^QW`!S)f(P936{Q{qH5q)Q6!T z#EvX(;$7S~gSvUtTTGzIti6QILOJ$Q`z==lSy+uHgIhV3#ZW?mYGRg?rpYhc@HeI} zON)ahHI~}Xsh?UYOvJAr%AHwzs>N^MA}0-`tR2kLYMS&C78XY8$AMugf@*HO8{5aX zLgLkOk|>oojb?|U^RIv-YfHGnj^wU<;#~E)0lD;sqKT+q#i!!Y^}CBvwX5G823~w- z;3=7KW%;cbDYrY@-DWio9WhM?Ykj>8v<5}^FA>mVSq}bE)mP1a#jo69NX^iUK_>j1m~?qS!04uH~4QVfz}~2dIj=&d*&4 z=Y-t=aH@c}Mjt>Q9^l>jomb5ciD0dc7ljk&!#3Jyc~75B(&J`Ts89dWeTSq+UfQ`| zr7cr#Mkr;Lx9Mx=iiRut$bG?bpU2hPQT7sNIbC@J3KoQBsEY5RKD4mU8wjG77KaQH z2Or(c^$gsDjPe|qL?I~K!1oexA`sg?u$%dqf!kO>0OK`z6?rnogUW!^g+H!5?TOTH zRa_peqR4Ar0=Xv2V;$KVKfv`eSD&Cy#2H`0REhOk7E>%w{4{aoZNjtvOE5J@bwoj{ z(O1QzdNt8jC?X;}1yEwB5+46DLx%3r>hP3|J7HcN3-v2;?z06w{e7|WBBE_ivo3z4 zy5%3*`dNteFNUk_-7YifZtvNsYQROJ2C51?8^2FWt2gmbmf~0U*VbO9tu@u$ z*~(>^-5K1_>nN)%j|=gp{~GtkkCYquEEgI}Nn?*wAPT$-rpV0`XQM?-Gh<=!{9!Gn z{{VzChx#X5UjTaVPN^NottqS+q2v?Z0ad&$P9ts_qODi&0`KhA(9pkRnU z4d7PH+oc*+>EfapWLOD<>vqtKivq)5y+6#rDLu7HZ1iZ{q*UnxTMEYNsMYpnabJ-Z z$I>-kfr~jEO|te0k59FpO$zi{*;rcO8mh><>w54a>mtuUrY-402%A!>YfBnKxVhV? zE5bS?OE~9-GfE9UI*Tp33)OsIt-G_$r%YeOuk;a6qQmiif@ybqjFR>J=yA7nZlNNZ z3T4r#)0o7xpI+Cnd}BT+@>YNP39-r|Ze3RXN^z;9lEyyu3z@YQmXUb;S+bLyzigQg z-0XNy-Cq{?D?f3imXwrGF*2Gm)8153iGKZB=)|d03FYON5t8rkhx&Bz%i9Jc&SM{^ zaT(;mU$ZQS|KPIna}lJsPTa&_cE7!UYS?`(X2Bp-?wrk~+bjfqHG0JE?)T2;)6Pix zli!!RfFn_obe5=iS3WbqHeI%An&WV}Agjtt=mH2#(9Fj~M4CLhxw*jcO2)8eY`B6$$c*BIYHdBd#@y$u9;}pd>E!>dmdpH7x&Bl z1o@vtf!|rv5dy9pJXgP*FkViFkw<}U!9`GwV*fG0krGWI0Sy)H$+;Hi55Em}3RSPY zq1#*If^XS}_`DiF{m2MkWc5zba|#(1b#Emxbe^cqzKd6ewu9){CN_ynOiaJ848tKI zt=dVnrLufc7adks@N<@Fh$;5mxg|8CX!qz~eJ+xV^78Rbl%7VTkO&Nc0^N$)J_nXt z?N^bz;P}6Qi0=y2vbL+E&)sdLDe~0QMyhz@7g@}TpPP?}&iFoR>iiBRJb?gpFdO2I zMNGI6dJl?HFD z0*r5-O}g^hN_6XI$-Y2*SNMMq?+(uqrI^p9)0IDo%@cX=X^U#_KeBi-GMwX3lHM?X zFSWM(LD15-X&d+k{%vlMrWZh^fg`e-V4vB7Kv8t&e#<#?Wq=$0!kzB4c((DCKcvIX z+bCa~$k9(G^)+vwgK`(lEC0Xo-4Vj-hkEH!e=o`6FjPUwP_25Ku7iZ(Y!r(S z$eBLz=OgH+K-)|JA_xJp!vS8na{ntkaDUFtw(uNlzs~Hq_4aW@J0lZL%fB<}no^_V z@=t+ClW{dKZ&OERrhkq}P?9)cinizFtt;)Q?`HQl1*1RY@lf_OH!o;a;+D^^1~l33 zm)i|#kgp&ovgK(P`rU?%%8rp|Z70~fzECfn{=S-69C_fXnJBBpdN!B|LT}CU z+WeFJdTdD=$NLuJ7-3|p#kNO8AYo9PJBgbKi+GUM7Z@PF^}f$5y^>2SCDe@V9)=wu z6W?cE605Hmv0&bxcydu>y?#l6DcJ6CD#rFvQbHQWpqJ49;e<2Z%6Vwo$b2-8}ZB6XbyJ{lcn9~~`@x@IO!c8*UsutanI{`%79kWw! z+do&9qqhc@2#;>vl4I(p#2FQb^H|a1mtP@Q@RIG({WCevdvg}jr;iduyK+eCx*{=XQQ8Tfh(nhRq#a{*U#bv>z zd1P08*$E7f%ne;3`3|Y_(AgRK?7J8rPV*98bTp(()L&DhV56QQJtZ*oZ#%i;Uy4A4xA4=rtlUsSVyT_DVZc^J?{UF@o8_spT31Zgxfn$K{AH)7 zyDf&Z{Vd{^QNu?K*WYY$_SvL|jv0Y#KZ=T)nF6XhawdrEf6X4`5qO98Rr|z!NEuP} z3v+5|em-|L{CQ_jPXdrd)sI$R((iJLh-d?ZoC7@c+0XjzlE~qU|1q)HH4tRkRU)2m z%spbE{sX2Nh<+w`#zY=&{sDA*dVY_c@D_cTK`w&&dx2nfKc3bHWdC{>Tqs0W{e%{t(@_vas?7M#0jN<2!9r zF;;^$Susfo+>|W782ycH?KXQ^_iyo3+ouns4N6JnX%>R-Bzd6K^Ub+UAi%RAKl^S+ z@0VIzo}`<05OLTY@8&bk%OohPc`G^oz5jGq$Y{1{TIhwk*^j0w(a^T@1@%uTOZc5U z8Eii55FV|l%Qg*1S+;Gprqe7h2k87$@i9md{pHCa8AL;RblXI^omW2iEM{GD>aWD^ z-`ua}KDVO|4b&E4m#xm*S})O(x9EpW@{T{qn;joBv4!T+9qW+)651g0xzq}JFx&0- zO50Vsa%O$J9B2x7FP+C^i`!2$=O~vynRpk*WXif`CVuvoTt?@R;)H^LkT%7qwU8%W zE<`;r@4-V8sq&%4ulhZkn0+OO_xJ9iQJkldmP*e+L(* ziBwfuFRq(T$%UPzV0iSCaUmu6s_K7hA~b>9z1Iyc`AqyxQUlQv&n?gqRTAZ?S@yRR zNqQ&wKN;#Qz79R>q`D&Rw_5(eWXoria;u^7els=vZEm6|IyySYDI`NGNtoQzNXX8# zI756Av#&2hd`LS@-r3c%69K1O$;DSNM~<;B`%|x+6M$7vvoy($zs*ciSNr94Dp@-L_%g{N%|HeNm(HORgMqamktB@xl58 zK9B5-AxK183lI;tsAN_ep}Oi&^)Quxw)J}J6c3N0$$irTzZz@iCGbXpBX$8&G{0Zz z#&y4vjtou>J1S10b^8-*Jyb8IQuAAlMToFe`(|BiAvlO=rs9;blyC2a(hFuSce=E&~eLM^8=2DGZvZp>*=O|E#S4j{YXq|0gseW$^GB zsAMMY=fGfH0VCU4MmKU>$rI6^X1(wzY3+0cIm%Y&*k#Z zpai~|S}k=MqI0+HY0xu%Sk%!WnkRq1mXYLF{+SD``$LkpX5Alj)6D+W67i!bue~{_ z#24xTH$~p3+k9V6OmRxE%R2&w;H&3R-5l9>u_=Rjff%h(wF8aBD!T)vp!?GHS3R_H z$?SQ?$dKoD#sQds77LTb(6zX4FlrK@4G#(mDtIRe5;+fiOZeTS(@5bC2`Y^j?wnjWt3CeU)!qO!Vwpw;Q#&Sq(;& z0)mut9*oS6etJQ%@qLJ7|NR>`kJ}}tY2sQemR((ydD9eH+~wi=ShPy+%?tr#6(DSM zQ-iYPA_AjB#hd4~24#tE>k44n5HVXjHgLS1E||Ld54?gg9?>Rk-iP}QPD~jw-|36K z5cnJr_w#%@RkA6zbaeLz@5{hJCo<0ng2(ovwa$zrlL|S*BE?AYg8(=&F)
    CkQ> zF_@Ok!IU0_eY}W2FAongGOohHxZvWN`8#jXCQ+jjUU&kzJbFzUd|+4IW7oaEynO4# z$sW6whrf=ez9HNF{I_}t)ah>dS;P#}r`!yO~M@EVe2HNnWk{BP*+mIGp0)r2> zrBPZGmvlb`FgxN1=QgjeH7i&rg}SCanIhY#VJUCdd%}-YyNhH?b@^hvmCMGqK<~D@ zC>yU}X1$)g0&#z9gSt^k>?L;nhK#os*VBCXZ?MxB4OS_=W29v59WuhLiV-?_8kaCK zs?|d$^WmGzh(#wULYv4y+z!d)MG7yR^cmir9uuvM#==wv(cQWc3Moy&G zJ}+(*pO1uRm;~>p$I|Rik;h#p!t!Zpt(*+nzOHqyJghlpey>GwLOYYjc-rduewsjj za*tz+VwQ^;T*qOq#AXq1gQY9Y5Yno1R(A>piO1~upwELc!S*-!AE4!mlqnJf6#i_lxUv4_){n&)!Zw;xmt9zUD4ps ze#VB|*a2nm&Rb31E7gJVo*z4~-m`cGlhA$~g8=1_WXt5+CyntyZ-Pwqq7P@NJ3n5F zEfY&j%|lZs6q{7Df(fX`3z83F($jlxrM43bS4Nweyl%<`;9@-1c|YEB=Sq3fV(zJx}e?)DzV|a&xxSq{I~iOzlP^jwh^H4Tq{n0EVhuaqYjj=M?HwNR_2t$(jYn zsn%YVG8l4V=034Z-N_57gXVr9gqlXrq3S!yFj$3C0)2fyNsi@dWg$l}ilvg0U+f&! zXZ*EmJkICq-nktx#M?{RqNE^*HBupf*z{!a$(!f6&o}G+=zCva{f1EofTcDqJ8zt$@19AQ0KZ_NeP&jl|xogV{qj2$d-9ahU~Y9f{P^wA&_rG*q54Ow7On_< z1#1s=`o|H+Ita-K+{KER0LUO}+Gppd*a@;-LF>?9lcVs?kNNE0G_S$>Ip96mjz?A} z6@FV2ei!b(=xX-K`)7&WJgb$Tf%)+0%0jZYCPGp@Tek_kWpezL2GQo^4b|`H1bfY(~^f zl%G9Hzb1b@|665>LTQU3UX8 zq@E2VUL*W5tP0G|E?{9jcH1Fqw!O0>zU~!{_UJNy+aw2zXc>BNv_jq`zc2{JgiEJ6 zAfoZUe}o@DF?K&U^ zOFECeq*B?Dq{+Ec7uxq4Ri&0v(p#*NO6yJOUms;XIZ!sqWDFX>_SZ}=mjAFL?1Wzmh7Y|p!Ku%xpDi6N8aZgiPkXR!jyZ3K+LBl5Z$UnvLFG#q>^Tm`MWy^f ziO=$Ofo?U-Wn#@=YOj*{_w=5VK8H(>=VNYrA~Q;_*wwLl1}GVB?siIRThjRp;XzNI zev$4z;s2N_nr`MwjR5|+gK)=&%p)1CQ0Fg&dd1-`SSI+>H}a3v`sBR67G#-mCHvXJ zm+JX@Ced9>CbFnpavQq?#prJn1DCub%AOS{F1%)6Q~1%Rp#Yx+!M#*D$tQXi6ayNc zV)Piud4dY!sENj^ba~AOm)?C3Qjwm{36QcJ%Z0v$#ftdvdC+rK=3{GtwL--79j@E! zF1P2oZ`@8nA&l-_2IY{va@*tQozL&>C4HM?mZ+}A%$dBjuNmy6A`lOJR+*TAi@c1x zU}OKgRk}~}m*4jvj}6?=gA7&7`j<*|ecbC@CrH%4E(10A0MXIW*fCv^aQ)U zyq2R|M0!<%4WrkE1h4DNxtO%_fHr|HGSUU#3Ttm~;SE)V^jQq5wc_iT`_`D}N0aDO zlbvspLKaqL)?Kx!%dNU>sUqxGkGa2iSpo~)IQ&*Q(;vN`ao^kB#Up7H$Sx$ndGw)Z zrr?$7r;X|dSU_MId=L2I*kHSa+Hv?Z~HH=W#r@LecQwl5_vQgR! zhP&F84yHkZj!@kq={ZKGdVo$M$!I)-4=E%rQC!zOy}QY}cO?GIO4efc13vMy&r{EN zncMBjI7>KMN_)P@F)qOLy)e6hPyr_7$J|+*k=G*5u^*5IdVcr11Gr2b>VE4D1Bs2y zXu-MkovO&QBZdQ;vj_H+);92>ougp}plAbunS%tD!G;IpGb~jd39Ei#E(|$u|0vWK z{;;DM+679(o;5Z4(=(%Q#&;%l1L!=rcD9jm6;4+&pX=wMUL3~icY(_!7Xv;(Z54GQ z`cAy9vp$mOeajO#j1CAaQauw6rf?(0a9S2HcOq2vAq5Y)J4hZIq|5c^Y4ySKE@br3 zlqlx$>mKBWsQI?>0)qHsNC}-nx>$G)WEOInsgA_VUP|(`zWZIhPYRv)RTe0<|6 z!Q$)FSNw47_1e8H2c5TmaVIbh1*HFdLEAVYLQ^>`N`-mvsPFm{f}?ma7hj$+c=>x9 z&RjhLWMS6fDvzebpfd@O29h}Un?fSC($r5wqf0x7xWa2W?UjJ)-8Xwpol0>hKAhLv zH>!6Fqn~DfXKCQ>)*^4#eg+pu;cg&Q>^h>Rp&`(Vf&JNeOAw_pOqiAiE5;;fuCle@ zArHPJ|5Dso7aQi>*n<=FM~Cb~ws2+B`(+_9#^KoFc(&`=u_a|zH*5?fu9@Cgxm?BL z*UxT3*%)GQ+rRdfjN8uDDDOu(PurXPR*~fPofxzZvH}1&K^5>ZgGs9B-7lcKl_AR< zny4Gdo&*4doaWZ9IE>`=p@FC7;Nd~lJp9Rn_$pq%er?g9__A-(CNYzeFTLJHRB(L9 z=-F~oSc}va?mT0HqrnzEO{Yh9hd5ki_DM@#*?mnIS{cs9r?pd>;*6c%q9tAZ#Bu9=rVUg6Moa#OcM1)p>pst&03Np~l{bBwEbOul2}fxa;WG z>zv+3Mcv)Ntn>*7llQ5wInW^OOl`I38E>Mfl&8F`;*?E0Oth4&G}r2IIykYZcq&$4 zkst^NkBrJ-o9{tcHM_;_94dmMfx&y9%iS3(YN4+{$TdKcE@E$-nm(_^Jp;ZBo*^$V zT(=06DWmB}ep0zWF8mMaG_gXjF=nY%AY5!VFob*FIln7VoK+P#WBK>j_f>me)P5w{ zKJIcTiX5DGDtU-b`Yd}4U4wib0sgB11D=yOjUs;Owk{dUuQ0Q= zHpV^7Eb7u$tu*VpKHCEU?`M-XVAEqP;@|^A{WMtg^!*uVh~;ag`DYZyFem6a324#m zg5y$Vkhq&zuC`-5i?jzjl1A2Dp~$q0B<^8(cn!vuO5y^1G1U)*A`;AvS;!BoD%hP} zuHc?%RPPQx+yVD;n>YB}b8ha@(Fu=JVW0B=&5`mXoOwfi`)_KA*aA~OMc$1D;UR5} zh}Mqu&-!S)s=Ce~l4)H@`FgiKE*m3$?Oo~2z7x@9D{y;#{p3-+&uXx*h^b zb$LyYFra|PkSl1*t{zTg6f?~dv5>5W|6fbFJCR_@pJf<~DmrPAA^q%F?nOGC$B!H3-)rNy1?$YJ?Rwx@eT?#7&<=<O3RLLktv|Q{u9qWNj)v8K~=J*z2F7oirh|^jr*4P2J5_Lp(o``=e zZE=>I_v1DsL*U6F333Ka6Zmg``g zqfxvU6g9Up;k7je)nTx?@J{haW|Y~;j@%v$ozklKPYS=EqlH83HjNlGVt z0XkfXzLxCE^S|C(gynO4V0BCE>%EpGwO$muK!Q=WQc5SXu+nb@(BS+7Ha8E%kq_o} z)e75zs@M>dFC)(qoSRh%&jrxyhO~EWD6W{wqIi@ZEx%raujLeZD+F;Pp&_=!OqcIz zmZ+Yk(wcwwdInME{fanIwS14wnQYwe8@S^=D{;oXlU(WWYv?G4j&Jwql+s-hubf3o zs0V)nD*xzs$mW@W)l+FRLRV`B2+#-7h{m0j#XMI*t=XgwM`PaX+SPr|c#r%-;zCk?p$T3Xkdv=~iY8lg9 z(V_^B+RLy*ma#R{6e(qOF9pD$q;5ZV9sM`vwta+K|K>LS z875f7wcTuGeKi_I*|>)9=ATY@BkF1@M)Ps&DVpZNa{b`sv$)u&{U0orHB4S#Ls4nR z{g5ENbtIEtaGX`^dt-~47))&9-ycss-dvG_kO5a@B6|6hA{g>XT3_^?ZjEM>nSto! z+59e^B9_a|=6<_g;J(!-3pD9y?}(2pw_pAJJ9<#G3vL}x8p#FL`DF|$@CVx0%`9wV zNhgux4Q*5V^Tcz_um8;1hj@||`fKt3EV?9vH8yo$k2uBg^(;p4hyFDMd~RbEH&N{I z&gf1*dCU_De9wtiF{U~$1bAKTprg{dDCc*S^e%6XwUw%cZGu9;F=V`|yi=p5`XCN= zvJ5*fgbEM99cnG`D&bx-R6Xc-MWaT8e8BdT(0CGQkYjT`Om-v>ecFyk9Kno&U^iob z_wgoAAhAl>!t@v1)8O_r*zqmtE&vNh>bDnofy}Y8DgMeY!Ag7J(n7H1)k096lFYAl zTONC3gu$qND&|tQeN=cHb|6+_vv4BR z2sA8g*#?BhBCjov2vz697?`beE1#T7AxBV(Q~#4GwWvcmN^AyA!$1=^8{f*dV|917 z-PbvUr^Rx%;@!^57JG+iTwyk|QFVn&>|jNwY6Zp$bv1>WTq5pZs<3%8||ea_4|sh#aJ{UGG{3C)a4<>TzRRvo=3Rcz}!gDi3$ zuVZ`0Ojubf!8!RzQ8Cb#Tq!3WsCi-OmJVB8vUoXJbCnyJX=%I!`>?riRa_E7vb}>! z%M*HK|LSep1Vf0rwr?d|nsQe!z1`##Y5%NbR+(@YsX+$5Q}M?|_-Q2TkX+FfTQj@2 zr--Ocl)=wWO+J0XaDaokn}xu|{H+=>Y%}jg1%+bTaKzR&9r7_nR8_#dzVHH)ObG!~ zU!x0hSby1S8xosB_Zbrt&3{`)W^(z}*N)z^0zf(cuKKGp6>QDRRw+c1vVT1yS*5g~ zVsv-a&i&Zt7V5Jt<$>$#LR~X@H0im=vJrPxb53VC;)AD}PxDVEWHy}0{SUimYnkuJ zNg2)gNgCN@C8DQ^=v^LsTwKbqqWU8BMKw*)j=f!j z-Z$z5!LR7oY7*WTrP&M}*LFf(ef~7;vKN(FQ!{bJ-YqO=snE;JgZPXxs#K9!*bb8}=%9M%VA&XiY^p z1b8piEDHP;i&+^nuU{kyKU`#F-c1osJ@q`cNLy#+n=u$Yz?(nk$|xpyF~`*Iry^%p z?GXn4#~fN>k{`ZQQ~A1;*-mGxY392PF_+RQj@=kIlU*?Jv|n2;0=(EltWUXtk>k*u z44#0BRZqTZkfjVd2N8!6@d45?NJfbfNN0d8Nt$C-SxAUol)Vf+XGw@nOmb`iEcH5z zyHm$iUFpasn(?XzBYOQ|Mor96LRX<#{fEbCymkqRGoqdDzvGE9H|L(i zysP~LfR3M2oybuctX2ir?nU+AZJ+{AK>U?{Bem^6KatRtxNi<`)nQHBeUDR%;Zlx7 z8ta)ZpL@e%QWH23-V4muktdmT`pVEAAqlpA3Cppo2hUC!GxUCM+#>0f=j;M+YJAMq zrp~%q%Yk}lNTcHkC}4291KF>967?=bOVpq3hr-EN&ZkB6D10``3I{9~#MAKXHOveDk<*t7(DNNZJdM?HA4= zLq>1c377q6sdSI^PrjE^GE$M}Uuo4`BdhIrt~8-Ae!^vt?AHd2>A+X!`TZ#wjoZqW z+yP;>=y!!@!KqgXtszy2e&NE_l+gq&B_)g7h*J>h2M7t?97h2@c8O%_paVwUcS<^1 z861hmcRsQ!E1s}*Ken}l&Ac|ST7UM@P{EmtKJ|Ru%9bgeZ&r_!_;tp7ny5v;X)jHW z?^OJgmipos2hUPwmf-fe6?}L*v7eL9e;agng*39yNQJFuHMrIKW7(>^u4 z!Pfq#;!54dFPiGmQo$Dd@**2lN3_nh!X8#YtzuinC~2A}!qj9>!m@y3MXQ!unlm=4 zqP#OX{N#)1AS>Jp_@B(wB46|m< zv*#?NH)dR;{}10+Q60c0%4Ym0U6@f}HWUUD0=*Ak+WI?Dq#3#MkN18j=KVzx?E}da z?wwS~){7A%U2b6c@~ZsTN);@n23gnooeK(&u1UL3iqH)iuJs>;N8agE87iS$g~e&Q#IP z@J;K8Ux9cC%==#RcbBbA{Iv|VbcyvZf#~jZjB41ou`01TDK1xA=Z9}9W+asN)3`od zC2@K~MzxUZ)E6SFN=jADj$~HoD{3k2Ub9G9s0u?qYO#lN>fJF}C2)=iWfbKN!fJt3`mRi` z+fD;q=6@P`;zc`E!&0W%uRgtQLG6x2Y|LNkZE)9}9~3guXnmE3oCEoK=H}0Woj~F9I&RKwDC12S*=?>dMNH?0${u`cJ0Y z^#rE>CfMr^Uq>4wi%q73FNwBz8i^vQIByujVr$~x^Ph4}dmyS)5#?5Rj$`Au-3RK+ ztgM0#(gHthc}RS*hQf{%TxRMO&vtyMl+PWndPQcBb-<^s9Q9>zVo>X4bvO}N-|9m4 z%ZXJW*{T~J|5d+}^mTqi#!7qF_clUVihBccwxI~ZDLi?BWE&h&%Rj>Wf3|fHn^E`C ztK)JfD+vz^t2)IGLv>mlX6pQezz7KjF%gvaPaFv=6jddVxX|W8=G`qrq@XbN&ErEs z{`Q?T2EV<=Bfe6actVe&G&6dhkZn zreI}3)_0dMxpHF(Uh&Hu5?Knk9o|eFqEzhpj5OGpJ6+o-B$dXFXpJ+ z)hou+lT9v8Ws&F>0PoRD$>K0{E(qRe8yHZ~oUC{O!nNHGo~{4h6~Unn(i)H^3^OTZ zBfU44R%S>GL}-!lk-jo*JlK}dA!JSk2~Iz4d3<;uhz@8I3&01B&|cd7jB^T4 z58s|D`9JGh$R{UznR^BiKK1 z?Y%B{VPZK;TTjH9%(lMNLC-lA-ruy@vJ7QBsvzhsUVw_S`X!BL57uk7_A6S3XY_Oc zwmf0A_WN1yW+k$iG~QDCk95gj+Psb0+(mWG|AtSwJoR!a57^lXMctIYbxKCYh&g>j z{K6j^fA;LVHKY5p1L(cOpbpkPtw+k45M30?a!Co`4_%>%L8s>Lu&#^KJx)wP>E>fu z*|0R`QDR(f0>ff>EH=q$!o-iK;;H9u2B3O?aDul{jeO1jU<-l8<68LJ%^|G;j897c zr2mG!{TfzcH4^qc+~gU>!Y>5X6WS*{~L-^PbPxU-!$Gp3LKJl^5 zP+Ud^er5(`bT`7IRxPR_<@E2?4o_GRF?ncM8qlnxC4hzGmu-VAj?ng}DOO7*ltH0X z`Vm})?N#&KRMd%l9rMRKz0s#6cnX7QA?^fnfnoUg49+QS{z=GlQZO0QtfOSP$R?0$ zl^;z@;QC5y30bmhuP@A}aCdC{x>-`g^&H=eudGxx9E!n{f&Cll(dLC8r49FFDfc<_ zzaMW_mx^F!ZaWZWmo#b@Hs4T;H?aQNCBsD9NvD7Ffi6*2(Rb_#O3LA5zJIKz-gZKV zz6Alnuo7+D01t;kz~$Y5-`9?pM{Kdhjn=TqRcO;B@Lg(Za3{ z-rn9ckXv2?B~jMf!}DQ46%;XZcVVLJIp$xOqA4Zs;?lf3qf$c|yxsz?$<+ebj87I8 z1JzVr@IgI0^U*#(7v$7^Or%IpACWadu{;vMjyJ&F*5;p47z#wiz1fs_c?xb7-iRrd z5ebdyKS!@$f)yj2&?3C7Fk1hIPOTeJ3mBVAw*3OZf?CV=7X_iA1^)+5ywW~F4T`!` z(8?$V1o9lgrSmZx4OrUgbuK%y?Yjq~f7Ut;kl2*buGN$T$ zU;J|!ASX&wha8y0fRKU7Gc4eP!5DCf<@5FR-OeU(D+3c)S2r3N(}y|lqTSW?(V$OH zoL<$}+QxmJ!)>bgR+9^^_Q!X9a9)3mg(`cy{p(MNy<<;epxd{XJ zc`9>CLodoALbfGl#f3>pNu_)+c15wbQe$GyL!nnHw-1B^9==o z9utuS{ZQPMN~!j*UaY0391*r|3@8lSp(QJzSp<~_1p3;1Z~(j~L%*hc{aj8U$Co%o z#wq(1#+FVQ-|_0Yg=PP>DX!{T7;)AmpN0Km=TM*a+F@M>U0q!Pp$a%E^I2k$VPOPb z{MiGALfr>yZ}S0S>pM9)DbU3VUb4LtA)8DW_T8>+j(X@}pHehf&oJR7-w zDscSrp)rN~nr{mg2M6-l*;%U%$sGY84S$=jL2RL9fS-P3f-gcGWQx%sEXb>Oda?~3 zVMfWPbJ2ikG4M;3a=Hd4-PYJqpnm?rJdSM@{)QA>;7bIXtYygPi} zoU{=q!sS&)Y*%-ec)<|+y1E$175nGqhDreqEMtMJ)yu2R!v-C%=vrfsTl+>Q5YB(A z;@m$>&P%Y2RAJn+0OFsG0<3#ZXR3;udRG_N(1|RP4sX$G0$2G5;2tcXJiU#MTt`Fb z&>php%U;X2{DEs#=Z#Kz_1d`+fwhcZ@xwfEP48pK&Yhazw%V)P^z3dIV`M*Xd~V}u z%h7H{6;ct6&*D~b3q2ma34?=!-fQ!V@jwcVStLvWB zN;ys8scwBEMMT(l(Qe6Vv2|@|&F4{|M+4&zY~b=6{*DF8vHR@jSAQ{0u(cf38jr;` zNBO;={Rq>}fc)+=;0INNbn7uc4u+IbUpoBu5!*>~E)Db@hZ3UCM83E%;?P;e(Ma;-F z2bQyQ!W3=c6R~ zzZOqb?rhBFRP%Pf9{c8!@n@*(r)^S0R!qO>zBAq@E=` z{iDkNzj&Sh2ZHt=>zR!f5_5jT zLx*u1Tu7bg`*=6Jbl}6nCY{x3>F7+9%wK+%nwm<*D)AOKhB@<4D@86h#Ry#HtYJ7y zo&6y+?Jb~@>kK?-q`qkkzg_xak3!?Xi7d;cx?5`!PEcq%L!eiOQ&O(tS zsPvlsBaS#g*ZC3gK@)fkAPRw%c+O@FM(5Wd4 z%ZK!@GQcXZ(dc3A{bwN|e2!tWz#E%L_1j+=+BE+;QWXUMf~R9=cNpOT69U5-gxYrx z-i_@a0}c=+#Rbr%xN*KwDB_@i_S3h}Q~n-UJ-Eh1Nh&~2#Hz5j{#Eyz zs~d>gYH!2r9zuej7htqRwD*6YW0qw&p)tr>_BX)6I{o9HkoUB0QvM5UGiQSHWm|$_AStw zP($b1{4rrEOFfQ?gq%vdy9kV~8W?MUjur%b3?_Ia@wI}21f;W=Kr48Fh)^U2B_#Sq zU^s;0Vc9OSCaqR%dNRn2nsX0LDT3*Y-X0qVhcH-tj$>9iPYYvS!eoakknX2ptUuP@ z|Imh%QFAQ?+P>$KjR+GLz?({pEF7PW(AW>aGNvH|EUDeaksJN|IC zw_=H9mkW%mbjhihw5{5ucIZua9|1RUVITb7bmoWW!V%jNN=8P;ETP0ndLSj=ud1k_ z>W%wSwFtaPEHOJW=IlEqb80*22LM)|EgjmwS#|a)7i476-wBzg*(|7L3?bchBs zgFo}-PaaMEsZ@h#{T5`C`Eu)DP=NWZkOa6EHX`rAkikW_jO4fAT&4OQNfHh8ppj@$ zMZiV8c>-fSQ3e-3Yz&%5Ah~A^%s`CjVMwhn!klb0ySyh3Y%pgQJJ0bz*F6!`a;P#_ z#JgxaJN@z`xY0C3W4T=b(u~Yfa>>aVBPsx=>u8FHF1_A6b8n_WAlB(d572#+ZHNE{ z?b3UQ9Q@AHp$ZZT8hv>JxO7~E>l#m4<%B_kw}4P+f5YD|pWPeS--*~s+Zf@FsgM;S zCN6E~j&?ash8x~Zs*E|wf9p&^Yvs@{$YOF52tSV+RGI<{2I!y$`rur_g(w&1Zo+rP zLQa%BZ=Z8Z(<2@-WinH3+R|q!@?M5nyyL549<(oWQ0OCV-_l9>JS0iKBTaV7CLQt( zuZ+!j_`Bf5j zR3O``-O*|%vv8V(ExULO@71yhMB}(859bGb{ZqT(JT9rx406dQhHekHK8lhc6RWUF zwCb!YtMiiIr3=aK1%vmSMd^c17W|Q*(O*_`knVGAhM&G>%hMYJzJTV~6d0|_-G%ft z@2VEAqL{&4aq-P6~N6=l(3Lzc?wmg!(R z{9N^F3a5-^dKf(F?fSujtC~5dUql`1OvmflvRyZze)&OZPJPu5y?+UPIZF3lO7%J@ zZJ63uMd>Q23k47=ra=z6+9lUzu2)&^aNfspaeL5NX326|8bcJ}hX7i;!}`hqahj}U zlF=UjNQomObzyJFveW~9&DQat}+40a2IjdiYyLk*ID%6=smo_Db`1?3n+1g4{`KhNY z=e|kdTP^V)2Ekc!ZSIuPojZ4$yRSJrUm80uT1=f&J#519DX z;>>5@EVQ_@^w}QH3gnhxq@*7;Tm>FJ-rZm=A&Q)wPtCk|_DtkA3|(k?$rpwQ5-@K8 zoHf5#BLXs6p|YiOGvE_!5S2=1%FDML^jNmE`?)NQ>b1`UnKu~lmqO~eRIgt7rk=^0 zju9g#wH+=m82E`X=Ds+;Wi2!KLa@GeobuMWC z*Tvfi<^MOhj!p`}XRIIjV2KC}vW%iY2_QQsw zo(qDl9UTItdQ3DqFuPVUL<5#($Vi|arbw`;0R#D*-rO;e{^+m?fG#qh_$H;4gKkQ( z`V+CzEKe1UPSC43oGSjYj)s=jtxY1{S1VG;1w1qImff_qBgO0nwMX3v)OgQ_t*}ev zpx1gf53LbCh{BN!&yiM%zyu*FkCVBIY9SyZS}46-aueN+cTESFaXM>jiuGsys=Ggm)18u3s~zKL+Cnmy{FW@il< zO+wJ7!|GJ4^TxbE=&#%c}(7aTHnAFV{%Qlvp;CKpZrtDI4o+1SgCX%R7q)~I2 zqX(q3d0tC~@PQNLkWr$BmKL5dm!+TyilsD&&~-5bbP7C z+BU*yrPb|srjI>~E?vG%4U+gvAjkjy1-Q$N;cT*Yn7TOyILtyNi3*7>J0oU0@L*8h z)kL0hZ>SrJ%xdhWAS?vLXywlDBm{~=mLhrY6KVcfC=_}cK~cRlO74AY<)5I$TB2$B$P~c6N4;m%waI z1S!JMLHZhgE*s%U#|D^_c|hTIY8CulDrZu>x*UN8?yYD6f>a%fHt_agM!0MnuU|z) z-ID@?5=y&kC=~VK!-s=_yFhs4c&O>p7(A7Ya7gO=4SaJkWH*`*rjM}xg$nSv9>#Rf zX%{r?)bgI|A{{C6!;Ub$hW%0lacCmW#8-zxP0kIDa@#4d*)pAMM=Ia%96+gLty8bh zbIF6cob>+Sb*W$GCP?~8e*IgGM+w>ljM%YbDQRhG*GO2@D>9y7ZfN`@}A3_5wF|?jHM@GjGdRK+VcM^fM#8r~J&!3iBtIk;F!YmeSK0_lC(;C(ajh zFoPjcn+k2-u$XoVNY?f%KhIf_eOiZ$y_2**_a-%O4$fz;bsP$$a^)B2{f*)gH}{!v zcKM;nt1OBqkA{SlV7xG>d!=a|fO#vsFhOxlQ6U`JW6J74r+RmD$q8?IW*L95zp2;= z(-e4Gxg#Zj@@C#&KG^U60(e&xluPv&n2v~(evEgZ1-F&KWvEt(f$;X@T!R!bA}bbU z$B_-Y+tAi+_`MG~)$)rd9^p5B^>H0c=!VOvk2VP`g`PI?*3m#Tt^dBYM=sM2Tkf_k$ zF2Ty}ot?a45?Z;mY%tJcN)0$u3kObCS(4`F=2)v?8g@A(#g1reSHfh(%08p_Ko{n6 z1EN0)D)y&slRfK$U>X^3O}k4|RT$2q&nYU}Q>jo}K_QMio`1n?F34E-I%-r_O`3E0 zlt&H>X8#$TL1sbQE>HS&;Bb+!W3|HUN+DlTW_I>GMuRNbG}M!Zcy4t~$C>_>?Js5( zZdS`{4ux6arMPcFTU%S++kK|ik?i_um0;eD9|LpAGF#SPckRIm?r}>Lg}VF_%6x1D->@%ndOcn`UjnH9d`795p+KX`XQMm;!@lwx*5g}j54ZnZ19Lk|VFp*%gG3P735g z=jQzprZpM@i_7{~nfSP=__UeLZ~c1)OdSCX*Ss+W)F1C-VoRlfMp9?lkICK15es^6xortk074A z)?YB1+h)0$evO3Ovq3d%>zq3jaFw;tp={{G8>8&{Uhmvq;wDon2T2}F`a1R;dDcy- z($(mBVhY(J$XJik!M>VXcx1d}ky7y;T74dlwoPo~?*6@DN3*q;8tDtf`$qRd87`=6 z&ykE_zd|)QemrOS(G4Q^5hv2Aeof-B6DNO+nTj=Tv`I*eq<6m?_0t{D9={ue{({z=N z?5ZT2lC~q=OU|57Ri0$hTNDj^m9szZ)^Hq9toooehBIx3>JB)b_4bQtB=WrBnUQ|p z-{1eqnh0}ibYXhE?~04*izn~2$JKppY-|h;vyn{O;H{%f7#$tG7P+?XRKO{9K|9aQ z#bw>vV=4UDSgoR^xSZS>-_)ce@9O2bxw&@X;ycZRRL|~ST=A%N^Xvmvv0uL}nmmd; zy5nMF={dSxC)8QZ$?Zk(tW?eTvdQ+`hjw;P_up@BZr+RJ+{GOZ2;i=@C4`-Pqr}|D z>oIP6tv|To<)VqV%)XB>M)P!wPk=P8KXr22mt3d#o`qP>-gKXs_@|LrgV!?}yDY_N z6+fq^@8cvMZQq|ytS9*9REsLIwom~Z*Y7^yLay%+3Dwh=B$Y_sVR?T|wU7w|lkg;)Bu0(NeW@SZGE|z%f z1$#>eaff-M6t6bVZIr7|A1AJ;$l-fB^s}(#x9-kPac@FQie$SjZs>Ia(Mq65C$yk+ zQo^=WbK&VmZpFKIk$IPIjG~vaFUiOx?D(}!uf7qYy;Ggru0EyT z*Fv!c_oNMZTO#^+;I+zBd!yy?@v|e>gI~K2eCTKl$m z>Dot`PuR$Xex3}THZeASs(b9g(!jt$k98WlAc%QQAoQQ#zkNJCS=#7^qSkBfU{Hn@ z*|?`sJ)%+y#l^m&5C`6?>u-Sg&D=9CPN2)eJ{ntyuuv)kJ7LV)PA67jdk zv-Dkvrb$`l$!|rsYAo0y{8r`F`P)4uWh63{d0BT^?%Tda(O>#yyb*J(^hNQ^yutxp z)J!dX^gg?>bApWHeD1=BOF0$V4hL}%9w^*d|_e1p_+XX!!>s+3V&g(ysT_GcG;%L-x73aPx**Hq!5t?uZH%gw>weCeAS8K0a?ZEi-7z8kBS^<^WFo_*s~ zfYXiW_1yZMt}e~l#~N`A)ot2Z8p&doGBW8LzQUvg@AP|#y*k0$M|NbH!_=d>Zb-j> zU<`k>J#VPwx)J@70r>%3ygcLt_)A@oM{?hY&P-zZ`%+P+ICzb-nTQa)X1dAr-~7f5 b@88GW(!Ly|y{d3M1ix-5Ybq69Gkx+O{6xq3 literal 47732 zcmb@ubySq$_dQC3G}6)_BHi683?W@IAfSlk(47h*D2)Qr-3$oQDN1)E-O}BCpTW;J z?jLvEyMF7=S}alDdFP!c&e>=0eV$M1st`Qv2iQnRNO+10&oq&cP?o@tAQlGr7w-p0 zvEUCeXSo;7T6S-p-CjAGA*sA_wzsi!wy}Io?`r1gWNBy1&&A8j#mhl&;p}YhB*x7R z`|m5b>>SOx9Vb!Gz+LXyE9g5RA(7ul{2-gB4uQ|AnOA%!tL>h&HG}C+qBD-U_r*vd z^Wo8~*KWh7mtS+}bvW{Zw}#&-LF4Ql7hjjLy;nsm-7cvKNrEI-6;`-He8=|i&!;)X+2yhQr0 z=Lcpr?)w=ZK73G^sda(P>U4H?I^JA)+{40({D?;tds*yya*?d)wf+A6`=K;hRC7Cg zjT*l;*!fj^Yis`b;o)Jn<4V_k8k%0W9+j-Dtd1I|)h_3h@y1R6fPieH2G2$KA3oc_ ztgMHh!oqY|@RE48&M%(^;pfvOId>dLiTPet54%OM+jOs{`f>Fp3dgG=!X6 z*H7AhUM6>o3Id;_u_(Dtur78u$%#yK2y}|5P`U_>J2@Aoz&&WAvQr3qAeaTMuo^12 z)DpbiENLpvS-K>t zZLG*{JimG9;UZ&tUQ!aZsLRIJg#MHl-@!n0{>1m2e~+1^3x{f!zLs+t4f=ust+PKC zvK)FCFKm}GAxy64G3R>;p9X7Us6dB3Upqfq?9bleQiSxz1%~HmQK!Fo*G@Ao#M5ZnDY?+#6e#~G zj7s>yQi;#yiGI@VI@iv5HneBB zDB^Qox6%1sr9&T<7}jL?o*p0TRP_-C8&})Qs-{S=@|*W*ON+bhe0ANPB)zyqOc+yy z#Oo^>mc)?}?U%DWjBf9~kTn`wj}@(NsW32-Utam0Oqg#tEVg{w;7V(feoEx#n>whz z-Y0c^q~)>KjHzRN8p-hV&0vPy{Ra;UavDx1tq*$m`xT6gQk1CRTDjp$y|RSiAzjk9 zS6V`_$*jftGLvq1qI>u5IbLp870OWSN&hhK*S&=hhL0T-6>*6AT$Fns_6l1`)r-pM zz@a*AVN@*qZ+jveyiN<|lRFZpJvZSSx^o3Og-H#p&z4`O-um5~LnOl~c>Uj3Q`N2K zB#vCxzwLeCpjT3V@Oj$r_L_~5PSi0=sbLVs!F>~6rEPuwD-cH~fp@W{rR5oE{qLLW z%as|oInG1ZJ;#yf#fh#{$aYGDx^~mGG8@?bu&ZxQ?KEzAbYX=n4K$LT+Kx_6I)#C* zfyA+P^)mSudY6?oGmVs>&9qv?&z*qxZ+x zM(=oIR412G^2Lvrm5pkZ&hx)YgH|$JaJwJp$Mg(&1#h>PhLAmQBJT*#@OkuoCWpn| zdg5R*#Bs!2v)|{HvbnC8aRhC9S&SGv~i#sfSP!blbN-f`2q9zuRh917mrwMVVP32Z@XN2kv zBHu5eJR}YDMSY>Gt{yp3V%+q{xd?Taw-RSc?*XK!-w;!kIw**MPKdVKojas4=+CXf z;rAAG)zzbRIzPi%-%Hr>t=CQsq#~01+E9AGh4^4CGlI;MZlCgIK7ioO`N7JDQ=R9r zmB&uaYF8yM2HWy_ZmN(;6&gDFdZqb&V&YN0{livrV-cSSy<$UsSoi6ipWTZeZ^P34 zZf`LB{QQ=LjOyJ7MRxa>+VR|GT;kG|V%XZ&rKF@j_9cr`k#QT@IM6u#?csG^8)%;2 z+_b(DFUoY;9Q!zKFUGDAK|86*)EzF7Mcl!HW&=`blmkBO|x!mZn68#FRm{&jFA%A@L4rz%GOq>aY zly~*3xkjJt9rF6{+!W7(0WS&)3JXN^TUhS=LN@NT*mCc8a{#7tQm(+9_hDGN<#0~h z@uJA?R8`3>S8{m&YkAXbxjMZxzb_~LeLFjINDFA#_PLDQj) zC~zXdJnYjEZ{oy%k)6yVM?`w_8|=TR-PMSGqe{2`fyw-1|)s zat#|T&>1bC_+d^pR$^>NA!2WmZn`~Puz7L7Dgnl zFOJ!TIK{=MsvDY{n?LsS^t?XAeITHq(k}JmBhy&Wwg0NsSdpbSm!M$9$myQn?PaRN zjqseemSoT=a%#=(&1q9$xY<^}fo+YPPs>sIWgvP#%gPD*!guTHAUaDRwoal;Ap z__V?EnC~lTz&_)w=O*V8Rk`KxqW~7nrK%a%#0$rPMngiwd5Ac}?S4BQ+uCp?O|uDj zv{6L^FMPVaj2yL8qqPu9mU1zd4HE1#2L>coR#vOzNkxL@uaz{D{a}D>4TV&LbLgz< zV<#)D$)Tszw#|w1xvjA-y*8C8ttl<1TlpFLtr4~ZKKEG z-;ZZ!?um_$x9=9ukpI5$tb68w&`5z!ceTwn&O)5AFRMe7^TtRg#k2Y9mN^T*h^5}4 zKx`C>p5Y(nY5|frhpD$DVvYL`DsY&u!nwnvQPjQuEtCBYrU;7L+5UdDZJg~Z|2c4I)xNMBDRv2==Xp!P?_|_~JN?X!B)~{!|%D0s$7eYpP z8j#(OVhA>$F8a0^UYn_@X;_Ke6sE|oMgIaLIx_D~33~Y-I=R!Qy1KEWQ^BVMQ)zGl za&33I?X>J5x(p@)(&jx?k$D$bquy`XtfeffUE& zCla`F4l+%#{ORso36y^ynKzJ_;Rb_lBzkMQMvIf2gvx~T>!Dt+p%y!q7M1-wwDhh)dIbIC8UgF%51xl)$*EMuUNp>A-w|K zcqjpGP*s3~`}`@7R=Hq{bJP5X>t;+k3gf3BgS~g>$b(p9D6PK>AIwQqw8<-%_ATP%FF5 z`BXWGp#5+LL?@2@0|?|Zg@=~8b^E-jLqkYcKg-L^lRA&2t=&)FSsvNX7^ z2NFRmsCo$=7mw4RQlK(AQT6HO4_Ey5eVd&{N;kIsGXPIW62ealvUF{ns7s}snyWE( zle7C9Sve`<3dz6jHk_s~NJ_p``E9-fAfr&*n-0UrV5)IGfpA%PyHs5zO#6F3V3%O~ zc;U3{HIFR@l>WJ$pd}^)bdlD3PxkjvwlWO6?-;*r@cw55-?PC+`mK9ygGwfziM=Bo zQB0!}7WddL%ra%afApwz+qLz^$E-VHHtUueVn>>-_xQ2@6z-GO>+^nLd1(zcXl=lv znqvlk-&R@r> z?UgOWs+=2(l>reqh_{39bmDhh#YwBpj&k0Arfm>f>Up{Y>*(+276LGHyG^El`fj^= z;4%<$V>f?Xmm}qZ_Hr8I5Pq(7k6m$@AP$Xl<{C!vn?~k=;QUGuw0`?NrVw-KjKga4 z90au9KK@s~&sFlt8l6dQIvwTcw%VWtUjemhy9k%}NblOQR-BO|S*P%i$PoMxd zEU`zB+G1Ve%*$Y{L8>8OpFryD=a$5mM=EUlB9%fE<2#NEb-hEp8V>}1&WB^i27F{6 z{`juUgvS=ypQ=2vrKN=gyQ_k3`pNn6xN-9R(4~MZPJ-@AVG3@~*F3O>K`wbErEVEo z>c6sk>DiTA6F=JjKm!aujFkx^wQY}^wue)4tL|maU~^3_miWA9ka=v`B+5onWp1$V zbhh{ENB2|q2r_~CqakHi%%w^jJ^cU&Lj|i698TI;`cy6JeBLVKfrDU)_UJJhQc|-4 zV(8Zj7$gN|33rfd`p|VSnY{xH0G7)(2M6m_^uaM7U5f zH!Jcv=7zZl$&+shnEM}CmsQCd9zT_dIOOI+vTx6M-THas+pp5>W_0_(X=`gRwvqw+ zAi#yduO>bB;yIOhim%hs+ooUc56xeX4HZS|n5?7p(sIBoWJAa}wXd~~u;lF~&JS0+ z{caBFEN5yfN-oH1K{QIB zUh{L>%OXD^bZnSzl%2X_+vnn_R3tUq(#&JK_#}S%hk3;Qt)$mUxri}{@;y?l{DcuDiea)5~JJG2%8<#0qRletzi2Hs3fqop<%aKC_l$}98aL}ZXaR2T=$PD); z>2k2#MQ%J&AL`cp{%YN2QzZh%3FARW8-5r4cBKZ4HO=eltnr)%uI&m-rbA05w#~c- zjBRDjU0C}KTfRW{dS&9JmzPmpq0b}u40)`PxxP_^Q9B_DfpYnf2Q!WNq(n};zP?eM zWRYmxtA}d<@)npPnN63MeKGdl5AKzs3Q` zn@(kLu@4-S#G*Zv1&y!ChU~*wo(H3y~$`NX}zfuD5QRt90e{znVq)S634LR#s=@EK* z15519Y|8kKRcdG1PDoijt)A4!t_vwq6cqf>2f?zNzE}H@QS5MWBMJzreNYdvW&5X! z)q%32-?OGfx9u}b-Y0}tiw#dg3N6_yPx8=vC!Md7;q5LSCHgwg53l~wXELo?JySUNdCChL*D9--ToAel3p7a037|`FT?3FZ7J+Uz zglHO(cs>+y+AEWiOYMQLN=g-}NM-Yj6vd&Tk~GsNhGffxFf1BsN>g#S6G8Nhe<@9` zKH{G^1g-1}sob`>*NQ|Li@`rMORVf?*Kh4s(Vw;6_>@o(Txv`Zm)tYt3ZWQ-{l5F(l*9^{yO19nOnv8Z@ujf8oJl_bUabnfg*lwCu`MU% z%yUb)q}i+9%CGVzuvV-$(C%j%6nNWUMX|X$taWU&Z8kRFGqw(d&|7>OBhS<+)OF1j zScilB*#>Ke`8g3LK@)Ui$9F>e-ze@u%LNFg|5l)Js{Z}Da$J4Dt37tA7~vEliq%$X zgGSroF?!G*^=t9-(H{b+61pZ1#|lOXv_dJY(Xs?^B^0|hFPu0Q7jrEN3&!q;H4?tS zfDT4g!3BTLzrKII_J>?KmP6yb&+J}!x3y2WSnTlu0)O}WD74Bdg~!%f-U}xWeI(|v z_-v30lq2?2H}5fNBXP*NUQAn7x1RQi@8;bu1?Z%(vaKM~WdOWne0>{lp28~h6{eBH zBIbPZ_>=4RCD*uhPAS|$oIks3$slU(t!>ky zhj&z3EAcgDmuX#Bt=)HcrhJ&@ZM(Sw)yF9kD6E{D8{*9IRGBN>a=Nn0Wh%-T2QV~B z>ESPORteC=J)2z3y&oz2BcZK(zVj9MyMzDQVjUF||J(H?Vy}pNC1CU`GYK+=*UvW>U~XmSLv?C*K`hh*@79%Cnn}Q*KHoI z&Hm_Q5psshOe{3Jry<;pFCLqWS-js<=5EOHH#gjK+4?hB4C5*}w%lcS@Xj9d7pY33 zkc!mR-qT}00UJC@zLsyFU*l_D@6oomw{JFbKG$MJ!JK7D^=35oJ^zT|`f;^8m@`>-+rl$X<9747aA!RW%oM^nc*kpVBGsZ_>Z1--W ztPM&d)YFkk*$E#q5ssKj9_yiD3K<6BANJ?-l@BhJ=j(JkZ zck$X}ruyY)9Mc9oDxv2UCnhuQy_IKkw>P=|65h@8wrD4kVsfwidJ=`z8o1s2InXsh z_FcJ-FMn%1Ut!WP+4$4-l`*~WIbcP(9E|^5dUn;)(ZD%Mlq34%m``WiQpL>a)yV9T zOVFPeV*ujT>~Bwe!7(N2`*Lw!5zK_Yf8)OL9f7L9P_^((R)?nC;DJabyWK1?6-SR;aW0xeF~Uk zO&aOiA-F`6dvlAN`8B3WS7TS_x{rWH55&)xF3v1yhleAZS`8h{Yq8$5YY zc9<{jV-u)>o)gH&@yT?8E`s8&TRixZMn{+d_TEQ>kTztp&-iWWqkk&|C_>>=nhy|x z_7QY4?hA{WkoehiMHk=hvX^5!i9;JnreB}I%7vYzDYTJM3rWm+z4At5v?$||AZYE8gDO*#QD-f?&qlV0k^{xakng76+R|t(v((Qc;fcQuR7$NPf#rK*Z0(b z3!8|=E9c0VCKJ<~-&MaP;l$3-8?YFagN=hiL}o=g0~NcH82$y+Bbsi!1ETk;5~<(PyIu9`?3AF= zCNC{q&M!7g%m78nCW*nxy+-oM6fOH{Z$-( zTn9O!#Rm5OPLmDL-elSkHocCGwfoQ9z984surDO?yh8 zfct9Z8-!if9Dmk$3*wYb^WW1uJb&GK89G8c&HFfCs$LMGqX^(lrTa)*nURJ3a`+_f zFLJuTMYgkOk&jP9IwK`87^p$8?-a475mnPFa)!sa#{C>0r zx{xGUf0Afsxa3iHGY#wX!edr+-0Zg%KK|2Y%9g^=TihOFGEcagS1M}oMBV}g&BL*z{?iYjh@pD603 zQbD7C?5pzo{$n1INtiMGgu$6HgT#>O@)67)O(~Hbo{W~{vHSfpNlBk9*`7UnL+;%5 z3vQg1O>c%$bri+r-xs4$@iR;>>vftINsM$B27Q>-?ag&tti84tYl%j1GM-=??{_@; z=g$;-0PrewCeK=D2*h50x(8TDaWAdTZeEqZ;NY#|=KJV=Lj|Z_H0!J*D;_@XzfiX{s#&YtmOIaZ3=}f zmjBiU@ss{Pi-fpsFQ;au|Y@HhY ztHp#zp->n81~Ae02?DR{(N~{`2R>r-%r>OlKER&OxheVXikc;;tP)pnL7$ zLGh8CDGecJ5_uT%+yGvz&(;b|TnkqnN2=FSJi5_Xr5kB=EpsT@IJ zFNi1W2`W4|cAFo-+(3%bKz0T!Y_?{$lH=aIEP_nX>AswA4nUM5_o=B3w{`(N!VWIs zpW^!`#5y#eVbXc0p*WQL+f9~tQaCrAErtvMev~x0>3GcG;Gmm3u9ibuS~_0ZuhG!& z`dBmlN#K*=$Vxto0eVsAwa6>5MEI;m@?QHz$Xo@I8x8D_B8DJJ*nSE>$!WDu5MwR} zJ_r8gpo_gE@751wN?p$G_I8MTD1`cY0 z`it7lVs;^fa2oaVr;Mn_K~_s^t0EYy?s6$OzXc$KdsU^qPKvy5&R3%&Bk2bR2Zu5t z1eOzJX7OWpYII8fT(Y$tlb(cxME#$mv^4sBy^@rTc@K~?`OudDzoS8_HL&R0%~|y= zd9|CnNn1K6a*_{-Bd?RC?lkIYF$y(BMU4mz4Gpr@!5quONz`n74YL)X6|nu0&wuLl z-VaP$7GR_c${C?sdL>4j09^3nPQ>N3%d_EXY~_ycugr!%4!#xq4@7l4`P8C#Z;|YgSXdmVrQ2 z#YO=MsTJ*!L`R8ZaYaEJ^I>!gl6-?0wQ#o<;2>DPqMy*!P0=`B17FD zvXB*!KX{0kn1gR7)X73bTa2XWnQzG|yw4pw`}$&QT{b62JcwCY6eML6IuV8L;#xws zsQvsqPmY9NPgblKegz&(*){23(Qp}7r^&zY0fMPccoW>w(Qz5nXEEJz+?!>+&*T_Z z@JJJZZQ!ulj0Zmx5}_Mm(V+VLxqwszejF~64*Qz zGHZwg#husE96Kw<2B>VSWn;LuCMyc90{eS=V*)S<3Y6F^Du9UC_&iYvPJM!*kJJ8X zeO?vI_Q}(yqiY2ZvRE0uS3WYPYrX_1{~{bz;$r7{3(5c9? zE@{x6(dEj&V*G?a%rT)hfLtpwZZZP$@55wqH(hcmxrJiP6X78xd!i3Mp)!+XB9ERt zc`|Z4_@wbo=Y1C<5MAu@SM7-c>53rYj~=O|*zEhXJT9Ri>6)I`65#^*pb#Y=!BlIb zmN#0vxjg5_h=K+p7Yxv$lyvv^mkdbcswPK+x;4MA6)V`G`CW(_EHUbJz!DMpr>QDC zy@)@L%JYBSTvqMQS&#*D2ef$4&Xi2Cath!s!00?S+R4MsimXAdVoM7Ei4+RZwUmqD(%BPhslO2W zU>qE8jPJV@^i-G5fPrK?#$MM1=8w_~1H&E@q(Tr1XDR0Of8e=auRYapPNk=4_pEbX z&zQwAq@Qn)`S}?q86@WY<7;7J{Hefgge1w?XnZE_+{1Nm$0#QDMrYD`=!9~|Y5@gv z>Z9zhN}Gv{zp`Bk@NTwRI{U%**P#xWJSG%l0VF~ai&vR0`0cfX^ZhB3Q`M>*nBh}mv^j4o4p5?y%+jXAh{3#qO zmTGb|w%yKRd#jh3#qrW)dU%2=G;U;~5ZojX>PWQ3kj8`AObopeSX&GJIYIr^v7FSO z9xcr4bVp~?U(eV_2b~S-p_Eijh{kY?#W`US^$h1BhRzeh53x_Y{2gRI-@70*$cV7L+i+4tml+dBVwj4a;X-r(kTo=;TsZ zBS5at^Or^&qCh6&%;}dn48?wj+`mvQGl_B#Twy|ZiAv7$2e5~~CSQgLh(!y3MklV?#N;Y2?A7(18hd!hI;6lvBNZLs>WERW?quE@xC zO!A@MMP?GlvtZ&}NTn*1=qy|!dt&y3uQ$s+?_e6PrMq|oT+)V2jKk7Ual7s34rd;PA_P-~==OhdQvSdE}{dLlJWV+>#r2wteHiCpEp_QvD zm{Tc&%B3_=Ku1ihwdSFHCcBA=!h$Dn!8w@_a)Z|)3Xy?l!_u-zk$;SJQsgsK@9{J! zzpFzZ-}KFNxSaPZ}jQhR9`D^{M&Xp zIYn@Rq+FM}qBJx5VrAo_EZq3n^X5|X-5XZ0_;~$w9f!-)J`uB!QC%WkQ`nEc0klEe z>f4@g>HgOlhl?-RxBHNZMCWpt$aTU+`lu|3gjpn7j8)pbsUlipE`_2u?Wd9)*eDy1 zI9U@)Lg;@@dhAv&;m^294$<^v*QH5hj}76xo8cqNMdRe~| zMNlN`UD5Uw#Cc~<&!{AgssyLhp**(ujwAa<3_nNM@WSY%8%LUczYNcq()?~jo zr5_u(BvV&siY}v76DJZ?dtv{sAi=*cq57Q= zdE}DE^V8GAn7zn@TU+MwW|JgCS|~Y{ux*i&J|p)7!ZpUA8jwL{m%gelhsMfbp)#t;HZU0?gQH_A0E~n>=YJK}WolILjHqHiv2na0L-9njfT~Jq znyQZ*7LI9tmXh^$jNl<@-DSfqUKpHujd$2aVj`7_H z0@k%6OD;MIK_t~$*ZT(c44?3cOPgFuo{-Byr_T9%ceAx$3 zIildq)|2zqRPICTyHN5agOx`votZkBaPU3J=40-$#zzotoDw>6jhQ+(J#y9RZ?d$a z8x@Jx%)h-tlRTdT1KgX^b*x1PWD}KI0wPhpR&3dVaFH(Z46Dx+*p4k>quP9qE8dQ> zf}}m|eTlPS%_5?tjsqdHt`PPG(+J5&3kwTvs{@STLr#+l@mRX*j3pMUokJMstQX=a z*%_7?Yu|$ADQV6LkzSSSC)LEQ`7QMVtqQl}J3r3k#CSfF}AEUof!c_ zNbf0dR;h%we6o~7b@;gkSMYjGuexl+Ga`v`Bs;U;3*jz6tkX}_4mmlIA32w}!a0|A zaGn-1si4CX-07l>ZSz}VSY*w0LS;Tvd*I&B!h5q67D@FPhu^vW*(ARIUIq=120LAV zK0_}$auq)t8K_7Pc_`c*{`rn2;#P%+5yco$&Ex;(6#k$6#>ozvi&}gumnx+N_*T%j z#UIcwl~3h)9lcI5tiJLfDx)DfWhze(%ip0}h915Ps^<>f4xOF+B#E^1t4vIFqjBrM z#T4w{rMgE06}JPh3|s((8kI5(APUSb)M4#<3Y7iQ)Z#BSRHnTci|2W4{(1nFH+L*1C8hWybCs)wC+@aUraYeIK-&8U zv%ptq*U+&bYT|Z9}(U5F{O$pWP)NX3;d4i(81C$Wiph_qM z)#}n5l)oR4Ejc*|Fa(~-zyp+`D}Wc&ZP)=uy@Ib~XhVn;PfiH1i@NMF5Wda+e1CFtV)<-; zsen{WTH3hcJ_G{k1|>u&XCHy|)%OdUX12fFuoO_}0+aqH?9Wk5_#MQ%WAMa;@cGGd z(%L*=SMTin!m=`u{uEF(=*QC^?h!91+BHU1SdS|Ko?^ra;eD{mm$&wK{ybD?YIjNN z3Lm8s4pgFom}}|VbG3As3=bU{Sx(9&Fr_Zv-t68k1FA!}>Hv88qpTN%MMPEL&+o26`1oH%z z_g9jvuNLpCPpIb?e@$+W7a27`w=vN+&SYg85mv{kH1x%RS{Liio}RD3L=U|d1G}W7 zf=4AV34^n(quF;g*)0oLC)ki0fRMBT3jD9YXP8I!_C>ixloIA3JrX+FAB$ry7%hC1 z0zR5kpbNJ;u1vsKoL^mKNl#C2O+kf=g$Pc~0aAui(zCn>NR=y}mQa``Nd!5!Q!KX!8n}GGM-rSO42>Uit^2xjM;(P1G;$IV_G)14p7Y*S<)KcCrDmJV< z_T9IDyOo!)br)3Hv0uNEAPnqbenAor%jVXkib{!?w|3*j$ps4js$VQv)s&Q?xf{=( zAU;?c7@Ix0*x3u!yRf^J=B{QI7|-mkTv39YuR41@BG6f*X-kylR1GFq~2ViJ62Ne5g8O7C8Qxl#JS&5 zA>XcCu1hEG`kXqy999S@Hzh4f;9f7N^dBL~Di^Sw{*yi>2O?VW`%!I)yoApUO!iRz zJ+&(NLB7?S#{1Ojk-0_YGT~Pi&hw$6q1rTzQOdTqv?Gj=Tt&1ID2tlGUs;&#$eMhg zkd<>L|1v_?*2;O#``0)zAZZu4ok^9LCMvAsHa_IXn4qBnMzdcx8&l{On)OhqwJRr+oSBnf@0b5|wpS)aO@kNeOD!z~<-BGL7-&Ne z^gF6q`$lAq0F3wlVZ+P+FI_YL2RHfDl!BwIs2C+^J=OsP0y99)+^3`CR#a5HPfE&+ z;9QGeim!t;d9*pM21f75Q1I?cPVMtj;C?LQdIKWk0kE)l)q5Nwf=PCtF^Elw6kyi% zm30UNhqB_?%nF+cP05|Prn!m4yOyG2l2AI3hEhB?SoOb~yuWAwj=+M^I*=%aiuCyr zT`E?{YO1QSp!-48eJ>3#P&ySUOLni$_7)C+-$)5W@*^)D+J7xZQLAiPJhT)PkCBm) z`9S9jrMPPm_Ahu~+7+NCV?ghWlB#O>8*}riI!aoA8WB^3c*hOMz1_g|&A`jc>uM)s zXqZ$FV$K1e@TVGZ)!ObbLx}_vCw)`S40YO~dPOUPn=Ndyy-;-=p zOjX2%{W@Hu+JFZ$+nKImh1K1T!ga7jKo0uGb-Tk}v6=5I?)PM&x z+n%gA0E&Em{XvA(g#xYnJaS@OYE%?M)6GR8zy^?kj)?gi;D690N2*kM{`~7`p`Nmh zO;JmGyK)ShP8J|S3YZEaBGhVZ?pSjVf}rgDt=R?x^!YiT(`3UMM>+Hy1QVyEVmX)* z=pdzSB+J@@Fc0*lRu(S-5W#uk22!3bd>h2bN-I%|{3g)SvjSw9uhY}huZIC2OOF6L z8^IfSuluu2ln+5&%fJJ~JU*!Tk8KLWKm=<4Q8I z4~%)zemC{l(TH+8CL}!V9FcKr9L*3*@fidHR1>s$u-2?DFHUz$#Nrlj&%8JYrh-EP zL1CyEh)qgeh8O#<>FvM-Bl%U-+4^il-78?61s3Zh&D$J=H>jPgU-C2wiqNESEiEnE zR{8n(a1rLNVOp;(-on_J)Kt3{jovjAuo*C;K$*d|@&L+(!ylm+L%>UX{PgZiK~C|L zfKDET@JByb0=weA-|f)w(BNVyodJTpl>}Dh26bx}d;9MQk7(duRy^p62Kmf%Wu^7F zs%`zg{PAr@{$o$ue=|Wh6R=NiD8D@a~a{- z_)}^+1QNPFP4{9na6%ZH!TB?JzJvbdmc4b|D+`F*5 zebk@h)qMuMt?&&Og#J}#(u@*>mH`dn$5=0YmI#W6SBD4O02u?7CV3TDm#N6*&*D4PcDBbrsVFGmMU^U|LNO;+G%!RR|6 zK6Y6x-*O^kJUwfZBs_+`{kFw%uCD)^V7>hFJ?a54Gy(sUgyzn_v740zYcf;AYM?BX zPI-&Q?lA|)bFgPK<(mv^o%0_|437Xr`$$w&)QXL8!HfgPdI#vUG5D|B zJUunJLKvqGcwDz$kYY(+Zv8-Lc=@{U1WyEj&+f7~{Ptw~5V&OvkO4j*Y@777@;zd36_x^`{95qRe<^~vKKlLev3PJ_tO6B6Z)9d*kc}aImPSoW zYkNgi?y45gCwBofMX9DHC>1^Jz_0 zQAFaX;hr&QtDvtkZt^ut$Q5{;cn12W^5)$ERyx?OnjG%gK%rpeVKK1-RF?@hE)ZN|KuMs8fU(#1 zcqsb-J=BgIn@lJ`Si6P(eEug3rxZ>HI7pP}=sFDJ#yhqmqI*X}vYJvHv$(3Fl6>W} z4`^K~G#d(l05-QH^RP-#vW!&5dp?I)QEIV~2@L(5e^~2OjiN_FsXM!cgHuZ)dZ2F4 zCP(2M-1v6fI=`MKHN-F9Yd%X71qu^B4Q(8Ue#T9NHa;R=tKqe<^t|9M-OrCsi3yoT z-ajPm>?T+2?TN}dvWe$mQh7Two`_r%H_WWkIySuq`la-%?bD~7hgHm8RyLtUf*f2KFfSX(cS(X8s=Gnz|7T^cV|zh9v;C6^zjIzuUmy5Ve1Ka1bV@lEpyN0vXD|K$?G}Q-1t}Rbxme<$dB# zXHvUtFU&sB}CyZs~0e8 zfaDZ5%wwSnlC5?mB7-ekszW3s9wIRnsd+l%I46HFZ*m1X%=m!x zGtuBxW$3kC(VBC*J`CyQ2LY}@PX+OQXNmz#mzF^XVjZo4C}PfqcaSS1b(L1*UIh81 z0(RU=oG}8Q#EZEUa2Qk;{c+}$F$UXqw(t>{WXF3H1G%!!)&F7TK20vUx9qyp4)fA^ z2+tJekH8J|fQXxY&4WJ4UVo^PtP3BLDi(^4V z)uI=rPe9`uE2xIG&KE)7-WMQVD67$s;xRvY8v$L;X?&&?XO+L1UNWOBO4O; z4elDyCE|8QwVwxq&T$YSXlg+FJK}vS111uJf;u2AzAR$O%>iA?;f9ik6&H1iz63lC zG3&#*-H1luMQ{9`$UC%l&}ib_7D~y7u>IVEYE%Jq7BchjsPB#fCIbXCNsu%FT!Bbt zSapGFoVBsBmfYswJF#l@wC-VJYl<#9I5>db&dk+}hWmy2i1&GL%_PEs6QzI_%nV}0 zwELPx~w#6D6t z7LkFN@et;yjgTgQ5D?hH_3xb}fGxtr8$i7chm);|p&BRi268pTTb6m9OLC&);&N#X zDy*_kGwwb@;`2_kJ4h*`OF&CKG7>-xOcZ2LlEM#wtd{X{Jw#({136-2JzLgr?x#e6 z`+LNXI>?Q^8NkB$0-6?EUt2zfz554#?XoV@q(SP^?Jqnmki3t1jTldWUaLZ-LBHw$ z;O|*J)r;4Na0DRzPy;Fp;&X{aSyfs|9?T$u3M;E(Mn*=KYygJP%GZB`$`@1~YX0_Z z7s&F=ItAKM@?j4YK~f!I^7Xq|&t+z3?Cog@deNnr2&ikOq@aI$?+A9aKE zp@SoO1cHTb(F+D}?7-cI`=1qjbH@!xfVOu;6PO}^zknCagIv)P5iuaT1rU&VSLffO zxbaAVcyoOAH_@&ObY0=7{NA-|IF7xz1+8egrv?8ePh5prKtP+1k8kb$__#B6W%^P*E(S8=W~te~cyKQ6xPVoEdAzyWq4RI6 zaRy;51>w^m%vuP03cy`rjyiYGYTM4xETXCP==rxaF&nFLn{QysI|nl%X23?&#uxnW z`zQdk%%Ya^6&V#HPv&aUx`b1xK`m9vDg_)eu+d@X9q9eH7h$+ZKBTQBy+9>DB?*Wr z>Zl9Y$0lXVYX>%>Z^^#@_W+EDtnOfqj4lB}wzo3$_Gnmb0~FBV|Le`5>$|W$&d5`7 zlw7#NWmDG!V7QT6(0`ZrpQ)by93jh&RQv_dhzO?i4G`uvf4o(Rc^aE zoDcMKHs~oM#&g*fU@_qskbMA86#3x;+NaN-H9Zzzy?PZ6O@{o}t~_vC{`K_9lPvT8 zlyF0I#AifJIa=@-(Q0aGnF0GynR!2L%>7|dc~T15e6#vg)ed&pod>l^6685Bq4Z!R z`NVfYPtBXLB7L&BQ4peHV(;o7ZLa2OL0NkflQiuDZNi3Q@nWTRceY`wpsBUiw(;B9HT*Sa= z)aYm_Dk*8o@*T7ys&i3LgaRr&4>Cdlp%4Fg444N5J5e|nwZRLjki@ZoCOP`qd@Yu9 zP|dD2F7`~v`nzB?Ekpr{!~+C7kGRirM^xZnD*m^jp3`-1C8L`PU`K)3 z$v=t%+m9r1w#v?UbkTR=7y2m5u7yGn;yi?9pf1D#3B?FdGX$Oe{cdNTh}(8Ftf%mY zPg`RV5D1G)GBRcicq7XHi1_$;-I65$McR^VvVM~UHV1Bumz4Ow=?RgrhwzP~L|n{w z1VN_!rX%w8u(0j4MH0b)CmuK|p&~yi z-O|J^Tu3ExTq7eha|ImY@)gW`;n3#}@mmnzN0tzV=g!t|-zb9qRn9dmGXV~0Igd|; z!5Js|00jzI4z-BjjcV=H>%K=*?>E>)v*grPc^|DQTfV`mVCL`@;j-GPKU9sm5A^9x zg;a3 zmj%JpE#h|Qk1I(6L4=4X1~h6`P?(}UFIU~2tWX}klE-lYT_d42i)KWCXOiEYZuF_U zbCQAsTAmCO&71C}13F6)cr05f)i0@wXq*v&F}OZr$C53F&>eJ>8-qHafc6-xMG){} zS`mo8^CU5suCpr;R}rUWa0x`b1;E6LCk4@J*s}@wehYNx#BKUon+ZOhTf^@3yTRjY zZ+~~6-u5vlEX3S)iV9-D@dBNKHSKx?7C)Fg8bjO)e4$c*HRC3E~3Ju*-ka*dIX+0LP1~gQOq~DEsWvzc> zZ?gEmNPEk$s@CprR1qvf0ZBo+OGTs`iG?7wKpH6#L8L@NP*6$f?v6z%l2W2HNQWq) zu;>Pnka)+$y`Sg*zUMvH`Ed3ZKkSRS=9=rC_dUk122bIsoGhZ`;{{d+Z-OtPMjTp= z;?f~^tAk_a`u#2%GeXCQ-uBoyeaxGp`H}ZIHecU1+%3e5x~JCs)=Hrdbg*THRxd&V z0yz60njZ%$VXHu-I-JTv&4Yc)VU<;Wqo)gwJ*}&6#8v>^PV2c*YTX?#kl(nK+W(-p zl(e+R?3Wq=@5>{WfAc?{1;~N{|OlqL zC$PHx;iMF-IcMzT-xr?xZ|U*>GwtdB{U*E=cs<}%4K@%%?zTg@CY&W9*dVgBlja;g zQRvDWvrN-h^w6VMou%_>FnRo2KGXtA60MLw+pDvWkk!ugf0s?R-;?gklP0UWFQw^! z_MFpi%W$>|B+)ADep_o-6zRDK;ZYH6(kDGNiczsIajz#{rSLo2_FXPnwMl2rG!F;~ z3Bld@gHnA|XP}6Rsf;8Lrb=Pu3t4D=KH9O~4`&mL>DsO5Kim&~3XOQNs*`B+wkNQi z^z_tV5mVF~!m!sSXrmj1$)o~>aMpJsZ-cCioS6}OSSwOlO!fGO+zb`Vu>kT1C6tPFyPb~lL)&-+HIY6ZRCmPQl2(8N0s@dsDzWQKj_QV zdnxxCH+RUe^f&8$Y-=w*hjpE`@07V(KDAN(W{>WsDe86R{aOE%JIPq=<;F8d9&$aJ ze4>0hI1Zc4-SpOVZxkl6AIi&WJ?mY$5b9lvz=y9db?b5-2gqgVuUPbX2R^5%7sODf z>9t-nF3v`?eC_B*s|m132xubo_UaVVi^ICf*(tM{=WYCy%%a1ojE4$ro&lek29TvD-K z-Kl+##);vIp8IF(LE6*Z-lp?H#Fg*({?7XFamJUt?22zC9>!@;UlYr|KRdHNd0YC) z&R~0QG^gU_6v0c(;Ss}q;*^q(GeP5l%#7&{jXwf!r^%ehWd8{E&q`B*zdEz`$Mso2 z!09{@PYFh0*VApcX20bBAqv@{?KX1@t*^(Xb)I+Sfi@9};tt3+oj$S-q#7o$E*t1!#WeyxUAcNjGlC8cTCMbMwL1|u`M zMqcT;uK($+yd*%o>FPeK^6~NZHR1*m>bYW^&GXWSBy+QDo^=un#P#{Jrw&*|#O}J? zC|ZEWe+QH807_IOM#E#HjjhCC8np{`RwjJ-Mvv@;Hba*&7I;J5ih&uSx#*Gh$JG6zOWN*OlJ8=rCFaBdB~_hlgB~x;(cXgkj?Juh3|vxiRU=t z7EEt`oX~M_z~5-em%~-gk0+a+R;OiQe0l2dDL=Ft@xOjK#fiEUo$qv68HK*S?mECB0}|X!b@*ShzQ$Yq?9*dTmr(exwnPK;+ThpIr<- zWvSl;e=5f*d(IRSnzz)Mg}0ML&_r|+2JBxG!5Bht`DT!vFBNjV5lRo@D3!@5?h%h) z>C~fV5b&pK;1Xwri5LhHvhq^V`#>+NcN0Q15F*~xt$tGI0SX3;&#VI7GiF54$|Tst z^Dp58h!K>%g*a=gb$?Wg45}CbDsljp9WxL~&Z9>;D?3dVDI>>r$LTQh)89-#KqjD; z89w_-(B^BZXE8*#Aqiy;KvyVcp`rr5qyRGvH6X@J)>C4SDuUy$$_DtAzSmaMuUE#Z z^R*viz5o2tH&TJZ{MwJd{sLB@ynv}tmwpPxap7SCP$l!VRLEf|$6Gv#$qRiJn~mTg zMFzI>ax(Fpuc%*Y#GpS7#NwKZukav+ONl#;_saq9wziIl6Oxj7Bwf;}iC2OYZrzF! zxv*ipVSqYG{IT zO-hXOM$gyS0EPYWT+mtMairBWXm&B8XRnWg)tPcybQ z4j==>b>K#G&GhBSJa`$oZ_uF_K0!pp*7&943kVz2AZzduM=d36XHOno(Qi&74oVSo z)ewK}>zj}S$cDO=L{QH;sK?#@wvqE^LEKk*?zMOPux4~}gNoQk`inguozZt8v49a*@)8h;QB z2>V_w7&_-?QsM5Rra*K28OW^}X*swBV`ndCI@}E%2)sEVq)bH6YzO4lE^x)>(4Ukz z&1)c{8hdI?&JWj0B!F)UI^h8`026lVpt1>5km^|5#CO4vJQsF-r5`#>8xY|Vy&MrX zw$SnG9_=V*i*lZg=OJ3_KcD%iAUiynNnuJO)5s|{zB|b%p(W^}>&mF`JP;9>XfT)2 z8-8#K1J_AA1j|?;PPh^j@O$PR(L5MUHt&tC#o-P_+5qs-fTN}ujso(cIMhpdEafbk zZ}#qJ%q|bH>sj8JpVH3Ij@G(Q!!zlOd8yP9uU^2vQeZ*@E}s3h&m~XSO`%ib)3VzK zVl)aM9AoPNlQE@-$NZxM-yM&7m6=CDWp0Pp#4oolR=AM6(7r$*?SGHWr>og$w5PBz z2jws`D{DI>HEQjC0t*F#u3ythEI6jVTb~)+ZVN5y=HZBr*9bo$NjI`L`Q=fZZ1PhU z<(st95+1=X@y;lQG^1HCGV5XvCAYWT8To3~ysZfpmLKDegWYTRFW|`gU_5(LcrvSD z9}!jzFz__e>Nl&q&`cw)tJYVS%7a9!Eogl9u+!7J+%B@^f2oE)ln3k7gVO!l&*ITh zRTqAdx_nHur?8_^$hqw_RO2)st_4YXt4#=~R>d93$K74;czTf0^QMw&V~$dWHL*3b zs+mAk5DnS_4?`OqAzk=Y;qB;|17xEZoB~|oFJKv_EK9q-v0uIMI)DQT3NkX>$z1*- z5Hc$42Rr!2B5&hHVr|cvn|p|H?w zheq|Ti#(GvsVhR+8SU-~Wd|>1=HLDB^y6mi z6-ug4#6s{kgF#Sf^%l|+RAr9VcEZp31B!{1CvuIPDmBad^L^Gqnm=Ewi0mCP5y??p+uh%oxwezI zZrW<3$cINJ^8{9;kT?z|_oj}1(7rX|DDtjnQ{Rjf+gzdDTHx-k9I+5KX&vh6O?;D& z`F`#{T2c7@e$fb1$1p-^G~m&|JKDa=IhwJ^i+o*tCB-k1>-<9PSvJDdd>1;uanFw&Mc)*2o(~w^ z=;GXU(r}p{uK7?&P2lXsz+rem$Abm2Dm^Vy#Ol0w`h{J9PWWCA-52nZnnTVQ$>qyj zBO@ctUGa)$4XIK9yj=aB7jb?nRT4SNSbm#gf)1Pt6K_{)%iMkN;q)tPU7YQh@O0A;ZX^zk8 zBC|j1xN&PEZb~)bhI6~EJyLP5clhS}uIe2e4i&pk4;$vNNU&&P6keg8zOA=@+1cG4 zGtg-uoD#{~OWfcfTm5Zghp=!O+JW199zIJVx|Lb9tiNpw^u8Y)7!(te7XD--?bCu) zmPxOWn8#7xdti6~)Gij1JHi6UxZ2?u3{as1&e<&yimQM4vRKLAiBVy6MI7XQKuU8; zN?z}w^kXw>3$HgL)k4~&&=ZnXp@|$hd|7$G#EDRcs<&Z%_|;zdI_oxC}kKAa`L)*Jdt-G0tPPj`L!JCAkaozl?=_kMB8)&86iHvn+sM#$Ho zoYNAIE(;IP=3EDxFR5g`m-1|C&ISN^#fSBee}0l7GemjsuRwg6kn?5a_}Nu6T`S+# z?r3JBz!wjp-}};=G~K(XANADNYb*_4=kprQ+1N4z$Eb1*_UFID1kT{+w8mo$tGxGr zwbIW|q(7})zi=o-uW9*6i%g~U8A#hC0q^I$`fMRDM@vC**+l%wQY`8SA*+~y{mfQL zu-q0QAtAvI7H!TR?aL@9T8e|cD?)_2zA zRE+l4O6PS*4U)qCfb)r7()Dy$jYn9bhYwhBj~R0L1@3cwchzpL7kM@pJK!plQ(ogE0%9|K)_^>EtsJ*Ghx_T0z3 zWG1w5&~n`<9{)9d8jDHz({Kxc+J(oB0oeHJoNM!XkZf&{D46oXWC96uV$unauAb z)zWpc2(HJ9TRVte9Wf%dQ!~;N;^S*dF9K?74sYspNZv3M@rSBeKaq_woxYwEqFiq; zI!yn4THN2B$gJ!;C1qM3>W|=3H^1!Y4k3!X>CCBHTaX|!pk6qlNOblW7iAR5Ns;8w zfA8wbB%h!Qdu>X*rNOHFb$1kvwby z>5YAne)Cw0+F#TLVx?a#`jgYXh=BF;%h%J8u2?9uoq~L@N}6nUFWE@roy@xE50(Zg zNlM%^^K5tx&k&&rwZf0i&Ia~GVG?R=@l2rqeuaFMz}}5~)_O4!MMB({y?g>xBd*tX zI|w&^eX*Ucd!f={P!-(nyZ9#G8n4iXapN6N71ZNtuHyE`>(FhAZ+} z9O>pBI?@-p+CO!FaFLcbP1+L#COqqTy5XVmX;;=olUtlR?}x|_GMcxU`D88*s!IyY zvF&YiS5L23eGeitY*mJelJzlc>YR7W2Hb+2$ld9WXn6W5_>8nNjO7(5BrbxKD=3K- zQC{Zon$Y@TSnop+^eL`5 z+W+sV+v4#XbBR2YmojWyj1^2zU74sg7cjuFc9^SVMRsHEE(ugecW?8ns>>lpuB7LRM{d(5U zmd)itPYyBRi{EUd?ycZK-F2Sjh$;uWS}SpdfIw4t>Iq3#_uCYM8ixyf?phs{mf#?% znqQ9E_)=S&mG(++2K%Bx==BRk{DV&JgS-rG+i#`U^ldW>Sb;Xgxi_%W{9OD?yntqR z#@^X>PV3EB1Rlp=y~UP*iW9Xsem<_jEjgG557ow_golKH(rp}mEy8SOjfcugRC!4U z{W$AawL&*wu%O`EAv^I%yK0KMrYix?h?KEGO*etUg}1i!e|9OIu7qVnJR*{Kfo^`w zUDl}V(-lVf0{<%yHnOT47Dx83!)D1-->`% zoJECUyu}tk_2kE7Sm&X;0+KtK3+e}7<=KYo_6nUKj0TsvQ+))X(Pb(8l&+-1*72B`AA-`(=f#@RNQhQ_|D3(Tykxh zZQRDslMUN@pY2PIZUe$>sibO@jEOV)-{4|l21*5S`%-o$#X@MJBhM6oWcL{-rTm&+ z^%L9on@!ElNDH6^)^coi=)@bb)$kvaIP>iPy`zMY%fanlv#VE$l!_my^P}pL`V=tp zk@d@{9@B%K3VrjD75=7dLK|H}iR2!`Pu}>$oE0n=kHGOX;swnVOD5tlvLMoZQy9U5NJ!saeD+mZOg0q>(cWU5m=U+ZDKo$P^Ey{7T4>(QH*o8v&%#0d4K6JtXfSh_~xZLosluc-scfY_WU3z5%NR$ zG8Jp*u2Ow1c}-xAmd(+bs&Xpk4^FgpGWXbDZ+Vk-02`QJJE^tJorKyZ8+ws+W%1pa zI*BW^ey8*bjvO5?yRLTelKjvS!iqZa>OS&#_{v7|k!#H*$DX}6f8S*VXG+`03}Tbp z@@OVN%5`q=_yzQffA-wKwjumEQbaGTeFaH?sX(OwB);^xZy{~kw8r(4>0~~^_SasM z0V)YkZ1G+9_q2@!M5fn@w$KZ$SiU390avoBaVE(f&GD-D`J;WkA2_c=OHnVI8bWUG z$S8Yi;Ft0PT?a-9bo`)+|EvRHb8nb>L~=4OQkVX1o5@A_g8w)r(TK$u6vVf3F^RUv zCKJ{=bUQzwERx?4vs2!jbiLg&b`F|r z2Fi)j2N4L44dB>dnxhit_0DeQCyN6aoCEXb#w*u*IyPe#wa-PQ(9yDkBb)i*1w-~5 z%VQRQSZ?y7$Eo|P8v9=6hI(T-6m_J7E32N<03>8$?uYw69kqLA3dCLg?qTp;M5i$ixqctWZxX zn=*Fl#7fiJFZF*7A=fXSWBTk7Nb5~J0DTfjqE}kXjcj=8J zsd*ZFDE%mYuFi7tg)FfjT^^VA2rRW(6|>iU|Qjc@I?J=ZJi(Q)_j_~o~y z46sAHp1o&TAi~XeR8#%6Gr;x2a>mhLe|1(l|@6Q%^@|5U0piQ&WOw;?3t0 z{rR?i+)pe7P+O9~se=P8>+=dfoE)wOH+UkV zNW@0r0rsI|xOcW}urpa)9~o7FM6FWuCZmom4!KvIh&sO7BHiuVwQMRvb2qm^N^bb9g*3;@ktwUtB#g>1Gm5kg?+zff5Lu zj{|T%h)#0+DN9zNB_Jb+qu{}4Zd>k!CnO_{X`(lQP$w);ijA@ zu?WQ6>bgO;c)Y^wDy%1jtL$dv+gf>PD78b*hmC0NRbn46Zg-EeKO{vr%9g5inU*tfbN@<;Wev7)tomjzV@*+(X|o3&av}nL(X?zN%xuL^DXISJ zioNPJdj7-qqq%>Z8i)+4*aG? zcdcq_-dg3zmUl$VT?qn?NLONSjtRtU(G0S$b5DAvs%OY=AgF`X!TWBB2!Om7k5y(H zCxl&hH|7!xHx5_N1gKOOIvVaiLD`|B8qsOiDCzrOyrUmGu6alkxz;gf-4Eo-zo-=& zBU$IdlyLOqaAu0LB9Yrsk+=)#Gp1i>G(QD_6f;-{AZLsPEFj?^*XjUkD9t-WE9>BB zrG7=yENQ=TOYYqR&Ef~Vt0Bvn(=qlIm1uVE{v*t434Fa#+(>n+-U(yU*3Yp)kWQ5k7t1%s&tE>4Ls_z}Yqjm2~(XIm= zlB*g&S3pIrRkKXmoq<6;^{XEpZ74%s*Ml?T_gAN1>9=Y7WVV;rF99)yNC*nHx_8Z~ z{>+OAC)}-Ex+d?q?_}yiWuwql@}541Hi<6G)B8D1r;g^j&Suls-pM5bcH_IrUb8&y zk=g0PXX9!7p?7I@Pixv5a?v5fZ-ir_E#$(MOTLtW&&g-UB!MRXdtjKA2d}gow@ksi zqj(E`F|!k{fe;eLqgSdXJ5DF!)LK+83HF|#%Yv#rzHna#bf`gEv6xKM#?N;JS&M3} z=D%7Vvgo4ec6Fc<*`4ow(N||=VXr{hdgjz#YOoMzcxvRm(E1dli5*9u*?*U*pFvrD zNObCCn1#g$tz9HR2hA{zIMN_TsTHi3$3E9e`N7C^HC+TnK8~mZl z@!=}T)IU`rH6A}3-yNL&{Y=!;(9#05x|?7jCrA`-M)B@Xc6(u)nyLd-9vT;icYesJ z&?0hKRi`1gGLxbkpgsC;$KVn(a7nYmrI>G(_xt^Kb(O-DC(bfCgr8FQ(WsaJQlS{o z*q|&erRU=jnIiE`E;NQ2f)zl8%m~t=Xo&6tkDtzf7_f+(PZ4piRuV-}hI|7^IV8S`8Jtp0GtwsGkYZN_ zRXBR&Gf)-c+RMGjUK{qlaMVRrds&i=ZL&X8z3i+%!8tXbJ%*Q3Zf&4u;4CyZ>7c%z zi+)2A^Xi1>w&ElY%5zV;tzWVb71Cnt`r_!e95?S*=zRZkH=s17MayZ+p zGCebc0uP%3giaA2l7cf5OoPb4vG?;_=2VDg(BaA27Nm2PY~8K7kxPBAob^dxl#PFF zETbW{d`Io61YLqUl4A3=%x=Nf@~_ffKH52MCb$EE#m#IX0Z2VeT1d6nsSi$IIT$G+M84{a6?QeT>n&d3ROAXu#T) zJ*s8mq{amnZtrxZW;_?`WG@FruL0K;PU85*V?1fdHel1jUCO5KaoTa@EHHvg> zR;M=P!BA$`S^X$;J94EuJWOfAS+`baC|Ub?lAy8nhw8bgAj2r6k4tw0#lo&ksZ;v< z?5gM<>@ONL8qep~Kp>Hz#L+?a-QDh*dX{2EQ~UmfFa-BN+kL;BTFL-mmASbxU^t@# zX~EQ4{ransYWI%iMXiju=v*dG24PvW$^$I{nKBRR{JhNB*4|2QYtGmUJX22(lPqA$ zmc^eyo??Y5%BXmfc|6)bS0`tnO2C{92RL-#7GMJZI(-V{C!BJmy*kG@u2F^R&~rGIp5 z5}%aKh^S5)JNB6O<8z6@qTrAI1}2&Gsneta%8a6@5pnndN@#4%MRhL|mqf*X-gH6I9wp&D#&S$8rA3Z~&?AN%i{7 zgNsd*#@pi8;D0~pK@q{v%zC|%34V)t4=f@Cd0EG9Q8>uq*t&(6U5xEl|`aZkH>&Jr@Zr0O+xvqlLG z2>fX5I7#DIbAh)r#kf+_u+%J+LUWO*DefR4fDC8wY!L0D4zZmNh6BoLe7#R!8aDBS zvAdjJqx~z9YQjTQ#^BGq4K6qkQkW0!$6nZkfp#SY;QvWF^H_>L7=l|?pNdVoU~x@F zX3XL^8COLWOt7601QLih89h*_A?@ku6siNM2^+6+k^G@nVz5S;eMl)T9#IWl!EsT) zHpz+k{3Gl|C8_s15@OK;t+1Xz{mqo$E^=4{nvXC1P~PMasWMeKt_19y7X3JVvwOZ( zYVgxa>WeL@A`jz1fGYh-c7J+?$hqU$ zTHIB!;b|J0HY&Yuud3w?!^*Pakl#Cp$$nmC%{lkge|I!*leNmEW9Jn<-gF>Fr}ztD8XDmy@-K~ z&pwEeb_TeCk@}~=V_MaoiC%}e$ov+dh>-6%0eYtP57vx*C(k^l->_j9hyk*hH`@|p z&CZ^&y?Vr7d?+c45PJcb6B}Ge}}SzU^~x z+YU+EOC45C+6@9Yiv;*b-8W_!s~HFAay|)TSF{uo+jZh3d1PM`gdr6h1XQtAjJxH( zV~!V|D_uiEs*i&G3{D}J@!i~pp6r=abaD(u=@c}YC26FR1hGR%UK#$9X{U-!2u z&q6IVW%hQ!tOEsRBjjvSx^pK6@i~t}&gl9!-<7FZEVdOQ#MwbVRHc=xCLK&C$qNpD z3*9mutC5l{w^6XXz88E|!*2EC1d&{p?BVKk>5})ai&5n1S$^iRSW^nz(W3D37bQNu zKN~azuHO}xS9qy7cvg&jU+c|T^$=p6Gl!qbU7kJiS4k2A-JSVjpkV&=MRnoQARQQ! z2+3y*py&l{`K}zI@J4+4rFIiW(-AqDnYq_r19nu{*i8PbNM5Q-DzUW;`Q5eWb!YV6 z;UCe3oO_^K)EhwiW7|%u5NG+~`=b`K6R$l=PwKT?TM|CK|8A)D(H4~-Y1Dl{K*4h2 z05LoWCDr8m>GS7lgm!{reEuD(_+vDj*3b0b!HKDR`@%KC%3|{>d9EkZZ|VL0&uX7s zujd)YD;zmGl)U@N2Vd_UtvF7Oz^NmrhnIL9S5`YnaxH%|oWA?Hk_a-L$+&KKT%4y4 zIwj%&2A3G50TzbIgw|vVi$Wrc2`qo4GUjgx6#bN4ARjrR^^tPs6AU^40?$>Bw zTibj@?Dzd!!1A2Q#;aSR7SIt`YHEiYxHpojg&pnu2T+Gn>ygni`m)kca4~ghLBbv4Yk46_;dy;Z25DrX z%sqGTdhzlYb@U614~^E2yeEkL{#2`R-K=^Y^{1;1b1;jQ^N7EJA!i8F5p9}$ANv15 z6cLxKs#eiCWq*HvBwwN(QZd;<&IG5q!9~Z*^SiG^3(q&6+!S9wm~iNR&T6??tfF7R zU$U84=h(|`$WoA6_AQn%z08VY*9*er-V4Tz?uZ5aLlQ_yQsZKGYSjnNLOGZ|vp87F z+3#KxTW=(NbHNvpn1@G1WPuqJVTX_$m3IXUIBST-=SyIpZz~(!E!O1hp`U^?B@ zP7{<7VTlYW0*!z3tQx61#3*A(**FW3=Bp5KKaf_ni0FGgcECi_0j%ZV5oB=i zf#D-c;fkx=;r`|wDwzKWJ{Ud{${o9|OT<;D;hSIZ*iET(x`j8^x+Q<{iu*Nm#kAS; zxb(tdJn8Mh*`n;gm#5#=l~n&628tNtxVe%@O4%Bw%n}5QBzYpPP>9ODEs@86KG4tLkb7NYQzvwB!@tI*!rWU(O zw=^4S5wW?x5@!pN7?DZE_0LJSKm-(AxgKoh7EJSoD>vJF_ZRO4lo5H5F+^vGYlEq@ zXA?3v1P|*WxZ1*ykXZPv<7IILRinBk(IhK^nu;sRp_#vy9CTemny*clT;cSpako6e z*6ew9h@%4n%b|2>gkF+C1n zX%Jd?{au7dC^o~HSopHz<*_q*RY$=*@Q~MPx1G|h2`kW9Z|UdzQUjc9DR~U3Z{-gO z_e(#Ne^N{DDA=r5D@Nec)k`5xv6>tONj}dd#1k4@+o0_fz~kU(CBAQc9tdpTmoLAH zeL3gjMWRmJY3pRzDmZ2xzWQv=qvWgOw*uX3S|4r(MDr46pih9_@~OXU31k2nIWLWR z+^8F{aDg4a-`Dx=vy4{TwUs!D*%3>Dbf>!{u8dC}urbcu{BYuiDm|?f*EK(DHY=s~ z3RBj^oYS~31U;>_F(!^bu0l`ZdaqYCu70-?8>^%%jr3~7e(zs(zh7UW3NzcKZ2u}| zsNsfs5Sy*>0R}EopV46CET;a{JSy)cIdfA)(a5uoMk0ngq4-Iu+=<|OQ?}Eyi%f;q zx2W`M*Zpo>sx=Y%gYY@=J``9`nY8n*4%fs7Ef4zTaTu9T*EyD0A60q3JB!LCxc=6H zF!+{~);P};L9V~pDaXsS_!NTfs$^Os(u>vD)mQorIj`$6Nf; zu^is~=k>NpC~G-s0(0+)buo;TiF@njyX$Ae@GF8M%)qGGDsI@;9=6=VXZU4XRSkXT z;^;R{F(D}tL8+_=tNCxa0i}6{c?LK?!cV(O>z0e4&$pMP>EWJoSpDE$zKCTRlqqtS zFQB7kwO!PnX`%{yuh%>Q!;?vJrfzD3lwbgCN-%#ACZreKQkF}eb;w8d+IP%Dw1TM8nyYa(F$0ShMglAIi)LJ40c8&t78LvTPSdn zbU;bTm?In$`{=f?^ohgnSC^`zH7ihf^L-}@$MGJpSy^zyvEurJp3i(aDI6PQ<2=m9 zz@tp~__HXBoa)(T*fNY4!~lnInCoE}JhxuGS{DHx;Z^|Fo!bN@A;rEkQCJf!(-|V+ zr63~$j+Gsx+K+cO;5TN4Jdg4w3+K5WX(TwVy4SJcTBEIdd|M%28(#VS*Q zD3Dc~-LK{RIWDMS@W`kR*pf`|M+4u+$rDmJJHNJ13@dya$a5;|QzDxmNDa^`NHxwP zRrm>yD8L2fN0V=mQ2pV_G6v_n=xU87Irm+)Y?FPE$iwUacBBSuSK_+Foft6(A1#mqoQPow_S$-g~2^rb_}lL2MJ=@(MScP@13H z=>lVxVXgA;56?`$#NAVBal`xEo-tkhthODL2%auvLaq35$4`cH+wL6hhv!;RB~p85 zaAv`U*#SY*k2>GUEp8z>=Qwjn3^R-#dg#k{frA_7({lDN!bl6dYP_WQunZKw_3n5z z;2cW^-+iM(2ne|;wGXZ9To%8Pv3xYY>*~@ZZPa>cHS(hC-K#~;qqkL5c={cn2h6+z zdCBvpNlD4BGgM?^Oifsrj8}r;j~wq{a-bN0j8A=P*yhba#rx?O@7u!d6wt$feDn=| z9VrR8kw>1l!uXVWluX*1WX11?OhKl!(bbe{+6(5F0g>5wYqAgS?}c~VgIbf5lUVSW zA_8NO4s+kEh13EPa&lHEU5+P7K-Ro$Oa+YejmrKXQxm8l3=i0m?+4WW?SVOC-tups z?)61^d#vUQ)S2;pb6KCDXRYaZyHUoavlbs{PqZOJ{m<+XFFjLj4p9ME}T*$~)>13lIS!TfH&@ zM!OokHz%zQt3qu$x0}0PWfCOB_{0QD+k`&SSkt@({Y+zmzHG0;i9DrBCU5*}HAgSP z7&L@C|A=h4@guhrY9hwOBx?zt$D@N)uR>n`yBc|NQV+{mzTL1jQs60uq{HhrU#~tG zuc@*Qliiiq3aEOzp587n0VyA27UcfzKGkF1U2PpP&vNizNAXdbvq^{dy|*<|O|363 zt=5~zl}w`6iuMcj(sauG@l}fM#WDKvsh){r8Rl)KYSAya zKvR`FVOPC;vuY|>bjc`Z$So=1Vb!-6<-CqOV{SYbl0?CBZeTNxL|Q;z6dz)i29#Ha z&_RnDcfGtcmejR_$I1ZLL@f+{4%`t>M(OJ%wsuLy7qNUv3lEcc@)fH;L{X`I>$$$r zo~ML;TB9k0{nQ|&=UMpbMyT%EFNp3IYHh!iH?UQeFn4GiJaeF`t=5UTVYK${vWE5CenMo)bzmCgTkBSH+XT%ufW0H-2HR2%~T`=>mXwgEf2BBaDi6Z?eFpcq>~M5A?+^z{Vpv@ zJu*^vdY#J+n_y>7Z6|)RFrXw}4?c7>mq9|yCF@u9FmTVBStSTpFM686cOT59I^A(k zdOEwhw98tJU8l*m{`e6yEqT6?n?CpA$#;<`^X>{V!oa_=iHE9eb{R-nP68u2&^y;Z$b2;Tl4hYc@1ck zv-Db0x%1_f1>`&A@NYc2m7&FW1<0kmY-BVI)XzD6cW&f5AbBM7kjn>0K=$p94hYl? z4fQU$kL>CYDvBXWWxo}hm{^e9vH_=Gg$#r+{o_w{lM-nMt3x}K)XJ(x>)jU5Q$m84 zw2iYbEUK9TX4Su)h>HB8b7r`KJ%)n{FDb?L#m9Q2Ht<*_+Tv8c`*=R-9U!!$Z)EkG zp`qYHhRcC9T>)Z93X(sk3gyaZL|bz{zeEw_=iil@yZEM^<@rY!aM*4@^3>N$xZgjB zWtP`~_8w!vv=CzUgI)#1m{|CKk!;CN({H~JQg_9h1Yumj5fRl_2ToNBF)JRjSb2*s1 zsx1N(TOX6kG>Ejcq<7$`9$alrN(SU|)Opwtk)f=ZGg92Ni`b=EFZrmFmOaRXzfnu< zAs7&C9X2jn03O*eSs#1stw#!;o~y zD4l%q{w!p0w|&aDz`u~B9Bm-#P%H4W;iQAh5odf0}?Ybwj4CE2xvor4Zs(CKfNWv`xHQFFCRaH@BQcU41QIOvRE!Ew^~R z6+enEDB5+&6tg7Zs`Mm9RqKV|xV}4QV&ss&y=$KBlpTXHca!ESg|4BrE) zDnMMAmKU_bN4B?l><4c3O-z?^k7VJjdXuk;=kO{g(td8G;NkPTx7<5I5<+?ROorY! zz55=wyH9WIEhoWDBCOHZ;+KA{Dn9m_XIY6OnWFG~uqPZ_lSqT?gx%Dy#Fi1_Zp#sO&fZvS_)Pwo{*Ka4sCYPk40%F1 z$qtFtBLJ?Q9@Lmwo9ujOb%6SvqNFt&m{=nQgKZ-I!K{9BF)%vhm>Bq|CPQFwG?I3! z4Q`+0{3pu6XPYIB?A0U;?bV9RXHia1L#>JV&hubFE2_cJYK-U+SLj9!Dz@DDaqB3{ z9N=)sf8ny!G;p2gy~f>8r5}dsWga?Z?}`M=7mtB7%T4E&u=wjG`p@Dd3QnztKUJqc zoQj^poLU1y(>aF#?fAIh2ZWn&-ow1#_AA;Nf#;J;&|d3F`#Mv>B#)3pZkkV!&$!2s zRNzr5O7K=5rw}JQANw1H@{X=A5k9lw=zzAO{1@Ajvzp8kYUSbHR?|a$Dv%n-4C=sW z*Bi>Xx{BhU9~(lzavPEVINzLyY9MrW#?PJVf4Z7Ts7O+S`Aa95{|~cv*SG-$U7>j! zp2FlOaA0L2mHsvHrnUvpbxBDYD=DNVCGed*Y}#yC7ZY&+6A9{I?s1MyPPO3=8h>T= z7ZpR3Z=}RThY#_ou=7mC`pWzdKoqm^*q-y3yvGugv zi`wC!sc)PJ$r#bt{k;B$#@=8j?bayGFTsZY3hMkgx0jh2H*i7E$-%+Ve4Y~_*SqoX z@Y>)6HwQu$!&;R>$K@zQDzbqHA3=SK(vS}Z#fx8kaqcz?AK#sx*uY+@44G5rQXLB< z^G&h|Q_=e)jlaCxS#=cN%JJA4w0^~>5C&!Ee!{F_Mfi>UD{VFU4td93tvwB;Je$~H zr-Iu_XQE&)Vf}zN1sYei2kf~#$;n1)=h7j8?7ID=T-9cmC^Fp5tuVmpIZ+(t)tS|= zxK$o3%2!o`?2L*~Dh)1nP+B1E*RbDS*S=r?zZhV;HvLAKJC-$+~EDCp==b?6F zPpZRek+iu$P)#7B(rduW`&XrRRuf{P+EY7;8z81E&--zqtR*zW7N1_8!;m3~nihs) z!iqqh);w`%gK)olBMvu=U_)aT1=S*A$)+ACQ z)w3;zFqii1X4^%gGSSNcM$hiQ7$x0fP4!@K_|&W0f8)M^PhXDW2!@87dFyyram^wu zBaeV2m{v%}nMRHWFyiJPzIJX-62!URo9eXL8wlJU(@%zzeHr%X5PbUn>yBD({?XRn z(W?>Xj)I9Eh9GCJ&s=JRMbzV9lv47&e?bpxPBfENaOJyr6atVbhq?9MLm&`sO$J8J zQ)4SpNOU#Khi9hXH3)@tp82x7oH;Zbi{FNR$3A7IuU)TJHB!nF=#g$)B3ByPh7$ zegSjWD(=7LuHcXJwU4D=2Z#zu^wghtpi-&E#K;!%F)lPT_4NmZ!%A`?d3pJEa90)K zY=W+2zgiy1kv)^X;lIoA}Q9w*Fy9hkzW ze4ZpOyTn$+`$eh)c!KDocrf1gvr58jDekjB^A8Z6n+rU|s}W*ycwsU+0FDo>hr*os znORv>J}BI2Bx(r~aJFb1o0Vq^8-#-2M4ZzJDjYjox1FRK$;JN1lntpgw3+&1r{45( zfxtjZpw2&62tp^^xc4ikE2*Q%e5A=#^f8)Q^tKP=hWEJpQP}ehm8V}+YCC!WsIgE5 z;g@Y~9J#%nCN)mu7lWpRq{v8l)`mpxvgX|XUVD4gqajr|+h^`ZF>#xIo}4y>Zjg)q z(hgMfk){&;Wr>7XEcGax^ZBy6zIyd4?>yIz^i}^|4A3b>s zZ2rJfe`rln6-ljyEJ>+*N7%oc1@)qUD{5-U3MruBSp1 zt<$e2WY4o$c4fT06LTMb6$%|LBzBbBsOB!-4Ax_vBW?o}3+QiG|MwJe|6q3axH>o^ z&3X#)#_#h|Xj#3bz{JJ|8~Jxd=bn|7+cR0amHm%Xx8YWtF#IO-b93&;N@2fARpt_n zDi<#zXqAB4%A<2wiCIQ0%rqYpYd*!HfX8kV?fBgWw3x--TO#d>$B-4;UdIoQiy4I+ zg0^g!o0MmI^YnW*Xi3#;WG)cT`!-OJsp!feFs(j@^A#AY$XwKnqWub5^yG}F{X!0$ zl58ZX?d-4M&P^PUn!_av@Pg zTcB>4jAV|j5Utbz=;Y-xYx{%invczRMYrjq``+c*}_xQLG1w;&dLN`K9NMx z(ALug&9$9PkpwVq5g!{l)nWveg}xg3$kN(xlW$JYM5*q>J+-nR_f*LrATmi2&ac&C zc#NyRlePN~u3n&G%bbpbT&qT<$+2+@3ka+V=2k5BsZO}%Ad{|3bnBa|JL&0&5{n}I zyvRhxs~W2$dosV_AdOApFk8wXcPcHFUd_+3k0Z1q0x+5MzPDH%ya?)YtrkdN914eW zt7x5!l%)<7@)A6MMOU> zF!JD4_JxPVuk$J_f6!qA09;QqC=X1$?3Z!4fB0U>fs>(7ZRC37HQ8tPA_YFbqS^3+ z4>~bH3Vw`%?Xz`sS5;5uV@6&#R8oBM#g!ec2dQP}giYNFJSrqCA+Z)Ko?=ZbnZW!X zxz{zS9wsIQ!e1xZAHKFe53S(XQ$e((#A%Q3;T_ihGTgG!bpH+ZM~}#A=_7KJX+M5} zCU-pn3FBvX%dZ5+oCdodk#mw0i+)+C%L7WDZ6cdy{2PpneA<0uS_C-g98?3nT1|pA zZF!KpFz%?PZ>gFjHtDfDvGAl&Tso8?1kc{gQ_fAvd2n{bL)_=Jag{*|i2igHE`1F8 zQayptcvB+qG6hxozL0?DD+)Uc?LjYNaH;&L%3z)7n%~4_^cyt@jIHNKwU{KfcBFkK zgv=S-ha3+HL3E#5Plp`mHL>C@ntA58H5oP>=5~WOUbUW6xUb?UJO%6%=Y_Wrh(Nsu zK{Ev|5*&bM3bBAGnuM5yUC=p~X)r@(%zFCdQRanFo zz)OJmE(GZE7cUa)vM~tAChNb|ButNnli79m^!nyj7)U-5d}_Ms=W-UPJ@qG41Y*Xm zY7(faKCKdPA(G?yu~H0%i3p@n-}sIGN}=TB%nj)!?lrslzqu3ji#y1CiOI8$@vrU) zafp*zX0{dlajQm4byer>ArVNJtWzwxAm^lg6gunM6>l{tg=pQG8A}ND8u4+~5y^5S zNR||;cTbko8s(Cyn7UnCwmhh$G2XD284 z%xdyK0~xuF=ZL)ql6-vgPlAi))^W15pQ1u;yhe}i`&5=u8ZqNg+2rDlmNCNB>;>Py z!w7=a>Ms71$S{NB00KGBx;jhc_7C{uDmIqnGxfFPHz#NfR{}>gknG3c$|f zhCV-+$Yp}Xb*xg-OR`7w*>)Fci)NTsn@vWQ^T^lXb!L=Mp+*OWwqA@NrU0#+V$kys z{p>CZ$AABS73Cbjm<^f#wh>QgKxi{@jalTMbST~X-#UUrAc|{Y<=jeuNvXi70Kg;P z>3f>tZbT0kyZ%>eXBib`+x30L77!7YMoH=K{`|rQ9=X(L0TFS zrE3Tg=?0}63F)Ep*(cZiUiY_mz3W{MA9S(AVdgx~V;{%f`@esGH86A=`F7OoGv6Zq z`|_e01A_VaxA`p_2%M8g&i#a6Nsg1go3O7R$_Pe<1NsR|<(LRdP^jaQHL?VWUI)LU zXvq5pKSlb^fdk%@tO!V>lW)+=4C>s2@nR7UF5lZM*k_(+k?Mxe)1SDD_Rmx*IK(Ow z@=^t0rCGRvd?bUM7P~kg+q!^G&`pTiV5o9zu7jSDq~wkPM8c!)EdV?+j)qS#3Md#p zxfZ0L(#5IA0R~e%?mHKe@juZIspBcIon-!(ljH&`5RuQe$ixNx^A27?NSqG(M{zNC zTuT_yX#xXl>}<@i&%ZwrdBy+`)G#w40AwQj=l>xSx#(RZ7-SH?M$lD?r`lkE`i4Ex zpTICf4%hD0x&|TSN`al$SL49|yAbluf4)U7i7K?6?*I1+A>Y0+yhZ>z(begV>Mdxo zOd%3?gmtSkt&y$O0HgG*P_?~XG&PGOZl#!%G7_vO6kIc?z$#|tqSGR$Lin>tr{-Ib z3d?!X)BKCs{5HeFlwhj_P95ZK!y_YlqKA79Q7F5ln#iPY`N#Xt4nj;J~+~J{RzsDML!&e5#`GpoayA_?^nzz$Zox3 z05;mMmnsx*b?=AmaE>7f5wAQ`r}bT07O5if99IT<`l?yHp?g934xnQZ_~xvu-CYZE z@51JQZH~$`_%pvPg%xZot2oCR@%_6c&XeWBGRrh$=i}YuZU`UJkm|4QFWesFQk;U` z-`qEN2x?#GFDPb%Inw-4nI@QosaeolwynY>cJ`_AbVwiMfG33rHBHi>ouZA54Nfe< zpnc&MBf>N*0ZdPzPx;*g&w|Ivf0}k2Dl+lUDs#l`@t!9WwS)ZcnfFx`Mw5ithebq~ zIF-OWkh%aq8o;&+hK5Oft~KAkKL#W4pCiEDzYdKNlWF1$7b1XON$ITzBle|dxz1c0 z2TM{)Z<)o44Z0|(&~8#39wI*R0NG`24NxlXf8`Cc|Epju@-W7EzV{nUyC)q~LTe3i zz^*-V<9s@O5<%)_t~})O*s^^}Qw2TPo>+1PKv=SXM11aWChx0p)71bx^m~Kml2o(K z`<4kuAjEmqG$q<~ZykLQ!~7aAm+i3t_k^88unT<*Xw0kA&~eM*vFcafzZM@p;(a(Y zUJh%E$4i4vr#7)3-oX)O>B@;OYF+cqq1S9zkKA9= z$sp&?q3jHycgVT+Zi=+Y*yE-4P7qz1GYv__>^+85&q5S=(Cvwe?!$OIt=322B&BNM zo5k<^hNdEd|9z91dB6H&L?LB3qUM zvIVXyy_piWukq1dvhouR_;E=aHr58RKedsu1++mTu2^kRn=c&f#ox=nEN?r$bk!Zp zYd-q$=t7w4CXH~T4>!{<9H3^{{i1Kz#`ZVIU01)Xp z1n8NzB~0xv1VNUbe!(aBA$+zY1C`-ARv){r;rzqhk;BIbt$G@`YxD^8bmtH;`Ixam zTFa}&0irfbtMl8`$aTA3v{;o<0*)$^AfZw@l%eEJ_D#5)prJ)v)AjMs>(>9!Oyx7O zp*hstl(%UUBl3?qRmtfhD@rJlZ;8`+@7ZCp1o-$gPE|8Ap@3V9QS1Fh-y4SnY>0-E zlOltq#yVl);VGxDRx4zH&llC*pEN{l4y%)LS5T4N8J+!8;p;Y^s}VsZ*bq&TRu?lJ z;VLC3OFp2*k6z>Eksuy+&3~+#{y)vJXr6b~zJ%F6@Nwd?>}~(YL~2qUtbE=>_o?>N z@TtPUvSOH6k-c~CS

    U`CiU-^pA<>F2VF(rpxXToRZr>lDhpb8I+EJjzpcPjr7Mv zqNVUQ_hds(w(!>@>bQTs9Gr&X?vVdItDd&iZ!=W_xPj9tKUC%>NzU@+v~-d-EtMGN>v!iB5{xpguWAnjXte?MGz24s0I zVgTU`@dGo$1T&5oMzBRj^f~L$`=CNXH5&9e$6pb6fwjZze9LvFUZ6aWfx!)rq_i~E z+fxf^5Lf(85nf&PT(Jl4JFl;=FU8w{?rwEcn@i!JV2`PxwkbF84-RH?3Z4OGngI+C zV|Tp(!`EgDVipRx@`1ZInVIXy_W(D~7f2^k+P>HS0K^(Z&36hy}MeNN7RxWYZmGijLw>Ej&B70n6Y6Ne&THn0YZ z!I`VKB2#;ajfdb>Xly5IakN|D0i=d7YNDqReNV?H89Z)96@GOn;_w6%nH>=@k@)`o zdkpT8y3zrxT<{GlPddmM;zU0IcOY#>MnEh)A^Db$p z`CbM9D9*9H5CD;KoG_SIzicgf^ytx?M}X(+h!qIcSU{REwq6B~q>2*~5)$2ofK}8) zyuAy1Li6)^o>o-p5>$Od=7D+ygQr21ff*lhXEfC#4#6jc8v(H24}PkLt1-fEslcR# zss@ltL~x3E8CMBJ_c0*fW9p;BXZQ7zBYOV=V-JW_et?wx(!0F@&h>tJ!^+0i$^rE9Gd2ag$eJ`;u^1}l)&7}5F)0QOeVKhR zuo(h-(q`PG6dSgj88I0U_1vike@?{TZO#?)ZIb-{B;#&<4p~*zP z7YFf%Q!H89232k{Rx!-@<;+}%!1*B=d*Ak1t;5bHXV_zK*FtZhKlK1Yga z$uCo0lq`E#2FKB!39v-L@Lti=(+d|0-g99IhS>7RH&^?u$i@dy`!qo`X%Lx`3h_G< z3k;PN4Z@T`>oARdF5t@=;i!Unf;`~bW>kVVD&Ph0KGKv%{1py&D-U6=@~rFr`tMp= zbK3F_IXw7>zAXn^zn5&|XNV~1k*>dmK$jW&vOw&EYNDhuzgKAg=)UB~S zZno`MTUw`H(O!!1_Ayq30@dkNPr-i%BXF`>fNiiNGl0kxBtUviOK+qwlJc*q>~bDL6sCNKej;nVB)0%l-p9}`Vr{UB`f z-fO{R#)KGJ`Ginsbh{^nk@R~s#>#LeX8~6kmFHL3`fW;QZy|to;gug`t;6a?R-`g# z%#$7Ir*}&Kej#&C(S>_>2$vqSJqaICikQN2wCXtKCV|xP-ZLp2z)w>@t5tf@zpkzh zA*W-WP`E~(p>wH1m5ze}VxU+tM;EeEhS$hva)N$^b`m2RI~B)*0+y@%N2!pH|353#dm0U=+vFVR9aQ-)O4)6 zpfa2sUXd_ckaX(=x!m1WZV-&G7klly@y?ucbt)0HI!57;KW*2Xa3$79{E;bSB|m=# zBR~$o4l^xY5Jw4mXtOk=)85e%B)p;zMV0fbmpcRqjhm5Sj2?r4aF}m`xdQ)ac$!0J|m#%^9A7#@j=f$E6i+6)4xur4buqj13f-FV&cxKfo zuNBO~qzGMytR_C59U14?5o5_}uY{T=r(;6Z>nkCIXJ>?8XtzZl?C&bT`-?3MQulfe zis%|lz!prVq}%V@e|v&h&=I^<-q{4aLVrmUNPm-XYGNj#wyH^pKA!US>6zrdY;Ar_ z|3z9JlA?R@qUZ7AAg<@++;|c7n0*Ii<^9 zZzNtRDLCI~%sZ@ygLa_L2?;|(!^GcyG|t&zRL-R$HT9Alt6XN$3x?|EG0$9 zvxVSqmE5G*8vD@CLh-^Iy+%J1vsw4`x3pJ1V3y@^Mje3{BmsZs6&aZt8yYz22s1J= zJlMzzv?~^ZK};V!P_rt z9|80wB{DXWhlkD%6=Ix{hNIK0jj#g;bx-%FPqCq}w}lUEgy8ex^xtiC{NbC%GK0|z z3n;s~<8t+05HK}aKg}q-xP0>IP8%c13Zv&fpJNA9s<(DZN_`7 zp4-HoG2~F1OhK^yN{irsZIXtV{XjbZ`HwkNfUTe7ky-^p?}1RLi5yueB0()vR9zI}Q^;{DS6h zFS@28k2LUjjCa;%L!+bT=35wIAvWgfLtJP1S`2{`XNXIM+#T#D7;<3@0pY}c2C=5{ zsLs_kVSJ#Q$M|a|9Ygo~>%bd9RKv(40-l)x{0%KF7;==@!PX}PF(eZDWVSmajDqMi z;+&etUz;WHmuA0P?Lh;izEeoq<@b8%1X8M$u+3fiofH3AyGZFY3Ef?o3kU^a&unfa zq~bT=lI0;beRKN#?*`N^7n&p9hEU!>oc?NcVOPzhZ|Q8)w!vZMMxf)@0EZfcR1yjc z8!U{{qe*s$2f3h`o|T#dFrwakI6g+~iqfw-BBvzSdYj-d%I50o3cF1wDKgOW{(C#X z$&u$hkb+a<6y^$xdb21ti{hz)S3Fm2UhMaS>C{t(BwZAo zrUcyc{AYj)DMra_@xIh&N-5}+y5Q-F$y_-5q*$N2PJb_$)4IbCntbINYSjGJ>aJgS zExJR%P$5G#M&qkrU+UR^B7ZTfp?ZU+pHf#VpMH- zumDo+$B26^khu!1wr1NCeGqHd(AZdA!!(W<_Q~5?hCvht-+rrv&RgWU}F`eK1t*k-$>ed30RqeJ89Mm*tDIZ?z*tpl&;Sv-1+`=v?|BQ(f@)pM!$Sz@8#d>>0sRFhZHHjC1>v_msQY!C#^;97 z%|54z)cQ-poa2Q9pVv!)rEljmv?K75=2Uhwz%<=14#pRoMkglJ%t|+WrV%=^M3cu< zxMp)});{1XwOaVS_FMvv0;omY*4EagNM4s^eYCZ;b>ccZ2=o-o5Fqbgj@U!!XBS%F zepUE&(ujHKv|OX4H17p&Q6l6gbDL@LUj3tIA;g~c_Vz}Db`Tbgd>%|N6&!$Ev>yka z(nK{d_h>T%%Y@5&*RR5au&Ep*ec5X`<+egu79ZJQLF z0}WvP^Vysb%|e;9u#&}Vs%^v|_vcC7fHU?jseZ zmj|uC@ZUt1qBgW_fZKxxcds}5K`^*52Bjw+m@uxL=K?RdebV%+Q zI5^rbl9Ov%x&xwt2-Xw_7}RDt_-7E3v!zds>WIhrRHI5fID(HY0)CYfk&m^L_d)X0 z{W=%+5kD|XRrNGCHa4b&(_k9G$FE;G&=Bhw`1pQp@uQ9#MuCPfdJN>976)NW5018@ zb*b0WJX=ZC<-i)@D)WTulOGU0ssFUbK0$eGeHApvmb-K;BNGQ?j0aBOdi(M~T1ajp z4Ip@Pn~(#LN2lel_pefYKG*p5t?hVSg~#TQGjT9DZCx#>fTVG@LIkoS#?w51yyOYp z*3bT2p&Q6mEIwB4Zq8~IAHHEk#S^~44-^W%y=k$-+gDGbZcx;b{K^B#;gtmXi>|XN zk==NgHZn4@Aa9?xM5n!S@O_gBR<+(#P>e*0e3=)xHh_k%1sxz<1HQ z{OLIN*7;RCh9F3_XEku~lRFfFVNGWyAkiuAn>9Q2eGPWpt4 zZ2N3AYU6E51lsaiDk#1Cw{+kk?{NTB2s@t`!JWbIsx>gfYD%@bM{i=SCk#hti=Bes zSnX7Y7p(EqoiFUPH#FsJ#h-7JUcOr3zG=O_eb-H=&q1_sTBHkmXM=K4@`FTsw(d|1>yw;;4mA{H(1FHy=P$KS(y(*d&Qf ziV*U@0Ltv=PITG$_Ux5-%4-K{wbNR14U+HMd5VO3CiLum)A0F_bnq{8F8 zC#OG6-embbN+HciZ7N~?r6f4NzRwS8A`;u(`qm+I>IubA&g(ubvKlby;v3C_@w7a{ zwi_h9j<}6Dvb8@ql%2bhowUk*Jj_hX%C%cWDg@8QH5+V{(!PLWe}DXHa@ufCb<_93HJTAO+A8_YZN0kK{Ec*iEe=&E4kb$`y^ zD;JIj7vwZ(xmBDd7JQtjMq1&!JcS1@l7Qj-Y7b@Z!7GJn!U^2{XRdi92u3h))I^$r ztuSVq$FUz>8w#eD>sX{#No=wPfevM*%~?e%xH|*`V4)vAqzm2_BxA$gwYp z?25u({1K%FsKYXX4@tB?d1q$Qo^r4^35d6~&1&0W_PyYLCLH37MeldRs^zF#$YOpt-zqBa~Hh`-PjW0e5K1T z$xxY>N1HD8oRGJ;@`L=-Y;L;(?G2)F-$JPTV;)d`BD$_G@g$`n&U53@y0H*g$zj|` zn$V{RKI6gTnXmHJli6)5aELi^+O1kPaO=7Y$nT{dXa&$}9bX@&Q+NOR^u4S*yN>1a zDr&uAv*SO$y?}-r_GAeh6xFIv6Zi@~J*dR(nwuLDp5PGOyuqxiXUpJqjNDH1MF3Tb zULvcbnk~ci7wQ-Lb<3C~2z4)ONmM(jS(K{qztARs@p@F6d~xsU<|@qxcdO?6TueXZ zI;YR5%FXX7K%-@|d^)zWhWXgYltFP->bq`@tnfgB{uA=tol!~sGEC+?`7ve zp0-~`bg=h7a-`!+;*I2FVlKG;3R2kZvN7tP5ce{{y}<>qWoipOsO6a3AtN?*44l{j zHT1HUR$V2T^?SUvERWlK;FHIk^9nayxighd1}zT`>f{y{UQG(s;d!l`GIR2eR1u|o zGO6i4(ZHqkWV^b>MeA<9_1xj>yULf*Ws;V!lT}g!m}hkv=9(((2K@}&7i)TYF#fiw zv38xJ7cN`~ZE0!YXcpz@cDexx1rn~{m;_P(yH z8pCb@i)bQ?zcA*)E&0)Pk88Wd`MDW2xvRZlp8c#*k?BK+pR@OuMIKl79asxY7>kaj5c7yCQ2D%}T44n^)# zR2`k2s<}l)Rb6D8L6T=wE(jH6d_3xzkbJ`F1}Y$tw4q z>n3Goy3}t*%cOMI-EeUpJ8QS2;^Hpk7#qa8Y%LqC3G(_j#)-2$%}Nc`3=ZgYo^3w( z9IugehTq^QoKnhv0BRdUVamz&WBL!mV9 zRL4Us_H}krC)bnrO4NP}=0)Y@I_X?;eQw#QT!5(AI_$P6(8kaiW@aQmZ4&d?YR7v$ z6?%XocR3s=g6iD?RQ{UxT>*D5)?$*k%x6o(?EYIJb|uM5q61%|%^RLncZV?L7B-=<2*ohBls>nA|FF`)0KXJ-=nTE5SQMi{($yISU>Z?(Vyl#>zOx zw#G*JtupCP9bLJ+RDpp{k~x1kdw9eMd+zW=MMs~nnVTfPq?njg-O-^U`&RbVvy4mk zL~oUptA~vi072j)DMMYJYQ|pS9{hS=^>Zc*Mbn2W<_R5UzLfa80a1H<6?u7NofmiB z6%2py^p|@4MwWtf;dhSStC?Y4$-(r8Rb-p+@UMwgn1+RgF`Sq64Q`|}eHAS!pga zABt=PZP}Eh+%pSzOBN)9sy7)6Md&6?HPoA(rr~a14-aw(%I7U)AS88(+i5u2Up`YJ z-KMuz-!yzy&iewLfFd1gyL$mc$%BY7iCTR4e)2ODIWR;(Yh` zxM6NZMO?Y(E+0N*erbuRK#MiG+tSL)b>^qP>APfiDfN%_1CUA!uNnmp>LuYExp=g& z6&_LL-iM+BVqzp=ZyTNi6wUoZP4ymu#p40eB@+Dc=oXXo|KhLdGBhCEP=5_>1Lrtesrw&5WB)KN;mhlS6wcBX7!u${YUg{$n!DApym^lF6l?V~>x) Omz str: "tab:richer", f"run_ids={all_ids}") +def _t6(rows: list[dict]) -> str: + """Adaptive halting (CRC guarantee) over seeds — H1 on a multi-seed footing (Phase 4l). + + Per task: the CRC-chosen agreement threshold, compute saved, halting risk vs the budget alpha, + end-task accuracy drop, and how many seeds keep the risk within budget — all mean$\\pm$std. + """ + cells = aggregate_halting(rows) + body, all_ids = [], [] + for task, a in cells.items(): + body.append([ + _esc(task), f"${a['alpha']}$", _pm(a["tau_hat"], 2), _pm(a["compute_saved"], 3), + _pm(a["halting_risk"], 3), _pm(a["accuracy_drop"], 3), + f"{a['seeds_within_budget']}/{a['n_seeds']}", + ]) + all_ids += a["run_ids"] + header = ["task", r"$\alpha$", r"$\hat\tau$", "compute saved", "halting risk", + "acc drop", "seeds $\\le\\alpha$"] + return _tabular( + header, body, "lcccccc", + "Adaptive halting (CRC), mean$\\pm$std over seeds: the basin-agreement threshold " + "$\\hat\\tau$ is calibrated on a disjoint fold so the halting risk (disagreement with the " + "full-budget answer) stays within the budget $\\alpha$. ``compute saved`` is the fraction " + "of descent steps skipped; ``seeds $\\le\\alpha$`` counts seeds with test risk in budget.", + "tab:halting", f"run_ids={all_ids}") + + +def _t7(rows: list[dict]) -> str: + """OOD stress over seeds — the guarantee holds ID and breaks under shift (Phase 4l). + + Per task: ID vs OOD accuracy, and the ID-calibrated LTT threshold's selective risk on the ID + fold vs the shifted fold, with a count of seeds where OOD risk stays within budget. + """ + body, all_ids = [], [] + for task in tasks_present(rows): + a = headline_cell(rows, task=task) + if not a.get("has_ood"): + continue + body.append([ + _esc(task), _pm(a["accuracy_id"], 2), _pm(a["accuracy_ood"], 2), + _pm(a["ltt_selective_risk"], 3), _pm(a["ood_selective_risk"], 3), + f"{a['seeds_ood_within_budget']}/{a['n_seeds']}", + ]) + all_ids += a["run_ids"] + header = ["task", "acc ID", "acc OOD", "sel.\\ risk ID", "sel.\\ risk OOD", "seeds OOD ok"] + return _tabular( + header, body, "lccccc", + "OOD stress (mean$\\pm$std over seeds): the ID-calibrated selective-prediction threshold " + "applied to a shifted fold. The selective-risk guarantee holds in-distribution but breaks " + "under shift (OOD risk exceeds the budget), motivating abstention; ``seeds OOD ok`` counts " + "seeds whose OOD risk still stays within budget.", + "tab:oodstress", f"run_ids={all_ids}") + + def _t3(rows: list[dict]) -> str: sel = selective_rows(rows) env = sel[-1]["env"] if sel else {} @@ -235,6 +290,12 @@ def _has_body(tex: str) -> bool: t5 = _t5(rows) # only when richer-geometry rows exist if _has_body(t5): outputs["T5_richer_geometry.tex"] = t5 + t6 = _t6(rows) # only when split='halting' rows exist + if _has_body(t6): + outputs["T6_halting.tex"] = t6 + t7 = _t7(rows) # only when include_ood rows exist + if _has_body(t7): + outputs["T7_ood_stress.tex"] = t7 for name, tex in outputs.items(): (TABLE_DIR / name).write_text(tex) print(f"[tables] wrote {TABLE_DIR / name}") diff --git a/configs/experiments/graph_halting.toml b/configs/experiments/graph_halting.toml new file mode 100644 index 0000000..08d7fd7 --- /dev/null +++ b/configs/experiments/graph_halting.toml @@ -0,0 +1,36 @@ +# Adaptive-halting experiment on graph shortest-path (Phase 4l — a 2nd task for the CRC guarantee). +# Same task/inference/train as the graph selective E2 config, but per-step decoding is recorded so +# the basin-agreement halting threshold can be CRC-calibrated. Folds kept modest to bound decode +# memory on CPU (mirrors arithmetic_halting). +[run] +task = "graph_planning" +seed = 0 +notes = "graph adaptive-halting run" + +[task.graph_planning] +n_nodes = 7 +ood_n_nodes = 10 +edge_prob = 0.4 +max_len = 4 + +[model] +latent_dim = 32 +hidden_dim = 128 +context_dim = 64 + +[inference] +k_restarts = 12 +steps = 50 + +[train] +epochs = 25 # graph needs more epochs than arithmetic to reach the 70-85% regime +n_train = 8000 +n_neg = 4 + +[eval] +n_eval = 800 # test fold (per-step decode kept small) + +[conformal] +alpha = 0.1 # halting error budget: E[early answer != full answer] <= alpha +delta = 0.05 +n_calib = 800 # calibration fold diff --git a/configs/sweeps/arith_halting_seeds.toml b/configs/sweeps/arith_halting_seeds.toml new file mode 100644 index 0000000..09f5a2f --- /dev/null +++ b/configs/sweeps/arith_halting_seeds.toml @@ -0,0 +1,9 @@ +# Phase 4l: 5-seed adaptive-halting (CRC) on arithmetic — the multi-seed version of the H1 headline +# (~58% compute saved at halting risk <= alpha), which was single-seed. Uses run_halting_sweep.py +# (each cell trains, CRC-calibrates, appends a split="halting" row). aggregate.aggregate_halting + +# make_tables T6 summarise the compute-saved / halting-risk mean±std. +[sweep] +base = "configs/experiments/arithmetic_halting.toml" + +[sweep.grid] +"run.seed" = [0, 1, 2, 3, 4] diff --git a/configs/sweeps/arith_ood_seeds.toml b/configs/sweeps/arith_ood_seeds.toml new file mode 100644 index 0000000..d53e3b9 --- /dev/null +++ b/configs/sweeps/arith_ood_seeds.toml @@ -0,0 +1,9 @@ +# Phase 4l: 5-seed FULL-FOLD arithmetic selective prediction with include_ood=true, so the OOD +# stress result (F6: ID risk vs OOD risk under the ID-calibrated threshold) gets a multi-seed +# mean±std instead of a single seed. Matches graph_cert_seeds so T7 is 5-seed on both tasks. +[sweep] +base = "configs/experiments/arithmetic_selective.toml" +include_ood = true + +[sweep.grid] +"run.seed" = [0, 1, 2, 3, 4] diff --git a/configs/sweeps/graph_cert_seeds.toml b/configs/sweeps/graph_cert_seeds.toml new file mode 100644 index 0000000..f6087ab --- /dev/null +++ b/configs/sweeps/graph_cert_seeds.toml @@ -0,0 +1,10 @@ +# Phase 4l THE CERT RUN: 5-seed FULL-FOLD graph selective prediction (n_calib=1500), so LTT +# abstention can actually certify nonzero coverage on graph — the reduced-fold seed sweeps +# (n_calib=800) certify zero. include_ood=true also yields 5-seed graph OOD stress (T7). No n_calib +# override: it inherits graph_selective.toml's full folds. Heaviest cell set in the project. +[sweep] +base = "configs/experiments/graph_selective.toml" +include_ood = true + +[sweep.grid] +"run.seed" = [0, 1, 2, 3, 4] diff --git a/configs/sweeps/graph_halting_seeds.toml b/configs/sweeps/graph_halting_seeds.toml new file mode 100644 index 0000000..b4d7fa1 --- /dev/null +++ b/configs/sweeps/graph_halting_seeds.toml @@ -0,0 +1,8 @@ +# Phase 4l: 5-seed adaptive-halting (CRC) on graph shortest-path — a 2nd task for the halting +# guarantee (secondary to the arithmetic headline). Uses run_halting_sweep.py -> split="halting" +# rows -> aggregate_halting -> T6. +[sweep] +base = "configs/experiments/graph_halting.toml" + +[sweep.grid] +"run.seed" = [0, 1, 2, 3, 4] diff --git a/docs/DECISIONS.md b/docs/DECISIONS.md index cc6d943..a41f1f7 100644 --- a/docs/DECISIONS.md +++ b/docs/DECISIONS.md @@ -2,6 +2,24 @@ Short, dated, append-only. Newest first. +## 2026-07-30 — Phase 4l: certificates on firm footing (multi-seed guarantees + graph certification) +**Finding/decision:** put the project's core contribution — the distribution-free certificates — on a +5-seed footing. (1) **Halting:** new `experiments/run_halting_sweep.py` + `aggregate.halting_rows`/ +`aggregate_halting_cell` (split="halting" was excluded from `selective_rows`). Arith saves +57.1%±1.7% (single-seed H1 was 57.8%), graph 80.8%±5.2%; both risk ≤ α, within budget 5/5 (T6). +(2) **OOD:** `[sweep] include_ood=true` in `run_sweep` + an OOD aggregation block; the selective +guarantee breaks under shift over 5 seeds (arith ID 0.077 → OOD 0.764, 0/5; graph 0.010 → 0.168, +4/5) (T7). (3) **Graph certification:** a full-fold run (n_calib=1500) certifies only 1.1% coverage +at α=0.1 but 22.5% at α=0.15 and 48.5% at α=0.20, with achieved risk ≤ target at every α — so the +old "certifies zero coverage" note was an **operating-point artifact** of graph's ~30% error floor, +not a broken certificate. **Bugs fixed:** (a) halting aggregation dedups by seed (the stale Phase-4b +H1 seed-0 row survived under a drifted config_hash, giving n=6); (b) `plotting._selective_row` and +`_halting_row` now prefer arithmetic, since the new full-fold (n_test=1500) graph rows would +otherwise tie/flip the arithmetic figures F2/F4/F6 to graph. **Why record it:** every guarantee +number the paper states is now multi-seed, and the graph certificate is shown valid at its honest α. +**Reversible?** Additive runner + aggregation + a sweep flag; default `include_ood=False` preserves +all prior sweeps. + ## 2026-07-29 — Phase 4k: the graph redundancy is a task property, not thin features **Finding/decision:** the 4j redundancy verdict rested on a thin curvature description (only `λmax`/`tr(H)`), so we added two opt-in richer feature groups — the **full Hessian spectrum** diff --git a/docs/EXPERIMENTS.md b/docs/EXPERIMENTS.md index 968c96d..41ed54c 100644 --- a/docs/EXPERIMENTS.md +++ b/docs/EXPERIMENTS.md @@ -232,6 +232,52 @@ contribute there, even with the richest geometry we can extract. The base rows r numbers exactly, confirming richer geometry is purely additive and non-breaking (default stays the canonical 14 features; aggregation is feature-set-aware so richer never pollutes T1/T2/T4). +### Certificates on firm footing — multi-seed guarantees + graph certification (Phase 4l → T6/T7) + +The certificates are the contribution, but three of their numbers were single-seed or vacuous. This +phase puts all three on a 5-seed footing (new `run_halting_sweep.py`; `[sweep] include_ood=true`; +objective/feature-set-style aggregation for halting + OOD; full-fold graph cert run). Additive — no +change to the geometry features or the falsification test. + +**Adaptive halting (CRC), 5 seeds — T6.** The H1 headline replicates and extends: + +| task | τ̂ | compute saved | halting risk (≤ α=0.1) | seeds ≤ α | +|------|-----|---------------|------------------------|-----------| +| arithmetic | $0.917$ | $0.571 \pm 0.017$ | $0.012 \pm 0.005$ | $5/5$ | +| graph | $0.85\pm0.03$ | $\mathbf{0.808 \pm 0.052}$ | $0.065 \pm 0.032$ | $5/5$ | + +The single-seed 57.8% arithmetic number is now $57.1\%\pm1.7\%$ over 5 seeds (within budget every +seed); on graph the CRC halt saves **~81%** of compute at risk $0.065 \le 0.1$ — the halting +guarantee holds on both tasks across seeds. + +**OOD stress, 5 seeds — T7.** The ID-calibrated selective threshold applied to the shifted fold: + +| task | acc ID | acc OOD | sel. risk ID | sel. risk OOD | seeds OOD ≤ α | +|------|--------|---------|--------------|---------------|---------------| +| arithmetic | $0.83$ | $0.21$ | $0.077$ | $\mathbf{0.764}$ | $0/5$ | +| graph | $0.70$ | $0.33$ | $0.010$ | $0.168$ | $4/5$ | + +The guarantee holds ID and **breaks under shift** on both tasks (arithmetic catastrophically: +$0.077\to0.764$; graph mildly: $0.010\to0.168$) — 5-seed confirmation of F6, and the argument for +treating an abstention as a routing signal. + +**Graph certification — the certificate is valid on graph, at its honest operating point.** At the +full fold ($n_\text{calib}=1500$) the graph LTT abstention still certifies almost nothing at +α=0.1 (coverage $0.011$), because graph's ~30% base error leaves no confident slice with true error +below 10%. But sweeping α (from the same `coverage_validity` block; mean over 5 seeds): + +| α | certified coverage | achieved risk | +|-----|--------------------|---------------| +| $0.10$ | $0.011$ | $0.010$ | +| $0.15$ | $0.225$ | $0.078$ | +| $0.20$ | $0.485$ | $0.157$ | +| $0.30$ | $0.895$ | $0.263$ | + +Achieved risk sits **at or below target at every α** — validity holds; the earlier "certifies zero +coverage" was an α=0.1 operating-point artifact, not a broken certificate. Graph certifies **22.5% +coverage at α=0.15** and **48.5% at α=0.20**, versus arithmetic's 74% at α=0.10 — the honest cost of +a harder task's higher error floor, not a failure. (F3 shows the arithmetic α-sweep on the diagonal.) + ## Operating regime Tune each task's difficulty so base accuracy is **70–85%**. Fully-solved tasks saturate AURC and diff --git a/docs/SESSION_HANDOFF.md b/docs/SESSION_HANDOFF.md index 3c43261..34fa828 100644 --- a/docs/SESSION_HANDOFF.md +++ b/docs/SESSION_HANDOFF.md @@ -2,7 +2,7 @@ **Start here.** Rolling state of the project between sessions. -## Current state (2026-07-29) +## Current state (2026-07-30) - **Phase 0 (skeleton): ✅** repo tree, `uv`/`pyproject.toml`, `Makefile`, CI, config system (`edc.config`), deterministic seeding (`edc.seeding`), append-only ledger (`edc.ledger`), @@ -99,9 +99,21 @@ geometry already wins (arith-IRED +0.007, 4/5, unchanged) and slightly worse on the fixed-bowl basin-center cells (extra features add noise). Aggregation is feature-set-aware (`feature_set_of`; `headline_cell`/`by_k` default `feature_set="base"`) so richer never pollutes T1/T2/T4. 109 tests. -- **Open question / next:** a 3rd/4th task (E3/E4, e.g. logic/Sudoku) to characterize *when* geometry - beats softmax now that feature poverty is ruled out; harden honesty gaps (multi-seed halting/OOD, - MC-dropout baseline, a full-fold graph run that actually certifies); Modal; scale-up. +- **Phase 4l (certificates on firm footing — multi-seed guarantees + graph certification): ✅** put + the project's core contribution (the certificates) on a 5-seed footing. New + `experiments/run_halting_sweep.py` + `aggregate.{halting_rows, aggregate_halting_cell}` → **T6**: + arith halting saves **57.1%±1.7%** (was single-seed 57.8%), graph **80.8%±5.2%**, both halting risk + ≤ α, within budget 5/5. `[sweep] include_ood=true` (run_sweep) + an OOD aggregation block → **T7**: + the selective guarantee breaks under shift over 5 seeds (arith ID sel-risk 0.077 → OOD 0.764, 0/5; + graph 0.010 → 0.168, 4/5). Full-fold graph cert run (`graph_cert_seeds.toml`, n_calib=1500): LTT + certifies only **1.1% coverage at α=0.1** but **22.5% at α=0.15, 48.5% at α=0.20; validity holds at + every α** — the old "certifies zero" was an operating-point artifact of graph's ~30% error floor, + not a broken certificate. Fixed a stale-row bug (halting dedups by seed, dropping the Phase-4b H1 + row under a drifted config_hash) and made `plotting._selective_row`/`_halting_row` prefer arithmetic + so the new full-fold (n_test=1500) graph rows don't flip F2/F4/F6. 115 tests. +- **Open question / next:** a 3rd/4th task (E3/E4, e.g. logic/parity) to characterize *when* geometry + beats softmax now that feature poverty is ruled out; a deep-ensemble baseline (the one remaining + deferred baseline); Modal; scale-up. ## What is real vs stub diff --git a/experiments/run_halting_sweep.py b/experiments/run_halting_sweep.py new file mode 100644 index 0000000..506b951 --- /dev/null +++ b/experiments/run_halting_sweep.py @@ -0,0 +1,71 @@ +"""Expand a halting sweep grid into many adaptive-halting runs. [Phase 4l] + + PYTHONPATH=src python experiments/run_halting_sweep.py configs/sweeps/arith_halting_seeds.toml + +The multi-seed counterpart of ``run_halting.py``: same sweep-TOML format as ``run_sweep`` (a +``base`` config + optional ``[sweep.override]`` + ``[sweep.grid]``), but each cell trains, CRC- +calibrates the halting threshold, and appends one ``split="halting"`` row via ``evaluate_halting`` — +so the compute-saved / halting-risk guarantee gets a multi-seed mean±std instead of a single seed. +``analysis/aggregate.halting_rows`` + ``aggregate_halting_cell`` summarise the rows (T6). +""" + +from __future__ import annotations + +import datetime as _dt +import sys +import uuid +from pathlib import Path + +from edc.config import load_from_dict +from edc.eval.evaluate_halting import evaluate_halting +from edc.ledger import RunRecord, append +from edc.registry import build_task + +try: # sys.path[0] is experiments/ when run as a script; fall back to package form. + from run_sweep import cell_config_dict, load_sweep +except ImportError: # pragma: no cover + from experiments.run_sweep import cell_config_dict, load_sweep + + +def run_halting_and_append(cfg, task) -> dict: + """Train + CRC-calibrate + evaluate halting for one cell, append a ``split='halting'`` row.""" + m = evaluate_halting(cfg, task) + return append(RunRecord( + run_id=uuid.uuid4().hex[:12], + timestamp=_dt.datetime.now(_dt.UTC).isoformat(), + resolved_config=cfg.to_dict(), + task=task.name, + split="halting", + seed=cfg.run.seed, + metrics=m, + )) + + +def main(argv: list[str]) -> int: + if not argv: + print("usage: run_halting_sweep.py ", file=sys.stderr) + return 2 + base, override, cells, _opts = load_sweep(argv[0]) + + print(f"[edc] halting sweep '{Path(argv[0]).name}': {len(cells)} cells " + f"(override={override or '{}'}) -> appending to results/ledger.jsonl") + for i, cell in enumerate(cells, 1): + d = cell_config_dict(base, override, cell) + cfg = load_from_dict(d) + task_kwargs = d.get("task", {}).get(cfg.run.task, {}) + task = build_task(cfg.run.task, **task_kwargs) + print(f"\n[edc] cell {i}/{len(cells)}: {cell}") + row = run_halting_and_append(cfg, task) + m = row["metrics"] + tau = m["tau_hat"] + tau_s = f"{tau:.3f}" if tau is not None else "None" + print(f"[edc] -> run_id={row['run_id']} tau_hat={tau_s} saved={m['compute_saved']:.3f} " + f"risk={m['halting_risk']:.3f} within_budget={m['risk_within_budget']}") + from edc.config import REPO_ROOT + ledger = REPO_ROOT / "results/ledger.jsonl" + print(f"\n[edc] halting sweep complete: {len(cells)} rows appended -> {ledger}") + return 0 + + +if __name__ == "__main__": + raise SystemExit(main(sys.argv[1:])) diff --git a/experiments/run_sweep.py b/experiments/run_sweep.py index a0562f7..4b9c5ed 100644 --- a/experiments/run_sweep.py +++ b/experiments/run_sweep.py @@ -51,14 +51,19 @@ def expand_grid(grid: dict[str, list]) -> list[dict]: return [dict(zip(keys, combo, strict=True)) for combo in combos] -def load_sweep(path: str | Path) -> tuple[dict, dict, list[dict]]: - """Return ``(base_resolved_dict, override, cells)`` from a sweep TOML.""" +def load_sweep(path: str | Path) -> tuple[dict, dict, list[dict], dict]: + """Return ``(base_resolved_dict, override, cells, opts)`` from a sweep TOML. + + ``opts`` carries sweep-level flags that are not config keys — currently ``include_ood`` (default + False), which turns the OOD fold back on for a sweep that needs multi-seed OOD/certification. + """ with open(path, "rb") as f: spec = tomllib.load(f)["sweep"] base = load_config(spec["base"]).to_dict() override = spec.get("override", {}) cells = expand_grid(spec.get("grid", {})) - return base, override, cells + opts = {"include_ood": bool(spec.get("include_ood", False))} + return base, override, cells, opts def cell_config_dict(base: dict, override: dict, cell: dict) -> dict: @@ -73,18 +78,19 @@ def main(argv: list[str]) -> int: if not argv: print("usage: run_sweep.py ", file=sys.stderr) return 2 - base, override, cells = load_sweep(argv[0]) + base, override, cells, opts = load_sweep(argv[0]) from edc.config import load_from_dict # local: keep module import surface small + include_ood = opts["include_ood"] print(f"[edc] sweep '{Path(argv[0]).name}': {len(cells)} cells " - f"(override={override or '{}'}) -> appending to results/ledger.jsonl") + f"(override={override or '{}'}, include_ood={include_ood}) -> results/ledger.jsonl") for i, cell in enumerate(cells, 1): d = cell_config_dict(base, override, cell) cfg = load_from_dict(d) task_kwargs = d.get("task", {}).get(cfg.run.task, {}) task = build_task(cfg.run.task, **task_kwargs) print(f"\n[edc] cell {i}/{len(cells)}: {cell}") - row = run_and_append(cfg, task, include_ood=False) + row = run_and_append(cfg, task, include_ood=include_ood) print(f"[edc] -> run_id={row['run_id']}") ledger = REPO_ROOT / "results/ledger.jsonl" print(f"\n[edc] sweep complete: {len(cells)} rows appended -> {ledger}") diff --git a/paper/main.tex b/paper/main.tex index 3886051..2b847d0 100644 --- a/paper/main.tex +++ b/paper/main.tex @@ -32,7 +32,8 @@ terminal energy used by Energy-Based Transformers as a nonconformity score (paired-bootstrap $\Delta$AURC CI excluding zero, $5/5$ seeds each), the advantage is caused by the restart geometry (it grows with the number of restarts; a single descent barely separates) and driven mainly by basin -agreement, and halting saves ${\sim}58\%$ of inference compute at a guaranteed error budget. Against a stronger, standard softmax-confidence baseline (maximum softmax probability, temperature +agreement, and adaptive halting saves $57\%$ (arithmetic) to $81\%$ (graph) of inference compute at a +guaranteed error budget, holding within budget on both tasks across $5$ seeds. Against a stronger, standard softmax-confidence baseline (maximum softmax probability, temperature scaling, entropy), geometry only \emph{ties} (arithmetic) or \emph{loses} (graph) under a supervised basin-center reasoner---but when the reasoner is a genuinely learned, multi-basin landscape, geometry \emph{beats} softmax confidence on arithmetic ($4/5$ seeds). This does not generalize: on diff --git a/paper/sections/experiments.tex b/paper/sections/experiments.tex index 54c5b1e..61f35cc 100644 --- a/paper/sections/experiments.tex +++ b/paper/sections/experiments.tex @@ -41,17 +41,30 @@ \section{Experiments} energy statistics the vector happens to contain. Figure~\ref{fig:f5} shows the corresponding per-feature separation. -\paragraph{The guarantees behave as designed (Figs.~\ref{fig:f3}--\ref{fig:f6}).} In-distribution, -Learn-then-Test selective prediction is valid: achieved selective risk sits on or below the nominal -target across the $\alpha$-sweep (Fig.~\ref{fig:f3}). Conformal-Risk-Control adaptive halting -(Fig.~\ref{fig:f4}) chooses a basin-agreement threshold that uses $42\%$ of the step budget---$58\%$ -compute saved---at a halting risk of $0.9\% \le \alpha{=}10\%$ with no end-task accuracy loss, and -the halting risk was empirically monotone in the threshold (CRC's assumption). Under distribution -shift the guarantee \emph{breaks by design} (Fig.~\ref{fig:f6}): the ID-calibrated threshold, applied -to OOD, achieves $76\%$ selective risk against a $10\%$ target---exchangeability is violated, so the -certificate is void. This is precisely the argument for abstention / shift detection in critical -systems: the guarantees are marginal under exchangeability, and we show them fail loudly when it -does not hold. +\paragraph{The guarantees behave as designed, over 5 seeds (Figs.~\ref{fig:f3}--\ref{fig:f6}, +Tables~\ref{tab:halting}--\ref{tab:oodstress}).} In-distribution, Learn-then-Test selective +prediction is valid: achieved selective risk sits on or below the nominal target across the +$\alpha$-sweep (Fig.~\ref{fig:f3}). Conformal-Risk-Control adaptive halting +(Fig.~\ref{fig:f4}, Table~\ref{tab:halting}) chooses a basin-agreement threshold that saves +$57.1\%\pm1.7\%$ of the step budget on arithmetic at a halting risk of $1.2\% \le \alpha{=}10\%$ +(within budget $5/5$ seeds, no end-task accuracy loss), and $80.8\%\pm5.2\%$ on graph at risk +$6.5\% \le 10\%$ ($5/5$)---the halting guarantee holds on both tasks across seeds, and the halting +risk was empirically monotone in the threshold (CRC's assumption). Under distribution shift the +selective guarantee \emph{breaks by design} over 5 seeds (Fig.~\ref{fig:f6}, +Table~\ref{tab:oodstress}): the ID-calibrated threshold applied to OOD achieves $76.4\%$ selective +risk against a $10\%$ target on arithmetic ($0/5$ within budget) and $16.8\%$ on graph +($4/5$)---exchangeability is violated, so the certificate is void. This is precisely the argument for +abstention / shift detection in critical systems: the guarantees are marginal under exchangeability, +and we show them fail loudly when it does not hold. + +On graph, the LTT abstention certifies almost no coverage at $\alpha{=}0.1$ (mean $1.1\%$ over 5 +seeds at the full $n_\text{calib}=1500$ fold): the task's ${\sim}30\%$ base error leaves no confident +slice with true error below $10\%$. This is an operating-point effect, not a broken certificate---% +validity holds at every $\alpha$ (achieved risk $\le$ target), and relaxing the budget certifies real +coverage: $22.5\%$ at $\alpha{=}0.15$ (achieved risk $0.078$) and $48.5\%$ at $\alpha{=}0.20$ +($0.157$), versus arithmetic's $74\%$ at $\alpha{=}0.10$. The honest reading is that a harder task's +higher error floor pushes the usable operating point to a larger $\alpha$, not that the guarantee +fails. \paragraph{Key caveat: geometry does not beat softmax confidence (Table~\ref{tab:main}).} When the baseline set is widened from scalar energy to a standard softmax-confidence score (MSP, @@ -113,6 +126,8 @@ \section{Experiments} well-calibrated softmax already captures what even the richest descent geometry would add. \input{tables/T1_main} +\input{tables/T6_halting} +\input{tables/T7_ood_stress} \input{tables/T4_complementarity} \input{tables/T5_richer_geometry} \input{tables/T2b_feature_ablation} @@ -138,8 +153,8 @@ \section{Experiments} \includegraphics[width=0.48\linewidth]{F3_coverage_validity.png}\hfill \includegraphics[width=0.48\linewidth]{F4_halting_pareto.png} \caption{\textbf{Left (F3):} LTT selective-risk validity---achieved risk on/below the diagonal. - \textbf{Right (F4):} CRC adaptive halting saves $\sim$58\% compute at a guaranteed risk with no - accuracy loss.} + \textbf{Right (F4):} CRC adaptive halting (arithmetic) saves ${\sim}57\%$ compute at a guaranteed + risk with no accuracy loss; the 5-seed numbers for both tasks are in Table~\ref{tab:halting}.} \label{fig:f3}\label{fig:f4} \end{figure} \begin{figure}[t]\centering diff --git a/paper/tables/T1_main.tex b/paper/tables/T1_main.tex index 20c45da..8572481 100644 --- a/paper/tables/T1_main.tex +++ b/paper/tables/T1_main.tex @@ -1,4 +1,4 @@ -% auto-generated by analysis/make_tables.py — do not edit. run_ids=['aefb55fc7987', '8cfc5a222d83', '8ab8e76e3b44', '7d6cba43dccf', '10f375729230', '7802e0276803', 'd80ff5b236eb', '198178fd3f60', '8867921abd20', 'c81e3990d838'] +% auto-generated by analysis/make_tables.py — do not edit. run_ids=['59f889f657d5', 'b04ffc85ccc3', '354783efb5d1', '5bd7ea239e4f', '52b5abb0b7f5', '2165fc5b4ecc', '496414a8f2cd', '9202e8561acd', '8bc3880fccd3', 'fadafc6f2be1'] \begin{table}[t] \centering \caption{Main selective-prediction results across reasoning families (mean$\pm$std over seeds): geometry's $\Delta$AURC over the best raw-energy baseline and over the best of all baselines (energy, MSP, temperature-scaled MSP, predictive entropy). $\Delta$AURC $>0$ with the CI clearing 0 means geometry wins; ``seeds win`` counts seeds beating the best overall baseline.} @@ -8,8 +8,8 @@ \toprule task & $K$ & acc & AURC geom & $\Delta$AURC vs energy & $\Delta$AURC vs best base & seeds win & LTT cov \\ \midrule - arithmetic & 12 & $0.83 \pm 0.02$ & $0.059 \pm 0.018$ & $0.085 \pm 0.036$ & $-0.011 \pm 0.006$ & 0/5 & $0.55 \pm 0.30$ \\ - graph\_planning & 12 & $0.70 \pm 0.02$ & $0.159 \pm 0.012$ & $0.111 \pm 0.054$ & $-0.034 \pm 0.008$ & 0/5 & $0.00 \pm 0.00$ \\ + arithmetic & 12 & $0.83 \pm 0.02$ & $0.059 \pm 0.018$ & $0.086 \pm 0.038$ & $-0.008 \pm 0.008$ & 0/5 & $0.74 \pm 0.08$ \\ + graph\_planning & 12 & $0.70 \pm 0.02$ & $0.159 \pm 0.016$ & $0.103 \pm 0.045$ & $-0.030 \pm 0.004$ & 0/5 & $0.01 \pm 0.02$ \\ \bottomrule \end{tabular}} \end{table} diff --git a/paper/tables/T2_kablation.tex b/paper/tables/T2_kablation.tex index 7f2cbc1..bfd7136 100644 --- a/paper/tables/T2_kablation.tex +++ b/paper/tables/T2_kablation.tex @@ -1,4 +1,4 @@ -% auto-generated by analysis/make_tables.py — do not edit. run_ids=['acd437459961', '852e712d4157', '8b6b1348cf28', '56ea729e5ea5', '8b8e1e4a7381', 'a4170c27c22b', 'af69f3e49ad2', '25320822fdfd', '9519934cb397', '978eda331152', 'b73a5b1750f2', 'f14b18cfaee4', '04b6026d0afc', '3dc77bab6f6f', '8c4a180e7c9e', 'db2094e35b89', '7a85d96687d2', '7f822ee20bdb', 'e7cfb7fd624f', '539e9c3f0431', '1bfe6b304e0f', 'f63036690d31', '46752558c5df', 'b8dc1a068df5', 'aa60425c10e1', 'c948f8130a3d', '4a5576e4c244', '70741801c041', 'b0d0d7f2f513', '45d9d36ce766', 'f301fe7ebb4d', '39101f30bcb7', 'aefb55fc7987', '8cfc5a222d83', '8ab8e76e3b44', '7d6cba43dccf', '10f375729230', '8d6b3f5b24a2', 'b5b8cfd55dee', '1a2ccd4ef091', 'ff953d4a0e15', 'e14af222ab26'] +% auto-generated by analysis/make_tables.py — do not edit. run_ids=['acd437459961', '852e712d4157', '8b6b1348cf28', '56ea729e5ea5', '8b8e1e4a7381', 'a4170c27c22b', 'af69f3e49ad2', '25320822fdfd', '9519934cb397', '978eda331152', 'b73a5b1750f2', 'f14b18cfaee4', '04b6026d0afc', '3dc77bab6f6f', '8c4a180e7c9e', 'db2094e35b89', '7a85d96687d2', '7f822ee20bdb', 'e7cfb7fd624f', '539e9c3f0431', '1bfe6b304e0f', 'f63036690d31', '46752558c5df', 'b8dc1a068df5', 'aa60425c10e1', 'c948f8130a3d', '4a5576e4c244', '70741801c041', 'b0d0d7f2f513', '45d9d36ce766', 'f301fe7ebb4d', '39101f30bcb7', 'aefb55fc7987', '8cfc5a222d83', '8ab8e76e3b44', '7d6cba43dccf', '10f375729230', '59f889f657d5', 'b04ffc85ccc3', '354783efb5d1', '5bd7ea239e4f', '52b5abb0b7f5', '8d6b3f5b24a2', 'b5b8cfd55dee', '1a2ccd4ef091', 'ff953d4a0e15', 'e14af222ab26'] \begin{table}[t] \centering \caption{Restart ablation: does landscape geometry across $K$ restarts add signal over a single descent ($K{=}1\approx$ EBT)? Lift is $\Delta$AURC $>0$ with the CI clearing 0.} @@ -12,7 +12,7 @@ 2 & $0.103 \pm 0.036$ & $0.146 \pm 0.048$ & $0.043 \pm 0.024$ & 3/5 & $0.18 \pm 0.36$ \\ 4 & $0.082 \pm 0.038$ & $0.144 \pm 0.046$ & $0.062 \pm 0.035$ & 4/5 & $0.29 \pm 0.36$ \\ 8 & $0.076 \pm 0.029$ & $0.145 \pm 0.053$ & $0.070 \pm 0.033$ & 5/5 & $0.54 \pm 0.29$ \\ - 12 & $0.049 \pm 0.026$ & $0.154 \pm 0.063$ & $0.105 \pm 0.053$ & 17/17 & $0.66 \pm 0.28$ \\ + 12 & $0.051 \pm 0.025$ & $0.152 \pm 0.061$ & $0.100 \pm 0.050$ & 22/22 & $0.68 \pm 0.25$ \\ 16 & $0.051 \pm 0.021$ & $0.134 \pm 0.050$ & $0.083 \pm 0.034$ & 5/5 & $0.68 \pm 0.11$ \\ \bottomrule \end{tabular}} diff --git a/paper/tables/T2b_feature_ablation.tex b/paper/tables/T2b_feature_ablation.tex index 282ff84..adcb3f7 100644 --- a/paper/tables/T2b_feature_ablation.tex +++ b/paper/tables/T2b_feature_ablation.tex @@ -8,11 +8,11 @@ \toprule subset (AURC, lower=better) & arithmetic & graph\_planning \\ \midrule - full & $0.059 \pm 0.018$ & $0.159 \pm 0.012$ \\ - drop\_basin & $0.110 \pm 0.030$ & $0.175 \pm 0.017$ \\ - drop\_energy & $0.062 \pm 0.018$ & $0.162 \pm 0.014$ \\ - drop\_curv & $0.063 \pm 0.020$ & $0.160 \pm 0.012$ \\ - drop\_dynamics & $0.064 \pm 0.024$ & $0.187 \pm 0.024$ \\ + full & $0.059 \pm 0.018$ & $0.159 \pm 0.016$ \\ + drop\_basin & $0.116 \pm 0.037$ & $0.175 \pm 0.016$ \\ + drop\_energy & $0.066 \pm 0.019$ & $0.163 \pm 0.019$ \\ + drop\_curv & $0.062 \pm 0.020$ & $0.160 \pm 0.018$ \\ + drop\_dynamics & $0.065 \pm 0.023$ & $0.190 \pm 0.028$ \\ \bottomrule \end{tabular}} \end{table} diff --git a/paper/tables/T3_repro.tex b/paper/tables/T3_repro.tex index 16b7979..281a3c4 100644 --- a/paper/tables/T3_repro.tex +++ b/paper/tables/T3_repro.tex @@ -1,4 +1,4 @@ -% auto-generated by analysis/make_tables.py — do not edit. n_rows=82 +% auto-generated by analysis/make_tables.py — do not edit. n_rows=92 \begin{table}[t] \centering \caption{Reproducibility appendix. Env: python 3.12.13, jax 0.11.0 (cpu). Every row regenerates from seed + config via \texttt{run\_sweep}.} @@ -41,6 +41,8 @@ c468dfa50c2c & 0 & 12 & 1e7f9a8e1d3b & 9367d60 \\ ec2457a78f8d & 0 & 12 & f821965bd47a & 9367d60 \\ abb4dcb4ea84 & 0 & 12 & b333edd88366 & 9367d60 \\ + 59f889f657d5 & 0 & 12 & 646eeb651c32 & 67dc7a2 \\ + 2165fc5b4ecc & 0 & 12 & c5e806bfe9f7 & 67dc7a2 \\ 452500308f3b & 1 & 12 & 60da27bc4d66 & 5207145 \\ b8dc1a068df5 & 1 & 12 & e65ef9e88f92 & 5207145 \\ b0d0d7f2f513 & 1 & 12 & 372c61dccb87 & 3e52940 \\ @@ -52,6 +54,8 @@ 69bcb88d896d & 1 & 12 & b199710e8a3d & 9367d60 \\ 34bc023358b8 & 1 & 12 & f71a346333e0 & 9367d60 \\ 6c7eb2cdb623 & 1 & 12 & a146d7ac1641 & 9367d60 \\ + b04ffc85ccc3 & 1 & 12 & c8e9150407ea & 67dc7a2 \\ + 496414a8f2cd & 1 & 12 & 593dfa402e05 & 67dc7a2 \\ b1e83969abe5 & 2 & 12 & ff22f3433ff5 & 5207145 \\ aa60425c10e1 & 2 & 12 & cacbb6ece843 & 5207145 \\ 45d9d36ce766 & 2 & 12 & c7aaaa71b4ba & 3e52940 \\ @@ -63,6 +67,8 @@ affc597f3e34 & 2 & 12 & cd311d01e1a0 & 9367d60 \\ 28ac1614501c & 2 & 12 & d133e8a33968 & 9367d60 \\ 9fdd2103f70b & 2 & 12 & a55d25d137ca & 9367d60 \\ + 354783efb5d1 & 2 & 12 & 3c2fcb2e5dde & 67dc7a2 \\ + 9202e8561acd & 2 & 12 & 882cd7965583 & 67dc7a2 \\ 70d97e02bed0 & 3 & 12 & fbed14eeb756 & 5207145 \\ c948f8130a3d & 3 & 12 & 054f01c6c637 & 5207145 \\ f301fe7ebb4d & 3 & 12 & 49d35c26cae8 & 3e52940 \\ @@ -74,6 +80,8 @@ 37776ebe2a5b & 3 & 12 & f586dabb6126 & 9367d60 \\ 39b5492337a7 & 3 & 12 & a93a0c192b97 & 9367d60 \\ cb3102b09662 & 3 & 12 & a0a621e5442c & 9367d60 \\ + 5bd7ea239e4f & 3 & 12 & 0c62cba9cde6 & 67dc7a2 \\ + 8bc3880fccd3 & 3 & 12 & e71a6b480861 & 67dc7a2 \\ 46bb116f8c7c & 4 & 12 & 609a4152cf1e & 5207145 \\ 4a5576e4c244 & 4 & 12 & 0b564d7f9e23 & 5207145 \\ 39101f30bcb7 & 4 & 12 & 83b44608e95e & 3e52940 \\ @@ -85,6 +93,8 @@ e7b91099dfc5 & 4 & 12 & d5fa1e138f15 & 9367d60 \\ 61d34e2aec3a & 4 & 12 & e76acc6a32ac & 9367d60 \\ a405cf7b867a & 4 & 12 & 23346bbdcf92 & 9367d60 \\ + 52b5abb0b7f5 & 4 & 12 & 4a05527e96db & 67dc7a2 \\ + fadafc6f2be1 & 4 & 12 & 68757964fe31 & 67dc7a2 \\ 8d6b3f5b24a2 & 0 & 16 & 3c20300a5e78 & 758d5f2 \\ b5b8cfd55dee & 1 & 16 & 22b25365d745 & 758d5f2 \\ 1a2ccd4ef091 & 2 & 16 & b6cea5e21b36 & 758d5f2 \\ diff --git a/paper/tables/T4_complementarity.tex b/paper/tables/T4_complementarity.tex index 5e07d78..e8ab310 100644 --- a/paper/tables/T4_complementarity.tex +++ b/paper/tables/T4_complementarity.tex @@ -1,4 +1,4 @@ -% auto-generated by analysis/make_tables.py — do not edit. run_ids=['aefb55fc7987', '8cfc5a222d83', '8ab8e76e3b44', '7d6cba43dccf', '10f375729230', '70741801c041', 'b0d0d7f2f513', '45d9d36ce766', 'f301fe7ebb4d', '39101f30bcb7', '7802e0276803', 'd80ff5b236eb', '198178fd3f60', '8867921abd20', 'c81e3990d838', '2a91bbeda6e4', 'a4a1a13b8234', 'a009b8f5ff16', 'c85bba3fffca', '6d2a4e347994'] +% auto-generated by analysis/make_tables.py — do not edit. run_ids=['59f889f657d5', 'b04ffc85ccc3', '354783efb5d1', '5bd7ea239e4f', '52b5abb0b7f5', '70741801c041', 'b0d0d7f2f513', '45d9d36ce766', 'f301fe7ebb4d', '39101f30bcb7', '2165fc5b4ecc', '496414a8f2cd', '9202e8561acd', '8bc3880fccd3', 'fadafc6f2be1', '2a91bbeda6e4', 'a4a1a13b8234', 'a009b8f5ff16', 'c85bba3fffca', '6d2a4e347994'] \begin{table}[t] \centering \caption{Complementarity (mean$\pm$std over seeds): head-to-head geometry vs the best softmax baseline, and whether geometry \emph{adds} conditional signal on top of a learned softmax (a mapper on [geometry $\cup$ softmax] vs [softmax]). $>0$ with the CI clearing 0 (``seeds add'') means geometry is complementary, even where it ties head-to-head.} @@ -8,9 +8,9 @@ \toprule task & reasoner & $\Delta$AURC vs softmax & $\Delta$AURC geom \emph{adds} & seeds add \\ \midrule - arithmetic & basin\_center & $-0.011 \pm 0.006$ & $0.001 \pm 0.002$ & 0/5 \\ + arithmetic & basin\_center & $-0.008 \pm 0.008$ & $0.004 \pm 0.002$ & 3/5 \\ arithmetic & ired & $0.007 \pm 0.005$ & $0.007 \pm 0.006$ & 4/5 \\ - graph\_planning & basin\_center & $-0.034 \pm 0.008$ & $-0.006 \pm 0.004$ & 0/5 \\ + graph\_planning & basin\_center & $-0.030 \pm 0.004$ & $-0.002 \pm 0.001$ & 0/5 \\ graph\_planning & ired & $0.001 \pm 0.003$ & $0.001 \pm 0.003$ & 0/5 \\ \bottomrule \end{tabular}} diff --git a/paper/tables/T5_richer_geometry.tex b/paper/tables/T5_richer_geometry.tex index e1554e5..1739122 100644 --- a/paper/tables/T5_richer_geometry.tex +++ b/paper/tables/T5_richer_geometry.tex @@ -1,4 +1,4 @@ -% auto-generated by analysis/make_tables.py — do not edit. run_ids=['aefb55fc7987', '8cfc5a222d83', '8ab8e76e3b44', '7d6cba43dccf', '10f375729230', 'afd5e1407276', 'eb3b1db3b4e5', '7c7410c6f66f', 'd117be743bac', '682227a040a9', '70741801c041', 'b0d0d7f2f513', '45d9d36ce766', 'f301fe7ebb4d', '39101f30bcb7', 'ec2457a78f8d', '34bc023358b8', '28ac1614501c', '39b5492337a7', '61d34e2aec3a', '7802e0276803', 'd80ff5b236eb', '198178fd3f60', '8867921abd20', 'c81e3990d838', 'c468dfa50c2c', '69bcb88d896d', 'affc597f3e34', '37776ebe2a5b', 'e7b91099dfc5', '2a91bbeda6e4', 'a4a1a13b8234', 'a009b8f5ff16', 'c85bba3fffca', '6d2a4e347994', 'abb4dcb4ea84', '6c7eb2cdb623', '9fdd2103f70b', 'cb3102b09662', 'a405cf7b867a'] +% auto-generated by analysis/make_tables.py — do not edit. run_ids=['59f889f657d5', 'b04ffc85ccc3', '354783efb5d1', '5bd7ea239e4f', '52b5abb0b7f5', 'afd5e1407276', 'eb3b1db3b4e5', '7c7410c6f66f', 'd117be743bac', '682227a040a9', '70741801c041', 'b0d0d7f2f513', '45d9d36ce766', 'f301fe7ebb4d', '39101f30bcb7', 'ec2457a78f8d', '34bc023358b8', '28ac1614501c', '39b5492337a7', '61d34e2aec3a', '2165fc5b4ecc', '496414a8f2cd', '9202e8561acd', '8bc3880fccd3', 'fadafc6f2be1', 'c468dfa50c2c', '69bcb88d896d', 'affc597f3e34', '37776ebe2a5b', 'e7b91099dfc5', '2a91bbeda6e4', 'a4a1a13b8234', 'a009b8f5ff16', 'c85bba3fffca', '6d2a4e347994', 'abb4dcb4ea84', '6c7eb2cdb623', '9fdd2103f70b', 'cb3102b09662', 'a405cf7b867a'] \begin{table}[t] \centering \caption{Richer geometry (mean$\pm$std over seeds): the base 14-feature vector vs the base plus the full Hessian spectrum and mode-connectivity barriers. If ``richer`` does not raise $\Delta$AURC vs softmax or the conditional ``geom adds`` on the graph cell, the graph redundancy is a task property, not an artifact of thin curvature features.} @@ -8,11 +8,11 @@ \toprule task & reasoner & features & AURC geom & $\Delta$AURC vs softmax & $\Delta$AURC geom \emph{adds} & seeds add \\ \midrule - arithmetic & basin\_center & base & $0.059 \pm 0.018$ & $-0.011 \pm 0.006$ & $0.001 \pm 0.002$ & 0/5 \\ + arithmetic & basin\_center & base & $0.059 \pm 0.018$ & $-0.008 \pm 0.008$ & $0.004 \pm 0.002$ & 3/5 \\ arithmetic & basin\_center & richer & $0.061 \pm 0.020$ & $-0.012 \pm 0.007$ & $-0.002 \pm 0.005$ & 1/5 \\ arithmetic & ired & base & $0.028 \pm 0.028$ & $0.007 \pm 0.005$ & $0.007 \pm 0.006$ & 4/5 \\ arithmetic & ired & richer & $0.028 \pm 0.029$ & $0.008 \pm 0.006$ & $0.007 \pm 0.006$ & 4/5 \\ - graph\_planning & basin\_center & base & $0.159 \pm 0.012$ & $-0.034 \pm 0.008$ & $-0.006 \pm 0.004$ & 0/5 \\ + graph\_planning & basin\_center & base & $0.159 \pm 0.016$ & $-0.030 \pm 0.004$ & $-0.002 \pm 0.001$ & 0/5 \\ graph\_planning & basin\_center & richer & $0.162 \pm 0.014$ & $-0.037 \pm 0.011$ & $-0.009 \pm 0.006$ & 0/5 \\ graph\_planning & ired & base & $0.056 \pm 0.005$ & $0.001 \pm 0.003$ & $0.001 \pm 0.003$ & 0/5 \\ graph\_planning & ired & richer & $0.055 \pm 0.005$ & $0.002 \pm 0.002$ & $-0.002 \pm 0.008$ & 0/5 \\ diff --git a/paper/tables/T6_halting.tex b/paper/tables/T6_halting.tex new file mode 100644 index 0000000..efba281 --- /dev/null +++ b/paper/tables/T6_halting.tex @@ -0,0 +1,15 @@ +% auto-generated by analysis/make_tables.py — do not edit. run_ids=['a6c36d85615f', '5b496a841123', '24d91758f427', '19cb5e9f64c6', '2a54cf585cb3', 'eaebf4a2b9f3', '20d1395ffa7a', '70cee2dae378', '470896cedd8b', 'ed919980f35e'] +\begin{table}[t] + \centering + \caption{Adaptive halting (CRC), mean$\pm$std over seeds: the basin-agreement threshold $\hat\tau$ is calibrated on a disjoint fold so the halting risk (disagreement with the full-budget answer) stays within the budget $\alpha$. ``compute saved`` is the fraction of descent steps skipped; ``seeds $\le\alpha$`` counts seeds with test risk in budget.} + \label{tab:halting} + \resizebox{\ifdim\width>\linewidth\linewidth\else\width\fi}{!}{% + \begin{tabular}{lcccccc} + \toprule + task & $\alpha$ & $\hat\tau$ & compute saved & halting risk & acc drop & seeds $\le\alpha$ \\ + \midrule + arithmetic & $0.1$ & $0.92 \pm 0.00$ & $0.571 \pm 0.017$ & $0.012 \pm 0.005$ & $-0.007 \pm 0.006$ & 5/5 \\ + graph\_planning & $0.1$ & $0.85 \pm 0.03$ & $0.808 \pm 0.052$ & $0.065 \pm 0.032$ & $0.011 \pm 0.007$ & 5/5 \\ + \bottomrule + \end{tabular}} +\end{table} diff --git a/paper/tables/T7_ood_stress.tex b/paper/tables/T7_ood_stress.tex new file mode 100644 index 0000000..e55f7ee --- /dev/null +++ b/paper/tables/T7_ood_stress.tex @@ -0,0 +1,15 @@ +% auto-generated by analysis/make_tables.py — do not edit. run_ids=['59f889f657d5', 'b04ffc85ccc3', '354783efb5d1', '5bd7ea239e4f', '52b5abb0b7f5', '2165fc5b4ecc', '496414a8f2cd', '9202e8561acd', '8bc3880fccd3', 'fadafc6f2be1'] +\begin{table}[t] + \centering + \caption{OOD stress (mean$\pm$std over seeds): the ID-calibrated selective-prediction threshold applied to a shifted fold. The selective-risk guarantee holds in-distribution but breaks under shift (OOD risk exceeds the budget), motivating abstention; ``seeds OOD ok`` counts seeds whose OOD risk still stays within budget.} + \label{tab:oodstress} + \resizebox{\ifdim\width>\linewidth\linewidth\else\width\fi}{!}{% + \begin{tabular}{lccccc} + \toprule + task & acc ID & acc OOD & sel.\ risk ID & sel.\ risk OOD & seeds OOD ok \\ + \midrule + arithmetic & $0.83 \pm 0.02$ & $0.21 \pm 0.01$ & $0.077 \pm 0.004$ & $0.764 \pm 0.012$ & 0/5 \\ + graph\_planning & $0.70 \pm 0.02$ & $0.33 \pm 0.05$ & $0.010 \pm 0.019$ & $0.168 \pm 0.336$ & 4/5 \\ + \bottomrule + \end{tabular}} +\end{table} diff --git a/pyproject.toml b/pyproject.toml index f12bb4c..71abfc5 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -1,6 +1,6 @@ [project] name = "energy-descent-certificates" -version = "0.0.4" +version = "0.0.5" description = "Distribution-free selective prediction and adaptive halting from the geometry of an energy-based reasoner's inference-time descent." readme = "README.md" requires-python = ">=3.12,<3.13" diff --git a/results/ledger.jsonl b/results/ledger.jsonl index 15ff1ff..969af07 100644 --- a/results/ledger.jsonl +++ b/results/ledger.jsonl @@ -112,3 +112,23 @@ {"run_id": "9fdd2103f70b", "timestamp": "2026-07-28T23:18:15.037116+00:00", "git_sha": "9367d60", "config_hash": "a55d25d137ca", "config": {"run": {"seed": 2, "task": "graph_planning", "notes": "stronger IRED learned-landscape reasoner on graph"}, "model": {"latent_dim": 48, "hidden_dim": 256, "context_dim": 64}, "inference": {"k_restarts": 12, "steps": 50, "step_size": 0.1, "temperature": 0.01, "init_scale": 1.0, "grad_tol": 0.001, "sampler": "annealed", "anneal_levels": 20, "anneal_steps_per_level": 10, "anneal_step_max": 0.5, "anneal_step_min": 0.003}, "train": {"epochs": 110, "batch_size": 128, "lr": 0.001, "n_train": 12000, "n_neg": 4, "neg_noise": 0.5, "objective": "ired", "ired_noise_min": 0.1, "ired_noise_max": 1.5, "ired_ridge": 0.2, "ired_decode_weight": 4.0, "ired_stat_weight": 8.0}, "eval": {"n_eval": 700, "richer_geometry": true}, "conformal": {"alpha": 0.1, "delta": 0.05, "n_calib": 800}, "task": {"graph_planning": {"n_nodes": 7, "ood_n_nodes": 10, "edge_prob": 0.4, "max_len": 4}}}, "task": "graph_planning", "split": "selective", "seed": 2, "metrics": {"n_fit": 700, "n_calib": 800, "n_test": 700, "k_restarts": 12, "objective": "ired", "sampler": "annealed", "feature_set": "richer", "accuracy_id": 0.8214285714285714, "base_error": 0.17857142857142858, "final_train_loss": 0.1651451736688614, "feature_names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual", "spectrum/lmin_mean", "spectrum/lmin_best", "spectrum/negfrac_mean", "spectrum/negfrac_best", "spectrum/effrank_mean", "spectrum/effrank_best", "spectrum/logdet_mean", "spectrum/logdet_best", "connect/barrier_mean", "connect/barrier_max", "connect/connected_frac"], "aurc": {"geometry": 0.060251437947040715, "rho_basin": 0.09443079176745177, "energy_min": 0.19082820106844728, "energy_mean": 0.1866157548243378, "energy_std": 0.17904397128304222, "msp": 0.07817313287375449, "temp_msp": 0.061490269632624606, "entropy": 0.07723791360586213, "softmax_learned": 0.06164043542026263, "geom_softmax": 0.061066118791798824}, "temperature": 7.072506528454141, "best_energy_baseline": "energy_std", "best_baseline": "temp_msp", "delta_aurc_vs_energy_min": [0.13057676312140656, 0.09222572919121763, 0.17134522211586267], "delta_aurc_vs_best_energy": [0.11879253333600151, 0.08170908894261716, 0.1550689073356496], "delta_aurc_vs_best_baseline": [0.001238831685583891, -0.01255199296098033, 0.014540953603940372], "delta_aurc_geom_adds": [0.0005743166284638071, -0.011961335862339472, 0.01325108482993284], "geometry_wins": true, "geometry_wins_vs_baseline": false, "geometry_adds_over_softmax": false, "risk_coverage": {"geometry": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.07142857142857142, 0.03571428571428571, 0.023809523809523808, 0.017857142857142856, 0.014285714285714285, 0.011904761904761902, 0.010204081632653062, 0.008928571428571428, 0.007936507936507936, 0.0071428571428571435, 0.006493506493506494, 0.005952380952380952, 0.005494505494505495, 0.00510204081632653, 0.0047619047619047615, 0.004464285714285714, 0.004201680672268907, 0.007936507936507934, 0.007518796992481203, 0.007142857142857143, 0.006802721088435373, 0.006493506493506494, 0.012422360248447204, 0.023809523809523805, 0.03428571428571429, 0.04120879120879121, 0.04497354497354497, 0.05102040816326531, 0.0541871921182266, 0.05714285714285714, 0.06451612903225806, 0.07142857142857142, 0.07575757575757576, 0.07983193277310924, 0.09183673469387754, 0.09920634920634919, 0.10617760617760617, 0.11466165413533834, 0.12637362637362637, 0.13035714285714287, 0.13588850174216024, 0.14285714285714282, 0.1461794019933555, 0.1525974025974026, 0.15555555555555556, 0.16304347826086957, 0.16413373860182368, 0.168154761904762, 0.17346938775510204, 0.17857142857142858]}, "rho_basin": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.07726269315673287, 0.07726269315673287, 0.07726269315673294, 0.07726269315673297, 0.077262693156733, 0.077262693156733, 0.07726269315673301, 0.07726269315673302, 0.07726269315673302, 0.07726269315673302, 0.07726269315673304, 0.07726269315673304, 0.07726269315673304, 0.07726269315673304, 0.07726269315673304, 0.07726269315673304, 0.07726269315673304, 0.07726269315673304, 0.07726269315673304, 0.07726269315673304, 0.07726269315673304, 0.07726269315673304, 0.07726269315673304, 0.07726269315673304, 0.07726269315673305, 0.07726269315673305, 0.07726269315673305, 0.07726269315673305, 0.07726269315673305, 0.07726269315673305, 0.07726269315673305, 0.07726269315673305, 0.08130081300813018, 0.08727881396461024, 0.09291521486643432, 0.09823848238482372, 0.10327400571302986, 0.10804450149764624, 0.11257035647279512, 0.11686991869918655, 0.1209597461828272, 0.12714609653385106, 0.133417708168538, 0.13940424654710287, 0.14512471655328707, 0.15167480035492367, 0.1598458532349103, 0.16721491228070112, 0.17346938775510137, 0.17857142857142796]}, "energy_min": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.5, 0.35714285714285715, 0.2857142857142857, 0.25, 0.24285714285714285, 0.23809523809523805, 0.21428571428571433, 0.21428571428571427, 0.20634920634920634, 0.1928571428571429, 0.19480519480519481, 0.19047619047619047, 0.1813186813186813, 0.17857142857142858, 0.1857142857142857, 0.17410714285714285, 0.17647058823529413, 0.1706349206349206, 0.16917293233082706, 0.17142857142857143, 0.16666666666666663, 0.1590909090909091, 0.16459627329192547, 0.16369047619047616, 0.16285714285714287, 0.16483516483516483, 0.164021164021164, 0.16326530612244897, 0.15763546798029554, 0.15714285714285714, 0.16129032258064516, 0.16071428571428573, 0.16233766233766234, 0.16596638655462184, 0.16530612244897958, 0.16666666666666663, 0.1640926640926641, 0.16541353383458646, 0.16483516483516483, 0.16428571428571428, 0.1655052264808362, 0.16496598639455792, 0.17109634551495018, 0.17694805194805194, 0.1761904761904762, 0.17391304347826086, 0.17629179331306988, 0.17559523809523805, 0.17346938775510204, 0.17857142857142858]}, "energy_mean": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.6428571428571429, 0.35714285714285715, 0.2619047619047619, 0.23214285714285715, 0.24285714285714285, 0.21428571428571425, 0.19387755102040818, 0.17857142857142858, 0.18253968253968253, 0.1857142857142856, 0.17532467532467533, 0.17857142857142858, 0.17032967032967034, 0.16326530612244897, 0.15238095238095237, 0.15178571428571427, 0.14705882352941177, 0.15476190476190474, 0.15413533834586465, 0.15357142857142858, 0.15646258503401358, 0.1525974025974026, 0.15527950310559005, 0.1607142857142857, 0.15428571428571428, 0.1565934065934066, 0.16137566137566137, 0.16071428571428573, 0.15517241379310343, 0.1523809523809524, 0.15668202764976957, 0.15848214285714285, 0.16017316017316016, 0.15756302521008403, 0.15918367346938772, 0.1607142857142857, 0.16216216216216217, 0.16165413533834586, 0.16117216117216118, 0.16428571428571428, 0.16550522648083635, 0.17006802721088432, 0.16943521594684385, 0.17045454545454544, 0.16666666666666666, 0.16770186335403728, 0.16717325227963523, 0.168154761904762, 0.17346938775510204, 0.17857142857142858]}, "energy_std": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.35714285714285715, 0.25, 0.19047619047619047, 0.21428571428571427, 0.18571428571428572, 0.19047619047619044, 0.17346938775510207, 0.17857142857142858, 0.18253968253968253, 0.1785714285714286, 0.18181818181818182, 0.17857142857142858, 0.17582417582417584, 0.1683673469387755, 0.1714285714285714, 0.17410714285714285, 0.1722689075630252, 0.17460317460317457, 0.17293233082706766, 0.17142857142857143, 0.17006802721088432, 0.16233766233766234, 0.15838509316770186, 0.16369047619047616, 0.16285714285714287, 0.16758241758241757, 0.1693121693121693, 0.17091836734693877, 0.17241379310344845, 0.1738095238095238, 0.17511520737327188, 0.171875, 0.16883116883116883, 0.16596638655462184, 0.16530612244897974, 0.16666666666666663, 0.16602316602316602, 0.16541353383458646, 0.1684981684981685, 0.16607142857142856, 0.16376306620209055, 0.16496598639455778, 0.16611295681063123, 0.1672077922077922, 0.16825396825396827, 0.16925465838509315, 0.16869300911854102, 0.1681547619047619, 0.1749271137026239, 0.17857142857142858]}, "msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.03571428571428571, 0.047619047619047616, 0.03571428571428571, 0.04285714285714286, 0.03571428571428582, 0.04081632653061225, 0.044642857142857144, 0.03968253968253968, 0.042857142857142864, 0.03896103896103896, 0.03571428571428571, 0.038461538461538464, 0.04081632653061224, 0.03809523809523809, 0.044642857142857144, 0.04201680672268908, 0.05158730158730158, 0.05263157894736842, 0.05714285714285714, 0.054421768707482984, 0.05194805194805195, 0.055900621118012424, 0.06249999999999999, 0.06, 0.07417582417582418, 0.07142857142857142, 0.07142857142857142, 0.07142857142857141, 0.0761904761904762, 0.07603686635944701, 0.078125, 0.08441558441558442, 0.09453781512605042, 0.09795918367346952, 0.10515873015873015, 0.11003861003861004, 0.11842105263157894, 0.12087912087912088, 0.125, 0.1289198606271778, 0.13265306122448992, 0.1362126245847176, 0.1396103896103896, 0.14603174603174604, 0.14906832298136646, 0.15653495440729495, 0.16369047619047628, 0.16909620991253643, 0.17857142857142858]}, "temp_msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.07142857142857142, 0.03571428571428571, 0.023809523809523808, 0.017857142857142856, 0.014285714285714285, 0.011904761904761902, 0.010204081632653062, 0.008928571428571428, 0.007936507936507936, 0.0071428571428571435, 0.006493506493506494, 0.005952380952380952, 0.005494505494505495, 0.00510204081632653, 0.0047619047619047615, 0.004464285714285714, 0.004201680672268907, 0.007936507936507934, 0.007518796992481203, 0.007142857142857143, 0.010204081632653059, 0.00974025974025974, 0.009316770186335404, 0.02678571428571428, 0.03428571428571429, 0.04120879120879121, 0.05291005291005291, 0.061224489795918366, 0.06403940886699507, 0.07142857142857142, 0.08525345622119816, 0.08482142857142858, 0.08874458874458875, 0.09033613445378151, 0.09591836734693891, 0.10317460317460315, 0.11196911196911197, 0.11654135338345864, 0.12087912087912088, 0.12678571428571428, 0.12717770034843218, 0.13605442176870747, 0.1362126245847176, 0.1396103896103896, 0.14603174603174604, 0.14751552795031056, 0.15501519756838916, 0.16369047619047616, 0.16909620991253643, 0.17857142857142858]}, "entropy": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.03571428571428571, 0.047619047619047616, 0.03571428571428571, 0.02857142857142857, 0.035714285714285705, 0.04081632653061225, 0.044642857142857144, 0.03968253968253968, 0.03571428571428572, 0.03896103896103896, 0.03571428571428571, 0.038461538461538464, 0.04081632653061224, 0.03809523809523809, 0.04017857142857143, 0.04201680672268908, 0.04761904761904761, 0.05263157894736842, 0.05357142857142857, 0.054421768707482984, 0.05194805194805195, 0.052795031055900624, 0.059523809523809514, 0.06, 0.06868131868131869, 0.07142857142857142, 0.07142857142857142, 0.07142857142857141, 0.0761904761904762, 0.07603686635944701, 0.078125, 0.08441558441558442, 0.09453781512605042, 0.09795918367346952, 0.10515873015873015, 0.11003861003861004, 0.11842105263157894, 0.12087912087912088, 0.12678571428571428, 0.13066202090592333, 0.13265306122448992, 0.1362126245847176, 0.1396103896103896, 0.14603174603174604, 0.14906832298136646, 0.15653495440729495, 0.16369047619047628, 0.16909620991253643, 0.17857142857142858]}, "softmax_learned": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.07142857142857142, 0.03571428571428571, 0.023809523809523808, 0.017857142857142856, 0.014285714285714285, 0.011904761904761902, 0.010204081632653062, 0.008928571428571428, 0.007936507936507936, 0.0071428571428571435, 0.006493506493506494, 0.005952380952380952, 0.005494505494505495, 0.00510204081632653, 0.0047619047619047615, 0.004464285714285714, 0.004201680672268907, 0.007936507936507934, 0.007518796992481203, 0.007142857142857143, 0.010204081632653059, 0.00974025974025974, 0.009316770186335404, 0.02678571428571428, 0.03428571428571429, 0.04120879120879121, 0.05291005291005291, 0.061224489795918366, 0.06403940886699507, 0.07142857142857142, 0.08525345622119816, 0.08482142857142858, 0.08874458874458875, 0.09033613445378151, 0.09591836734693891, 0.10317460317460315, 0.11196911196911197, 0.11654135338345864, 0.12087912087912088, 0.12678571428571428, 0.12717770034843218, 0.13605442176870747, 0.1362126245847176, 0.1396103896103896, 0.14603174603174604, 0.14751552795031056, 0.15501519756838902, 0.1651785714285714, 0.1749271137026239, 0.17857142857142858]}, "geom_softmax": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.03571428571428571, 0.023809523809523808, 0.017857142857142856, 0.014285714285714285, 0.011904761904761902, 0.010204081632653062, 0.008928571428571428, 0.007936507936507936, 0.0071428571428571435, 0.006493506493506494, 0.005952380952380952, 0.005494505494505495, 0.00510204081632653, 0.0047619047619047615, 0.004464285714285714, 0.004201680672268907, 0.007936507936507934, 0.007518796992481203, 0.007142857142857143, 0.006802721088435373, 0.006493506493506494, 0.012422360248447204, 0.020833333333333447, 0.03428571428571429, 0.04120879120879121, 0.05026455026455026, 0.05612244897959184, 0.059113300492611015, 0.06904761904761905, 0.07142857142857142, 0.08258928571428571, 0.09307359307359307, 0.09663865546218488, 0.10000000000000014, 0.10515873015873015, 0.11003861003861004, 0.11842105263157894, 0.12454212454212454, 0.13035714285714287, 0.13240418118466896, 0.13775510204081629, 0.14285714285714285, 0.14935064935064934, 0.15396825396825398, 0.15372670807453417, 0.16109422492401212, 0.16964285714285712, 0.17346938775510204, 0.17857142857142858]}}, "ltt": {"alpha": 0.1, "delta": 0.05, "lambda_hat": 0.16349649886293133, "coverage": 0.5314285714285715, "abstain_rate": 0.4685714285714286, "selective_risk": 0.0456989247311828, "selective_accuracy": 0.9543010752688172, "risk_within_budget": true}, "coverage_validity": {"alpha_grid": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "target": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "achieved_risk": [0.0, 0.006993006993006993, 0.0456989247311828, 0.1053639846743295, 0.16566265060240964, 0.17621776504297995], "coverage": [0.0, 0.4085714285714286, 0.5314285714285715, 0.7457142857142857, 0.9485714285714286, 0.9971428571428571]}, "feature_diagnostics": {"names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual", "spectrum/lmin_mean", "spectrum/lmin_best", "spectrum/negfrac_mean", "spectrum/negfrac_best", "spectrum/effrank_mean", "spectrum/effrank_best", "spectrum/logdet_mean", "spectrum/logdet_best", "connect/barrier_mean", "connect/barrier_max", "connect/connected_frac"], "auroc": {"basin/rho": 0.7423443478260869, "basin/entropy": 0.2577530434782609, "basin/dispersion": 0.48076521739130434, "energy/mean": 0.45377391304347825, "energy/min": 0.4752, "energy/std": 0.46964869565217393, "curv/lmax_mean": 0.49575652173913043, "curv/lmax_best": 0.4976904347826087, "curv/trace_mean": 0.5, "curv/trace_best": 0.5, "dynamics/steps": 0.5, "dynamics/monotonic": 0.5954643478260869, "dynamics/drop": 0.7773495652173913, "dynamics/residual": 0.4896973913043478, "spectrum/lmin_mean": 0.5, "spectrum/lmin_best": 0.5, "spectrum/negfrac_mean": 0.5, "spectrum/negfrac_best": 0.5, "spectrum/effrank_mean": 0.5, "spectrum/effrank_best": 0.5, "spectrum/logdet_mean": 0.5, "spectrum/logdet_best": 0.5, "connect/barrier_mean": 0.4774539130434783, "connect/barrier_max": 0.47738434782608696, "connect/connected_frac": 0.5224904347826087}, "hist": {"basin/rho": {"edges": [0.4166666666666667, 0.44583333333333336, 0.47500000000000003, 0.5041666666666667, 0.5333333333333333, 0.5625, 0.5916666666666667, 0.6208333333333333, 0.65, 0.6791666666666667, 0.7083333333333333, 0.7375, 0.7666666666666666, 0.7958333333333334, 0.825, 0.8541666666666666, 0.8833333333333333, 0.9125, 0.9416666666666667, 0.9708333333333332, 1.0], "correct_counts": [0, 0, 3, 0, 0, 5, 0, 0, 10, 0, 0, 13, 0, 0, 38, 0, 0, 88, 0, 418], "incorrect_counts": [1, 0, 2, 0, 0, 3, 0, 0, 9, 0, 0, 15, 0, 0, 25, 0, 0, 35, 0, 35]}, "basin/entropy": {"edges": [0.0, 0.05387781635334003, 0.10775563270668007, 0.16163344906002008, 0.21551126541336013, 0.2693890817667002, 0.32326689812004017, 0.3771447144733802, 0.43102253082672026, 0.4849003471800603, 0.5387781635334004, 0.5926559798867403, 0.6465337962400803, 0.7004116125934204, 0.7542894289467604, 0.8081672453001005, 0.8620450616534405, 0.9159228780067805, 0.9698006943601206, 1.0236785107134607, 1.0775563270668007], "correct_counts": [418, 0, 0, 0, 0, 88, 0, 0, 36, 0, 13, 10, 6, 2, 0, 0, 0, 1, 1, 0], "incorrect_counts": [35, 0, 0, 0, 0, 35, 0, 0, 24, 0, 14, 8, 5, 1, 0, 2, 0, 0, 0, 1]}, "basin/dispersion": {"edges": [1.98346868203938, 2.1969464865932773, 2.4104242911471747, 2.623902095701072, 2.8373799002549696, 3.050857704808867, 3.2643355093627644, 3.477813313916662, 3.6912911184705592, 3.9047689230244567, 4.118246727578354, 4.3317245321322515, 4.545202336686149, 4.758680141240046, 4.972157945793944, 5.185635750347841, 5.399113554901739, 5.612591359455636, 5.8260691640095335, 6.039546968563431, 6.253024773117328], "correct_counts": [312, 249, 1, 1, 0, 1, 0, 0, 0, 1, 2, 0, 2, 2, 0, 1, 1, 1, 0, 1], "incorrect_counts": [62, 53, 0, 0, 0, 1, 2, 0, 0, 2, 0, 2, 1, 1, 1, 0, 0, 0, 0, 0]}, "energy/mean": {"edges": [-2.283770283063253, -2.1011410554250083, -1.9185118277867637, -1.735882600148519, -1.5532533725102744, -1.3706241448720298, -1.1879949172337851, -1.0053656895955405, -0.8227364619572959, -0.6401072343190513, -0.4574780066808066, -0.274848779042562, -0.09221955140431737, 0.09040967623392726, 0.2730389038721719, 0.4556681315104165, 0.6382973591486611, 0.8209265867869058, 1.0035558144251504, 1.186185042063395, 1.3688142697016399], "correct_counts": [1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 6, 514, 44, 1, 5, 2, 1, 1], "incorrect_counts": [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 8, 96, 10, 3, 4, 3, 0, 0]}, "energy/min": {"edges": [-2.531198501586914, -2.3657448172569273, -2.200291132926941, -2.0348374485969543, -1.8693837642669677, -1.7039300799369812, -1.5384763956069945, -1.373022711277008, -1.2075690269470214, -1.0421153426170349, -0.8766616582870483, -0.7112079739570616, -0.5457542896270751, -0.38030060529708853, -0.21484692096710178, -0.04939323663711548, 0.11606044769287127, 0.281514132022858, 0.4469678163528443, 0.6124215006828311, 0.7778751850128174], "correct_counts": [1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 320, 236, 15, 1, 0], "incorrect_counts": [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 2, 60, 55, 2, 1, 4]}, "energy/std": {"edges": [0.020960167890197088, 0.10303025054062304, 0.185100333191049, 0.26717041584147494, 0.3492404984919009, 0.43131058114232684, 0.5133806637927528, 0.5954507464431787, 0.6775208290936047, 0.7595909117440306, 0.8416609943944565, 0.9237310770448826, 1.0058011596953085, 1.0878712423457346, 1.1699413249961605, 1.2520114076465865, 1.3340814902970124, 1.4161515729474383, 1.4982216555978642, 1.5802917382482902, 1.662361820898716], "correct_counts": [562, 1, 2, 1, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 3, 1, 1, 0, 1, 1], "incorrect_counts": [117, 0, 3, 0, 0, 1, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0]}, "curv/lmax_mean": {"edges": [0.1999999893208345, 0.19999999056259793, 0.19999999180436134, 0.19999999304612479, 0.19999999428788823, 0.19999999552965164, 0.19999999677141508, 0.19999999801317853, 0.19999999925494194, 0.20000000049670538, 0.20000000173846882, 0.20000000298023224, 0.20000000422199568, 0.2000000054637591, 0.20000000670552254, 0.20000000794728598, 0.2000000091890494, 0.20000001043081284, 0.20000001167257628, 0.2000000129143397, 0.20000001415610313], "correct_counts": [5, 0, 3, 25, 0, 12, 69, 0, 74, 140, 0, 58, 52, 51, 38, 18, 14, 9, 1, 6], "incorrect_counts": [0, 0, 1, 4, 0, 4, 21, 0, 9, 29, 0, 14, 9, 18, 8, 3, 4, 1, 0, 0]}, "curv/lmax_best": {"edges": [0.19999995827674866, 0.19999996200203896, 0.19999996572732925, 0.19999996945261955, 0.19999997317790985, 0.19999997690320015, 0.19999998062849045, 0.19999998435378075, 0.19999998807907104, 0.19999999180436134, 0.19999999552965164, 0.19999999925494194, 0.20000000298023224, 0.20000000670552254, 0.20000001043081284, 0.20000001415610313, 0.20000001788139343, 0.20000002160668373, 0.20000002533197403, 0.20000002905726433, 0.20000003278255463], "correct_counts": [2, 0, 0, 0, 5, 0, 0, 0, 221, 0, 0, 0, 149, 0, 0, 0, 186, 0, 0, 12], "incorrect_counts": [0, 0, 0, 0, 1, 0, 0, 0, 48, 0, 0, 0, 32, 0, 0, 0, 43, 0, 0, 1]}, "curv/trace_mean": {"edges": [9.600000381469727, 9.650000381469727, 9.700000381469726, 9.750000381469727, 9.800000381469726, 9.850000381469727, 9.900000381469727, 9.950000381469726, 10.000000381469727, 10.050000381469726, 10.100000381469727, 10.150000381469727, 10.200000381469726, 10.250000381469727, 10.300000381469726, 10.350000381469727, 10.400000381469727, 10.450000381469726, 10.500000381469727, 10.550000381469726, 10.600000381469727], "correct_counts": [575, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [125, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "curv/trace_best": {"edges": [9.600000381469727, 9.650000381469727, 9.700000381469726, 9.750000381469727, 9.800000381469726, 9.850000381469727, 9.900000381469727, 9.950000381469726, 10.000000381469727, 10.050000381469726, 10.100000381469727, 10.150000381469727, 10.200000381469726, 10.250000381469727, 10.300000381469726, 10.350000381469727, 10.400000381469727, 10.450000381469726, 10.500000381469727, 10.550000381469726, 10.600000381469727], "correct_counts": [575, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [125, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "dynamics/steps": {"edges": [1.0, 1.05, 1.1, 1.15, 1.2, 1.25, 1.3, 1.35, 1.4, 1.45, 1.5, 1.55, 1.6, 1.65, 1.7000000000000002, 1.75, 1.8, 1.85, 1.9, 1.9500000000000002, 2.0], "correct_counts": [575, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [125, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "dynamics/monotonic": {"edges": [0.5341666666666667, 0.5398958333333334, 0.545625, 0.5513541666666667, 0.5570833333333334, 0.5628124999999999, 0.5685416666666666, 0.5742708333333333, 0.58, 0.5857291666666666, 0.5914583333333333, 0.5971875, 0.6029166666666667, 0.6086458333333333, 0.614375, 0.6201041666666667, 0.6258333333333332, 0.6315624999999999, 0.6372916666666666, 0.6430208333333333, 0.6487499999999999], "correct_counts": [1, 9, 47, 95, 139, 148, 80, 38, 13, 2, 2, 0, 0, 0, 0, 0, 0, 0, 0, 1], "incorrect_counts": [3, 5, 12, 25, 31, 31, 14, 1, 2, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0]}, "dynamics/drop": {"edges": [5.8192191918691, 6.054377051194509, 6.289534910519918, 6.524692769845327, 6.759850629170735, 6.995008488496144, 7.230166347821553, 7.4653242071469625, 7.700482066472372, 7.93563992579778, 8.17079778512319, 8.405955644448598, 8.641113503774008, 8.876271363099416, 9.111429222424825, 9.346587081750235, 9.581744941075643, 9.816902800401053, 10.052060659726461, 10.28721851905187, 10.52237637837728], "correct_counts": [12, 12, 29, 49, 54, 41, 29, 18, 4, 5, 5, 18, 40, 53, 55, 63, 42, 31, 12, 3], "incorrect_counts": [7, 8, 13, 28, 18, 19, 19, 4, 4, 1, 1, 1, 0, 1, 0, 0, 1, 0, 0, 0]}, "dynamics/residual": {"edges": [0.280464556068182, 0.2857741178944707, 0.2910836797207594, 0.2963932415470481, 0.30170280337333677, 0.3070123651996255, 0.3123219270259142, 0.3176314888522029, 0.3229410506784916, 0.32825061250478027, 0.333560174331069, 0.3388697361573577, 0.3441792979836464, 0.3494888598099351, 0.35479842163622377, 0.3601079834625125, 0.3654175452888012, 0.3707271071150899, 0.3760366689413786, 0.38134623076766727, 0.386655792593956], "correct_counts": [2, 9, 38, 63, 99, 132, 113, 66, 36, 11, 1, 2, 1, 0, 0, 0, 1, 0, 0, 1], "incorrect_counts": [0, 2, 3, 15, 23, 32, 27, 12, 6, 4, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0]}, "spectrum/lmin_mean": {"edges": [0.20000000298023224, 0.2500000029802322, 0.3000000029802322, 0.35000000298023226, 0.40000000298023225, 0.45000000298023224, 0.5000000029802323, 0.5500000029802323, 0.6000000029802323, 0.6500000029802322, 0.7000000029802322, 0.7500000029802323, 0.8000000029802323, 0.8500000029802323, 0.9000000029802323, 0.9500000029802322, 1.0000000029802323, 1.0500000029802323, 1.1000000029802321, 1.1500000029802324, 1.2000000029802322], "correct_counts": [575, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [125, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "spectrum/lmin_best": {"edges": [0.20000000298023224, 0.2500000029802322, 0.3000000029802322, 0.35000000298023226, 0.40000000298023225, 0.45000000298023224, 0.5000000029802323, 0.5500000029802323, 0.6000000029802323, 0.6500000029802322, 0.7000000029802322, 0.7500000029802323, 0.8000000029802323, 0.8500000029802323, 0.9000000029802323, 0.9500000029802322, 1.0000000029802323, 1.0500000029802323, 1.1000000029802321, 1.1500000029802324, 1.2000000029802322], "correct_counts": [575, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [125, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "spectrum/negfrac_mean": {"edges": [0.0, 0.05, 0.1, 0.15000000000000002, 0.2, 0.25, 0.30000000000000004, 0.35000000000000003, 0.4, 0.45, 0.5, 0.55, 0.6000000000000001, 0.65, 0.7000000000000001, 0.75, 0.8, 0.8500000000000001, 0.9, 0.9500000000000001, 1.0], "correct_counts": [575, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [125, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "spectrum/negfrac_best": {"edges": [0.0, 0.05, 0.1, 0.15000000000000002, 0.2, 0.25, 0.30000000000000004, 0.35000000000000003, 0.4, 0.45, 0.5, 0.55, 0.6000000000000001, 0.65, 0.7000000000000001, 0.75, 0.8, 0.8500000000000001, 0.9, 0.9500000000000001, 1.0], "correct_counts": [575, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [125, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "spectrum/effrank_mean": {"edges": [47.99999975000001, 48.049999750000005, 48.09999975000001, 48.149999750000006, 48.19999975000001, 48.24999975000001, 48.299999750000005, 48.34999975000001, 48.399999750000006, 48.44999975000001, 48.49999975000001, 48.549999750000005, 48.59999975000001, 48.649999750000006, 48.69999975000001, 48.74999975000001, 48.799999750000005, 48.84999975000001, 48.899999750000006, 48.94999975000001, 48.99999975000001], "correct_counts": [575, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [125, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "spectrum/effrank_best": {"edges": [47.99999975000001, 48.049999750000005, 48.09999975000001, 48.149999750000006, 48.19999975000001, 48.24999975000001, 48.299999750000005, 48.34999975000001, 48.399999750000006, 48.44999975000001, 48.49999975000001, 48.549999750000005, 48.59999975000001, 48.649999750000006, 48.69999975000001, 48.74999975000001, 48.799999750000005, 48.84999975000001, 48.899999750000006, 48.94999975000001, 48.99999975000001], "correct_counts": [575, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [125, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "spectrum/logdet_mean": {"edges": [-77.25301668158117, -77.20301668158118, -77.15301668158118, -77.10301668158117, -77.05301668158117, -77.00301668158117, -76.95301668158118, -76.90301668158118, -76.85301668158117, -76.80301668158117, -76.75301668158117, -76.70301668158118, -76.65301668158118, -76.60301668158117, -76.55301668158117, -76.50301668158117, -76.45301668158118, -76.40301668158118, -76.35301668158117, -76.30301668158117, -76.25301668158117], "correct_counts": [575, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [125, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "spectrum/logdet_best": {"edges": [-77.25301668158119, -77.20301668158119, -77.15301668158119, -77.10301668158118, -77.05301668158118, -77.00301668158119, -76.95301668158119, -76.90301668158119, -76.85301668158118, -76.80301668158118, -76.75301668158119, -76.70301668158119, -76.65301668158119, -76.60301668158118, -76.55301668158118, -76.50301668158119, -76.45301668158119, -76.40301668158119, -76.35301668158118, -76.30301668158118, -76.25301668158119], "correct_counts": [575, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [125, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "connect/barrier_mean": {"edges": [0.0, 0.055023192275654186, 0.11004638455130837, 0.16506957682696255, 0.22009276910261674, 0.27511596137827093, 0.3301391536539251, 0.3851623459295793, 0.4401855382052335, 0.49520873048088765, 0.5502319227565419, 0.6052551150321961, 0.6602783073078502, 0.7153014995835044, 0.7703246918591586, 0.8253478841348127, 0.880371076410467, 0.9353942686861212, 0.9904174609617753, 1.0454406532374296, 1.1004638455130837], "correct_counts": [567, 3, 0, 1, 0, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [119, 2, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1]}, "connect/barrier_max": {"edges": [0.0, 0.06335251331329346, 0.12670502662658692, 0.1900575399398804, 0.25341005325317384, 0.3167625665664673, 0.3801150798797608, 0.44346759319305423, 0.5068201065063477, 0.5701726198196412, 0.6335251331329346, 0.6968776464462281, 0.7602301597595216, 0.823582673072815, 0.8869351863861085, 0.9502876996994019, 1.0136402130126954, 1.0769927263259889, 1.1403452396392824, 1.2036977529525756, 1.2670502662658691], "correct_counts": [564, 1, 0, 2, 0, 1, 0, 0, 0, 2, 0, 1, 0, 3, 0, 0, 0, 1, 0, 0], "incorrect_counts": [117, 0, 0, 1, 1, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 1, 2, 0, 0, 1]}, "connect/connected_frac": {"edges": [0.09090909090909091, 0.13636363636363635, 0.18181818181818182, 0.22727272727272727, 0.2727272727272727, 0.31818181818181823, 0.36363636363636365, 0.40909090909090906, 0.4545454545454546, 0.5, 0.5454545454545455, 0.5909090909090909, 0.6363636363636364, 0.6818181818181819, 0.7272727272727273, 0.7727272727272728, 0.8181818181818182, 0.8636363636363636, 0.9090909090909092, 0.9545454545454546, 1.0], "correct_counts": [0, 0, 0, 0, 0, 0, 0, 2, 0, 1, 0, 0, 1, 0, 2, 0, 3, 2, 0, 564], "incorrect_counts": [1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 2, 1, 0, 117]}}}, "feature_ablation": {"full": 0.060251437947040715, "drop_basin": 0.06621243031414509, "basin_only": 0.09591115960898094, "drop_energy": 0.06067653000248025, "energy_only": 0.18899204074897005, "drop_curv": 0.05970888894980699, "curv_only": 0.17780760616067234, "drop_dynamics": 0.11094637528088547, "dynamics_only": 0.06495557418904481, "drop_spectrum": 0.060251437947040715, "spectrum_only": 0.1785714285714291, "drop_connect": 0.059672699284536494, "connect_only": 0.17190996073703077}, "ece_geometry": 0.04131107511114491}, "env": {"python": "3.12.13", "platform": "macOS-26.5.2-arm64-arm-64bit", "jax": "0.11.0", "jax_backend": "cpu"}, "artifact_paths": []} {"run_id": "cb3102b09662", "timestamp": "2026-07-28T23:18:50.308360+00:00", "git_sha": "9367d60", "config_hash": "a0a621e5442c", "config": {"run": {"seed": 3, "task": "graph_planning", "notes": "stronger IRED learned-landscape reasoner on graph"}, "model": {"latent_dim": 48, "hidden_dim": 256, "context_dim": 64}, "inference": {"k_restarts": 12, "steps": 50, "step_size": 0.1, "temperature": 0.01, "init_scale": 1.0, "grad_tol": 0.001, "sampler": "annealed", "anneal_levels": 20, "anneal_steps_per_level": 10, "anneal_step_max": 0.5, "anneal_step_min": 0.003}, "train": {"epochs": 110, "batch_size": 128, "lr": 0.001, "n_train": 12000, "n_neg": 4, "neg_noise": 0.5, "objective": "ired", "ired_noise_min": 0.1, "ired_noise_max": 1.5, "ired_ridge": 0.2, "ired_decode_weight": 4.0, "ired_stat_weight": 8.0}, "eval": {"n_eval": 700, "richer_geometry": true}, "conformal": {"alpha": 0.1, "delta": 0.05, "n_calib": 800}, "task": {"graph_planning": {"n_nodes": 7, "ood_n_nodes": 10, "edge_prob": 0.4, "max_len": 4}}}, "task": "graph_planning", "split": "selective", "seed": 3, "metrics": {"n_fit": 700, "n_calib": 800, "n_test": 700, "k_restarts": 12, "objective": "ired", "sampler": "annealed", "feature_set": "richer", "accuracy_id": 0.8285714285714286, "base_error": 0.17142857142857143, "final_train_loss": 0.16593526303768158, "feature_names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual", "spectrum/lmin_mean", "spectrum/lmin_best", "spectrum/negfrac_mean", "spectrum/negfrac_best", "spectrum/effrank_mean", "spectrum/effrank_best", "spectrum/logdet_mean", "spectrum/logdet_best", "connect/barrier_mean", "connect/barrier_max", "connect/connected_frac"], "aurc": {"geometry": 0.0533799136280693, "rho_basin": 0.11253061398624546, "energy_min": 0.21096449789187388, "energy_mean": 0.21638804114554994, "energy_std": 0.16204516813051956, "msp": 0.14599198362345006, "temp_msp": 0.057174909505143715, "entropy": 0.14494194070453414, "softmax_learned": 0.05731160565959381, "geom_softmax": 0.05430276688183326}, "temperature": 7.043475628034012, "best_energy_baseline": "energy_std", "best_baseline": "temp_msp", "delta_aurc_vs_energy_min": [0.1575845842638046, 0.1181093627407647, 0.19629026325032864], "delta_aurc_vs_best_energy": [0.10866525450245026, 0.07712510933316896, 0.1425734539285489], "delta_aurc_vs_best_baseline": [0.0037949958770744155, -0.0030469980404765977, 0.010411379788991282], "delta_aurc_geom_adds": [0.00300883877776055, -0.0031222374730943523, 0.008727410350044167], "geometry_wins": true, "geometry_wins_vs_baseline": false, "geometry_adds_over_softmax": false, "risk_coverage": {"geometry": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.01020408163265306, 0.009523809523809523, 0.008928571428571428, 0.008403361344537815, 0.007936507936507934, 0.007518796992481203, 0.007142857142857143, 0.006802721088435373, 0.006493506493506494, 0.006211180124223602, 0.014880952380952493, 0.022857142857142857, 0.027472527472527472, 0.037037037037037035, 0.04336734693877551, 0.04926108374384236, 0.05952380952380952, 0.06451612903225806, 0.08035714285714286, 0.08441558441558442, 0.08823529411764706, 0.09387755102040815, 0.0912698412698414, 0.10617760617760617, 0.10714285714285714, 0.11721611721611722, 0.12321428571428572, 0.12369337979094075, 0.13095238095238093, 0.13953488372093023, 0.1444805194805195, 0.14603174603174604, 0.14596273291925466, 0.1474164133738603, 0.1562500000000001, 0.16326530612244897, 0.17142857142857143]}, "rho_basin": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.10282258064516127, 0.10282258064516134, 0.10282258064516135, 0.10282258064516127, 0.10282258064516123, 0.10282258064516127, 0.10282258064516137, 0.10282258064516143, 0.10282258064516149, 0.10282258064516153, 0.10282258064516156, 0.10282258064516145, 0.10282258064516135, 0.10282258064516127, 0.1028225806451612, 0.10282258064516113, 0.10282258064516107, 0.10282258064516102, 0.10282258064516098, 0.10282258064516094, 0.1028225806451609, 0.10282258064516087, 0.10282258064516082, 0.1028225806451608, 0.10282258064516077, 0.10282258064516075, 0.10282258064516073, 0.1028225806451607, 0.10282258064516069, 0.10282258064516067, 0.10282258064516066, 0.10282258064516063, 0.10282258064516062, 0.1028225806451606, 0.10282258064516059, 0.10582595870206428, 0.11085864625687578, 0.1156264555193288, 0.12014976174268162, 0.12444690265486667, 0.12853442693718908, 0.13242730720606752, 0.13613912327639346, 0.14047805642633135, 0.14540229885057374, 0.15011244377810992, 0.15462215700660203, 0.15968406593406492, 0.16472303206996985, 0.17142857142857038]}, "energy_min": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.35714285714285715, 0.39285714285714285, 0.3333333333333333, 0.32142857142857145, 0.2714285714285714, 0.23809523809523817, 0.23469387755102047, 0.24107142857142858, 0.23015873015873015, 0.22142857142857145, 0.21428571428571427, 0.2261904761904762, 0.22527472527472528, 0.2193877551020408, 0.2095238095238095, 0.20535714285714285, 0.21428571428571427, 0.21825396825396823, 0.20676691729323307, 0.20357142857142857, 0.19727891156462582, 0.19805194805194806, 0.19875776397515527, 0.1964285714285714, 0.19428571428571428, 0.19230769230769232, 0.1931216931216931, 0.18877551020408162, 0.18965517241379307, 0.1880952380952381, 0.18663594470046083, 0.19196428571428573, 0.18831168831168832, 0.18277310924369747, 0.17755102040816323, 0.17460317460317457, 0.17567567567567569, 0.17481203007518797, 0.17032967032967034, 0.17142857142857143, 0.1689895470383275, 0.16496598639455778, 0.1644518272425249, 0.16233766233766234, 0.16031746031746033, 0.15683229813664595, 0.15349544072948326, 0.150297619047619, 0.16034985422740525, 0.17142857142857143]}, "energy_mean": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.5, 0.39285714285714285, 0.40476190476190477, 0.4107142857142857, 0.35714285714285715, 0.3214285714285714, 0.28571428571428575, 0.25892857142857145, 0.25396825396825395, 0.24285714285714288, 0.22727272727272727, 0.20833333333333334, 0.1978021978021978, 0.19387755102040816, 0.19999999999999998, 0.19196428571428573, 0.20588235294117646, 0.19841269841269837, 0.20300751879699247, 0.19642857142857142, 0.2040816326530612, 0.19805194805194806, 0.1956521739130435, 0.1964285714285714, 0.2, 0.20054945054945056, 0.19576719576719576, 0.19642857142857142, 0.18965517241379326, 0.1880952380952381, 0.18202764976958524, 0.18080357142857142, 0.18614718614718614, 0.18277310924369747, 0.18163265306122447, 0.17857142857142855, 0.17374517374517376, 0.17293233082706766, 0.17032967032967034, 0.16964285714285715, 0.1655052264808362, 0.16496598639455778, 0.16279069767441862, 0.1590909090909091, 0.15714285714285714, 0.15527950310559005, 0.15197568389057747, 0.1532738095238095, 0.16034985422740525, 0.17142857142857143]}, "energy_std": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.21428571428571427, 0.17857142857142858, 0.19047619047619047, 0.19642857142857142, 0.18571428571428572, 0.16666666666666663, 0.15306122448979595, 0.13392857142857142, 0.12698412698412698, 0.1285714285714286, 0.12987012987012986, 0.13690476190476192, 0.13736263736263737, 0.15306122448979592, 0.1571428571428571, 0.15625, 0.1638655462184874, 0.16269841269841284, 0.16165413533834586, 0.16071428571428573, 0.15986394557823141, 0.16558441558441558, 0.16770186335403728, 0.17559523809523817, 0.17714285714285713, 0.17582417582417584, 0.1746031746031746, 0.17091836734693877, 0.1650246305418719, 0.16666666666666666, 0.16820276497695852, 0.16741071428571427, 0.1645021645021645, 0.16176470588235295, 0.16326530612244894, 0.16666666666666663, 0.16795366795366795, 0.16729323308270677, 0.16666666666666666, 0.16607142857142856, 0.16724738675958187, 0.16666666666666663, 0.16611295681063123, 0.16233766233766234, 0.16031746031746033, 0.16459627329192547, 0.16261398176291791, 0.16220238095238093, 0.1661807580174927, 0.17142857142857143]}, "msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.14285714285714285, 0.25, 0.23809523809523808, 0.19642857142857142, 0.17142857142857143, 0.15476190476190474, 0.15306122448979595, 0.16071428571428573, 0.15079365079365079, 0.1642857142857143, 0.16233766233766234, 0.15476190476190477, 0.15934065934065933, 0.16326530612244897, 0.15238095238095237, 0.14732142857142858, 0.14285714285714285, 0.13888888888888887, 0.14285714285714285, 0.1357142857142857, 0.13605442176870747, 0.13636363636363635, 0.13664596273291926, 0.1369047619047619, 0.13142857142857142, 0.12912087912087913, 0.13227513227513227, 0.13520408163265307, 0.13300492610837436, 0.12857142857142856, 0.1313364055299539, 0.13392857142857142, 0.13203463203463203, 0.13025210084033614, 0.1285714285714287, 0.12698412698412695, 0.12934362934362933, 0.12969924812030076, 0.13003663003663005, 0.13035714285714287, 0.13414634146341475, 0.13775510204081629, 0.14285714285714285, 0.1444805194805195, 0.1492063492063492, 0.15062111801242237, 0.15349544072948326, 0.15922619047619044, 0.1661807580174927, 0.17142857142857143]}, "temp_msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0047619047619047615, 0.004464285714285714, 0.004201680672268907, 0.003968253968253967, 0.0037593984962406013, 0.0035714285714285713, 0.006802721088435373, 0.00974025974025974, 0.018633540372670808, 0.02083333333333333, 0.022857142857142857, 0.03571428571428571, 0.047619047619047616, 0.061224489795918366, 0.07389162561576373, 0.08095238095238096, 0.08294930875576037, 0.08482142857142858, 0.09307359307359307, 0.09663865546218488, 0.1020408163265306, 0.11111111111111124, 0.11969111969111969, 0.11842105263157894, 0.1227106227106227, 0.12321428571428572, 0.12543554006968638, 0.13095238095238093, 0.1378737541528239, 0.1396103896103896, 0.14761904761904762, 0.15217391304347827, 0.15501519756838902, 0.1607142857142857, 0.16472303206997085, 0.17142857142857143]}, "entropy": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.21428571428571427, 0.21428571428571427, 0.21428571428571427, 0.19642857142857142, 0.17142857142857143, 0.15476190476190474, 0.14285714285714288, 0.16071428571428573, 0.15873015873015872, 0.1642857142857143, 0.15584415584415584, 0.15476190476190477, 0.15934065934065933, 0.15306122448979592, 0.15238095238095237, 0.14732142857142858, 0.14285714285714285, 0.13888888888888887, 0.13909774436090225, 0.1357142857142857, 0.12925170068027222, 0.1331168831168831, 0.13043478260869565, 0.1339285714285714, 0.13142857142857142, 0.12912087912087913, 0.12962962962962962, 0.1326530612244898, 0.13300492610837436, 0.12857142857142856, 0.12903225806451613, 0.13392857142857142, 0.13203463203463203, 0.13025210084033614, 0.12857142857142853, 0.12698412698412695, 0.12934362934362933, 0.12781954887218044, 0.13003663003663005, 0.13035714285714287, 0.1324041811846691, 0.13775510204081629, 0.14119601328903655, 0.1444805194805195, 0.1492063492063492, 0.15062111801242237, 0.15349544072948326, 0.15922619047619044, 0.1661807580174927, 0.17142857142857143]}, "softmax_learned": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0047619047619047615, 0.004464285714285714, 0.004201680672268907, 0.003968253968253967, 0.0037593984962406013, 0.0035714285714285713, 0.006802721088435373, 0.00974025974025974, 0.018633540372670808, 0.02083333333333333, 0.022857142857142857, 0.03571428571428571, 0.047619047619047616, 0.061224489795918366, 0.07389162561576373, 0.08095238095238096, 0.08294930875576037, 0.08482142857142858, 0.09307359307359307, 0.09663865546218488, 0.1020408163265306, 0.11111111111111124, 0.11969111969111969, 0.11842105263157894, 0.1227106227106227, 0.12321428571428572, 0.12543554006968638, 0.13095238095238093, 0.1378737541528239, 0.14123376623376624, 0.15079365079365079, 0.15527950310559005, 0.1565349544072948, 0.1607142857142857, 0.16472303206997085, 0.17142857142857143]}, "geom_softmax": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.004464285714285714, 0.008403361344537815, 0.007936507936507934, 0.007518796992481203, 0.007142857142857143, 0.006802721088435373, 0.006493506493506494, 0.006211180124223602, 0.008928571428571426, 0.02, 0.03021978021978022, 0.042328042328042326, 0.04846938775510204, 0.0566502463054189, 0.06904761904761905, 0.07603686635944701, 0.08258928571428571, 0.08874458874458875, 0.09243697478991597, 0.09795918367346952, 0.10912698412698411, 0.10810810810810811, 0.11466165413533834, 0.11721611721611722, 0.12321428571428572, 0.12543554006968652, 0.13265306122448992, 0.1345514950166113, 0.13636363636363635, 0.1365079365079365, 0.14130434782608695, 0.14893617021276606, 0.1562500000000001, 0.16034985422740525, 0.17142857142857143]}}, "ltt": {"alpha": 0.1, "delta": 0.05, "lambda_hat": 0.11405049303851267, "coverage": 0.5271428571428571, "abstain_rate": 0.47285714285714286, "selective_risk": 0.02981029810298103, "selective_accuracy": 0.9701897018970189, "risk_within_budget": true}, "coverage_validity": {"alpha_grid": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "target": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "achieved_risk": [0.0, 0.0, 0.02981029810298103, 0.07623318385650224, 0.14660493827160495, 0.17142857142857143], "coverage": [0.0, 0.0, 0.5271428571428571, 0.6371428571428571, 0.9257142857142857, 1.0]}, "feature_diagnostics": {"names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual", "spectrum/lmin_mean", "spectrum/lmin_best", "spectrum/negfrac_mean", "spectrum/negfrac_best", "spectrum/effrank_mean", "spectrum/effrank_best", "spectrum/logdet_mean", "spectrum/logdet_best", "connect/barrier_mean", "connect/barrier_max", "connect/connected_frac"], "auroc": {"basin/rho": 0.6801724137931034, "basin/entropy": 0.3199640804597701, "basin/dispersion": 0.5014798850574713, "energy/mean": 0.5422413793103448, "energy/min": 0.541882183908046, "energy/std": 0.475617816091954, "curv/lmax_mean": 0.5129310344827587, "curv/lmax_best": 0.500926724137931, "curv/trace_mean": 0.5, "curv/trace_best": 0.5, "dynamics/steps": 0.5, "dynamics/monotonic": 0.5025215517241379, "dynamics/drop": 0.7858045977011494, "dynamics/residual": 0.5214655172413794, "spectrum/lmin_mean": 0.5, "spectrum/lmin_best": 0.5, "spectrum/negfrac_mean": 0.5, "spectrum/negfrac_best": 0.5, "spectrum/effrank_mean": 0.5, "spectrum/effrank_best": 0.5, "spectrum/logdet_mean": 0.5, "spectrum/logdet_best": 0.5, "connect/barrier_mean": 0.48670258620689655, "connect/barrier_max": 0.4867744252873563, "connect/connected_frac": 0.5132686781609196}, "hist": {"basin/rho": {"edges": [0.5, 0.525, 0.55, 0.575, 0.6, 0.625, 0.65, 0.675, 0.7, 0.725, 0.75, 0.775, 0.8, 0.825, 0.8500000000000001, 0.875, 0.9, 0.925, 0.95, 0.9750000000000001, 1.0], "correct_counts": [2, 0, 0, 1, 0, 0, 8, 0, 0, 0, 7, 0, 0, 37, 0, 0, 80, 0, 0, 445], "incorrect_counts": [1, 0, 0, 4, 0, 0, 4, 0, 0, 0, 6, 0, 0, 21, 0, 0, 33, 0, 0, 51]}, "basin/entropy": {"edges": [0.0, 0.041197960825054114, 0.08239592165010823, 0.12359388247516234, 0.16479184330021646, 0.20598980412527057, 0.24718776495032468, 0.2883857257753788, 0.3295836866004329, 0.370781647425487, 0.41197960825054114, 0.45317756907559525, 0.49437552990064937, 0.5355734907257035, 0.5767714515507576, 0.6179694123758117, 0.6591673732008658, 0.7003653340259199, 0.741563294850974, 0.7827612556760282, 0.8239592165010823], "correct_counts": [445, 0, 0, 0, 0, 0, 80, 0, 0, 0, 36, 0, 0, 8, 0, 8, 3, 0, 0, 0], "incorrect_counts": [51, 0, 0, 0, 0, 0, 33, 0, 0, 0, 21, 0, 0, 6, 0, 3, 5, 0, 0, 1]}, "basin/dispersion": {"edges": [1.9691264667768913, 2.1556350537882016, 2.342143640799512, 2.5286522278108228, 2.7151608148221333, 2.901669401833444, 3.0881779888447545, 3.274686575856065, 3.4611951628673756, 3.647703749878686, 3.8342123368899967, 4.020720923901307, 4.207229510912617, 4.3937380979239276, 4.580246684935238, 4.766755271946549, 4.953263858957859, 5.13977244596917, 5.32628103298048, 5.512789619991791, 5.6992982070031015], "correct_counts": [154, 409, 8, 0, 1, 0, 2, 1, 0, 0, 0, 2, 2, 0, 0, 0, 0, 0, 0, 1], "incorrect_counts": [42, 67, 4, 0, 1, 0, 0, 2, 0, 0, 1, 0, 2, 0, 0, 0, 1, 0, 0, 0]}, "energy/mean": {"edges": [0.05226107438405355, 0.1975594937801361, 0.3428579131762186, 0.4881563325723011, 0.6334547519683837, 0.7787531713644663, 0.9240515907605488, 1.0693500101566313, 1.2146484295527138, 1.3599468489487962, 1.505245268344879, 1.6505436877409614, 1.7958421071370438, 1.9411405265331265, 2.0864389459292094, 2.2317373653252917, 2.3770357847213743, 2.522334204117457, 2.6676326235135392, 2.812931042909622, 2.9582294623057046], "correct_counts": [230, 333, 7, 2, 1, 1, 2, 1, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1], "incorrect_counts": [60, 40, 3, 3, 4, 0, 0, 4, 1, 1, 0, 1, 0, 0, 1, 0, 0, 0, 1, 1]}, "energy/min": {"edges": [-0.02320098876953125, 0.11365594863891601, 0.25051288604736327, 0.3873698234558105, 0.5242267608642578, 0.6610836982727051, 0.7979406356811523, 0.9347975730895995, 1.0716545104980468, 1.208511447906494, 1.3453683853149414, 1.4822253227233886, 1.6190822601318358, 1.7559391975402832, 1.8927961349487303, 2.0296530723571777, 2.166510009765625, 2.303366947174072, 2.4402238845825193, 2.577080821990967, 2.713937759399414], "correct_counts": [179, 378, 17, 0, 2, 1, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0], "incorrect_counts": [46, 55, 2, 4, 3, 1, 0, 1, 3, 1, 0, 1, 0, 0, 1, 1, 0, 0, 0, 1]}, "energy/std": {"edges": [0.01770560819923182, 0.11276130864966308, 0.20781700910009432, 0.3028727095505256, 0.39792841000095686, 0.4929841104513881, 0.5880398109018193, 0.6830955113522507, 0.7781512118026819, 0.8732069122531131, 0.9682626127035444, 1.0633183131539756, 1.1583740136044067, 1.253429714054838, 1.3484854145052694, 1.4435411149557005, 1.5385968154061318, 1.6336525158565631, 1.7287082163069942, 1.8237639167574256, 1.9188196172078569], "correct_counts": [571, 0, 2, 0, 1, 1, 1, 0, 0, 1, 0, 0, 1, 0, 1, 0, 0, 0, 0, 1], "incorrect_counts": [114, 0, 1, 1, 0, 2, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "curv/lmax_mean": {"edges": [0.1999999905625979, 0.19999999174227318, 0.19999999292194842, 0.1999999941016237, 0.19999999528129894, 0.19999999646097422, 0.19999999764064946, 0.19999999882032474, 0.19999999999999998, 0.20000000117967526, 0.2000000023593505, 0.20000000353902578, 0.20000000471870105, 0.2000000058983763, 0.20000000707805157, 0.20000000825772682, 0.2000000094374021, 0.20000001061707734, 0.2000000117967526, 0.20000001297642786, 0.20000001415610313], "correct_counts": [1, 3, 7, 16, 20, 36, 45, 67, 70, 70, 62, 61, 40, 31, 28, 15, 4, 3, 0, 1], "incorrect_counts": [0, 2, 1, 1, 5, 8, 14, 6, 20, 17, 13, 11, 5, 8, 3, 2, 3, 0, 1, 0]}, "curv/lmax_best": {"edges": [0.19999995827674866, 0.19999996274709703, 0.19999996721744537, 0.19999997168779374, 0.19999997615814208, 0.19999998062849045, 0.19999998509883882, 0.19999998956918716, 0.19999999403953553, 0.19999999850988387, 0.20000000298023224, 0.2000000074505806, 0.20000001192092895, 0.20000001639127732, 0.20000002086162566, 0.20000002533197403, 0.2000000298023224, 0.20000003427267074, 0.2000000387430191, 0.20000004321336745, 0.20000004768371582], "correct_counts": [1, 0, 0, 15, 0, 0, 216, 0, 0, 0, 160, 0, 0, 179, 0, 0, 7, 0, 0, 2], "incorrect_counts": [0, 0, 0, 2, 0, 0, 46, 0, 0, 0, 35, 0, 0, 34, 0, 0, 3, 0, 0, 0]}, "curv/trace_mean": {"edges": [9.600000381469727, 9.650000381469727, 9.700000381469726, 9.750000381469727, 9.800000381469726, 9.850000381469727, 9.900000381469727, 9.950000381469726, 10.000000381469727, 10.050000381469726, 10.100000381469727, 10.150000381469727, 10.200000381469726, 10.250000381469727, 10.300000381469726, 10.350000381469727, 10.400000381469727, 10.450000381469726, 10.500000381469727, 10.550000381469726, 10.600000381469727], "correct_counts": [580, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [120, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "curv/trace_best": {"edges": [9.600000381469727, 9.650000381469727, 9.700000381469726, 9.750000381469727, 9.800000381469726, 9.850000381469727, 9.900000381469727, 9.950000381469726, 10.000000381469727, 10.050000381469726, 10.100000381469727, 10.150000381469727, 10.200000381469726, 10.250000381469727, 10.300000381469726, 10.350000381469727, 10.400000381469727, 10.450000381469726, 10.500000381469727, 10.550000381469726, 10.600000381469727], "correct_counts": [580, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [120, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "dynamics/steps": {"edges": [1.0, 1.05, 1.1, 1.15, 1.2, 1.25, 1.3, 1.35, 1.4, 1.45, 1.5, 1.55, 1.6, 1.65, 1.7000000000000002, 1.75, 1.8, 1.85, 1.9, 1.9500000000000002, 2.0], "correct_counts": [580, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [120, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "dynamics/monotonic": {"edges": [0.5375, 0.5408124999999999, 0.544125, 0.5474375, 0.55075, 0.5540624999999999, 0.557375, 0.5606875, 0.564, 0.5673124999999999, 0.5706249999999999, 0.5739375, 0.5772499999999999, 0.5805624999999999, 0.5838749999999999, 0.5871875, 0.5904999999999999, 0.5938124999999999, 0.5971249999999999, 0.6004375, 0.6037499999999999], "correct_counts": [2, 3, 15, 24, 59, 66, 78, 108, 91, 57, 44, 20, 9, 1, 2, 1, 0, 0, 0, 0], "incorrect_counts": [1, 4, 6, 0, 7, 18, 17, 26, 12, 10, 7, 1, 3, 2, 0, 1, 1, 1, 0, 3]}, "dynamics/drop": {"edges": [5.1263887484868365, 5.376814742883046, 5.627240737279256, 5.877666731675466, 6.128092726071675, 6.378518720467885, 6.628944714864096, 6.879370709260305, 7.129796703656515, 7.380222698052725, 7.6306486924489345, 7.881074686845144, 8.131500681241354, 8.381926675637564, 8.632352670033773, 8.882778664429983, 9.133204658826193, 9.383630653222403, 9.634056647618612, 9.884482642014822, 10.134908636411032], "correct_counts": [1, 0, 0, 2, 8, 27, 45, 52, 44, 40, 20, 13, 15, 53, 64, 68, 64, 31, 23, 10], "incorrect_counts": [0, 0, 0, 3, 2, 18, 17, 30, 18, 12, 10, 7, 1, 0, 0, 2, 0, 0, 0, 0]}, "dynamics/residual": {"edges": [0.2833298866947492, 0.2860603225727876, 0.28879075845082597, 0.2915211943288644, 0.2942516302069028, 0.2969820660849412, 0.2997125019629796, 0.302442937841018, 0.30517337371905645, 0.30790380959709485, 0.31063424547513324, 0.31336468135317164, 0.3160951172312101, 0.3188255531092485, 0.3215559889872869, 0.3242864248653253, 0.32701686074336367, 0.3297472966214021, 0.3324777324994405, 0.3352081683774789, 0.3379386042555173], "correct_counts": [2, 2, 3, 11, 16, 30, 39, 49, 59, 71, 63, 60, 54, 42, 34, 24, 10, 6, 3, 2], "incorrect_counts": [0, 1, 1, 2, 2, 6, 8, 16, 14, 15, 12, 7, 12, 10, 3, 4, 2, 2, 2, 1]}, "spectrum/lmin_mean": {"edges": [0.20000000298023224, 0.2500000029802322, 0.3000000029802322, 0.35000000298023226, 0.40000000298023225, 0.45000000298023224, 0.5000000029802323, 0.5500000029802323, 0.6000000029802323, 0.6500000029802322, 0.7000000029802322, 0.7500000029802323, 0.8000000029802323, 0.8500000029802323, 0.9000000029802323, 0.9500000029802322, 1.0000000029802323, 1.0500000029802323, 1.1000000029802321, 1.1500000029802324, 1.2000000029802322], "correct_counts": [580, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [120, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "spectrum/lmin_best": {"edges": [0.20000000298023224, 0.2500000029802322, 0.3000000029802322, 0.35000000298023226, 0.40000000298023225, 0.45000000298023224, 0.5000000029802323, 0.5500000029802323, 0.6000000029802323, 0.6500000029802322, 0.7000000029802322, 0.7500000029802323, 0.8000000029802323, 0.8500000029802323, 0.9000000029802323, 0.9500000029802322, 1.0000000029802323, 1.0500000029802323, 1.1000000029802321, 1.1500000029802324, 1.2000000029802322], "correct_counts": [580, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [120, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "spectrum/negfrac_mean": {"edges": [0.0, 0.05, 0.1, 0.15000000000000002, 0.2, 0.25, 0.30000000000000004, 0.35000000000000003, 0.4, 0.45, 0.5, 0.55, 0.6000000000000001, 0.65, 0.7000000000000001, 0.75, 0.8, 0.8500000000000001, 0.9, 0.9500000000000001, 1.0], "correct_counts": [580, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [120, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "spectrum/negfrac_best": {"edges": [0.0, 0.05, 0.1, 0.15000000000000002, 0.2, 0.25, 0.30000000000000004, 0.35000000000000003, 0.4, 0.45, 0.5, 0.55, 0.6000000000000001, 0.65, 0.7000000000000001, 0.75, 0.8, 0.8500000000000001, 0.9, 0.9500000000000001, 1.0], "correct_counts": [580, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [120, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "spectrum/effrank_mean": {"edges": [47.99999975000001, 48.049999750000005, 48.09999975000001, 48.149999750000006, 48.19999975000001, 48.24999975000001, 48.299999750000005, 48.34999975000001, 48.399999750000006, 48.44999975000001, 48.49999975000001, 48.549999750000005, 48.59999975000001, 48.649999750000006, 48.69999975000001, 48.74999975000001, 48.799999750000005, 48.84999975000001, 48.899999750000006, 48.94999975000001, 48.99999975000001], "correct_counts": [580, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [120, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "spectrum/effrank_best": {"edges": [47.99999975000001, 48.049999750000005, 48.09999975000001, 48.149999750000006, 48.19999975000001, 48.24999975000001, 48.299999750000005, 48.34999975000001, 48.399999750000006, 48.44999975000001, 48.49999975000001, 48.549999750000005, 48.59999975000001, 48.649999750000006, 48.69999975000001, 48.74999975000001, 48.799999750000005, 48.84999975000001, 48.899999750000006, 48.94999975000001, 48.99999975000001], "correct_counts": [580, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [120, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "spectrum/logdet_mean": {"edges": [-77.25301668158117, -77.20301668158118, -77.15301668158118, -77.10301668158117, -77.05301668158117, -77.00301668158117, -76.95301668158118, -76.90301668158118, -76.85301668158117, -76.80301668158117, -76.75301668158117, -76.70301668158118, -76.65301668158118, -76.60301668158117, -76.55301668158117, -76.50301668158117, -76.45301668158118, -76.40301668158118, -76.35301668158117, -76.30301668158117, -76.25301668158117], "correct_counts": [580, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [120, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "spectrum/logdet_best": {"edges": [-77.25301668158119, -77.20301668158119, -77.15301668158119, -77.10301668158118, -77.05301668158118, -77.00301668158119, -76.95301668158119, -76.90301668158119, -76.85301668158118, -76.80301668158118, -76.75301668158119, -76.70301668158119, -76.65301668158119, -76.60301668158118, -76.55301668158118, -76.50301668158119, -76.45301668158119, -76.40301668158119, -76.35301668158118, -76.30301668158118, -76.25301668158119], "correct_counts": [580, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [120, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "connect/barrier_mean": {"edges": [0.0, 0.05873445651747964, 0.11746891303495928, 0.17620336955243893, 0.23493782606991856, 0.2936722825873982, 0.35240673910487785, 0.41114119562235746, 0.4698756521398371, 0.5286101086573167, 0.5873445651747964, 0.646079021692276, 0.7048134782097557, 0.7635479347272354, 0.8222823912447149, 0.8810168477621946, 0.9397513042796742, 0.9984857607971539, 1.0572202173146334, 1.115954673832113, 1.1746891303495928], "correct_counts": [577, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [115, 2, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1]}, "connect/barrier_max": {"edges": [0.0, 0.06394378542900085, 0.1278875708580017, 0.19183135628700254, 0.2557751417160034, 0.3197189271450043, 0.3836627125740051, 0.44760649800300595, 0.5115502834320068, 0.5754940688610076, 0.6394378542900085, 0.7033816397190094, 0.7673254251480102, 0.8312692105770111, 0.8952129960060119, 0.9591567814350128, 1.0231005668640136, 1.0870443522930144, 1.1509881377220152, 1.2149319231510163, 1.278875708580017], "correct_counts": [572, 0, 4, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0], "incorrect_counts": [115, 0, 0, 0, 0, 0, 0, 2, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1]}, "connect/connected_frac": {"edges": [0.0, 0.05, 0.1, 0.15000000000000002, 0.2, 0.25, 0.30000000000000004, 0.35000000000000003, 0.4, 0.45, 0.5, 0.55, 0.6000000000000001, 0.65, 0.7000000000000001, 0.75, 0.8, 0.8500000000000001, 0.9, 0.9500000000000001, 1.0], "correct_counts": [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 3, 0, 4, 571], "incorrect_counts": [1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 2, 0, 0, 115]}}}, "feature_ablation": {"full": 0.0533799136280693, "drop_basin": 0.05512073495619045, "basin_only": 0.12163708444636095, "drop_energy": 0.0540551948406156, "energy_only": 0.21717465689606066, "drop_curv": 0.05368458197556656, "curv_only": 0.1688173605877835, "drop_dynamics": 0.12406978604857082, "dynamics_only": 0.05622078593396844, "drop_spectrum": 0.0533799136280693, "spectrum_only": 0.17142857142857193, "drop_connect": 0.05324936333102736, "connect_only": 0.1676576914051575}, "ece_geometry": 0.042492555746301765}, "env": {"python": "3.12.13", "platform": "macOS-26.5.2-arm64-arm-64bit", "jax": "0.11.0", "jax_backend": "cpu"}, "artifact_paths": []} {"run_id": "a405cf7b867a", "timestamp": "2026-07-28T23:19:30.542724+00:00", "git_sha": "9367d60", "config_hash": "23346bbdcf92", "config": {"run": {"seed": 4, "task": "graph_planning", "notes": "stronger IRED learned-landscape reasoner on graph"}, "model": {"latent_dim": 48, "hidden_dim": 256, "context_dim": 64}, "inference": {"k_restarts": 12, "steps": 50, "step_size": 0.1, "temperature": 0.01, "init_scale": 1.0, "grad_tol": 0.001, "sampler": "annealed", "anneal_levels": 20, "anneal_steps_per_level": 10, "anneal_step_max": 0.5, "anneal_step_min": 0.003}, "train": {"epochs": 110, "batch_size": 128, "lr": 0.001, "n_train": 12000, "n_neg": 4, "neg_noise": 0.5, "objective": "ired", "ired_noise_min": 0.1, "ired_noise_max": 1.5, "ired_ridge": 0.2, "ired_decode_weight": 4.0, "ired_stat_weight": 8.0}, "eval": {"n_eval": 700, "richer_geometry": true}, "conformal": {"alpha": 0.1, "delta": 0.05, "n_calib": 800}, "task": {"graph_planning": {"n_nodes": 7, "ood_n_nodes": 10, "edge_prob": 0.4, "max_len": 4}}}, "task": "graph_planning", "split": "selective", "seed": 4, "metrics": {"n_fit": 700, "n_calib": 800, "n_test": 700, "k_restarts": 12, "objective": "ired", "sampler": "annealed", "feature_set": "richer", "accuracy_id": 0.8285714285714286, "base_error": 0.17142857142857143, "final_train_loss": 0.1806420087814331, "feature_names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual", "spectrum/lmin_mean", "spectrum/lmin_best", "spectrum/negfrac_mean", "spectrum/negfrac_best", "spectrum/effrank_mean", "spectrum/effrank_best", "spectrum/logdet_mean", "spectrum/logdet_best", "connect/barrier_mean", "connect/barrier_max", "connect/connected_frac"], "aurc": {"geometry": 0.0499999909178207, "rho_basin": 0.10276071740829948, "energy_min": 0.14142815403252237, "energy_mean": 0.13715414350531202, "energy_std": 0.16317105464147816, "msp": 0.09655790013556867, "temp_msp": 0.053993209301107216, "entropy": 0.09625911916651533, "softmax_learned": 0.053985762260104715, "geom_softmax": 0.04996327543640174}, "temperature": 7.166651351401238, "best_energy_baseline": "energy_mean", "best_baseline": "temp_msp", "delta_aurc_vs_energy_min": [0.09142816311470167, 0.05939135802345537, 0.12372412125974012], "delta_aurc_vs_best_energy": [0.08715415258749132, 0.05679752406767961, 0.1182631632770357], "delta_aurc_vs_best_baseline": [0.003993218383286513, -0.0029830289649739033, 0.010870978934784942], "delta_aurc_geom_adds": [0.004022486823702973, -0.00434027542554494, 0.012002982463869206], "geometry_wins": true, "geometry_wins_vs_baseline": false, "geometry_adds_over_softmax": false, "risk_coverage": {"geometry": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.004201680672268907, 0.003968253968253967, 0.0037593984962406013, 0.0035714285714285713, 0.0034013605442176865, 0.006493506493506494, 0.009316770186335404, 0.008928571428571426, 0.02, 0.02197802197802198, 0.0291005291005291, 0.04081632653061224, 0.046798029556650425, 0.05476190476190476, 0.06451612903225806, 0.06696428571428571, 0.07575757575757576, 0.07983193277310924, 0.08367346938775509, 0.0853174603174603, 0.0945945945945946, 0.09962406015037593, 0.1043956043956044, 0.11428571428571428, 0.1219512195121951, 0.12585034013605453, 0.13122923588039867, 0.1396103896103896, 0.14603174603174604, 0.14906832298136646, 0.15501519756838916, 0.15624999999999997, 0.16472303206997085, 0.17142857142857143]}, "rho_basin": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.09110169491525424, 0.09110169491525429, 0.0911016949152543, 0.0911016949152543, 0.0911016949152543, 0.0911016949152543, 0.09110169491525431, 0.09110169491525431, 0.09110169491525431, 0.09110169491525431, 0.09110169491525431, 0.09110169491525431, 0.09110169491525424, 0.09110169491525413, 0.09110169491525402, 0.09110169491525393, 0.09110169491525386, 0.09110169491525377, 0.09110169491525372, 0.09110169491525366, 0.0911016949152536, 0.09110169491525355, 0.09110169491525351, 0.09110169491525347, 0.09110169491525344, 0.09110169491525341, 0.09110169491525338, 0.09110169491525334, 0.09110169491525331, 0.0911016949152533, 0.09110169491525327, 0.09110169491525325, 0.09110169491525323, 0.09250700280111943, 0.0972448979591827, 0.10171957671957575, 0.10595238095237998, 0.10996240601503664, 0.11376678876678782, 0.11738095238095152, 0.12081881533100979, 0.12409297052154146, 0.12922442433268372, 0.13490819525302244, 0.1403393541324572, 0.14553437566930785, 0.15012554513017012, 0.15424430641821898, 0.16107871720116568, 0.17142857142857088]}, "energy_min": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.35714285714285715, 0.21428571428571427, 0.16666666666666666, 0.17857142857142858, 0.14285714285714285, 0.13095238095238093, 0.12244897959183655, 0.125, 0.1111111111111111, 0.1214285714285713, 0.12987012987012986, 0.125, 0.12087912087912088, 0.11734693877551021, 0.11428571428571425, 0.11160714285714286, 0.12605042016806722, 0.12698412698412695, 0.12406015037593984, 0.12142857142857143, 0.11904761904761915, 0.12012987012987013, 0.12422360248447205, 0.12202380952380962, 0.12285714285714286, 0.12912087912087913, 0.12962962962962962, 0.125, 0.12807881773399013, 0.1261904761904762, 0.12211981566820276, 0.13169642857142858, 0.13203463203463203, 0.13655462184873948, 0.14081632653061238, 0.14285714285714282, 0.14092664092664092, 0.15225563909774437, 0.15567765567765568, 0.15178571428571427, 0.15679442508710797, 0.15476190476190474, 0.15282392026578073, 0.1525974025974026, 0.1523809523809524, 0.15062111801242237, 0.1519756838905776, 0.1577380952380952, 0.16326530612244897, 0.17142857142857143]}, "energy_mean": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.21428571428571427, 0.14285714285714285, 0.14285714285714285, 0.14285714285714285, 0.14285714285714285, 0.14285714285714282, 0.14285714285714288, 0.13392857142857142, 0.1349206349206349, 0.1285714285714286, 0.12337662337662338, 0.11904761904761904, 0.11538461538461539, 0.11734693877551021, 0.11428571428571425, 0.10714285714285714, 0.1092436974789916, 0.11904761904761903, 0.12406015037593984, 0.12142857142857143, 0.1258503401360544, 0.1266233766233766, 0.13043478260869565, 0.12797619047619044, 0.12285714285714286, 0.12362637362637363, 0.12433862433862433, 0.11989795918367346, 0.12561576354679801, 0.1261904761904762, 0.12903225806451613, 0.12946428571428573, 0.1277056277056277, 0.13025210084033614, 0.13265306122448978, 0.13690476190476206, 0.14285714285714285, 0.14285714285714285, 0.14285714285714285, 0.14642857142857144, 0.14634146341463425, 0.1462585034013605, 0.14950166112956811, 0.15097402597402598, 0.15396825396825398, 0.15217391304347827, 0.15501519756838902, 0.1577380952380952, 0.16326530612244897, 0.17142857142857143]}, "energy_std": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.2857142857142857, 0.21428571428571427, 0.23809523809523808, 0.21428571428571427, 0.18571428571428572, 0.17857142857142855, 0.15306122448979595, 0.16071428571428573, 0.16666666666666666, 0.15714285714285717, 0.16233766233766234, 0.17261904761904762, 0.17582417582417584, 0.17346938775510204, 0.16666666666666663, 0.16071428571428573, 0.15966386554621848, 0.15079365079365076, 0.15413533834586465, 0.15, 0.1462585034013605, 0.14285714285714285, 0.14906832298136646, 0.15476190476190474, 0.15142857142857144, 0.1510989010989011, 0.15079365079365079, 0.14795918367346939, 0.14532019704433494, 0.14761904761904762, 0.15668202764976957, 0.15848214285714285, 0.16017316017316016, 0.15966386554621848, 0.15918367346938772, 0.1607142857142857, 0.16216216216216217, 0.16917293233082706, 0.17032967032967034, 0.16607142857142856, 0.16376306620209055, 0.16156462585034012, 0.16279069767441862, 0.16233766233766234, 0.16031746031746033, 0.16770186335403728, 0.16413373860182368, 0.16369047619047616, 0.1661807580174927, 0.17142857142857143]}, "msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.07142857142857142, 0.07142857142857142, 0.07142857142857142, 0.05357142857142857, 0.04285714285714286, 0.04761904761904773, 0.07142857142857144, 0.07142857142857142, 0.07142857142857142, 0.07142857142857144, 0.06493506493506493, 0.07142857142857142, 0.08241758241758242, 0.08163265306122448, 0.07619047619047636, 0.08482142857142858, 0.08823529411764706, 0.08333333333333331, 0.08270676691729323, 0.08214285714285714, 0.08163265306122447, 0.07792207792207792, 0.08074534161490683, 0.08333333333333331, 0.08285714285714285, 0.08791208791208792, 0.09788359788359788, 0.09438775510204081, 0.09852216748768472, 0.09523809523809523, 0.0944700460829493, 0.09598214285714286, 0.09956709956709957, 0.10294117647058823, 0.11020408163265305, 0.11507936507936506, 0.11776061776061776, 0.11654135338345864, 0.11904761904761904, 0.11964285714285715, 0.12543554006968652, 0.12925170068027222, 0.1378737541528239, 0.13636363636363635, 0.13968253968253969, 0.14285714285714285, 0.1458966565349545, 0.1532738095238095, 0.16326530612244897, 0.17142857142857143]}, "temp_msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0037593984962406013, 0.0035714285714285713, 0.006802721088435373, 0.006493506493506494, 0.006211180124223602, 0.008928571428571426, 0.008571428571428572, 0.016483516483516484, 0.023809523809523808, 0.04846938775510204, 0.0566502463054189, 0.06904761904761905, 0.07834101382488479, 0.08705357142857142, 0.09740259740259741, 0.09873949579831932, 0.10816326530612244, 0.10912698412698411, 0.11583011583011583, 0.12030075187969924, 0.12087912087912088, 0.12142857142857143, 0.12717770034843204, 0.13435374149659862, 0.13953488372093023, 0.14123376623376624, 0.13968253968253969, 0.14596273291925466, 0.14893617021276592, 0.15624999999999997, 0.1618075801749271, 0.17142857142857143]}, "entropy": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.07142857142857142, 0.07142857142857142, 0.07142857142857142, 0.05357142857142857, 0.04285714285714286, 0.04761904761904773, 0.07142857142857144, 0.07142857142857142, 0.07142857142857142, 0.07142857142857144, 0.06493506493506493, 0.07142857142857142, 0.08241758241758242, 0.08163265306122448, 0.07619047619047618, 0.08482142857142858, 0.08823529411764706, 0.08333333333333331, 0.08270676691729323, 0.08214285714285714, 0.08163265306122447, 0.07792207792207792, 0.08074534161490683, 0.08333333333333331, 0.08285714285714285, 0.08791208791208792, 0.09788359788359788, 0.09438775510204081, 0.09852216748768472, 0.09523809523809523, 0.0944700460829493, 0.09598214285714286, 0.09956709956709957, 0.10084033613445378, 0.1081632653061226, 0.11507936507936506, 0.11776061776061776, 0.11654135338345864, 0.11904761904761904, 0.11964285714285715, 0.12543554006968652, 0.12925170068027222, 0.1378737541528239, 0.137987012987013, 0.13968253968253969, 0.14285714285714285, 0.1458966565349545, 0.15327380952380965, 0.16326530612244897, 0.17142857142857143]}, "softmax_learned": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0037593984962406013, 0.0035714285714285713, 0.006802721088435373, 0.006493506493506494, 0.006211180124223602, 0.008928571428571426, 0.008571428571428572, 0.016483516483516484, 0.023809523809523808, 0.04846938775510204, 0.0566502463054189, 0.06904761904761905, 0.07834101382488479, 0.08705357142857142, 0.09740259740259741, 0.09873949579831932, 0.10816326530612244, 0.10912698412698411, 0.11583011583011583, 0.12030075187969924, 0.12087912087912088, 0.12142857142857143, 0.12717770034843204, 0.13435374149659862, 0.13953488372093023, 0.14123376623376624, 0.13968253968253969, 0.14596273291925466, 0.1474164133738603, 0.15624999999999997, 0.16326530612244897, 0.17142857142857143]}, "geom_softmax": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0047619047619047615, 0.008928571428571428, 0.008403361344537815, 0.007936507936507934, 0.007518796992481203, 0.010714285714285714, 0.010204081632653059, 0.012987012987012988, 0.015527950310559006, 0.014880952380952378, 0.02, 0.024725274725274724, 0.03439153439153439, 0.04336734693877551, 0.04679802955665024, 0.05238095238095238, 0.055299539170506916, 0.06696428571428571, 0.06926406926406926, 0.07352941176470588, 0.07959183673469386, 0.0853174603174603, 0.08687258687258688, 0.09398496240601503, 0.10073260073260074, 0.10892857142857143, 0.11846689895470382, 0.12074829931972787, 0.12458471760797342, 0.13474025974025974, 0.14126984126984127, 0.14751552795031056, 0.1534954407294834, 0.15922619047619044, 0.16472303206997085, 0.17142857142857143]}}, "ltt": {"alpha": 0.1, "delta": 0.05, "lambda_hat": 0.16840793323263656, "coverage": 0.5757142857142857, "abstain_rate": 0.42428571428571427, "selective_risk": 0.04714640198511166, "selective_accuracy": 0.9528535980148883, "risk_within_budget": true}, "coverage_validity": {"alpha_grid": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "target": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "achieved_risk": [0.0, 0.0, 0.04714640198511166, 0.0706401766004415, 0.14664586583463338, 0.17167381974248927], "coverage": [0.0, 0.0, 0.5757142857142857, 0.6471428571428571, 0.9157142857142857, 0.9985714285714286]}, "feature_diagnostics": {"names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual", "spectrum/lmin_mean", "spectrum/lmin_best", "spectrum/negfrac_mean", "spectrum/negfrac_best", "spectrum/effrank_mean", "spectrum/effrank_best", "spectrum/logdet_mean", "spectrum/logdet_best", "connect/barrier_mean", "connect/barrier_max", "connect/connected_frac"], "auroc": {"basin/rho": 0.7098060344827586, "basin/entropy": 0.2908979885057471, "basin/dispersion": 0.4298419540229885, "energy/mean": 0.3820402298850575, "energy/min": 0.39098419540229884, "energy/std": 0.4610488505747126, "curv/lmax_mean": 0.47841954022988503, "curv/lmax_best": 0.484683908045977, "curv/trace_mean": 0.5, "curv/trace_best": 0.5, "dynamics/steps": 0.5, "dynamics/monotonic": 0.5555316091954023, "dynamics/drop": 0.7991379310344827, "dynamics/residual": 0.4871551724137931, "spectrum/lmin_mean": 0.5, "spectrum/lmin_best": 0.5, "spectrum/negfrac_mean": 0.5, "spectrum/negfrac_best": 0.5, "spectrum/effrank_mean": 0.5, "spectrum/effrank_best": 0.5, "spectrum/logdet_mean": 0.5, "spectrum/logdet_best": 0.5, "connect/barrier_mean": 0.47932471264367815, "connect/barrier_max": 0.479382183908046, "connect/connected_frac": 0.5207183908045977}, "hist": {"basin/rho": {"edges": [0.5, 0.525, 0.55, 0.575, 0.6, 0.625, 0.65, 0.675, 0.7, 0.725, 0.75, 0.775, 0.8, 0.825, 0.8500000000000001, 0.875, 0.9, 0.925, 0.95, 0.9750000000000001, 1.0], "correct_counts": [0, 0, 0, 2, 0, 0, 9, 0, 0, 0, 15, 0, 0, 36, 0, 0, 89, 0, 0, 429], "incorrect_counts": [2, 0, 0, 5, 0, 0, 9, 0, 0, 0, 8, 0, 0, 22, 0, 0, 31, 0, 0, 43]}, "basin/entropy": {"edges": [0.0, 0.03607318433462558, 0.07214636866925116, 0.10821955300387673, 0.1442927373385023, 0.1803659216731279, 0.21643910600775346, 0.25251229034237904, 0.2885854746770046, 0.3246586590116302, 0.3607318433462558, 0.3968050276808814, 0.4328782120155069, 0.4689513963501325, 0.5050245806847581, 0.5410977650193837, 0.5771709493540093, 0.6132441336886348, 0.6493173180232604, 0.685390502357886, 0.7214636866925116], "correct_counts": [429, 0, 0, 0, 0, 0, 0, 89, 0, 0, 0, 0, 35, 0, 0, 13, 0, 9, 2, 3], "incorrect_counts": [43, 0, 0, 0, 0, 0, 0, 31, 0, 0, 0, 0, 21, 0, 0, 9, 0, 9, 5, 2]}, "basin/dispersion": {"edges": [2.006589668932686, 2.1440741312539653, 2.281558593575245, 2.419043055896524, 2.5565275182178033, 2.6940119805390825, 2.831496442860362, 2.9689809051816414, 3.1064653675029206, 3.2439498298242, 3.3814342921454794, 3.5189187544667586, 3.6564032167880383, 3.7938876791093175, 3.9313721414305967, 4.068856603751876, 4.206341066073156, 4.343825528394435, 4.481309990715714, 4.618794453036994, 4.756278915358273], "correct_counts": [141, 392, 38, 0, 1, 0, 0, 2, 2, 2, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0], "incorrect_counts": [23, 68, 21, 0, 0, 0, 1, 0, 3, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 2]}, "energy/mean": {"edges": [-1.6460671226183574, -1.4175087143977483, -1.1889503061771394, -0.9603918979565302, -0.7318334897359212, -0.5032750815153122, -0.27471667329470306, -0.04615826507409415, 0.18240014314651498, 0.4109585513671241, 0.639516959587733, 0.8680753678083419, 1.0966337760289513, 1.3251921842495602, 1.5537505924701691, 1.7823090006907785, 2.0108674089113876, 2.2394258171319965, 2.4679842253526054, 2.6965426335732143, 2.9251010417938232], "correct_counts": [1, 0, 0, 0, 0, 0, 0, 439, 127, 5, 0, 3, 0, 1, 1, 0, 2, 1, 0, 0], "incorrect_counts": [0, 0, 0, 0, 0, 0, 1, 72, 33, 2, 1, 3, 1, 1, 0, 1, 1, 2, 0, 2]}, "energy/min": {"edges": [-2.360767364501953, -2.1069081664085387, -1.8530489683151246, -1.5991897702217104, -1.345330572128296, -1.0914713740348816, -0.8376121759414674, -0.5837529778480532, -0.32989377975463885, -0.07603458166122445, 0.17782461643218994, 0.4316838145256039, 0.6855430126190183, 0.9394022107124327, 1.1932614088058466, 1.447120606899261, 1.7009798049926754, 1.9548390030860894, 2.208698201179504, 2.462557399272918, 2.716416597366333], "correct_counts": [1, 0, 0, 0, 0, 0, 0, 0, 0, 553, 15, 4, 2, 1, 0, 2, 1, 1, 0, 0], "incorrect_counts": [0, 0, 0, 0, 0, 0, 0, 0, 1, 97, 9, 3, 2, 2, 1, 1, 1, 1, 0, 2]}, "energy/std": {"edges": [0.019892553899172424, 0.06891081471924504, 0.11792907553931765, 0.1669473363593903, 0.2159655971794629, 0.2649838579995355, 0.31400211881960816, 0.36302037963968076, 0.41203864045975336, 0.46105690127982596, 0.5100751620998986, 0.5590934229199712, 0.6081116837400439, 0.6571299445601164, 0.7061482053801891, 0.7551664662002616, 0.8041847270203343, 0.8532029878404069, 0.9022212486604795, 0.9512395094805521, 1.0002577703006248], "correct_counts": [560, 11, 1, 0, 2, 0, 2, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2], "incorrect_counts": [110, 4, 1, 2, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0]}, "curv/lmax_mean": {"edges": [0.1999999893208345, 0.19999999050050976, 0.199999991680185, 0.19999999285986025, 0.19999999403953553, 0.1999999952192108, 0.19999999639888605, 0.1999999975785613, 0.19999999875823657, 0.19999999993791184, 0.2000000011175871, 0.20000000229726234, 0.2000000034769376, 0.20000000465661288, 0.20000000583628813, 0.20000000701596338, 0.20000000819563865, 0.20000000937531393, 0.20000001055498917, 0.20000001173466442, 0.2000000129143397], "correct_counts": [1, 3, 1, 7, 13, 20, 28, 51, 65, 66, 76, 76, 55, 44, 30, 22, 14, 5, 1, 2], "incorrect_counts": [0, 0, 1, 1, 1, 6, 1, 9, 13, 17, 20, 13, 9, 14, 10, 4, 1, 0, 0, 0]}, "curv/lmax_best": {"edges": [0.19999995827674866, 0.19999996274709703, 0.19999996721744537, 0.19999997168779374, 0.19999997615814208, 0.19999998062849045, 0.19999998509883882, 0.19999998956918716, 0.19999999403953553, 0.19999999850988387, 0.20000000298023224, 0.2000000074505806, 0.20000001192092895, 0.20000001639127732, 0.20000002086162566, 0.20000002533197403, 0.2000000298023224, 0.20000003427267074, 0.2000000387430191, 0.20000004321336745, 0.20000004768371582], "correct_counts": [2, 0, 0, 8, 0, 0, 233, 0, 0, 0, 139, 0, 0, 185, 0, 0, 12, 0, 0, 1], "incorrect_counts": [0, 0, 0, 2, 0, 0, 44, 0, 0, 0, 31, 0, 0, 41, 0, 0, 2, 0, 0, 0]}, "curv/trace_mean": {"edges": [9.600000381469727, 9.650000381469727, 9.700000381469726, 9.750000381469727, 9.800000381469726, 9.850000381469727, 9.900000381469727, 9.950000381469726, 10.000000381469727, 10.050000381469726, 10.100000381469727, 10.150000381469727, 10.200000381469726, 10.250000381469727, 10.300000381469726, 10.350000381469727, 10.400000381469727, 10.450000381469726, 10.500000381469727, 10.550000381469726, 10.600000381469727], "correct_counts": [580, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [120, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "curv/trace_best": {"edges": [9.600000381469727, 9.650000381469727, 9.700000381469726, 9.750000381469727, 9.800000381469726, 9.850000381469727, 9.900000381469727, 9.950000381469726, 10.000000381469727, 10.050000381469726, 10.100000381469727, 10.150000381469727, 10.200000381469726, 10.250000381469727, 10.300000381469726, 10.350000381469727, 10.400000381469727, 10.450000381469726, 10.500000381469727, 10.550000381469726, 10.600000381469727], "correct_counts": [580, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [120, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "dynamics/steps": {"edges": [1.0, 1.05, 1.1, 1.15, 1.2, 1.25, 1.3, 1.35, 1.4, 1.45, 1.5, 1.55, 1.6, 1.65, 1.7000000000000002, 1.75, 1.8, 1.85, 1.9, 1.9500000000000002, 2.0], "correct_counts": [580, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [120, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "dynamics/monotonic": {"edges": [0.5308333333333334, 0.5355000000000001, 0.5401666666666667, 0.5448333333333334, 0.5495, 0.5541666666666667, 0.5588333333333334, 0.5635, 0.5681666666666667, 0.5728333333333333, 0.5775, 0.5821666666666667, 0.5868333333333333, 0.5915, 0.5961666666666666, 0.6008333333333333, 0.6055, 0.6101666666666666, 0.6148333333333333, 0.6194999999999999, 0.6241666666666666], "correct_counts": [0, 2, 5, 20, 61, 102, 134, 108, 76, 42, 22, 4, 1, 1, 0, 0, 1, 0, 0, 1], "incorrect_counts": [1, 0, 0, 7, 17, 24, 27, 22, 8, 10, 2, 2, 0, 0, 0, 0, 0, 0, 0, 0]}, "dynamics/drop": {"edges": [5.520331422487895, 5.784909941752752, 6.049488461017609, 6.314066980282466, 6.5786454995473225, 6.84322401881218, 7.107802538077037, 7.372381057341894, 7.636959576606751, 7.901538095871608, 8.166116615136465, 8.430695134401322, 8.69527365366618, 8.959852172931036, 9.224430692195892, 9.48900921146075, 9.753587730725606, 10.018166249990465, 10.28274476925532, 10.547323288520179, 10.811901807785034], "correct_counts": [0, 4, 2, 16, 36, 51, 49, 40, 18, 16, 15, 33, 58, 61, 73, 53, 29, 14, 7, 5], "incorrect_counts": [2, 1, 3, 8, 16, 29, 27, 11, 10, 8, 2, 2, 0, 1, 0, 0, 0, 0, 0, 0]}, "dynamics/residual": {"edges": [0.28685005381703377, 0.29140506281207007, 0.2959600718071063, 0.3005150808021426, 0.3050700897971789, 0.30962509879221517, 0.3141801077872515, 0.3187351167822878, 0.323290125777324, 0.3278451347723603, 0.33240014376739657, 0.3369551527624329, 0.3415101617574692, 0.3460651707525054, 0.3506201797475417, 0.35517518874257803, 0.3597301977376143, 0.3642852067326506, 0.3688402157276869, 0.37339522472272313, 0.37795023371775943], "correct_counts": [10, 24, 45, 83, 110, 108, 99, 60, 31, 2, 5, 0, 2, 0, 0, 0, 0, 0, 0, 1], "incorrect_counts": [0, 4, 13, 17, 22, 21, 17, 14, 8, 2, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0]}, "spectrum/lmin_mean": {"edges": [0.20000000298023224, 0.2500000029802322, 0.3000000029802322, 0.35000000298023226, 0.40000000298023225, 0.45000000298023224, 0.5000000029802323, 0.5500000029802323, 0.6000000029802323, 0.6500000029802322, 0.7000000029802322, 0.7500000029802323, 0.8000000029802323, 0.8500000029802323, 0.9000000029802323, 0.9500000029802322, 1.0000000029802323, 1.0500000029802323, 1.1000000029802321, 1.1500000029802324, 1.2000000029802322], "correct_counts": [580, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [120, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "spectrum/lmin_best": {"edges": [0.20000000298023224, 0.2500000029802322, 0.3000000029802322, 0.35000000298023226, 0.40000000298023225, 0.45000000298023224, 0.5000000029802323, 0.5500000029802323, 0.6000000029802323, 0.6500000029802322, 0.7000000029802322, 0.7500000029802323, 0.8000000029802323, 0.8500000029802323, 0.9000000029802323, 0.9500000029802322, 1.0000000029802323, 1.0500000029802323, 1.1000000029802321, 1.1500000029802324, 1.2000000029802322], "correct_counts": [580, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [120, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "spectrum/negfrac_mean": {"edges": [0.0, 0.05, 0.1, 0.15000000000000002, 0.2, 0.25, 0.30000000000000004, 0.35000000000000003, 0.4, 0.45, 0.5, 0.55, 0.6000000000000001, 0.65, 0.7000000000000001, 0.75, 0.8, 0.8500000000000001, 0.9, 0.9500000000000001, 1.0], "correct_counts": [580, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [120, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "spectrum/negfrac_best": {"edges": [0.0, 0.05, 0.1, 0.15000000000000002, 0.2, 0.25, 0.30000000000000004, 0.35000000000000003, 0.4, 0.45, 0.5, 0.55, 0.6000000000000001, 0.65, 0.7000000000000001, 0.75, 0.8, 0.8500000000000001, 0.9, 0.9500000000000001, 1.0], "correct_counts": [580, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [120, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "spectrum/effrank_mean": {"edges": [47.99999975000001, 48.049999750000005, 48.09999975000001, 48.149999750000006, 48.19999975000001, 48.24999975000001, 48.299999750000005, 48.34999975000001, 48.399999750000006, 48.44999975000001, 48.49999975000001, 48.549999750000005, 48.59999975000001, 48.649999750000006, 48.69999975000001, 48.74999975000001, 48.799999750000005, 48.84999975000001, 48.899999750000006, 48.94999975000001, 48.99999975000001], "correct_counts": [580, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [120, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "spectrum/effrank_best": {"edges": [47.99999975000001, 48.049999750000005, 48.09999975000001, 48.149999750000006, 48.19999975000001, 48.24999975000001, 48.299999750000005, 48.34999975000001, 48.399999750000006, 48.44999975000001, 48.49999975000001, 48.549999750000005, 48.59999975000001, 48.649999750000006, 48.69999975000001, 48.74999975000001, 48.799999750000005, 48.84999975000001, 48.899999750000006, 48.94999975000001, 48.99999975000001], "correct_counts": [580, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [120, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "spectrum/logdet_mean": {"edges": [-77.25301668158117, -77.20301668158118, -77.15301668158118, -77.10301668158117, -77.05301668158117, -77.00301668158117, -76.95301668158118, -76.90301668158118, -76.85301668158117, -76.80301668158117, -76.75301668158117, -76.70301668158118, -76.65301668158118, -76.60301668158117, -76.55301668158117, -76.50301668158117, -76.45301668158118, -76.40301668158118, -76.35301668158117, -76.30301668158117, -76.25301668158117], "correct_counts": [580, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [120, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "spectrum/logdet_best": {"edges": [-77.25301668158119, -77.20301668158119, -77.15301668158119, -77.10301668158118, -77.05301668158118, -77.00301668158119, -76.95301668158119, -76.90301668158119, -76.85301668158118, -76.80301668158118, -76.75301668158119, -76.70301668158119, -76.65301668158119, -76.60301668158118, -76.55301668158118, -76.50301668158119, -76.45301668158119, -76.40301668158119, -76.35301668158118, -76.30301668158118, -76.25301668158119], "correct_counts": [580, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [120, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "connect/barrier_mean": {"edges": [0.0, 0.06370989842848344, 0.1274197968569669, 0.19112969528545032, 0.2548395937139338, 0.31854949214241723, 0.38225939057090064, 0.4459692889993841, 0.5096791874278676, 0.573389085856351, 0.6370989842848345, 0.7008088827133179, 0.7645187811418013, 0.8282286795702848, 0.8919385779987682, 0.9556484764272517, 1.019358374855735, 1.0830682732842185, 1.146778171712702, 1.2104880701411855, 1.274197968569669], "correct_counts": [578, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1], "incorrect_counts": [118, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "connect/barrier_max": {"edges": [0.0, 0.06671607494354248, 0.13343214988708496, 0.20014822483062744, 0.2668642997741699, 0.3335803747177124, 0.4002964496612549, 0.46701252460479736, 0.5337285995483398, 0.6004446744918823, 0.6671607494354248, 0.7338768243789673, 0.8005928993225098, 0.8673089742660522, 0.9340250492095947, 1.0007411241531372, 1.0674571990966797, 1.1341732740402222, 1.2008893489837646, 1.2676054239273071, 1.3343214988708496], "correct_counts": [576, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1], "incorrect_counts": [115, 1, 0, 1, 2, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "connect/connected_frac": {"edges": [0.0, 0.05, 0.1, 0.15000000000000002, 0.2, 0.25, 0.30000000000000004, 0.35000000000000003, 0.4, 0.45, 0.5, 0.55, 0.6000000000000001, 0.65, 0.7000000000000001, 0.75, 0.8, 0.8500000000000001, 0.9, 0.9500000000000001, 1.0], "correct_counts": [1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 3, 575], "incorrect_counts": [0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 2, 0, 2, 114]}}}, "feature_ablation": {"full": 0.0499999909178207, "drop_basin": 0.05322165715841172, "basin_only": 0.10375895158755924, "drop_energy": 0.049739566895181775, "energy_only": 0.13920001795584305, "drop_curv": 0.05021424448265369, "curv_only": 0.17253571329725742, "drop_dynamics": 0.09321471121667474, "dynamics_only": 0.05282089227280329, "drop_spectrum": 0.0499999909178207, "spectrum_only": 0.17142857142857193, "drop_connect": 0.049927555157745766, "connect_only": 0.16551065695700648}, "ece_geometry": 0.040849909596787144}, "env": {"python": "3.12.13", "platform": "macOS-26.5.2-arm64-arm-64bit", "jax": "0.11.0", "jax_backend": "cpu"}, "artifact_paths": []} +{"run_id": "a6c36d85615f", "timestamp": "2026-07-30T11:17:03.568524+00:00", "git_sha": "67dc7a2", "config_hash": "f7c0c557ea97", "config": {"run": {"seed": 0, "task": "arithmetic", "notes": "F4 adaptive-halting run"}, "model": {"latent_dim": 32, "hidden_dim": 128, "context_dim": 64}, "inference": {"k_restarts": 12, "steps": 50, "step_size": 0.1, "temperature": 0.05, "init_scale": 1.0, "grad_tol": 0.001, "sampler": "langevin", "anneal_levels": 10, "anneal_steps_per_level": 8, "anneal_step_max": 0.2, "anneal_step_min": 0.01}, "train": {"epochs": 7, "batch_size": 128, "lr": 0.001, "n_train": 6000, "n_neg": 4, "neg_noise": 0.5, "objective": "basin_center", "ired_noise_min": 0.1, "ired_noise_max": 1.5, "ired_ridge": 0.05, "ired_decode_weight": 1.0, "ired_stat_weight": 1.0}, "eval": {"n_eval": 800, "richer_geometry": false}, "conformal": {"alpha": 0.1, "delta": 0.05, "n_calib": 800}, "task": {"arithmetic": {"modular": false, "n_operands": 6, "ood_n_operands": 6, "max_operand": 7, "ood_max_operand": 9}}}, "task": "arithmetic", "split": "halting", "seed": 0, "metrics": {"k_restarts": 12, "steps": 50, "alpha": 0.1, "n_calib": 800, "n_test": 800, "final_train_loss": 1.010847568511963, "tau_hat": 0.9167, "compute_used": 0.42152500000000004, "compute_saved": 0.578475, "halting_risk": 0.00875, "risk_within_budget": true, "full_accuracy": 0.80375, "halted_accuracy": 0.8075, "accuracy_drop": -0.003750000000000031, "risk_monotone_in_tau": true, "calib_risk_by_tau": [0.9025, 0.845, 0.845, 0.71, 0.48375, 0.48375, 0.37625, 0.19, 0.19, 0.13875, 0.00875, 0.00875], "calib_compute_by_tau": [0.0, 0.013049999999999886, 0.013049999999999886, 0.03759999999999957, 0.09074999999999989, 0.09074999999999989, 0.1236749999999998, 0.20574999999999993, 0.20574999999999993, 0.2681749999999995, 0.4182, 0.4182], "tau_sweep": {"taus": [0.0833, 0.1667, 0.25, 0.3333, 0.4167, 0.5, 0.5833, 0.6667, 0.75, 0.8333, 0.9167, 1.0], "compute_used": [0.0, 0.013375000000000001, 0.013375000000000001, 0.0383, 0.09365000000000002, 0.09365000000000002, 0.12260000000000001, 0.20592500000000002, 0.20592500000000002, 0.25880000000000003, 0.42152500000000004, 0.42152500000000004], "accuracy": [0.0875, 0.14375, 0.14375, 0.2375, 0.455, 0.455, 0.56, 0.69125, 0.69125, 0.7425, 0.8075, 0.8075], "disagreement": [0.91625, 0.85625, 0.85625, 0.7425, 0.4975, 0.4975, 0.37375, 0.20875, 0.20875, 0.14875, 0.00875, 0.00875]}}, "env": {"python": "3.12.13", "platform": "macOS-26.5.2-arm64-arm-64bit", "jax": "0.11.0", "jax_backend": "cpu"}, "artifact_paths": []} +{"run_id": "5b496a841123", "timestamp": "2026-07-30T11:17:05.368848+00:00", "git_sha": "67dc7a2", "config_hash": "b6cd6bcd6e19", "config": {"run": {"seed": 1, "task": "arithmetic", "notes": "F4 adaptive-halting run"}, "model": {"latent_dim": 32, "hidden_dim": 128, "context_dim": 64}, "inference": {"k_restarts": 12, "steps": 50, "step_size": 0.1, "temperature": 0.05, "init_scale": 1.0, "grad_tol": 0.001, "sampler": "langevin", "anneal_levels": 10, "anneal_steps_per_level": 8, "anneal_step_max": 0.2, "anneal_step_min": 0.01}, "train": {"epochs": 7, "batch_size": 128, "lr": 0.001, "n_train": 6000, "n_neg": 4, "neg_noise": 0.5, "objective": "basin_center", "ired_noise_min": 0.1, "ired_noise_max": 1.5, "ired_ridge": 0.05, "ired_decode_weight": 1.0, "ired_stat_weight": 1.0}, "eval": {"n_eval": 800, "richer_geometry": false}, "conformal": {"alpha": 0.1, "delta": 0.05, "n_calib": 800}, "task": {"arithmetic": {"modular": false, "n_operands": 6, "ood_n_operands": 6, "max_operand": 7, "ood_max_operand": 9}}}, "task": "arithmetic", "split": "halting", "seed": 1, "metrics": {"k_restarts": 12, "steps": 50, "alpha": 0.1, "n_calib": 800, "n_test": 800, "final_train_loss": 0.9732173085212708, "tau_hat": 0.9167, "compute_used": 0.45080000000000003, "compute_saved": 0.5491999999999999, "halting_risk": 0.0175, "risk_within_budget": true, "full_accuracy": 0.80375, "halted_accuracy": 0.8175, "accuracy_drop": -0.01375000000000004, "risk_monotone_in_tau": true, "calib_risk_by_tau": [0.88625, 0.84625, 0.84625, 0.73375, 0.465, 0.465, 0.35125, 0.21875, 0.21875, 0.12625, 0.015, 0.015], "calib_compute_by_tau": [0.0, 0.009049999999999954, 0.009049999999999954, 0.03192499999999969, 0.09209999999999983, 0.09209999999999983, 0.12292499999999983, 0.20739999999999958, 0.20739999999999958, 0.27014999999999956, 0.43754999999999983, 0.43754999999999983], "tau_sweep": {"taus": [0.0833, 0.1667, 0.25, 0.3333, 0.4167, 0.5, 0.5833, 0.6667, 0.75, 0.8333, 0.9167, 1.0], "compute_used": [0.0, 0.010074999999999999, 0.010074999999999999, 0.0333, 0.09310000000000002, 0.09310000000000002, 0.12170000000000002, 0.21402500000000002, 0.21402500000000002, 0.27865, 0.45080000000000003, 0.45080000000000003], "accuracy": [0.095, 0.135, 0.135, 0.235, 0.4575, 0.4575, 0.5575, 0.7175, 0.7175, 0.76625, 0.8175, 0.8175], "disagreement": [0.9025, 0.8575, 0.8575, 0.74625, 0.50375, 0.50375, 0.38875, 0.2125, 0.2125, 0.1475, 0.0175, 0.0175]}}, "env": {"python": "3.12.13", "platform": "macOS-26.5.2-arm64-arm-64bit", "jax": "0.11.0", "jax_backend": "cpu"}, "artifact_paths": []} +{"run_id": "24d91758f427", "timestamp": "2026-07-30T11:17:07.226291+00:00", "git_sha": "67dc7a2", "config_hash": "4b65b46d70e6", "config": {"run": {"seed": 2, "task": "arithmetic", "notes": "F4 adaptive-halting run"}, "model": {"latent_dim": 32, "hidden_dim": 128, "context_dim": 64}, "inference": {"k_restarts": 12, "steps": 50, "step_size": 0.1, "temperature": 0.05, "init_scale": 1.0, "grad_tol": 0.001, "sampler": "langevin", "anneal_levels": 10, "anneal_steps_per_level": 8, "anneal_step_max": 0.2, "anneal_step_min": 0.01}, "train": {"epochs": 7, "batch_size": 128, "lr": 0.001, "n_train": 6000, "n_neg": 4, "neg_noise": 0.5, "objective": "basin_center", "ired_noise_min": 0.1, "ired_noise_max": 1.5, "ired_ridge": 0.05, "ired_decode_weight": 1.0, "ired_stat_weight": 1.0}, "eval": {"n_eval": 800, "richer_geometry": false}, "conformal": {"alpha": 0.1, "delta": 0.05, "n_calib": 800}, "task": {"arithmetic": {"modular": false, "n_operands": 6, "ood_n_operands": 6, "max_operand": 7, "ood_max_operand": 9}}}, "task": "arithmetic", "split": "halting", "seed": 2, "metrics": {"k_restarts": 12, "steps": 50, "alpha": 0.1, "n_calib": 800, "n_test": 800, "final_train_loss": 0.9444285035133362, "tau_hat": 0.9167, "compute_used": 0.417825, "compute_saved": 0.582175, "halting_risk": 0.0175, "risk_within_budget": true, "full_accuracy": 0.86125, "halted_accuracy": 0.875, "accuracy_drop": -0.01375000000000004, "risk_monotone_in_tau": true, "calib_risk_by_tau": [0.91375, 0.8575, 0.8575, 0.71875, 0.4675, 0.4675, 0.35125, 0.20625, 0.20625, 0.13625, 0.01375, 0.01375], "calib_compute_by_tau": [0.0, 0.015349999999999841, 0.015349999999999841, 0.042149999999999584, 0.09782499999999984, 0.09782499999999984, 0.13077499999999975, 0.2109499999999998, 0.2109499999999998, 0.27007499999999957, 0.41997499999999993, 0.41997499999999993], "tau_sweep": {"taus": [0.0833, 0.1667, 0.25, 0.3333, 0.4167, 0.5, 0.5833, 0.6667, 0.75, 0.8333, 0.9167, 1.0], "compute_used": [0.0, 0.0146, 0.0146, 0.03995, 0.097375, 0.097375, 0.12955000000000003, 0.20885, 0.20885, 0.266425, 0.417825, 0.417825], "accuracy": [0.0775, 0.13875, 0.13875, 0.27375, 0.5125, 0.5125, 0.6025, 0.73, 0.73, 0.78, 0.875, 0.875], "disagreement": [0.9225, 0.86625, 0.86625, 0.7275, 0.45, 0.45, 0.3525, 0.23625, 0.23625, 0.1675, 0.0175, 0.0175]}}, "env": {"python": "3.12.13", "platform": "macOS-26.5.2-arm64-arm-64bit", "jax": "0.11.0", "jax_backend": "cpu"}, "artifact_paths": []} +{"run_id": "19cb5e9f64c6", "timestamp": "2026-07-30T11:17:08.943784+00:00", "git_sha": "67dc7a2", "config_hash": "e5e215c688c3", "config": {"run": {"seed": 3, "task": "arithmetic", "notes": "F4 adaptive-halting run"}, "model": {"latent_dim": 32, "hidden_dim": 128, "context_dim": 64}, "inference": {"k_restarts": 12, "steps": 50, "step_size": 0.1, "temperature": 0.05, "init_scale": 1.0, "grad_tol": 0.001, "sampler": "langevin", "anneal_levels": 10, "anneal_steps_per_level": 8, "anneal_step_max": 0.2, "anneal_step_min": 0.01}, "train": {"epochs": 7, "batch_size": 128, "lr": 0.001, "n_train": 6000, "n_neg": 4, "neg_noise": 0.5, "objective": "basin_center", "ired_noise_min": 0.1, "ired_noise_max": 1.5, "ired_ridge": 0.05, "ired_decode_weight": 1.0, "ired_stat_weight": 1.0}, "eval": {"n_eval": 800, "richer_geometry": false}, "conformal": {"alpha": 0.1, "delta": 0.05, "n_calib": 800}, "task": {"arithmetic": {"modular": false, "n_operands": 6, "ood_n_operands": 6, "max_operand": 7, "ood_max_operand": 9}}}, "task": "arithmetic", "split": "halting", "seed": 3, "metrics": {"k_restarts": 12, "steps": 50, "alpha": 0.1, "n_calib": 800, "n_test": 800, "final_train_loss": 0.943260669708252, "tau_hat": 0.9167, "compute_used": 0.447575, "compute_saved": 0.5524249999999999, "halting_risk": 0.01, "risk_within_budget": true, "full_accuracy": 0.84, "halted_accuracy": 0.83875, "accuracy_drop": 0.0012499999999999734, "risk_monotone_in_tau": true, "calib_risk_by_tau": [0.92375, 0.83875, 0.83875, 0.73, 0.48375, 0.48375, 0.37625, 0.2, 0.2, 0.12625, 0.01375, 0.01375], "calib_compute_by_tau": [0.0, 0.02069999999999977, 0.02069999999999977, 0.04744999999999969, 0.10197499999999976, 0.10197499999999976, 0.13589999999999983, 0.22402499999999953, 0.22402499999999953, 0.28324999999999995, 0.4625250000000001, 0.4625250000000001], "tau_sweep": {"taus": [0.0833, 0.1667, 0.25, 0.3333, 0.4167, 0.5, 0.5833, 0.6667, 0.75, 0.8333, 0.9167, 1.0], "compute_used": [0.0, 0.021325000000000004, 0.021325000000000004, 0.046150000000000004, 0.09740000000000001, 0.09740000000000001, 0.12895, 0.215375, 0.215375, 0.28347500000000003, 0.447575, 0.447575], "accuracy": [0.0875, 0.18, 0.18, 0.30625, 0.49625, 0.49625, 0.61125, 0.75875, 0.75875, 0.78125, 0.83875, 0.83875], "disagreement": [0.89875, 0.8075, 0.8075, 0.6825, 0.47125, 0.47125, 0.3475, 0.19625, 0.19625, 0.135, 0.01, 0.01]}}, "env": {"python": "3.12.13", "platform": "macOS-26.5.2-arm64-arm-64bit", "jax": "0.11.0", "jax_backend": "cpu"}, "artifact_paths": []} +{"run_id": "2a54cf585cb3", "timestamp": "2026-07-30T11:17:10.656072+00:00", "git_sha": "67dc7a2", "config_hash": "315846db967f", "config": {"run": {"seed": 4, "task": "arithmetic", "notes": "F4 adaptive-halting run"}, "model": {"latent_dim": 32, "hidden_dim": 128, "context_dim": 64}, "inference": {"k_restarts": 12, "steps": 50, "step_size": 0.1, "temperature": 0.05, "init_scale": 1.0, "grad_tol": 0.001, "sampler": "langevin", "anneal_levels": 10, "anneal_steps_per_level": 8, "anneal_step_max": 0.2, "anneal_step_min": 0.01}, "train": {"epochs": 7, "batch_size": 128, "lr": 0.001, "n_train": 6000, "n_neg": 4, "neg_noise": 0.5, "objective": "basin_center", "ired_noise_min": 0.1, "ired_noise_max": 1.5, "ired_ridge": 0.05, "ired_decode_weight": 1.0, "ired_stat_weight": 1.0}, "eval": {"n_eval": 800, "richer_geometry": false}, "conformal": {"alpha": 0.1, "delta": 0.05, "n_calib": 800}, "task": {"arithmetic": {"modular": false, "n_operands": 6, "ood_n_operands": 6, "max_operand": 7, "ood_max_operand": 9}}}, "task": "arithmetic", "split": "halting", "seed": 4, "metrics": {"k_restarts": 12, "steps": 50, "alpha": 0.1, "n_calib": 800, "n_test": 800, "final_train_loss": 0.8476110100746155, "tau_hat": 0.9167, "compute_used": 0.407875, "compute_saved": 0.592125, "halting_risk": 0.00625, "risk_within_budget": true, "full_accuracy": 0.84, "halted_accuracy": 0.84625, "accuracy_drop": -0.006249999999999978, "risk_monotone_in_tau": true, "calib_risk_by_tau": [0.90375, 0.83, 0.83, 0.72375, 0.485, 0.485, 0.3775, 0.18, 0.18, 0.1225, 0.01125, 0.01125], "calib_compute_by_tau": [0.0, 0.01557499999999983, 0.01557499999999983, 0.03847499999999953, 0.08722499999999994, 0.08722499999999994, 0.11397499999999977, 0.19922499999999954, 0.19922499999999954, 0.2511749999999996, 0.39365, 0.39365], "tau_sweep": {"taus": [0.0833, 0.1667, 0.25, 0.3333, 0.4167, 0.5, 0.5833, 0.6667, 0.75, 0.8333, 0.9167, 1.0], "compute_used": [0.0, 0.015425, 0.015425, 0.0375, 0.08922500000000001, 0.08922500000000001, 0.12187500000000001, 0.20292500000000005, 0.20292500000000005, 0.2681, 0.407875, 0.407875], "accuracy": [0.07875, 0.14875, 0.14875, 0.2475, 0.44875, 0.44875, 0.535, 0.71625, 0.71625, 0.7725, 0.84625, 0.84625], "disagreement": [0.91375, 0.8425, 0.8425, 0.7175, 0.50125, 0.50125, 0.40375, 0.19375, 0.19375, 0.12375, 0.00625, 0.00625]}}, "env": {"python": "3.12.13", "platform": "macOS-26.5.2-arm64-arm-64bit", "jax": "0.11.0", "jax_backend": "cpu"}, "artifact_paths": []} +{"run_id": "eaebf4a2b9f3", "timestamp": "2026-07-30T11:17:52.284763+00:00", "git_sha": "67dc7a2", "config_hash": "fccc394c29a9", "config": {"run": {"seed": 0, "task": "graph_planning", "notes": "graph adaptive-halting run"}, "model": {"latent_dim": 32, "hidden_dim": 128, "context_dim": 64}, "inference": {"k_restarts": 12, "steps": 50, "step_size": 0.1, "temperature": 0.05, "init_scale": 1.0, "grad_tol": 0.001, "sampler": "langevin", "anneal_levels": 10, "anneal_steps_per_level": 8, "anneal_step_max": 0.2, "anneal_step_min": 0.01}, "train": {"epochs": 25, "batch_size": 128, "lr": 0.001, "n_train": 8000, "n_neg": 4, "neg_noise": 0.5, "objective": "basin_center", "ired_noise_min": 0.1, "ired_noise_max": 1.5, "ired_ridge": 0.05, "ired_decode_weight": 1.0, "ired_stat_weight": 1.0}, "eval": {"n_eval": 800, "richer_geometry": false}, "conformal": {"alpha": 0.1, "delta": 0.05, "n_calib": 800}, "task": {"graph_planning": {"n_nodes": 7, "ood_n_nodes": 10, "edge_prob": 0.4, "max_len": 4}}}, "task": "graph_planning", "split": "halting", "seed": 0, "metrics": {"k_restarts": 12, "steps": 50, "alpha": 0.1, "n_calib": 800, "n_test": 800, "final_train_loss": 0.227946937084198, "tau_hat": 0.8333, "compute_used": 0.168025, "compute_saved": 0.831975, "halting_risk": 0.0825, "risk_within_budget": true, "full_accuracy": 0.715, "halted_accuracy": 0.70625, "accuracy_drop": 0.008749999999999925, "risk_monotone_in_tau": true, "calib_risk_by_tau": [0.4575, 0.4575, 0.4575, 0.455, 0.3575, 0.3575, 0.2775, 0.14, 0.14, 0.08875, 0.005, 0.005], "calib_compute_by_tau": [0.0, 0.0, 0.0, 0.0005750000000000002, 0.025024999999999804, 0.025024999999999804, 0.05142499999999985, 0.12042500000000007, 0.12042500000000007, 0.16520000000000032, 0.28642499999999965, 0.28642499999999965], "tau_sweep": {"taus": [0.0833, 0.1667, 0.25, 0.3333, 0.4167, 0.5, 0.5833, 0.6667, 0.75, 0.8333, 0.9167, 1.0], "compute_used": [0.0, 0.0, 0.0, 0.000625, 0.024675000000000002, 0.024675000000000002, 0.0509, 0.11850000000000001, 0.11850000000000001, 0.168025, 0.28750000000000003, 0.28750000000000003], "accuracy": [0.53625, 0.53625, 0.53625, 0.53625, 0.58625, 0.58625, 0.61875, 0.6825, 0.6825, 0.70625, 0.71625, 0.71625], "disagreement": [0.4375, 0.4375, 0.4375, 0.43625, 0.34625, 0.34625, 0.2725, 0.12125, 0.12125, 0.0825, 0.005, 0.005]}}, "env": {"python": "3.12.13", "platform": "macOS-26.5.2-arm64-arm-64bit", "jax": "0.11.0", "jax_backend": "cpu"}, "artifact_paths": []} +{"run_id": "59f889f657d5", "timestamp": "2026-07-30T11:17:53.773421+00:00", "git_sha": "67dc7a2", "config_hash": "646eeb651c32", "config": {"run": {"seed": 0, "task": "arithmetic", "notes": "E1 selective-prediction main run"}, "model": {"latent_dim": 32, "hidden_dim": 128, "context_dim": 64}, "inference": {"k_restarts": 12, "steps": 50, "step_size": 0.1, "temperature": 0.05, "init_scale": 1.0, "grad_tol": 0.001, "sampler": "langevin", "anneal_levels": 10, "anneal_steps_per_level": 8, "anneal_step_max": 0.2, "anneal_step_min": 0.01}, "train": {"epochs": 7, "batch_size": 128, "lr": 0.001, "n_train": 6000, "n_neg": 4, "neg_noise": 0.5, "objective": "basin_center", "ired_noise_min": 0.1, "ired_noise_max": 1.5, "ired_ridge": 0.05, "ired_decode_weight": 1.0, "ired_stat_weight": 1.0}, "eval": {"n_eval": 1500, "richer_geometry": false}, "conformal": {"alpha": 0.1, "delta": 0.05, "n_calib": 1500}, "task": {"arithmetic": {"modular": false, "n_operands": 6, "ood_n_operands": 6, "max_operand": 7, "ood_max_operand": 9}}}, "task": "arithmetic", "split": "selective", "seed": 0, "metrics": {"n_fit": 1500, "n_calib": 1500, "n_test": 1500, "k_restarts": 12, "objective": "basin_center", "sampler": "langevin", "feature_set": "base", "accuracy_id": 0.81, "base_error": 0.19, "final_train_loss": 1.010847568511963, "feature_names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual"], "aurc": {"geometry": 0.0711432604851103, "rho_basin": 0.1192048662021191, "energy_min": 0.1447332231633481, "energy_mean": 0.14862598663924864, "energy_std": 0.18501378708598182, "msp": 0.06933811307974048, "temp_msp": 0.08261243554210007, "entropy": 0.07082020691013495, "softmax_learned": 0.06450094181149578, "geom_softmax": 0.05673560229226548}, "temperature": 0.4068744349087813, "best_energy_baseline": "energy_min", "best_baseline": "msp", "delta_aurc_vs_energy_min": [0.07358996267823781, 0.05406001270195325, 0.09343612485819916], "delta_aurc_vs_best_energy": [0.07358996267823781, 0.05406001270195325, 0.09343612485819916], "delta_aurc_vs_best_baseline": [-0.001805147405369814, -0.015438010117724282, 0.011245405317148913], "delta_aurc_geom_adds": [0.007765339519230301, 0.003572430837825001, 0.012225885708889614], "geometry_wins": true, "geometry_wins_vs_baseline": false, "geometry_adds_over_softmax": true, "risk_coverage": {"geometry": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.03333333333333333, 0.05, 0.03333333333333333, 0.025, 0.02666666666666667, 0.033333333333333326, 0.052380952380952396, 0.05, 0.04814814814814815, 0.046666666666666676, 0.04242424242424243, 0.041666666666666664, 0.04358974358974359, 0.04047619047619048, 0.04444444444444444, 0.04375, 0.043137254901960784, 0.04444444444444444, 0.043859649122807015, 0.045, 0.04285714285714285, 0.04242424242424243, 0.04057971014492753, 0.04166666666666666, 0.044, 0.04358974358974359, 0.04938271604938271, 0.05238095238095238, 0.0586206896551724, 0.06444444444444444, 0.06451612903225806, 0.06458333333333334, 0.06666666666666667, 0.07058823529411765, 0.07714285714285712, 0.08148148148148163, 0.08558558558558559, 0.08947368421052632, 0.10170940170940171, 0.11, 0.11788617886178875, 0.12380952380952392, 0.12945736434108526, 0.13636363636363635, 0.14444444444444443, 0.1492753623188406, 0.15673758865248238, 0.16805555555555565, 0.1782312925170068, 0.19]}, "rho_basin": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.10957642725598533, 0.10957642725598518, 0.10957642725598526, 0.1095764272559854, 0.10957642725598545, 0.10957642725598521, 0.10957642725598504, 0.10957642725598492, 0.10957642725598482, 0.10957642725598482, 0.10957642725598507, 0.10957642725598529, 0.10957642725598546, 0.10957642725598561, 0.10957642725598575, 0.10957642725598586, 0.10957642725598597, 0.10957642725598606, 0.10957642725598614, 0.1095764272559862, 0.10957642725598626, 0.10957642725598632, 0.10957642725598638, 0.10957642725598643, 0.10957642725598647, 0.10957642725598653, 0.10957642725598656, 0.1095764272559866, 0.10957642725598664, 0.10957642725598667, 0.1095764272559867, 0.10957642725598672, 0.10957642725598675, 0.10957642725598678, 0.10957642725598679, 0.10957642725598682, 0.11300813008130228, 0.11709456568250055, 0.12097144048363749, 0.12465447154471758, 0.12967479674796975, 0.1350907029478481, 0.14025470653377864, 0.14618844696969926, 0.15196759259259482, 0.15961352657005046, 0.1675650118203331, 0.17561111111111322, 0.18333333333333537, 0.19000000000000197]}, "energy_min": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.03333333333333333, 0.16666666666666666, 0.16666666666666666, 0.15833333333333333, 0.16, 0.15555555555555553, 0.14285714285714268, 0.1375, 0.14074074074074075, 0.1399999999999999, 0.13636363636363635, 0.1361111111111111, 0.13333333333333333, 0.13095238095238096, 0.12888888888888886, 0.13541666666666666, 0.13725490196078433, 0.13518518518518516, 0.13859649122807016, 0.14333333333333334, 0.14126984126984124, 0.1409090909090909, 0.14202898550724638, 0.1444444444444444, 0.14533333333333334, 0.14743589743589744, 0.14938271604938272, 0.15, 0.14712643678160917, 0.15, 0.14623655913978495, 0.14375, 0.1404040404040404, 0.13725490196078433, 0.13619047619047617, 0.13425925925925924, 0.13423423423423422, 0.13508771929824562, 0.13931623931623932, 0.1425, 0.14552845528455283, 0.14841269841269839, 0.15426356589147286, 0.15757575757575756, 0.16296296296296298, 0.16666666666666666, 0.17163120567375897, 0.17638888888888887, 0.18503401360544217, 0.19]}, "energy_mean": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.1, 0.15, 0.17777777777777778, 0.14166666666666666, 0.15333333333333332, 0.14999999999999997, 0.14761904761904765, 0.14166666666666666, 0.14814814814814814, 0.1466666666666667, 0.14545454545454545, 0.1388888888888889, 0.13333333333333333, 0.1357142857142857, 0.13999999999999999, 0.14375, 0.1411764705882353, 0.1444444444444444, 0.14385964912280702, 0.14333333333333334, 0.13968253968253966, 0.15, 0.15072463768115943, 0.15138888888888885, 0.14933333333333335, 0.15512820512820513, 0.15432098765432098, 0.15119047619047618, 0.14942528735632182, 0.14777777777777779, 0.14731182795698924, 0.14270833333333333, 0.1404040404040404, 0.13823529411764707, 0.1352380952380952, 0.13425925925925924, 0.13603603603603603, 0.13771929824561405, 0.14017094017094017, 0.14083333333333334, 0.1422764227642276, 0.146031746031746, 0.15193798449612403, 0.1590909090909091, 0.16296296296296298, 0.17028985507246377, 0.1730496453900709, 0.17638888888888898, 0.18435374149659864, 0.19]}, "energy_std": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.23333333333333334, 0.18333333333333332, 0.16666666666666666, 0.19166666666666668, 0.16666666666666666, 0.1611111111111111, 0.16666666666666669, 0.15833333333333333, 0.15925925925925927, 0.16666666666666669, 0.1696969696969697, 0.18611111111111112, 0.18461538461538463, 0.19523809523809524, 0.1911111111111111, 0.19166666666666668, 0.18627450980392157, 0.18888888888888886, 0.18947368421052632, 0.18833333333333332, 0.18888888888888886, 0.19242424242424241, 0.19130434782608696, 0.18888888888888886, 0.192, 0.18974358974358974, 0.18518518518518517, 0.18214285714285713, 0.18160919540229883, 0.17888888888888888, 0.18494623655913978, 0.18645833333333334, 0.18686868686868688, 0.18627450980392157, 0.1876190476190476, 0.18703703703703717, 0.1873873873873874, 0.18859649122807018, 0.18803418803418803, 0.1925, 0.19430894308943086, 0.19444444444444442, 0.1937984496124031, 0.19242424242424241, 0.18962962962962962, 0.18840579710144928, 0.18936170212765954, 0.19097222222222218, 0.19115646258503402, 0.19]}, "msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.011111111111111112, 0.03333333333333333, 0.03333333333333333, 0.033333333333333326, 0.03333333333333334, 0.029166666666666667, 0.02962962962962963, 0.026666666666666672, 0.024242424242424242, 0.03611111111111111, 0.03333333333333333, 0.03333333333333333, 0.033333333333333326, 0.0375, 0.0392156862745098, 0.04074074074074074, 0.042105263157894736, 0.04833333333333333, 0.04761904761904761, 0.051515151515151514, 0.04927536231884058, 0.049999999999999996, 0.052, 0.052564102564102565, 0.05308641975308642, 0.055952380952380955, 0.06436781609195401, 0.06777777777777778, 0.07311827956989247, 0.07395833333333333, 0.08282828282828283, 0.08529411764705883, 0.08952380952380966, 0.09537037037037036, 0.1009009009009009, 0.10526315789473684, 0.1094017094017094, 0.11583333333333333, 0.12357723577235771, 0.12619047619047616, 0.13023255813953488, 0.13787878787878788, 0.1437037037037037, 0.1471014492753623, 0.15602836879432622, 0.1659722222222222, 0.17755102040816326, 0.19]}, "temp_msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.016666666666666666, 0.03333333333333333, 0.041666666666666664, 0.03333333333333333, 0.03888888888888888, 0.038095238095238106, 0.0375, 0.04814814814814815, 0.046666666666666676, 0.06363636363636363, 0.06111111111111111, 0.05897435897435897, 0.06190476190476191, 0.06222222222222221, 0.06458333333333334, 0.06666666666666667, 0.06296296296296294, 0.06666666666666667, 0.06666666666666667, 0.06349206349206347, 0.06363636363636363, 0.06521739130434782, 0.06805555555555554, 0.068, 0.07179487179487179, 0.07407407407407407, 0.07976190476190476, 0.08390804597701168, 0.08555555555555555, 0.08602150537634409, 0.08645833333333333, 0.0898989898989899, 0.0892156862745098, 0.09333333333333348, 0.10092592592592592, 0.1045045045045045, 0.10789473684210527, 0.11196581196581197, 0.11583333333333333, 0.1219512195121951, 0.1269841269841271, 0.1325581395348837, 0.14166666666666666, 0.1474074074074074, 0.15434782608695652, 0.16099290780141856, 0.16736111111111118, 0.1782312925170068, 0.19]}, "entropy": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.011111111111111112, 0.008333333333333333, 0.02666666666666667, 0.02222222222222222, 0.02380952380952381, 0.025, 0.025925925925925925, 0.023333333333333338, 0.03636363636363636, 0.03333333333333333, 0.03333333333333333, 0.0380952380952381, 0.039999999999999994, 0.041666666666666664, 0.041176470588235294, 0.04444444444444444, 0.043859649122807015, 0.05, 0.057142857142857134, 0.06060606060606061, 0.06231884057971015, 0.06666666666666665, 0.068, 0.06923076923076923, 0.07037037037037037, 0.07261904761904762, 0.07241379310344827, 0.07111111111111111, 0.07526881720430108, 0.07604166666666666, 0.07575757575757576, 0.0784313725490196, 0.0809523809523811, 0.08425925925925924, 0.08828828828828829, 0.09385964912280702, 0.10085470085470086, 0.10583333333333333, 0.11707317073170745, 0.123015873015873, 0.13178294573643412, 0.14166666666666666, 0.1474074074074074, 0.15869565217391304, 0.1638297872340425, 0.17013888888888887, 0.17959183673469387, 0.19]}, "softmax_learned": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.011111111111111112, 0.016666666666666666, 0.02666666666666667, 0.033333333333333326, 0.028571428571428577, 0.025, 0.025925925925925925, 0.026666666666666672, 0.024242424242424242, 0.022222222222222223, 0.02564102564102564, 0.023809523809523808, 0.02888888888888888, 0.03125, 0.03333333333333333, 0.03518518518518518, 0.03508771929824561, 0.03666666666666667, 0.04126984126984126, 0.03939393939393939, 0.042028985507246375, 0.04305555555555555, 0.044, 0.04230769230769231, 0.04691358024691358, 0.04880952380952381, 0.0528735632183908, 0.058888888888888886, 0.06344086021505377, 0.06666666666666667, 0.07474747474747474, 0.0784313725490196, 0.0819047619047619, 0.09074074074074073, 0.0945945945945946, 0.10087719298245613, 0.10598290598290598, 0.11416666666666667, 0.12032520325203251, 0.1238095238095238, 0.1310077519379845, 0.14015151515151514, 0.14148148148148149, 0.15, 0.15673758865248225, 0.16736111111111118, 0.17687074829931973, 0.19]}, "geom_softmax": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.02, 0.02222222222222222, 0.019047619047619053, 0.016666666666666666, 0.022222222222222223, 0.026666666666666672, 0.02727272727272727, 0.025, 0.023076923076923078, 0.02142857142857143, 0.019999999999999997, 0.01875, 0.0196078431372549, 0.018518518518518514, 0.021052631578947368, 0.023333333333333334, 0.023809523809523937, 0.02878787878787879, 0.02753623188405797, 0.029166666666666664, 0.032, 0.03333333333333333, 0.037037037037037035, 0.04404761904761905, 0.04712643678160919, 0.04777777777777778, 0.04946236559139785, 0.05520833333333333, 0.05858585858585859, 0.06470588235294118, 0.0676190476190476, 0.07592592592592591, 0.08378378378378379, 0.09473684210526316, 0.10085470085470086, 0.10666666666666667, 0.11382113821138223, 0.11984126984126983, 0.12868217054263567, 0.13484848484848486, 0.13925925925925925, 0.1463768115942029, 0.15744680851063841, 0.16736111111111118, 0.17959183673469387, 0.19]}}, "ltt": {"alpha": 0.1, "delta": 0.05, "lambda_hat": 0.19374220945381396, "coverage": 0.6933333333333334, "abstain_rate": 0.30666666666666664, "selective_risk": 0.075, "selective_accuracy": 0.925, "risk_within_budget": true}, "coverage_validity": {"alpha_grid": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "target": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "achieved_risk": [0.0, 0.0, 0.075, 0.1295577967416602, 0.17525773195876287, 0.1879598662207358], "coverage": [0.0, 0.0, 0.6933333333333334, 0.8593333333333333, 0.97, 0.9966666666666667]}, "feature_diagnostics": {"names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual"], "auroc": {"basin/rho": 0.7041700960219479, "basin/entropy": 0.2946790845426323, "basin/dispersion": 0.5298130098909826, "energy/mean": 0.3736336726590138, "energy/min": 0.36821890116237094, "energy/std": 0.49345462421485814, "curv/lmax_mean": 0.4698808750270739, "curv/lmax_best": 0.4712165186629124, "curv/trace_mean": 0.4452761533463288, "curv/trace_best": 0.4445397444227854, "dynamics/steps": 0.5, "dynamics/monotonic": 0.341111833080644, "dynamics/drop": 0.3597834091401343, "dynamics/residual": 0.5355483358602267}, "hist": {"basin/rho": {"edges": [0.3333333333333333, 0.36666666666666664, 0.4, 0.43333333333333335, 0.4666666666666667, 0.5, 0.5333333333333333, 0.5666666666666667, 0.6000000000000001, 0.6333333333333333, 0.6666666666666667, 0.7000000000000001, 0.7333333333333334, 0.7666666666666667, 0.8, 0.8333333333333335, 0.8666666666666667, 0.9000000000000001, 0.9333333333333333, 0.9666666666666668, 1.0], "correct_counts": [2, 0, 0, 0, 0, 14, 0, 22, 0, 28, 0, 0, 38, 0, 54, 0, 0, 90, 0, 967], "incorrect_counts": [1, 0, 2, 0, 0, 14, 0, 28, 0, 32, 0, 0, 26, 0, 30, 0, 0, 33, 0, 119]}, "basin/entropy": {"edges": [0.0, 0.06789889274936621, 0.13579778549873242, 0.20369667824809862, 0.27159557099746484, 0.33949446374683107, 0.40739335649619723, 0.47529224924556346, 0.5431911419949297, 0.6110900347442959, 0.6789889274936621, 0.7468878202430284, 0.8147867129923945, 0.8826856057417607, 0.9505844984911269, 1.0184833912404931, 1.0863822839898594, 1.1542811767392256, 1.2221800694885918, 1.290078962237958, 1.3579778549873243], "correct_counts": [967, 0, 0, 0, 90, 0, 53, 0, 37, 25, 35, 0, 3, 1, 1, 1, 1, 0, 0, 1], "incorrect_counts": [119, 0, 0, 0, 33, 0, 26, 0, 27, 29, 35, 0, 3, 2, 6, 3, 2, 0, 0, 0]}, "basin/dispersion": {"edges": [1.592531476672375, 1.614116358729778, 1.6357012407871807, 1.6572861228445837, 1.6788710049019866, 1.7004558869593893, 1.7220407690167923, 1.7436256510741952, 1.7652105331315981, 1.7867954151890009, 1.8083802972464038, 1.8299651793038068, 1.8515500613612095, 1.8731349434186124, 1.8947198254760154, 1.9163047075334183, 1.937889589590821, 1.959474471648224, 1.981059353705627, 2.0026442357630296, 2.0242291178204326], "correct_counts": [3, 2, 8, 9, 23, 47, 59, 82, 96, 177, 167, 139, 126, 93, 70, 52, 31, 17, 10, 4], "incorrect_counts": [0, 0, 2, 1, 5, 11, 19, 21, 37, 36, 38, 29, 30, 20, 17, 12, 5, 1, 1, 0]}, "energy/mean": {"edges": [-0.30860641847054165, -0.24559052648643653, -0.18257463450233144, -0.11955874251822635, -0.056542850534121225, 0.006473041449983896, 0.06948893343408896, 0.13250482541819408, 0.1955207174022992, 0.25853660938640427, 0.32155250137050945, 0.3845683933546145, 0.4475842853387196, 0.5106001773228248, 0.5736160693069299, 0.6366319612910349, 0.69964785327514, 0.7626637452592451, 0.8256796372433501, 0.8886955292274554, 0.9517114212115606], "correct_counts": [10, 53, 216, 277, 172, 82, 46, 41, 37, 37, 39, 36, 43, 41, 43, 21, 18, 2, 0, 1], "incorrect_counts": [2, 11, 35, 49, 32, 8, 2, 2, 4, 7, 14, 6, 21, 33, 19, 24, 12, 3, 1, 0]}, "energy/min": {"edges": [-0.6984671354293823, -0.628644996881485, -0.5588228583335877, -0.48900071978569026, -0.41917858123779295, -0.34935644268989563, -0.27953430414199826, -0.20971216559410089, -0.13989002704620357, -0.07006788849830625, -0.00024574995040893555, 0.06957638859748849, 0.1393985271453858, 0.20922066569328313, 0.27904280424118055, 0.34886494278907776, 0.4186870813369752, 0.4885092198848726, 0.5583313584327698, 0.6281534969806672, 0.6979756355285645], "correct_counts": [5, 25, 85, 195, 264, 157, 92, 53, 48, 43, 42, 45, 44, 51, 33, 14, 13, 6, 0, 0], "incorrect_counts": [0, 3, 21, 25, 47, 32, 10, 2, 5, 6, 12, 17, 21, 28, 22, 19, 11, 3, 0, 1]}, "energy/std": {"edges": [0.058812152065186904, 0.07638791223625852, 0.09396367240733014, 0.11153943257840175, 0.12911519274947336, 0.14669095292054496, 0.1642667130916166, 0.1818424732626882, 0.1994182334337598, 0.21699399360483146, 0.23456975377590306, 0.25214551394697465, 0.2697212741180463, 0.2872970342891179, 0.3048727944601895, 0.32244855463126115, 0.34002431480233275, 0.35760007497340435, 0.375175835144476, 0.3927515953155476, 0.4103273554866192], "correct_counts": [3, 6, 20, 52, 77, 135, 180, 211, 168, 133, 90, 68, 39, 16, 9, 5, 1, 0, 1, 1], "incorrect_counts": [0, 3, 4, 12, 10, 38, 42, 44, 44, 39, 19, 10, 16, 1, 1, 1, 0, 1, 0, 0]}, "curv/lmax_mean": {"edges": [1.0000369946161907, 1.0007940803964934, 1.001551166176796, 1.0023082519570987, 1.0030653377374015, 1.003822423517704, 1.0045795092980068, 1.0053365950783095, 1.006093680858612, 1.0068507666389148, 1.0076078524192176, 1.0083649381995201, 1.0091220239798229, 1.0098791097601254, 1.0106361955404282, 1.011393281320731, 1.0121503671010335, 1.0129074528813362, 1.013664538661639, 1.0144216244419415, 1.0151787102222443], "correct_counts": [617, 173, 79, 69, 52, 45, 45, 33, 29, 19, 20, 8, 6, 5, 6, 2, 6, 0, 0, 1], "incorrect_counts": [119, 74, 29, 22, 10, 10, 4, 5, 3, 4, 3, 1, 0, 0, 0, 1, 0, 0, 0, 0]}, "curv/lmax_best": {"edges": [1.0, 1.0017322301864624, 1.0034644603729248, 1.0051966905593872, 1.0069289207458496, 1.008661150932312, 1.0103933811187744, 1.0121256113052368, 1.0138578414916992, 1.0155900716781616, 1.017322301864624, 1.0190545320510864, 1.0207867622375488, 1.0225189924240112, 1.0242512226104736, 1.025983452796936, 1.0277156829833984, 1.0294479131698608, 1.0311801433563232, 1.0329123735427856, 1.034644603729248], "correct_counts": [934, 93, 63, 35, 27, 11, 9, 7, 9, 7, 3, 3, 2, 5, 3, 2, 1, 0, 0, 1], "incorrect_counts": [213, 31, 16, 6, 6, 3, 3, 3, 2, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0]}, "curv/trace_mean": {"edges": [32.00269635518392, 32.00753688812256, 32.0123774210612, 32.01721795399984, 32.02205848693848, 32.02689901987712, 32.03173955281576, 32.036580085754395, 32.04142061869303, 32.046261151631676, 32.05110168457031, 32.05594221750895, 32.06078275044759, 32.06562328338623, 32.07046381632487, 32.075304349263504, 32.08014488220215, 32.084985415140785, 32.08982594807942, 32.094666481018066, 32.0995070139567], "correct_counts": [175, 275, 151, 82, 84, 56, 49, 58, 42, 42, 57, 38, 42, 21, 20, 15, 4, 2, 1, 1], "incorrect_counts": [9, 50, 53, 31, 37, 20, 19, 17, 9, 8, 7, 13, 4, 5, 2, 0, 1, 0, 0, 0]}, "curv/trace_best": {"edges": [32.0009880065918, 32.010338592529294, 32.0196891784668, 32.029039764404295, 32.0383903503418, 32.0477409362793, 32.057091522216794, 32.0664421081543, 32.075792694091795, 32.0851432800293, 32.0944938659668, 32.103844451904294, 32.1131950378418, 32.122545623779295, 32.1318962097168, 32.1412467956543, 32.150597381591794, 32.1599479675293, 32.169298553466795, 32.1786491394043, 32.1879997253418], "correct_counts": [402, 281, 150, 106, 83, 59, 46, 28, 17, 9, 10, 6, 3, 5, 2, 2, 2, 3, 0, 1], "incorrect_counts": [66, 67, 55, 31, 17, 13, 8, 8, 7, 6, 3, 2, 0, 1, 1, 0, 0, 0, 0, 0]}, "dynamics/steps": {"edges": [1.0, 1.05, 1.1, 1.15, 1.2, 1.25, 1.3, 1.35, 1.4, 1.45, 1.5, 1.55, 1.6, 1.65, 1.7000000000000002, 1.75, 1.8, 1.85, 1.9, 1.9500000000000002, 2.0], "correct_counts": [1215, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [285, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "dynamics/monotonic": {"edges": [0.7133333333333333, 0.7193333333333333, 0.7253333333333333, 0.7313333333333333, 0.7373333333333333, 0.7433333333333333, 0.7493333333333333, 0.7553333333333333, 0.7613333333333333, 0.7673333333333333, 0.7733333333333334, 0.7793333333333334, 0.7853333333333334, 0.7913333333333334, 0.7973333333333334, 0.8033333333333335, 0.8093333333333335, 0.8153333333333335, 0.8213333333333335, 0.8273333333333335, 0.8333333333333335], "correct_counts": [1, 0, 2, 23, 32, 66, 131, 145, 240, 187, 167, 125, 55, 25, 12, 3, 1, 0, 0, 0], "incorrect_counts": [1, 0, 0, 3, 8, 14, 10, 17, 33, 38, 46, 39, 19, 20, 14, 11, 6, 3, 2, 1]}, "dynamics/drop": {"edges": [63.47809874266386, 76.39078911232451, 89.30347948198516, 102.21616985164583, 115.12886022130648, 128.04155059096712, 140.9542409606278, 153.86693133028845, 166.7796216999491, 179.69231206960976, 192.6050024392704, 205.51769280893106, 218.43038317859174, 231.3430735482524, 244.25576391791304, 257.1684542875737, 270.08114465723435, 282.993835026895, 295.90652539655565, 308.8192157662163, 321.73190613587695], "correct_counts": [10, 31, 108, 203, 266, 281, 168, 89, 28, 12, 10, 3, 1, 2, 2, 0, 0, 1, 0, 0], "incorrect_counts": [1, 7, 23, 29, 45, 36, 31, 27, 17, 10, 15, 12, 4, 11, 5, 4, 0, 4, 1, 3]}, "dynamics/residual": {"edges": [1.1409447093804677, 1.1559197435776392, 1.1708947777748107, 1.1858698119719822, 1.2008448461691539, 1.2158198803663254, 1.2307949145634969, 1.2457699487606684, 1.2607449829578399, 1.2757200171550114, 1.2906950513521829, 1.3056700855493546, 1.320645119746526, 1.3356201539436976, 1.350595188140869, 1.3655702223380406, 1.380545256535212, 1.3955202907323838, 1.4104953249295553, 1.4254703591267268, 1.4404453933238983], "correct_counts": [2, 4, 6, 19, 33, 43, 85, 102, 112, 156, 148, 145, 122, 97, 61, 43, 24, 7, 4, 2], "incorrect_counts": [0, 2, 2, 1, 11, 8, 21, 34, 36, 27, 34, 40, 24, 26, 10, 6, 3, 0, 0, 0]}}}, "feature_ablation": {"full": 0.0711432604851103, "drop_basin": 0.11290811853904308, "basin_only": 0.13271270080190306, "drop_energy": 0.08423482839718313, "energy_only": 0.14821380110222312, "drop_curv": 0.07388292606950923, "curv_only": 0.14046656429951282, "drop_dynamics": 0.07747284186509383, "dynamics_only": 0.15568347575014724}, "ece_geometry": 0.03153101157848922, "accuracy_ood": 0.20533333333333334, "aurc_ood": {"geometry": 0.8035703588171388, "rho_basin": 0.7533412924431383, "energy_min": 0.7509755512657853, "energy_mean": 0.740760170602411, "energy_std": 0.7992759846101417, "msp": 0.7106997166086145, "temp_msp": 0.709349201412843, "entropy": 0.7172185327581629, "softmax_learned": 0.7078058975428306, "geom_softmax": 0.7203531164404758}, "ood_ltt": {"alpha": 0.1, "lambda_hat": 0.19374220945381396, "selective_risk": 0.7619532044760936, "coverage": 0.6553333333333333, "risk_within_budget": false}, "ood_validity": {"target": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "id_risk": [0.0, 0.0, 0.075, 0.1295577967416602, 0.17525773195876287, 0.1879598662207358], "ood_risk": [0.0, 0.0, 0.7619532044760936, 0.7915726109857035, 0.7924914675767918, 0.7949231796927188], "id_coverage": [0.0, 0.0, 0.6933333333333334, 0.8593333333333333, 0.97, 0.9966666666666667], "ood_coverage": [0.0, 0.0, 0.6553333333333333, 0.886, 0.9766666666666667, 0.998]}}, "env": {"python": "3.12.13", "platform": "macOS-26.5.2-arm64-arm-64bit", "jax": "0.11.0", "jax_backend": "cpu"}, "artifact_paths": []} +{"run_id": "2165fc5b4ecc", "timestamp": "2026-07-30T11:17:57.810793+00:00", "git_sha": "67dc7a2", "config_hash": "c5e806bfe9f7", "config": {"run": {"seed": 0, "task": "graph_planning", "notes": "E2 graph selective-prediction"}, "model": {"latent_dim": 32, "hidden_dim": 128, "context_dim": 64}, "inference": {"k_restarts": 12, "steps": 50, "step_size": 0.1, "temperature": 0.05, "init_scale": 1.0, "grad_tol": 0.001, "sampler": "langevin", "anneal_levels": 10, "anneal_steps_per_level": 8, "anneal_step_max": 0.2, "anneal_step_min": 0.01}, "train": {"epochs": 25, "batch_size": 128, "lr": 0.001, "n_train": 8000, "n_neg": 4, "neg_noise": 0.5, "objective": "basin_center", "ired_noise_min": 0.1, "ired_noise_max": 1.5, "ired_ridge": 0.05, "ired_decode_weight": 1.0, "ired_stat_weight": 1.0}, "eval": {"n_eval": 1500, "richer_geometry": false}, "conformal": {"alpha": 0.1, "delta": 0.05, "n_calib": 1500}, "task": {"graph_planning": {"n_nodes": 7, "ood_n_nodes": 10, "edge_prob": 0.4, "max_len": 4}}}, "task": "graph_planning", "split": "selective", "seed": 0, "metrics": {"n_fit": 1500, "n_calib": 1500, "n_test": 1500, "k_restarts": 12, "objective": "basin_center", "sampler": "langevin", "feature_set": "base", "accuracy_id": 0.7153333333333334, "base_error": 0.2846666666666667, "final_train_loss": 0.227946937084198, "feature_names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual"], "aurc": {"geometry": 0.14828452971658584, "rho_basin": 0.22590771079860786, "energy_min": 0.37002240610013937, "energy_mean": 0.38689280922170594, "energy_std": 0.2757969578417908, "msp": 0.11876501845457642, "temp_msp": 0.11800947069600645, "entropy": 0.11733858388676534, "softmax_learned": 0.118432975750678, "geom_softmax": 0.1213234913882991}, "temperature": 3.0183027222710757, "best_energy_baseline": "energy_std", "best_baseline": "entropy", "delta_aurc_vs_energy_min": [0.22173787638355352, 0.18892196301826353, 0.2546275576199255], "delta_aurc_vs_best_energy": [0.12751242812520497, 0.09889745209899486, 0.15448230768652413], "delta_aurc_vs_best_baseline": [-0.030945945829820506, -0.04266843039923084, -0.020834827103470434], "delta_aurc_geom_adds": [-0.002890515637621105, -0.006592549629494021, 0.0005597097649683881], "geometry_wins": true, "geometry_wins_vs_baseline": false, "geometry_adds_over_softmax": false, "risk_coverage": {"geometry": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.06666666666666667, 0.05, 0.044444444444444446, 0.05, 0.04666666666666667, 0.06666666666666678, 0.07142857142857144, 0.07083333333333333, 0.08518518518518518, 0.08666666666666668, 0.07878787878787878, 0.08888888888888889, 0.10256410256410256, 0.10952380952380952, 0.11111111111111109, 0.11041666666666666, 0.11568627450980393, 0.12037037037037035, 0.1280701754385965, 0.13, 0.13015873015873014, 0.13484848484848486, 0.13478260869565217, 0.138888888888889, 0.14533333333333334, 0.14358974358974358, 0.1469135802469136, 0.14642857142857144, 0.14827586206896548, 0.15333333333333332, 0.15591397849462366, 0.1625, 0.1717171717171717, 0.1784313725490196, 0.18952380952380948, 0.1916666666666668, 0.1963963963963964, 0.2, 0.20598290598290597, 0.21166666666666667, 0.2203252032520325, 0.23015873015873012, 0.2364341085271318, 0.24166666666666667, 0.2511111111111111, 0.2565217391304348, 0.2645390070921985, 0.27083333333333326, 0.27755102040816326, 0.2846666666666667]}, "rho_basin": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.21985233798195233, 0.21985233798195258, 0.2198523379819524, 0.2198523379819521, 0.21985233798195206, 0.21985233798195253, 0.21985233798195283, 0.21985233798195308, 0.21985233798195328, 0.21985233798195325, 0.21985233798195275, 0.2198523379819523, 0.21985233798195197, 0.21985233798195167, 0.2198523379819514, 0.21985233798195117, 0.21985233798195095, 0.21985233798195078, 0.2198523379819506, 0.21985233798195086, 0.2198523379819514, 0.21985233798195186, 0.2198523379819523, 0.21985233798195272, 0.21985233798195308, 0.21985233798195342, 0.21985233798195372, 0.21985233798195403, 0.2198523379819543, 0.21985233798195455, 0.21985233798195478, 0.219852337981955, 0.21985233798195522, 0.21985233798195541, 0.2198523379819556, 0.21985233798195578, 0.21985233798195594, 0.2198523379819561, 0.21985233798195614, 0.21985233798195555, 0.2225706542779741, 0.22974300831443945, 0.23658176448874357, 0.24213688610240544, 0.2467177522349954, 0.25484057971014695, 0.26473758865248487, 0.27293836805555816, 0.2785034013605472, 0.2846666666666696]}, "energy_min": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.4, 0.48333333333333334, 0.4666666666666667, 0.48333333333333334, 0.47333333333333333, 0.44999999999999996, 0.44285714285714295, 0.42916666666666664, 0.4148148148148148, 0.4133333333333334, 0.4, 0.4027777777777778, 0.3974358974358974, 0.3952380952380952, 0.3955555555555555, 0.3875, 0.3862745098039216, 0.38888888888888884, 0.38070175438596493, 0.38, 0.3714285714285714, 0.3696969696969697, 0.36666666666666664, 0.3666666666666666, 0.364, 0.3628205128205128, 0.3567901234567901, 0.3595238095238095, 0.36436781609195396, 0.3611111111111111, 0.3548387096774194, 0.346875, 0.3434343434343434, 0.3392156862745098, 0.33619047619047615, 0.3324074074074074, 0.3279279279279279, 0.32456140350877194, 0.3213675213675214, 0.32083333333333336, 0.3186991869918699, 0.3142857142857144, 0.31085271317829455, 0.30757575757575756, 0.30444444444444446, 0.29927536231884055, 0.2964539007092198, 0.2930555555555555, 0.2891156462585034, 0.2846666666666667]}, "energy_mean": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.5, 0.5333333333333333, 0.5333333333333333, 0.49166666666666664, 0.4866666666666667, 0.48888888888888893, 0.49047619047619034, 0.4625, 0.4444444444444444, 0.4366666666666667, 0.43636363636363634, 0.425, 0.41794871794871796, 0.4023809523809524, 0.4066666666666668, 0.3958333333333333, 0.4, 0.3981481481481481, 0.3929824561403509, 0.39166666666666666, 0.38730158730158726, 0.39090909090909093, 0.3927536231884058, 0.38888888888888884, 0.38666666666666666, 0.37948717948717947, 0.37777777777777777, 0.3726190476190476, 0.36896551724137927, 0.3622222222222222, 0.35698924731182796, 0.35104166666666664, 0.3434343434343434, 0.34019607843137256, 0.33809523809523806, 0.3342592592592592, 0.3324324324324324, 0.3307017543859649, 0.32564102564102565, 0.3225, 0.3227642276422765, 0.3166666666666666, 0.31317829457364343, 0.3090909090909091, 0.30592592592592593, 0.30144927536231886, 0.29574468085106376, 0.2909722222222222, 0.28775510204081634, 0.2846666666666667]}, "energy_std": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.26666666666666666, 0.26666666666666666, 0.26666666666666666, 0.25833333333333336, 0.28, 0.27222222222222214, 0.2666666666666667, 0.2791666666666667, 0.2777777777777778, 0.2733333333333332, 0.2818181818181818, 0.2833333333333333, 0.28205128205128205, 0.28095238095238095, 0.2888888888888888, 0.28125, 0.27450980392156865, 0.274074074074074, 0.2736842105263158, 0.28, 0.2793650793650793, 0.2787878787878788, 0.27391304347826084, 0.26944444444444454, 0.268, 0.2717948717948718, 0.2654320987654321, 0.26666666666666666, 0.2701149425287358, 0.27, 0.26881720430107525, 0.26979166666666665, 0.2696969696969697, 0.2735294117647059, 0.27238095238095233, 0.2703703703703703, 0.27297297297297296, 0.27631578947368424, 0.27863247863247864, 0.2775, 0.2796747967479674, 0.2793650793650793, 0.2813953488372093, 0.2825757575757576, 0.2837037037037037, 0.28405797101449276, 0.2843971631205675, 0.28472222222222227, 0.282312925170068, 0.2846666666666667]}, "msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.011111111111111112, 0.016666666666666666, 0.02666666666666667, 0.02222222222222222, 0.02380952380952381, 0.020833333333333332, 0.02962962962962963, 0.030000000000000002, 0.03636363636363636, 0.03611111111111111, 0.046153846153846156, 0.047619047619047616, 0.05333333333333333, 0.05625, 0.056862745098039215, 0.0574074074074074, 0.06315789473684211, 0.06666666666666667, 0.07777777777777777, 0.08787878787878788, 0.08985507246376812, 0.09722222222222221, 0.108, 0.11794871794871795, 0.1271604938271605, 0.13214285714285715, 0.1333333333333333, 0.14333333333333334, 0.15053763440860216, 0.16041666666666668, 0.16363636363636364, 0.1676470588235294, 0.17238095238095236, 0.17777777777777792, 0.1837837837837838, 0.193859649122807, 0.20256410256410257, 0.20916666666666667, 0.21300813008130093, 0.22222222222222232, 0.2310077519379845, 0.23712121212121212, 0.24814814814814815, 0.25507246376811593, 0.2645390070921985, 0.26736111111111105, 0.2755102040816326, 0.2846666666666667]}, "temp_msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.011111111111111112, 0.025, 0.02666666666666667, 0.02222222222222222, 0.02380952380952381, 0.029166666666666667, 0.03333333333333333, 0.030000000000000002, 0.030303030303030304, 0.041666666666666664, 0.046153846153846156, 0.05476190476190476, 0.055555555555555546, 0.05625, 0.054901960784313725, 0.059259259259259255, 0.06666666666666667, 0.07333333333333333, 0.08253968253968266, 0.0893939393939394, 0.09420289855072464, 0.09999999999999999, 0.10666666666666667, 0.11025641025641025, 0.12345679012345678, 0.12857142857142856, 0.13448275862068962, 0.1411111111111111, 0.14516129032258066, 0.14895833333333333, 0.15858585858585858, 0.16568627450980392, 0.1714285714285714, 0.1759259259259259, 0.1837837837837838, 0.19210526315789472, 0.19914529914529913, 0.20833333333333334, 0.21463414634146352, 0.21746031746031758, 0.2255813953488372, 0.23257575757575757, 0.23925925925925925, 0.24782608695652175, 0.25673758865248236, 0.26597222222222233, 0.2761904761904762, 0.2846666666666667]}, "entropy": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.011111111111111112, 0.016666666666666666, 0.02666666666666667, 0.02222222222222222, 0.02380952380952381, 0.020833333333333332, 0.02962962962962963, 0.030000000000000002, 0.03636363636363636, 0.03611111111111111, 0.046153846153846156, 0.047619047619047616, 0.05333333333333333, 0.05625, 0.056862745098039215, 0.055555555555555546, 0.06315789473684211, 0.06833333333333333, 0.07777777777777777, 0.08787878787878788, 0.08840579710144927, 0.09583333333333333, 0.10666666666666667, 0.11666666666666667, 0.1259259259259259, 0.13095238095238096, 0.13563218390804596, 0.14, 0.14838709677419354, 0.159375, 0.1606060606060606, 0.16470588235294117, 0.17047619047619045, 0.17777777777777776, 0.1864864864864865, 0.19210526315789472, 0.19658119658119658, 0.2025, 0.20813008130081298, 0.21825396825396837, 0.22325581395348837, 0.23106060606060605, 0.23703703703703705, 0.24710144927536232, 0.2574468085106382, 0.2666666666666668, 0.2761904761904762, 0.2846666666666667]}, "softmax_learned": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.011111111111111112, 0.025, 0.02666666666666667, 0.02222222222222222, 0.02380952380952381, 0.029166666666666667, 0.03333333333333333, 0.030000000000000002, 0.030303030303030304, 0.041666666666666664, 0.046153846153846156, 0.05476190476190476, 0.055555555555555546, 0.05625, 0.054901960784313725, 0.059259259259259255, 0.06666666666666667, 0.075, 0.08095238095238108, 0.09090909090909091, 0.09710144927536232, 0.1000000000000001, 0.10666666666666667, 0.11025641025641025, 0.12222222222222222, 0.12738095238095237, 0.13218390804597718, 0.1388888888888889, 0.14408602150537633, 0.15104166666666666, 0.15757575757575756, 0.16470588235294117, 0.1714285714285714, 0.1759259259259259, 0.18288288288288287, 0.18859649122807018, 0.20085470085470086, 0.21, 0.2138211382113822, 0.21825396825396837, 0.22790697674418606, 0.2356060606060606, 0.24222222222222223, 0.2514492753623188, 0.2595744680851065, 0.2666666666666668, 0.2761904761904762, 0.2846666666666667]}, "geom_softmax": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.03333333333333333, 0.03333333333333333, 0.025, 0.02666666666666667, 0.03888888888888888, 0.03333333333333334, 0.03333333333333333, 0.03333333333333333, 0.036666666666666674, 0.04242424242424243, 0.044444444444444446, 0.04871794871794872, 0.05238095238095238, 0.048888888888888885, 0.04791666666666667, 0.052941176470588235, 0.06296296296296294, 0.06666666666666667, 0.07833333333333334, 0.08412698412698412, 0.08636363636363636, 0.09565217391304348, 0.10138888888888888, 0.10666666666666667, 0.11153846153846154, 0.12098765432098765, 0.1261904761904762, 0.13103448275862067, 0.14, 0.14623655913978495, 0.15208333333333332, 0.15858585858585858, 0.16176470588235295, 0.16952380952380966, 0.17499999999999996, 0.18468468468468469, 0.19210526315789472, 0.19743589743589743, 0.2075, 0.21626016260162598, 0.2214285714285714, 0.2317829457364341, 0.23939393939393938, 0.25037037037037035, 0.25579710144927537, 0.2652482269503547, 0.27083333333333326, 0.27687074829931974, 0.2846666666666667]}}, "ltt": {"alpha": 0.1, "delta": 0.05, "lambda_hat": null, "coverage": 0.0, "abstain_rate": 1.0, "selective_risk": 0.0, "selective_accuracy": null, "risk_within_budget": true}, "coverage_validity": {"alpha_grid": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "target": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "achieved_risk": [0.0, 0.0, 0.0, 0.07346938775510205, 0.15090090090090091, 0.2565217391304348], "coverage": [0.0, 0.0, 0.0, 0.16333333333333333, 0.592, 0.92]}, "feature_diagnostics": {"names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual"], "auroc": {"basin/rho": 0.6313428828974335, "basin/entropy": 0.36828280270903224, "basin/dispersion": 0.4898869636009263, "energy/mean": 0.656499429252397, "energy/min": 0.6382769751904857, "energy/std": 0.4808925052000236, "curv/lmax_mean": 0.32037165163225084, "curv/lmax_best": 0.37439951459171356, "curv/trace_mean": 0.31464016709918347, "curv/trace_best": 0.34222484618188403, "dynamics/steps": 0.5, "dynamics/monotonic": 0.6361162098867017, "dynamics/drop": 0.6684338380211755, "dynamics/residual": 0.47719956086264737}, "hist": {"basin/rho": {"edges": [0.3333333333333333, 0.36666666666666664, 0.4, 0.43333333333333335, 0.4666666666666667, 0.5, 0.5333333333333333, 0.5666666666666667, 0.6000000000000001, 0.6333333333333333, 0.6666666666666667, 0.7000000000000001, 0.7333333333333334, 0.7666666666666667, 0.8, 0.8333333333333335, 0.8666666666666667, 0.9000000000000001, 0.9333333333333333, 0.9666666666666668, 1.0], "correct_counts": [0, 0, 2, 0, 0, 9, 0, 14, 0, 11, 0, 0, 14, 0, 32, 0, 0, 40, 0, 951], "incorrect_counts": [1, 0, 0, 0, 0, 15, 0, 16, 0, 21, 0, 0, 36, 0, 26, 0, 0, 44, 0, 268]}, "basin/entropy": {"edges": [0.0, 0.06648306744273791, 0.13296613488547582, 0.19944920232821373, 0.26593226977095163, 0.33241533721368954, 0.39889840465642745, 0.46538147209916536, 0.5318645395419033, 0.5983476069846412, 0.6648306744273791, 0.731313741870117, 0.7977968093128549, 0.8642798767555928, 0.9307629441983307, 0.9972460116410686, 1.0637290790838065, 1.1302121465265444, 1.1966952139692824, 1.2631782814120203, 1.3296613488547582], "correct_counts": [951, 0, 0, 0, 40, 0, 31, 0, 14, 11, 23, 0, 0, 0, 1, 1, 1, 0, 0, 0], "incorrect_counts": [268, 0, 0, 0, 44, 0, 25, 0, 34, 17, 28, 0, 3, 5, 1, 1, 0, 0, 0, 1]}, "basin/dispersion": {"edges": [1.6005140780864282, 1.6224848783774093, 1.6444556786683906, 1.6664264789593717, 1.688397279250353, 1.710368079541334, 1.7323388798323154, 1.7543096801232965, 1.7762804804142778, 1.7982512807052589, 1.8202220809962402, 1.8421928812872213, 1.8641636815782023, 1.8861344818691836, 1.9081052821601647, 1.930076082451146, 1.9520468827421271, 1.9740176830331084, 1.9959884833240895, 2.017959283615071, 2.039930083906052], "correct_counts": [0, 4, 7, 11, 23, 46, 61, 84, 113, 167, 135, 114, 105, 69, 50, 41, 20, 17, 4, 2], "incorrect_counts": [2, 0, 5, 3, 8, 14, 28, 30, 47, 54, 54, 56, 39, 35, 29, 16, 3, 2, 2, 0]}, "energy/mean": {"edges": [0.5080794406433901, 0.5479599892472228, 0.5878405378510555, 0.6277210864548882, 0.6676016350587209, 0.7074821836625537, 0.7473627322663863, 0.787243280870219, 0.8271238294740517, 0.8670043780778844, 0.9068849266817172, 0.9467654752855499, 0.9866460238893826, 1.0265265724932153, 1.066407121097048, 1.1062876697008808, 1.1461682183047135, 1.1860487669085462, 1.225929315512379, 1.2658098641162117, 1.3056904127200444], "correct_counts": [2, 1, 2, 6, 18, 39, 57, 97, 121, 125, 130, 149, 116, 82, 52, 39, 17, 14, 4, 2], "incorrect_counts": [1, 1, 5, 6, 20, 34, 44, 48, 64, 69, 40, 40, 27, 15, 3, 2, 4, 3, 1, 0]}, "energy/min": {"edges": [0.20125961303710938, 0.2415154755115509, 0.2817713379859924, 0.322027200460434, 0.3622830629348755, 0.402538925409317, 0.4427947878837586, 0.4830506503582001, 0.5233065128326416, 0.5635623753070831, 0.6038182377815247, 0.6440741002559662, 0.6843299627304078, 0.7245858252048493, 0.7648416876792908, 0.8050975501537323, 0.8453534126281739, 0.8856092751026154, 0.9258651375770569, 0.9661210000514985, 1.00637686252594], "correct_counts": [2, 3, 12, 10, 23, 44, 81, 91, 107, 115, 120, 113, 109, 99, 49, 39, 25, 18, 8, 5], "incorrect_counts": [2, 1, 8, 14, 23, 30, 46, 53, 52, 51, 50, 29, 33, 17, 9, 4, 3, 2, 0, 0]}, "energy/std": {"edges": [0.06508645292670491, 0.08258405015316497, 0.10008164737962504, 0.1175792446060851, 0.13507684183254515, 0.15257443905900522, 0.1700720362854653, 0.18756963351192535, 0.2050672307383854, 0.22256482796484545, 0.24006242519130555, 0.2575600224177656, 0.27505761964422565, 0.2925552168706857, 0.3100528140971458, 0.32755041132360585, 0.3450480085500659, 0.36254560577652595, 0.380043203002986, 0.3975408002294461, 0.41503839745590615], "correct_counts": [3, 8, 30, 47, 85, 138, 177, 162, 170, 93, 57, 54, 33, 6, 6, 3, 0, 0, 0, 1], "incorrect_counts": [1, 3, 11, 15, 35, 59, 64, 53, 69, 43, 33, 21, 8, 4, 2, 4, 0, 1, 0, 1]}, "curv/lmax_mean": {"edges": [0.9988154868284861, 0.9990696673591931, 0.9993238478899001, 0.9995780284206072, 0.9998322089513143, 1.0000863894820213, 1.0003405700127284, 1.0005947505434354, 1.0008489310741424, 1.0011031116048494, 1.0013572921355565, 1.0016114726662635, 1.0018656531969705, 1.0021198337276775, 1.0023740142583848, 1.0026281947890918, 1.0028823753197988, 1.0031365558505059, 1.003390736381213, 1.00364491691192, 1.003899097442627], "correct_counts": [4, 18, 97, 334, 405, 123, 49, 25, 8, 4, 2, 2, 1, 0, 0, 1, 0, 0, 0, 0], "incorrect_counts": [1, 2, 18, 62, 138, 113, 51, 19, 11, 5, 4, 1, 1, 0, 0, 0, 0, 0, 0, 1]}, "curv/lmax_best": {"edges": [0.9955618977546692, 0.99611676633358, 0.9966716349124909, 0.9972265034914016, 0.9977813720703125, 0.9983362406492233, 0.9988911092281342, 0.999445977807045, 1.0000008463859558, 1.0005557149648667, 1.0011105835437775, 1.0016654521226882, 1.0022203207015992, 1.00277518928051, 1.0033300578594209, 1.0038849264383316, 1.0044397950172423, 1.0049946635961533, 1.005549532175064, 1.006104400753975, 1.0066592693328857], "correct_counts": [1, 0, 2, 5, 11, 19, 79, 651, 258, 21, 14, 8, 3, 0, 1, 0, 0, 0, 0, 0], "incorrect_counts": [1, 0, 1, 0, 1, 4, 19, 183, 173, 23, 12, 6, 0, 1, 0, 0, 0, 2, 0, 1]}, "curv/trace_mean": {"edges": [31.957900683085125, 31.961792182922366, 31.965683682759604, 31.969575182596845, 31.973466682434083, 31.977358182271324, 31.98124968210856, 31.985141181945803, 31.98903268178304, 31.992924181620282, 31.99681568145752, 32.000707181294764, 32.004598681132, 32.00849018096924, 32.01238168080648, 32.01627318064372, 32.02016468048096, 32.0240561803182, 32.027947680155435, 32.03183917999268, 32.03573067982992], "correct_counts": [3, 3, 7, 25, 38, 49, 99, 128, 127, 145, 164, 120, 66, 46, 20, 12, 9, 6, 4, 2], "incorrect_counts": [0, 1, 0, 0, 1, 4, 16, 25, 45, 51, 54, 72, 61, 46, 25, 16, 9, 0, 1, 0]}, "curv/trace_best": {"edges": [31.9183292388916, 31.92680730819702, 31.935285377502442, 31.94376344680786, 31.952241516113283, 31.9607195854187, 31.96919765472412, 31.97767572402954, 31.98615379333496, 31.994631862640382, 32.0031099319458, 32.01158800125122, 32.02006607055664, 32.02854413986206, 32.03702220916748, 32.0455002784729, 32.05397834777832, 32.06245641708374, 32.07093448638916, 32.07941255569458, 32.087890625], "correct_counts": [1, 0, 3, 4, 11, 21, 52, 128, 281, 358, 140, 45, 19, 4, 2, 2, 1, 0, 0, 1], "incorrect_counts": [0, 0, 0, 0, 1, 4, 5, 19, 77, 159, 98, 42, 15, 6, 0, 0, 1, 0, 0, 0]}, "dynamics/steps": {"edges": [1.0, 1.05, 1.1, 1.15, 1.2, 1.25, 1.3, 1.35, 1.4, 1.45, 1.5, 1.55, 1.6, 1.65, 1.7000000000000002, 1.75, 1.8, 1.85, 1.9, 1.9500000000000002, 2.0], "correct_counts": [1073, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [427, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "dynamics/monotonic": {"edges": [0.6383333333333333, 0.645, 0.6516666666666666, 0.6583333333333333, 0.6649999999999999, 0.6716666666666666, 0.6783333333333333, 0.6849999999999999, 0.6916666666666667, 0.6983333333333333, 0.705, 0.7116666666666667, 0.7183333333333333, 0.725, 0.7316666666666666, 0.7383333333333333, 0.745, 0.7516666666666666, 0.7583333333333333, 0.7649999999999999, 0.7716666666666666], "correct_counts": [0, 2, 6, 12, 52, 94, 96, 161, 117, 142, 138, 65, 79, 37, 37, 21, 7, 5, 1, 1], "incorrect_counts": [1, 3, 8, 16, 36, 47, 54, 77, 64, 43, 40, 18, 11, 3, 3, 1, 1, 1, 0, 0]}, "dynamics/drop": {"edges": [13.273132185141245, 17.449723252654074, 21.626314320166905, 25.802905387679733, 29.97949645519256, 34.15608752270539, 38.33267859021822, 42.509269657731046, 46.68586072524388, 50.86245179275671, 55.03904286026954, 59.21563392778236, 63.392224995295194, 67.56881606280803, 71.74540713032086, 75.92199819783369, 80.09858926534652, 84.27518033285935, 88.45177140037218, 92.62836246788501, 96.80495353539784], "correct_counts": [35, 165, 236, 186, 132, 96, 75, 48, 29, 25, 19, 5, 10, 1, 3, 4, 2, 1, 0, 1], "incorrect_counts": [52, 120, 93, 70, 48, 20, 15, 1, 1, 2, 2, 2, 0, 0, 0, 1, 0, 0, 0, 0]}, "dynamics/residual": {"edges": [1.1351244548956554, 1.1498689969380698, 1.1646135389804841, 1.1793580810228985, 1.1941026230653127, 1.208847165107727, 1.2235917071501414, 1.2383362491925558, 1.2530807912349702, 1.2678253332773846, 1.282569875319799, 1.297314417362213, 1.3120589594046275, 1.3268035014470418, 1.3415480434894562, 1.3562925855318706, 1.371037127574285, 1.3857816696166991, 1.4005262116591135, 1.4152707537015279, 1.4300152957439423], "correct_counts": [2, 0, 5, 14, 20, 39, 68, 81, 95, 120, 140, 113, 122, 101, 58, 54, 17, 16, 4, 4], "incorrect_counts": [2, 1, 0, 5, 8, 16, 17, 34, 38, 34, 59, 59, 43, 41, 32, 16, 14, 6, 1, 1]}}}, "feature_ablation": {"full": 0.14828452971658584, "drop_basin": 0.1651095026755807, "basin_only": 0.21983508381035266, "drop_energy": 0.1476320101970345, "energy_only": 0.18812416405697532, "drop_curv": 0.1481249587041615, "curv_only": 0.17486262874155167, "drop_dynamics": 0.1564704847843828, "dynamics_only": 0.17978757695863193}, "ece_geometry": 0.048452947907751065, "accuracy_ood": 0.3546666666666667, "aurc_ood": {"geometry": 0.634405040392116, "rho_basin": 0.6277002900945788, "energy_min": 0.7263364569173539, "energy_mean": 0.728125986539371, "energy_std": 0.6284172651454782, "msp": 0.605943901894947, "temp_msp": 0.5966869880651136, "entropy": 0.6037165831832931, "softmax_learned": 0.5966414374218213, "geom_softmax": 0.6175133507470368}, "ood_ltt": {"alpha": 0.1, "lambda_hat": null, "selective_risk": 0.0, "coverage": 0.0, "risk_within_budget": true}, "ood_validity": {"target": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "id_risk": [0.0, 0.0, 0.0, 0.07346938775510205, 0.15090090090090091, 0.2565217391304348], "ood_risk": [0.0, 0.0, 0.0, 0.6439169139465876, 0.6160804020100502, 0.635846372688478], "id_coverage": [0.0, 0.0, 0.0, 0.16333333333333333, 0.592, 0.92], "ood_coverage": [0.0, 0.0, 0.0, 0.22466666666666665, 0.6633333333333333, 0.9373333333333334]}}, "env": {"python": "3.12.13", "platform": "macOS-26.5.2-arm64-arm-64bit", "jax": "0.11.0", "jax_backend": "cpu"}, "artifact_paths": []} +{"run_id": "20d1395ffa7a", "timestamp": "2026-07-30T11:17:58.833773+00:00", "git_sha": "67dc7a2", "config_hash": "1c0de51a1390", "config": {"run": {"seed": 1, "task": "graph_planning", "notes": "graph adaptive-halting run"}, "model": {"latent_dim": 32, "hidden_dim": 128, "context_dim": 64}, "inference": {"k_restarts": 12, "steps": 50, "step_size": 0.1, "temperature": 0.05, "init_scale": 1.0, "grad_tol": 0.001, "sampler": "langevin", "anneal_levels": 10, "anneal_steps_per_level": 8, "anneal_step_max": 0.2, "anneal_step_min": 0.01}, "train": {"epochs": 25, "batch_size": 128, "lr": 0.001, "n_train": 8000, "n_neg": 4, "neg_noise": 0.5, "objective": "basin_center", "ired_noise_min": 0.1, "ired_noise_max": 1.5, "ired_ridge": 0.05, "ired_decode_weight": 1.0, "ired_stat_weight": 1.0}, "eval": {"n_eval": 800, "richer_geometry": false}, "conformal": {"alpha": 0.1, "delta": 0.05, "n_calib": 800}, "task": {"graph_planning": {"n_nodes": 7, "ood_n_nodes": 10, "edge_prob": 0.4, "max_len": 4}}}, "task": "graph_planning", "split": "halting", "seed": 1, "metrics": {"k_restarts": 12, "steps": 50, "alpha": 0.1, "n_calib": 800, "n_test": 800, "final_train_loss": 0.2080264687538147, "tau_hat": 0.9167, "compute_used": 0.29487500000000005, "compute_saved": 0.705125, "halting_risk": 0.0025, "risk_within_budget": true, "full_accuracy": 0.68875, "halted_accuracy": 0.68875, "accuracy_drop": 0.0, "risk_monotone_in_tau": true, "calib_risk_by_tau": [0.47875, 0.47875, 0.47875, 0.4775, 0.4025, 0.4025, 0.335, 0.1975, 0.1975, 0.115, 0.01, 0.01], "calib_compute_by_tau": [0.0, 0.0, 0.0, 0.00027499999999999996, 0.016949999999999875, 0.016949999999999875, 0.039174999999999655, 0.11260000000000017, 0.11260000000000017, 0.16875000000000026, 0.31139999999999984, 0.31139999999999984], "tau_sweep": {"taus": [0.0833, 0.1667, 0.25, 0.3333, 0.4167, 0.5, 0.5833, 0.6667, 0.75, 0.8333, 0.9167, 1.0], "compute_used": [0.0, 0.0, 0.0, 0.000275, 0.017725, 0.017725, 0.03785, 0.11257500000000001, 0.11257500000000001, 0.165775, 0.29487500000000005, 0.29487500000000005], "accuracy": [0.47625, 0.47625, 0.47625, 0.47625, 0.5175, 0.5175, 0.56875, 0.64125, 0.64125, 0.66625, 0.68875, 0.68875], "disagreement": [0.47125, 0.47125, 0.47125, 0.47125, 0.405, 0.405, 0.33125, 0.1675, 0.1675, 0.08125, 0.0025, 0.0025]}}, "env": {"python": "3.12.13", "platform": "macOS-26.5.2-arm64-arm-64bit", "jax": "0.11.0", "jax_backend": "cpu"}, "artifact_paths": []} +{"run_id": "b04ffc85ccc3", "timestamp": "2026-07-30T11:17:59.455068+00:00", "git_sha": "67dc7a2", "config_hash": "c8e9150407ea", "config": {"run": {"seed": 1, "task": "arithmetic", "notes": "E1 selective-prediction main run"}, "model": {"latent_dim": 32, "hidden_dim": 128, "context_dim": 64}, "inference": {"k_restarts": 12, "steps": 50, "step_size": 0.1, "temperature": 0.05, "init_scale": 1.0, "grad_tol": 0.001, "sampler": "langevin", "anneal_levels": 10, "anneal_steps_per_level": 8, "anneal_step_max": 0.2, "anneal_step_min": 0.01}, "train": {"epochs": 7, "batch_size": 128, "lr": 0.001, "n_train": 6000, "n_neg": 4, "neg_noise": 0.5, "objective": "basin_center", "ired_noise_min": 0.1, "ired_noise_max": 1.5, "ired_ridge": 0.05, "ired_decode_weight": 1.0, "ired_stat_weight": 1.0}, "eval": {"n_eval": 1500, "richer_geometry": false}, "conformal": {"alpha": 0.1, "delta": 0.05, "n_calib": 1500}, "task": {"arithmetic": {"modular": false, "n_operands": 6, "ood_n_operands": 6, "max_operand": 7, "ood_max_operand": 9}}}, "task": "arithmetic", "split": "selective", "seed": 1, "metrics": {"n_fit": 1500, "n_calib": 1500, "n_test": 1500, "k_restarts": 12, "objective": "basin_center", "sampler": "langevin", "feature_set": "base", "accuracy_id": 0.8093333333333333, "base_error": 0.19066666666666668, "final_train_loss": 0.9732173085212708, "feature_names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual"], "aurc": {"geometry": 0.07534184983249899, "rho_basin": 0.10651266796902968, "energy_min": 0.19613627208106035, "energy_mean": 0.18655715870175713, "energy_std": 0.1850079661273594, "msp": 0.056516158696690806, "temp_msp": 0.06264464287256487, "entropy": 0.07361179491321607, "softmax_learned": 0.05595378790404033, "geom_softmax": 0.0522583408533078}, "temperature": 0.3758208374823171, "best_energy_baseline": "energy_std", "best_baseline": "msp", "delta_aurc_vs_energy_min": [0.12079442224856136, 0.09713330903618891, 0.14596535431672475], "delta_aurc_vs_best_energy": [0.10966611629486041, 0.08345583515537251, 0.135750761601739], "delta_aurc_vs_best_baseline": [-0.018825691135808183, -0.031267766857238775, -0.007356400682959861], "delta_aurc_geom_adds": [0.003695447050732527, 0.0006793574386630384, 0.006965369236499613], "geometry_wins": true, "geometry_wins_vs_baseline": false, "geometry_adds_over_softmax": true, "risk_coverage": {"geometry": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.06666666666666667, 0.03333333333333333, 0.022222222222222223, 0.03333333333333333, 0.04666666666666667, 0.03888888888888888, 0.038095238095238106, 0.05, 0.05555555555555555, 0.06333333333333335, 0.05757575757575758, 0.05277777777777778, 0.05128205128205128, 0.05, 0.05111111111111111, 0.052083333333333336, 0.052941176470588235, 0.05185185185185184, 0.05263157894736842, 0.051666666666666666, 0.055555555555555546, 0.05454545454545454, 0.05217391304347826, 0.05138888888888888, 0.04933333333333333, 0.05384615384615385, 0.05308641975308642, 0.055952380952380955, 0.0586206896551724, 0.06444444444444444, 0.06774193548387097, 0.06979166666666667, 0.07171717171717172, 0.07450980392156863, 0.07619047619047618, 0.08055555555555555, 0.0891891891891892, 0.09298245614035087, 0.10341880341880341, 0.11, 0.11626016260162614, 0.11984126984126996, 0.1263565891472868, 0.13636363636363635, 0.1474074074074074, 0.15434782608695652, 0.1617021276595746, 0.17013888888888898, 0.18095238095238095, 0.19066666666666668]}, "rho_basin": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.09414024975984632, 0.09414024975984626, 0.09414024975984618, 0.09414024975984615, 0.09414024975984613, 0.09414024975984611, 0.0941402497598461, 0.09414024975984608, 0.09414024975984608, 0.09414024975984607, 0.09414024975984607, 0.09414024975984607, 0.09414024975984607, 0.09414024975984606, 0.09414024975984606, 0.09414024975984606, 0.09414024975984606, 0.09414024975984606, 0.09414024975984606, 0.09414024975984606, 0.09414024975984606, 0.09414024975984606, 0.09414024975984604, 0.09414024975984604, 0.09414024975984604, 0.09414024975984604, 0.09414024975984604, 0.09414024975984604, 0.09414024975984604, 0.09414024975984604, 0.09414024975984604, 0.09414024975984604, 0.09414024975984604, 0.09414024975984604, 0.09550476190476165, 0.09988888888888856, 0.1040360360360356, 0.10796491228070125, 0.11169230769230713, 0.11560307017543801, 0.12016474112109492, 0.12450918964076814, 0.13166112956810566, 0.139383116883116, 0.14806314511232435, 0.15803041102399493, 0.16711125224939005, 0.17513474295190584, 0.18291074005359584, 0.19066666666666512]}, "energy_min": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.26666666666666666, 0.2833333333333333, 0.24444444444444444, 0.20833333333333334, 0.19333333333333333, 0.19444444444444442, 0.20000000000000004, 0.2, 0.1962962962962963, 0.1866666666666667, 0.18787878787878787, 0.2, 0.19743589743589743, 0.19285714285714287, 0.19555555555555554, 0.18958333333333333, 0.19215686274509805, 0.18888888888888886, 0.18596491228070175, 0.18833333333333332, 0.18888888888888886, 0.19545454545454546, 0.19130434782608696, 0.19305555555555554, 0.19333333333333333, 0.191025641025641, 0.18888888888888888, 0.18571428571428572, 0.18275862068965515, 0.18555555555555556, 0.1827956989247312, 0.18125, 0.18585858585858586, 0.18529411764705883, 0.18190476190476204, 0.1851851851851853, 0.18468468468468469, 0.1850877192982456, 0.18888888888888888, 0.18833333333333332, 0.18699186991869915, 0.18968253968253965, 0.18992248062015504, 0.19090909090909092, 0.19037037037037038, 0.19130434782608696, 0.19219858156028366, 0.19166666666666676, 0.19115646258503402, 0.19066666666666668]}, "energy_mean": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.16666666666666666, 0.15, 0.17777777777777778, 0.20833333333333334, 0.18, 0.17777777777777776, 0.17619047619047623, 0.19583333333333333, 0.1962962962962963, 0.20000000000000004, 0.19393939393939394, 0.18333333333333332, 0.1794871794871795, 0.17857142857142858, 0.18000000000000013, 0.17708333333333334, 0.1843137254901961, 0.1833333333333333, 0.18771929824561404, 0.19, 0.18888888888888886, 0.19545454545454546, 0.1927536231884058, 0.19166666666666665, 0.18666666666666668, 0.18461538461538463, 0.18518518518518517, 0.18571428571428572, 0.18965517241379307, 0.18555555555555556, 0.1870967741935484, 0.184375, 0.18282828282828284, 0.18529411764705883, 0.1876190476190476, 0.1898148148148148, 0.19099099099099098, 0.19210526315789472, 0.19145299145299147, 0.19, 0.1926829268292684, 0.19047619047619044, 0.1883720930232558, 0.18712121212121213, 0.18814814814814815, 0.19202898550724637, 0.19007092198581557, 0.19027777777777774, 0.19047619047619047, 0.19066666666666668]}, "energy_std": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.13333333333333333, 0.11666666666666667, 0.12222222222222222, 0.16666666666666666, 0.19333333333333333, 0.19444444444444442, 0.20952380952380958, 0.20833333333333334, 0.1962962962962963, 0.1833333333333332, 0.18484848484848485, 0.18333333333333332, 0.18974358974358974, 0.18095238095238095, 0.1844444444444444, 0.1875, 0.19019607843137254, 0.188888888888889, 0.18771929824561404, 0.185, 0.18730158730158727, 0.18636363636363637, 0.1855072463768116, 0.1847222222222222, 0.18666666666666668, 0.18846153846153846, 0.18888888888888888, 0.19166666666666668, 0.19195402298850572, 0.19111111111111112, 0.1913978494623656, 0.19375, 0.19494949494949496, 0.19411764705882353, 0.1933333333333333, 0.19259259259259257, 0.1900900900900901, 0.1868421052631579, 0.18717948717948718, 0.1875, 0.18780487804878046, 0.188888888888889, 0.19069767441860466, 0.1893939393939394, 0.18888888888888888, 0.1891304347826087, 0.18865248226950362, 0.1895833333333334, 0.19183673469387755, 0.19066666666666668]}, "msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.005555555555555555, 0.009523809523809526, 0.008333333333333333, 0.007407407407407408, 0.010000000000000002, 0.012121212121212121, 0.013888888888888888, 0.015384615384615385, 0.016666666666666666, 0.015555555555555552, 0.01875, 0.01764705882352941, 0.02037037037037037, 0.021052631578947368, 0.021666666666666667, 0.02222222222222222, 0.024242424242424242, 0.02608695652173913, 0.027777777777777773, 0.029333333333333333, 0.03333333333333333, 0.0345679012345679, 0.039285714285714285, 0.041379310344827766, 0.04888888888888889, 0.053763440860215055, 0.057291666666666664, 0.06363636363636363, 0.06862745098039216, 0.07809523809523808, 0.08425925925925924, 0.0936936936936937, 0.10087719298245613, 0.10598290598290598, 0.1125, 0.1195121951219512, 0.12857142857142853, 0.13953488372093023, 0.1462121212121212, 0.15333333333333332, 0.16014492753623188, 0.16595744680851074, 0.17500000000000007, 0.18299319727891156, 0.19066666666666668]}, "temp_msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.016666666666666666, 0.011111111111111112, 0.008333333333333333, 0.006666666666666667, 0.005555555555555555, 0.004761904761904763, 0.008333333333333333, 0.007407407407407408, 0.013333333333333336, 0.012121212121212121, 0.013888888888888888, 0.017948717948717947, 0.01904761904761905, 0.026666666666666665, 0.03125, 0.03529411764705882, 0.03518518518518518, 0.03684210526315789, 0.035, 0.03809523809523809, 0.03939393939393939, 0.042028985507246375, 0.041666666666666775, 0.04533333333333334, 0.047435897435897434, 0.05308641975308642, 0.05357142857142857, 0.06091954022988504, 0.06444444444444444, 0.06451612903225806, 0.065625, 0.07373737373737374, 0.0784313725490196, 0.0857142857142857, 0.0898148148148148, 0.0972972972972973, 0.10087719298245613, 0.10683760683760683, 0.11166666666666666, 0.11707317073170745, 0.12380952380952392, 0.1310077519379845, 0.14015151515151514, 0.14814814814814814, 0.1572463768115942, 0.1652482269503547, 0.1722222222222223, 0.1816326530612245, 0.19066666666666668]}, "entropy": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.005555555555555555, 0.014285714285714289, 0.016666666666666666, 0.014814814814814815, 0.020000000000000004, 0.024242424242424242, 0.027777777777777776, 0.03333333333333333, 0.04047619047619048, 0.039999999999999994, 0.041666666666666664, 0.047058823529411764, 0.04814814814814814, 0.04912280701754386, 0.05333333333333334, 0.053968253968253964, 0.05303030303030303, 0.05217391304347826, 0.056944444444444554, 0.06266666666666666, 0.06538461538461539, 0.06666666666666667, 0.06785714285714285, 0.07586206896551742, 0.08555555555555555, 0.08494623655913978, 0.09166666666666666, 0.09494949494949495, 0.1, 0.10285714285714284, 0.10648148148148147, 0.10990990990990991, 0.11578947368421053, 0.12051282051282051, 0.1275, 0.1333333333333333, 0.13571428571428581, 0.14573643410852713, 0.15, 0.1562962962962963, 0.16304347826086957, 0.16666666666666663, 0.17430555555555552, 0.18231292517006803, 0.19066666666666668]}, "softmax_learned": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.005555555555555555, 0.004761904761904763, 0.008333333333333333, 0.007407407407407408, 0.013333333333333336, 0.012121212121212121, 0.013888888888888888, 0.01282051282051282, 0.016666666666666666, 0.015555555555555552, 0.01875, 0.0196078431372549, 0.02037037037037037, 0.021052631578947368, 0.023333333333333334, 0.02222222222222222, 0.025757575757575757, 0.02608695652173913, 0.027777777777777773, 0.032, 0.03333333333333333, 0.03950617283950617, 0.04404761904761905, 0.045977011494252866, 0.04888888888888889, 0.05161290322580645, 0.058333333333333334, 0.06262626262626263, 0.06862745098039216, 0.07333333333333332, 0.08055555555555555, 0.0891891891891892, 0.09385964912280702, 0.09829059829059829, 0.11, 0.1227642276422764, 0.12936507936507935, 0.13488372093023257, 0.1409090909090909, 0.1511111111111111, 0.15942028985507245, 0.1659574468085106, 0.1736111111111112, 0.18231292517006803, 0.19066666666666668]}, "geom_softmax": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.003703703703703704, 0.003333333333333334, 0.006060606060606061, 0.008333333333333333, 0.010256410256410256, 0.014285714285714285, 0.015555555555555552, 0.016666666666666666, 0.01764705882352941, 0.016666666666666663, 0.01929824561403509, 0.02, 0.02063492063492063, 0.019696969696969695, 0.01884057971014493, 0.024999999999999998, 0.028, 0.029487179487179487, 0.02962962962962963, 0.030952380952380953, 0.03563218390804597, 0.043333333333333335, 0.046236559139784944, 0.052083333333333336, 0.0595959595959596, 0.06470588235294118, 0.06857142857142873, 0.07777777777777777, 0.08558558558558559, 0.09210526315789473, 0.09914529914529914, 0.10666666666666667, 0.11707317073170731, 0.12698412698412695, 0.1325581395348837, 0.13787878787878788, 0.14814814814814814, 0.15579710144927536, 0.16382978723404265, 0.17013888888888887, 0.17959183673469387, 0.19066666666666668]}}, "ltt": {"alpha": 0.1, "delta": 0.05, "lambda_hat": 0.1463238254140143, "coverage": 0.638, "abstain_rate": 0.362, "selective_risk": 0.07001044932079414, "selective_accuracy": 0.9299895506792059, "risk_within_budget": true}, "coverage_validity": {"alpha_grid": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "target": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "achieved_risk": [0.0, 0.0, 0.07001044932079414, 0.11437908496732026, 0.15036496350364964, 0.18904475617902472], "coverage": [0.0, 0.0, 0.638, 0.816, 0.9133333333333333, 0.998]}, "feature_diagnostics": {"names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual"], "auroc": {"basin/rho": 0.7400893998917063, "basin/entropy": 0.25900623264708933, "basin/dispersion": 0.49065390951717147, "energy/mean": 0.4921515881153443, "energy/min": 0.4963422080390779, "energy/std": 0.49589866476192673, "curv/lmax_mean": 0.5486183914931856, "curv/lmax_best": 0.5253496503496503, "curv/trace_mean": 0.5459758528127556, "curv/trace_best": 0.5448453934862502, "dynamics/steps": 0.5, "dynamics/monotonic": 0.42523991659082266, "dynamics/drop": 0.39914862731996176, "dynamics/residual": 0.5023185216760175}, "hist": {"basin/rho": {"edges": [0.3333333333333333, 0.36666666666666664, 0.4, 0.43333333333333335, 0.4666666666666667, 0.5, 0.5333333333333333, 0.5666666666666667, 0.6000000000000001, 0.6333333333333333, 0.6666666666666667, 0.7000000000000001, 0.7333333333333334, 0.7666666666666667, 0.8, 0.8333333333333335, 0.8666666666666667, 0.9000000000000001, 0.9333333333333333, 0.9666666666666668, 1.0], "correct_counts": [1, 0, 0, 0, 0, 14, 0, 30, 0, 24, 0, 0, 37, 0, 53, 0, 0, 112, 0, 943], "incorrect_counts": [0, 0, 1, 0, 0, 19, 0, 37, 0, 37, 0, 0, 33, 0, 23, 0, 0, 38, 0, 98]}, "basin/entropy": {"edges": [0.0, 0.05493061443340548, 0.10986122886681096, 0.16479184330021646, 0.21972245773362192, 0.2746530721670274, 0.3295836866004329, 0.3845143010338384, 0.43944491546724385, 0.4943755299006493, 0.5493061443340548, 0.6042367587674603, 0.6591673732008658, 0.7140979876342712, 0.7690286020676768, 0.8239592165010822, 0.8788898309344877, 0.9338204453678932, 0.9887510598012986, 1.0436816742347041, 1.0986122886681096], "correct_counts": [943, 0, 0, 0, 0, 112, 0, 0, 53, 0, 37, 23, 38, 0, 0, 1, 5, 1, 0, 1], "incorrect_counts": [98, 0, 0, 0, 0, 38, 0, 0, 22, 0, 32, 33, 49, 2, 0, 4, 4, 1, 2, 1]}, "basin/dispersion": {"edges": [1.6209795073409872, 1.6420014502019364, 1.663023393062886, 1.6840453359238352, 1.7050672787847847, 1.726089221645734, 1.7471111645066832, 1.7681331073676327, 1.789155050228582, 1.8101769930895315, 1.8311989359504808, 1.8522208788114303, 1.8732428216723795, 1.8942647645333288, 1.9152867073942783, 1.9363086502552276, 1.9573305931161769, 1.9783525359771263, 1.9993744788380756, 2.020396421699025, 2.0414183645599744], "correct_counts": [2, 8, 14, 32, 47, 65, 98, 126, 152, 145, 133, 112, 97, 76, 42, 37, 11, 9, 6, 2], "incorrect_counts": [1, 2, 1, 6, 13, 15, 17, 35, 31, 30, 45, 28, 18, 18, 10, 6, 4, 3, 2, 1]}, "energy/mean": {"edges": [-0.2875423381725947, -0.26474122231205305, -0.2419401064515114, -0.21913899059096975, -0.19633787473042807, -0.1735367588698864, -0.15073564300934475, -0.1279345271488031, -0.10513341128826142, -0.08233229542771975, -0.0595311795671781, -0.03673006370663645, -0.013928947846094775, 0.008872168014446902, 0.03167328387498852, 0.0544743997355302, 0.07727551559607188, 0.10007663145661355, 0.12287774731715523, 0.14567886317769685, 0.16847997903823853], "correct_counts": [7, 20, 30, 58, 83, 135, 149, 163, 127, 126, 94, 86, 52, 33, 25, 14, 8, 4, 0, 0], "incorrect_counts": [1, 4, 6, 16, 21, 25, 40, 32, 31, 35, 26, 11, 19, 6, 6, 4, 2, 0, 0, 1]}, "energy/min": {"edges": [-0.7031644582748413, -0.6723987430334091, -0.641633027791977, -0.6108673125505447, -0.5801015973091126, -0.5493358820676804, -0.5185701668262481, -0.487804451584816, -0.45703873634338377, -0.4262730211019516, -0.3955073058605194, -0.36474159061908723, -0.33397587537765505, -0.3032101601362229, -0.27244444489479064, -0.24167872965335846, -0.21091301441192628, -0.1801472991704941, -0.14938158392906187, -0.11861586868762974, -0.08785015344619751], "correct_counts": [1, 3, 16, 12, 41, 73, 86, 119, 149, 155, 154, 128, 105, 75, 38, 31, 20, 3, 4, 1], "incorrect_counts": [0, 3, 4, 7, 8, 13, 20, 30, 32, 36, 31, 33, 27, 21, 11, 5, 0, 3, 2, 0]}, "energy/std": {"edges": [0.07042911362104956, 0.08610065979939398, 0.1017722059777384, 0.1174437521560828, 0.1331152983344272, 0.14878684451277163, 0.16445839069111604, 0.18012993686946044, 0.19580148304780487, 0.21147302922614927, 0.2271445754044937, 0.2428161215828381, 0.2584876677611825, 0.2741592139395269, 0.28983076011787134, 0.3055023062962158, 0.3211738524745602, 0.33684539865290464, 0.35251694483124896, 0.3681884910095934, 0.3838600371879378], "correct_counts": [3, 14, 19, 49, 89, 135, 163, 181, 155, 142, 98, 72, 49, 20, 3, 8, 8, 2, 1, 3], "incorrect_counts": [0, 3, 4, 9, 30, 27, 34, 45, 44, 20, 30, 17, 11, 8, 3, 0, 0, 1, 0, 0]}, "curv/lmax_mean": {"edges": [1.0000324149926503, 1.0006146644552547, 1.0011969139178594, 1.0017791633804638, 1.0023614128430685, 1.002943662305673, 1.0035259117682773, 1.004108161230882, 1.0046904106934864, 1.005272660156091, 1.0058549096186955, 1.0064371590813, 1.0070194085439046, 1.007601658006509, 1.0081839074691137, 1.0087661569317181, 1.0093484063943226, 1.0099306558569272, 1.0105129053195316, 1.0110951547821363, 1.0116774042447407], "correct_counts": [561, 240, 130, 87, 59, 46, 26, 18, 12, 9, 4, 7, 8, 1, 2, 2, 0, 0, 1, 1], "incorrect_counts": [154, 60, 27, 18, 11, 6, 4, 1, 1, 1, 2, 1, 0, 0, 0, 0, 0, 0, 0, 0]}, "curv/lmax_best": {"edges": [1.0, 1.0020527243614197, 1.0041054487228394, 1.006158173084259, 1.0082108974456787, 1.0102636218070984, 1.012316346168518, 1.0143690705299377, 1.0164217948913574, 1.018474519252777, 1.0205272436141968, 1.0225799679756165, 1.0246326923370361, 1.0266854166984558, 1.0287381410598755, 1.0307908654212952, 1.0328435897827148, 1.0348963141441345, 1.0369490385055542, 1.0390017628669739, 1.0410544872283936], "correct_counts": [1013, 104, 42, 23, 9, 5, 3, 4, 3, 2, 2, 2, 1, 0, 0, 0, 0, 0, 0, 1], "incorrect_counts": [254, 19, 7, 3, 2, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "curv/trace_mean": {"edges": [32.001986503601074, 32.00630712509155, 32.01062774658203, 32.01494836807251, 32.01926898956299, 32.02358961105347, 32.027910232543945, 32.032230854034424, 32.0365514755249, 32.04087209701538, 32.04519271850586, 32.04951333999634, 32.053833961486816, 32.058154582977295, 32.06247520446777, 32.06679582595825, 32.07111644744873, 32.07543706893921, 32.07975769042969, 32.084078311920166, 32.088398933410645], "correct_counts": [158, 234, 163, 124, 97, 88, 75, 63, 68, 49, 38, 15, 18, 8, 9, 2, 2, 0, 1, 2], "incorrect_counts": [35, 70, 54, 25, 19, 20, 20, 15, 10, 9, 2, 2, 2, 1, 1, 0, 0, 0, 0, 1]}, "curv/trace_best": {"edges": [32.00096893310547, 32.00789184570313, 32.01481475830078, 32.02173767089844, 32.02866058349609, 32.03558349609375, 32.04250640869141, 32.04942932128906, 32.05635223388672, 32.06327514648437, 32.07019805908203, 32.07712097167969, 32.08404388427734, 32.090966796875, 32.09788970947265, 32.10481262207031, 32.11173553466797, 32.118658447265624, 32.12558135986328, 32.132504272460935, 32.139427185058594], "correct_counts": [338, 281, 171, 124, 100, 56, 36, 33, 22, 17, 9, 7, 3, 5, 4, 1, 4, 2, 0, 1], "incorrect_counts": [81, 87, 43, 31, 12, 11, 5, 6, 4, 1, 1, 1, 2, 1, 0, 0, 0, 0, 0, 0]}, "dynamics/steps": {"edges": [1.0, 1.05, 1.1, 1.15, 1.2, 1.25, 1.3, 1.35, 1.4, 1.45, 1.5, 1.55, 1.6, 1.65, 1.7000000000000002, 1.75, 1.8, 1.85, 1.9, 1.9500000000000002, 2.0], "correct_counts": [1214, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [286, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "dynamics/monotonic": {"edges": [0.7166666666666667, 0.7220833333333333, 0.7275, 0.7329166666666667, 0.7383333333333334, 0.74375, 0.7491666666666668, 0.7545833333333334, 0.7600000000000001, 0.7654166666666667, 0.7708333333333335, 0.7762500000000001, 0.7816666666666667, 0.7870833333333335, 0.7925000000000001, 0.7979166666666668, 0.8033333333333335, 0.8087500000000002, 0.8141666666666668, 0.8195833333333336, 0.8250000000000002], "correct_counts": [1, 2, 5, 25, 29, 69, 103, 144, 199, 162, 148, 138, 95, 45, 25, 11, 6, 3, 2, 2], "incorrect_counts": [0, 0, 2, 7, 9, 13, 14, 39, 30, 31, 29, 22, 25, 20, 17, 7, 10, 5, 4, 2]}, "dynamics/drop": {"edges": [62.5166342407465, 77.29895427798232, 92.08127431521814, 106.86359435245394, 121.64591438968976, 136.4282344269256, 151.2105544641614, 165.9928745013972, 180.77519453863303, 195.55751457586885, 210.33983461310467, 225.1221546503405, 239.90447468757628, 254.6867947248121, 269.4691147620479, 284.25143479928374, 299.03375483651956, 313.8160748737554, 328.5983949109912, 343.380714948227, 358.16303498546284], "correct_counts": [17, 72, 212, 278, 288, 178, 67, 37, 30, 12, 6, 4, 2, 6, 2, 2, 0, 1, 0, 0], "incorrect_counts": [2, 9, 39, 66, 39, 38, 24, 11, 4, 7, 10, 8, 5, 7, 4, 7, 2, 1, 2, 1]}, "dynamics/residual": {"edges": [1.1406713624795277, 1.1573103273908296, 1.1739492923021315, 1.1905882572134334, 1.2072272221247355, 1.2238661870360374, 1.2405051519473393, 1.2571441168586412, 1.2737830817699431, 1.290422046681245, 1.307061011592547, 1.323699976503849, 1.340338941415151, 1.3569779063264529, 1.3736168712377548, 1.3902558361490567, 1.4068948010603588, 1.4235337659716607, 1.4401727308829626, 1.4568116957942645, 1.4734506607055664], "correct_counts": [1, 6, 14, 25, 53, 88, 111, 153, 153, 165, 149, 107, 96, 53, 25, 8, 5, 1, 0, 1], "incorrect_counts": [0, 1, 0, 4, 10, 23, 35, 34, 38, 36, 45, 21, 17, 7, 9, 4, 1, 0, 1, 0]}}}, "feature_ablation": {"full": 0.07534184983249899, "drop_basin": 0.15063825991789662, "basin_only": 0.10286261887290694, "drop_energy": 0.08021799152104715, "energy_only": 0.19386361374456082, "drop_curv": 0.07785615271208078, "curv_only": 0.16390990398067096, "drop_dynamics": 0.08761931255420763, "dynamics_only": 0.15113347189816922}, "ece_geometry": 0.027494786480900396, "accuracy_ood": 0.20933333333333334, "aurc_ood": {"geometry": 0.7968967969651003, "rho_basin": 0.7527208068366574, "energy_min": 0.7530102393631208, "energy_mean": 0.7621969907261918, "energy_std": 0.7900061103185938, "msp": 0.6957160882123623, "temp_msp": 0.6999530402417576, "entropy": 0.698412184022287, "softmax_learned": 0.6954866419202574, "geom_softmax": 0.7147067020732829}, "ood_ltt": {"alpha": 0.1, "lambda_hat": 0.1463238254140143, "selective_risk": 0.75, "coverage": 0.616, "risk_within_budget": false}, "ood_validity": {"target": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "id_risk": [0.0, 0.0, 0.07001044932079414, 0.11437908496732026, 0.15036496350364964, 0.18904475617902472], "ood_risk": [0.0, 0.0, 0.75, 0.7721311475409836, 0.7871428571428571, 0.7906666666666666], "id_coverage": [0.0, 0.0, 0.638, 0.816, 0.9133333333333333, 0.998], "ood_coverage": [0.0, 0.0, 0.616, 0.8133333333333334, 0.9333333333333333, 1.0]}}, "env": {"python": "3.12.13", "platform": "macOS-26.5.2-arm64-arm-64bit", "jax": "0.11.0", "jax_backend": "cpu"}, "artifact_paths": []} +{"run_id": "354783efb5d1", "timestamp": "2026-07-30T11:18:04.745595+00:00", "git_sha": "67dc7a2", "config_hash": "3c2fcb2e5dde", "config": {"run": {"seed": 2, "task": "arithmetic", "notes": "E1 selective-prediction main run"}, "model": {"latent_dim": 32, "hidden_dim": 128, "context_dim": 64}, "inference": {"k_restarts": 12, "steps": 50, "step_size": 0.1, "temperature": 0.05, "init_scale": 1.0, "grad_tol": 0.001, "sampler": "langevin", "anneal_levels": 10, "anneal_steps_per_level": 8, "anneal_step_max": 0.2, "anneal_step_min": 0.01}, "train": {"epochs": 7, "batch_size": 128, "lr": 0.001, "n_train": 6000, "n_neg": 4, "neg_noise": 0.5, "objective": "basin_center", "ired_noise_min": 0.1, "ired_noise_max": 1.5, "ired_ridge": 0.05, "ired_decode_weight": 1.0, "ired_stat_weight": 1.0}, "eval": {"n_eval": 1500, "richer_geometry": false}, "conformal": {"alpha": 0.1, "delta": 0.05, "n_calib": 1500}, "task": {"arithmetic": {"modular": false, "n_operands": 6, "ood_n_operands": 6, "max_operand": 7, "ood_max_operand": 9}}}, "task": "arithmetic", "split": "selective", "seed": 2, "metrics": {"n_fit": 1500, "n_calib": 1500, "n_test": 1500, "k_restarts": 12, "objective": "basin_center", "sampler": "langevin", "feature_set": "base", "accuracy_id": 0.8713333333333333, "base_error": 0.12866666666666668, "final_train_loss": 0.9444285035133362, "feature_names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual"], "aurc": {"geometry": 0.024274182941440476, "rho_basin": 0.05598821987398549, "energy_min": 0.04412798365125574, "energy_mean": 0.042606873866458325, "energy_std": 0.1207663666916667, "msp": 0.024891583299795343, "temp_msp": 0.027860819750315315, "entropy": 0.032137841946629477, "softmax_learned": 0.02315351961620342, "geom_softmax": 0.021598978976821472}, "temperature": 0.2805001582613758, "best_energy_baseline": "energy_mean", "best_baseline": "msp", "delta_aurc_vs_energy_min": [0.019853800709815263, 0.00990417624353011, 0.032291575515518305], "delta_aurc_vs_best_energy": [0.01833269092501785, 0.007161408501466717, 0.03183435108421057], "delta_aurc_vs_best_baseline": [0.0006174003583548672, -0.003978080304717403, 0.004496550989570287], "delta_aurc_geom_adds": [0.0015545406393819468, -1.6751835250081416e-05, 0.003164056614505232], "geometry_wins": true, "geometry_wins_vs_baseline": false, "geometry_adds_over_softmax": false, "risk_coverage": {"geometry": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.011111111111111112, 0.008333333333333333, 0.006666666666666667, 0.005555555555555555, 0.004761904761904763, 0.004166666666666667, 0.003703703703703704, 0.003333333333333334, 0.006060606060606061, 0.005555555555555556, 0.007692307692307693, 0.007142857142857143, 0.006666666666666666, 0.00625, 0.00784313725490196, 0.007407407407407407, 0.007017543859649123, 0.006666666666666667, 0.006349206349206348, 0.006060606060606061, 0.008695652173913044, 0.008333333333333331, 0.008, 0.008974358974358974, 0.008641975308641974, 0.008333333333333333, 0.008045977011494251, 0.008888888888888889, 0.00967741935483871, 0.0125, 0.014141414141414142, 0.014705882352941176, 0.016190476190476345, 0.018518518518518514, 0.021621621621621623, 0.02894736842105263, 0.03247863247863248, 0.04, 0.04552845528455284, 0.054761904761904755, 0.06434108527131784, 0.07348484848484849, 0.08148148148148149, 0.09130434782608696, 0.10283687943262422, 0.11180555555555555, 0.11972789115646258, 0.12866666666666668]}, "rho_basin": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0470588235294118, 0.04705882352941174, 0.047058823529411695, 0.047058823529411674, 0.04705882352941166, 0.047058823529411646, 0.04705882352941164, 0.04705882352941163, 0.04705882352941163, 0.047058823529411625, 0.047058823529411625, 0.047058823529411625, 0.047058823529411625, 0.047058823529411625, 0.04705882352941162, 0.04705882352941162, 0.04705882352941162, 0.04705882352941162, 0.04705882352941162, 0.04705882352941162, 0.04705882352941161, 0.04705882352941161, 0.04705882352941162, 0.04705882352941162, 0.04705882352941162, 0.04705882352941162, 0.04705882352941162, 0.04705882352941161, 0.04705882352941161, 0.04705882352941161, 0.04705882352941161, 0.04705882352941161, 0.04705882352941161, 0.04705882352941161, 0.04705882352941161, 0.04705882352941161, 0.047714381047714215, 0.05152696556205304, 0.05514403292181039, 0.05858024691357986, 0.0618488407106289, 0.06746031746031694, 0.07403100775193731, 0.08030303030302952, 0.08962962962962887, 0.09840579710144855, 0.10482269503546043, 0.11097222222222178, 0.11687074829931939, 0.1286666666666665]}, "energy_min": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.06666666666666667, 0.03333333333333333, 0.022222222222222223, 0.016666666666666666, 0.02, 0.027777777777777773, 0.03333333333333334, 0.029166666666666667, 0.02962962962962963, 0.026666666666666672, 0.024242424242424242, 0.022222222222222223, 0.02564102564102564, 0.023809523809523808, 0.031111111111111103, 0.029166666666666667, 0.027450980392156862, 0.027777777777777773, 0.028070175438596492, 0.028333333333333332, 0.028571428571428567, 0.030303030303030304, 0.030434782608695653, 0.03055555555555555, 0.030666666666666665, 0.03205128205128205, 0.03333333333333333, 0.03214285714285714, 0.033333333333333326, 0.03666666666666667, 0.03978494623655914, 0.03854166666666667, 0.03939393939393939, 0.04215686274509804, 0.041904761904761896, 0.04444444444444444, 0.047747747747747746, 0.05087719298245614, 0.0547008547008547, 0.06, 0.06422764227642275, 0.06587301587301586, 0.06976744186046512, 0.07803030303030303, 0.08518518518518518, 0.09565217391304348, 0.10212765957446807, 0.10972222222222232, 0.11836734693877551, 0.12866666666666668]}, "energy_mean": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.03333333333333333, 0.016666666666666666, 0.022222222222222223, 0.025, 0.02666666666666667, 0.027777777777777773, 0.028571428571428577, 0.025, 0.022222222222222223, 0.023333333333333338, 0.024242424242424242, 0.022222222222222223, 0.023076923076923078, 0.02142857142857143, 0.019999999999999997, 0.022916666666666665, 0.021568627450980392, 0.02037037037037037, 0.021052631578947368, 0.02, 0.02063492063492063, 0.022727272727272728, 0.02463768115942029, 0.027777777777777773, 0.02666666666666667, 0.026923076923076925, 0.027160493827160494, 0.030952380952380953, 0.032183908045977004, 0.03111111111111111, 0.03225806451612903, 0.035416666666666666, 0.03636363636363636, 0.0392156862745098, 0.04095238095238095, 0.043518518518518665, 0.04594594594594595, 0.04912280701754386, 0.05042735042735043, 0.058333333333333334, 0.06341463414634145, 0.06825396825396837, 0.07364341085271318, 0.0803030303030303, 0.08444444444444445, 0.09202898550724638, 0.09929078014184409, 0.10624999999999998, 0.11904761904761904, 0.12866666666666668]}, "energy_std": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.1, 0.15, 0.13333333333333333, 0.1, 0.12, 0.11111111111111109, 0.11428571428571431, 0.11666666666666667, 0.11851851851851852, 0.11333333333333334, 0.11818181818181818, 0.11944444444444445, 0.12564102564102564, 0.12142857142857143, 0.12222222222222219, 0.12291666666666666, 0.11960784313725491, 0.11666666666666665, 0.11228070175438597, 0.11, 0.11111111111111122, 0.11212121212121212, 0.11014492753623188, 0.11527777777777777, 0.12266666666666666, 0.12307692307692308, 0.12345679012345678, 0.12380952380952381, 0.12298850574712641, 0.12555555555555556, 0.12365591397849462, 0.12395833333333334, 0.12323232323232323, 0.12254901960784313, 0.12285714285714285, 0.12314814814814813, 0.12162162162162163, 0.12192982456140351, 0.12478632478632479, 0.12583333333333332, 0.12439024390243901, 0.1238095238095238, 0.12713178294573643, 0.1303030303030303, 0.1288888888888889, 0.1289855072463768, 0.1283687943262411, 0.12777777777777774, 0.12857142857142856, 0.12866666666666668]}, "msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.002222222222222222, 0.0020833333333333333, 0.00196078431372549, 0.0018518518518518517, 0.0017543859649122807, 0.0016666666666666668, 0.001587301587301587, 0.0015151515151515152, 0.0014492753623188406, 0.0027777777777777775, 0.004, 0.0038461538461538464, 0.0049382716049382715, 0.007142857142857143, 0.008045977011494251, 0.008888888888888889, 0.00967741935483871, 0.016666666666666666, 0.022222222222222223, 0.024509803921568627, 0.030476190476190473, 0.03703703703703703, 0.043243243243243246, 0.04736842105263158, 0.05042735042735043, 0.059166666666666666, 0.060162601626016256, 0.06507936507936507, 0.07131782945736434, 0.07803030303030303, 0.08592592592592592, 0.09492753623188406, 0.10070921985815602, 0.10833333333333343, 0.11904761904761904, 0.12866666666666668]}, "temp_msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.004761904761904565, 0.004166666666666667, 0.003703703703703704, 0.006666666666666668, 0.006060606060606061, 0.008333333333333333, 0.010256410256410256, 0.009523809523809525, 0.01111111111111111, 0.010416666666666666, 0.00980392156862745, 0.009259259259259257, 0.008771929824561403, 0.008333333333333333, 0.007936507936507934, 0.007575757575757576, 0.008695652173913044, 0.008333333333333331, 0.008, 0.010256410256410256, 0.013580246913580247, 0.015476190476190477, 0.018390804597701337, 0.02, 0.02258064516129032, 0.022916666666666665, 0.024242424242424242, 0.024509803921568627, 0.026666666666666665, 0.029629629629629627, 0.03423423423423423, 0.03859649122807018, 0.04358974358974359, 0.0525, 0.05609756097560988, 0.06269841269841268, 0.06666666666666667, 0.07424242424242425, 0.0837037037037037, 0.0927536231884058, 0.10496453900709218, 0.11597222222222221, 0.11972789115646258, 0.12866666666666668]}, "entropy": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.004761904761904762, 0.004444444444444444, 0.010416666666666666, 0.00980392156862745, 0.009259259259259257, 0.010526315789473684, 0.013333333333333334, 0.012698412698412697, 0.015151515151515152, 0.014492753623188406, 0.016666666666666663, 0.018666666666666668, 0.02435897435897436, 0.030864197530864196, 0.034523809523809526, 0.03448275862068965, 0.03666666666666667, 0.04086021505376344, 0.041666666666666664, 0.04040404040404041, 0.041176470588235294, 0.04476190476190475, 0.045370370370370366, 0.04864864864864865, 0.05, 0.05128205128205128, 0.056666666666666664, 0.060162601626016256, 0.06825396825396837, 0.07441860465116279, 0.07803030303030303, 0.08592592592592592, 0.09130434782608696, 0.09716312056737587, 0.10763888888888888, 0.11700680272108843, 0.12866666666666668]}, "softmax_learned": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.8462688879574732e-16, 0.0020833333333333333, 0.00196078431372549, 0.0018518518518518517, 0.0017543859649122807, 0.0016666666666666668, 0.001587301587301587, 0.0015151515151515152, 0.0014492753623188406, 0.0013888888888888887, 0.0026666666666666666, 0.0038461538461538464, 0.003703703703703704, 0.005952380952380952, 0.005747126436781608, 0.006666666666666667, 0.007526881720430108, 0.010416666666666666, 0.01616161616161616, 0.018627450980392157, 0.023809523809523805, 0.0287037037037037, 0.03153153153153153, 0.03508771929824561, 0.042735042735042736, 0.05, 0.05609756097560975, 0.06587301587301586, 0.06976744186046512, 0.07803030303030303, 0.08592592592592592, 0.09420289855072464, 0.10141843971631216, 0.11249999999999999, 0.12108843537414966, 0.12866666666666668]}, "geom_softmax": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.002380952380952381, 0.002222222222222222, 0.0020833333333333333, 0.00196078431372549, 0.0037037037037037034, 0.0035087719298245615, 0.0033333333333333335, 0.003174603174603174, 0.0030303030303030303, 0.004347826086956522, 0.004166666666666666, 0.004, 0.0038461538461538464, 0.0049382716049382715, 0.005952380952380952, 0.005747126436781608, 0.006666666666666667, 0.0064516129032258064, 0.007291666666666667, 0.010101010101010102, 0.012745098039215686, 0.016190476190476186, 0.01851851851851867, 0.024324324324324326, 0.028070175438596492, 0.03247863247863248, 0.04416666666666667, 0.05040650406504078, 0.05873015873015872, 0.06511627906976744, 0.07196969696969698, 0.08074074074074074, 0.09130434782608696, 0.10283687943262422, 0.11041666666666665, 0.12040816326530612, 0.12866666666666668]}}, "ltt": {"alpha": 0.1, "delta": 0.05, "lambda_hat": 0.42350354070874574, "coverage": 0.8913333333333333, "abstain_rate": 0.10866666666666666, "selective_risk": 0.07778608825729244, "selective_accuracy": 0.9222139117427075, "risk_within_budget": true}, "coverage_validity": {"alpha_grid": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "target": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "achieved_risk": [0.0, 0.03186907838070629, 0.07778608825729244, 0.12750333778371162, 0.12750333778371162, 0.12750333778371162], "coverage": [0.0, 0.774, 0.8913333333333333, 0.9986666666666667, 0.9986666666666667, 0.9986666666666667]}, "feature_diagnostics": {"names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual"], "auroc": {"basin/rho": 0.7907897292775846, "basin/entropy": 0.2093569500220019, "basin/dispersion": 0.4859485195301505, "energy/mean": 0.16196962549206942, "energy/min": 0.17429465096273156, "energy/std": 0.47604964896075735, "curv/lmax_mean": 0.43994671973550153, "curv/lmax_best": 0.5017086156249133, "curv/trace_mean": 0.4422876420707946, "curv/trace_best": 0.4850169077625064, "dynamics/steps": 0.5, "dynamics/monotonic": 0.2952654300676707, "dynamics/drop": 0.2783695604774609, "dynamics/residual": 0.46715374765610446}, "hist": {"basin/rho": {"edges": [0.4166666666666667, 0.44583333333333336, 0.47500000000000003, 0.5041666666666667, 0.5333333333333333, 0.5625, 0.5916666666666667, 0.6208333333333333, 0.65, 0.6791666666666667, 0.7083333333333333, 0.7375, 0.7666666666666666, 0.7958333333333334, 0.825, 0.8541666666666666, 0.8833333333333333, 0.9125, 0.9416666666666667, 0.9708333333333332, 1.0], "correct_counts": [0, 0, 4, 0, 0, 27, 0, 0, 33, 0, 0, 29, 0, 0, 52, 0, 0, 109, 0, 1053], "incorrect_counts": [2, 0, 16, 0, 0, 18, 0, 0, 22, 0, 0, 29, 0, 0, 28, 0, 0, 26, 0, 52]}, "basin/entropy": {"edges": [0.0, 0.05387781635334003, 0.10775563270668007, 0.16163344906002008, 0.21551126541336013, 0.2693890817667002, 0.32326689812004017, 0.3771447144733802, 0.43102253082672026, 0.4849003471800603, 0.5387781635334004, 0.5926559798867403, 0.6465337962400803, 0.7004116125934204, 0.7542894289467604, 0.8081672453001005, 0.8620450616534405, 0.9159228780067805, 0.9698006943601206, 1.0236785107134607, 1.0775563270668007], "correct_counts": [1053, 0, 0, 0, 0, 109, 0, 0, 50, 0, 30, 30, 28, 1, 0, 3, 3, 0, 0, 0], "incorrect_counts": [52, 0, 0, 0, 0, 26, 0, 0, 27, 0, 28, 20, 32, 2, 0, 1, 1, 1, 1, 2]}, "basin/dispersion": {"edges": [1.5319602421529375, 1.557597310257325, 1.5832343783617129, 1.6088714464661005, 1.6345085145704883, 1.6601455826748759, 1.6857826507792635, 1.7114197188836513, 1.737056786988039, 1.7626938550924267, 1.7883309231968143, 1.813967991301202, 1.8396050594055897, 1.8652421275099773, 1.8908791956143651, 1.9165162637187527, 1.9421533318231403, 1.9677903999275281, 1.9934274680319157, 2.0190645361363035, 2.044701604240691], "correct_counts": [0, 0, 0, 4, 13, 13, 50, 72, 111, 158, 164, 177, 192, 126, 117, 61, 28, 16, 2, 3], "incorrect_counts": [1, 0, 0, 0, 2, 3, 3, 11, 21, 19, 23, 25, 26, 27, 10, 9, 6, 4, 2, 1]}, "energy/mean": {"edges": [-0.30340700844923657, -0.2387943930923939, -0.1741817777355512, -0.10956916237870853, -0.044956547021865856, 0.01965606833497685, 0.0842686836918195, 0.14888129904866215, 0.21349391440550486, 0.27810652976234757, 0.34271914511919027, 0.40733176047603287, 0.4719443758328756, 0.5365569911897183, 0.6011696065465608, 0.6657822219034035, 0.7303948372602462, 0.7950074526170889, 0.8596200679739316, 0.9242326833307744, 0.988845298687617], "correct_counts": [2, 39, 170, 267, 275, 152, 74, 55, 57, 38, 33, 34, 23, 33, 19, 19, 10, 5, 1, 1], "incorrect_counts": [1, 0, 5, 5, 10, 9, 10, 8, 8, 13, 13, 14, 10, 21, 15, 16, 21, 11, 2, 1]}, "energy/min": {"edges": [-0.6361721754074097, -0.5738012850284576, -0.5114303946495056, -0.4490595042705536, -0.3866886138916016, -0.32431772351264954, -0.2619468331336975, -0.1995759427547455, -0.13720505237579345, -0.07483416199684145, -0.012463271617889404, 0.049907618761062644, 0.11227850914001469, 0.17464939951896674, 0.23702028989791868, 0.2993911802768707, 0.3617620706558228, 0.4241329610347748, 0.48650385141372676, 0.5488747417926789, 0.6112456321716309], "correct_counts": [10, 45, 106, 187, 224, 203, 140, 82, 53, 42, 45, 31, 38, 17, 25, 23, 16, 10, 3, 7], "incorrect_counts": [0, 2, 1, 5, 8, 11, 10, 7, 8, 9, 14, 8, 8, 18, 23, 13, 18, 17, 8, 5]}, "energy/std": {"edges": [0.09172678671306, 0.10609920482521892, 0.12047162293737784, 0.13484404104953673, 0.14921645916169568, 0.16358887727385457, 0.1779612953860135, 0.1923337134981724, 0.20670613161033133, 0.22107854972249025, 0.23545096783464917, 0.24982338594680809, 0.264195804058967, 0.27856822217112587, 0.2929406402832848, 0.30731305839544376, 0.32168547650760265, 0.33605789461976154, 0.3504303127319205, 0.3648027308440794, 0.37917514895623833], "correct_counts": [11, 25, 52, 107, 109, 155, 166, 148, 140, 133, 93, 55, 50, 25, 15, 12, 5, 3, 1, 2], "incorrect_counts": [0, 6, 6, 12, 19, 19, 20, 28, 17, 23, 17, 9, 5, 5, 3, 0, 2, 1, 0, 1]}, "curv/lmax_mean": {"edges": [1.0000314265489578, 1.0005539747575918, 1.001076522966226, 1.00159907117486, 1.002121619383494, 1.002644167592128, 1.0031667158007622, 1.0036892640093962, 1.0042118122180304, 1.0047343604266643, 1.0052569086352983, 1.0057794568439324, 1.0063020050525664, 1.0068245532612006, 1.0073471014698345, 1.0078696496784687, 1.0083921978871027, 1.0089147460957368, 1.0094372943043708, 1.009959842513005, 1.010482390721639], "correct_counts": [106, 240, 245, 181, 115, 97, 84, 74, 42, 38, 24, 22, 11, 10, 5, 4, 6, 1, 1, 1], "incorrect_counts": [15, 28, 24, 21, 25, 22, 14, 12, 9, 7, 3, 3, 3, 1, 2, 2, 0, 0, 1, 1]}, "curv/lmax_best": {"edges": [0.9999086856842041, 1.0016137421131135, 1.0033187985420227, 1.005023854970932, 1.0067289113998412, 1.0084339678287506, 1.01013902425766, 1.0118440806865692, 1.0135491371154786, 1.0152541935443877, 1.0169592499732971, 1.0186643064022065, 1.0203693628311157, 1.022074419260025, 1.0237794756889342, 1.0254845321178436, 1.027189588546753, 1.0288946449756622, 1.0305997014045716, 1.0323047578334807, 1.0340098142623901], "correct_counts": [815, 192, 101, 73, 40, 24, 16, 15, 5, 8, 5, 3, 1, 1, 3, 1, 1, 0, 1, 2], "incorrect_counts": [114, 36, 14, 7, 9, 2, 2, 2, 2, 1, 1, 1, 0, 0, 1, 0, 1, 0, 0, 0]}, "curv/trace_mean": {"edges": [31.999862353007, 32.003960879643756, 32.008059406280516, 32.012157932917276, 32.016256459554036, 32.020354986190796, 32.024453512827556, 32.028552039464316, 32.032650566101076, 32.036749092737836, 32.040847619374595, 32.044946146011355, 32.049044672648115, 32.053143199284875, 32.05724172592163, 32.06134025255839, 32.06543877919515, 32.06953730583191, 32.07363583246867, 32.07773435910543, 32.08183288574219], "correct_counts": [1, 2, 41, 105, 113, 142, 143, 135, 129, 115, 82, 90, 71, 49, 50, 16, 10, 7, 3, 3], "incorrect_counts": [1, 4, 10, 4, 18, 13, 15, 16, 19, 13, 16, 20, 10, 14, 5, 9, 4, 2, 0, 0]}, "curv/trace_best": {"edges": [31.99859619140625, 32.006431579589844, 32.01426696777344, 32.02210235595703, 32.029937744140625, 32.03777313232422, 32.04560852050781, 32.053443908691406, 32.061279296875, 32.069114685058594, 32.07695007324219, 32.08478546142578, 32.092620849609375, 32.10045623779297, 32.10829162597656, 32.116127014160156, 32.12396240234375, 32.131797790527344, 32.13963317871094, 32.14746856689453, 32.155303955078125], "correct_counts": [38, 192, 243, 218, 177, 139, 87, 80, 37, 35, 10, 14, 13, 9, 7, 2, 2, 1, 1, 2], "incorrect_counts": [9, 27, 33, 30, 22, 19, 14, 9, 7, 9, 5, 2, 3, 1, 0, 3, 0, 0, 0, 0]}, "dynamics/steps": {"edges": [1.0, 1.05, 1.1, 1.15, 1.2, 1.25, 1.3, 1.35, 1.4, 1.45, 1.5, 1.55, 1.6, 1.65, 1.7000000000000002, 1.75, 1.8, 1.85, 1.9, 1.9500000000000002, 2.0], "correct_counts": [1307, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [193, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "dynamics/monotonic": {"edges": [0.7233333333333333, 0.7286666666666666, 0.734, 0.7393333333333333, 0.7446666666666666, 0.75, 0.7553333333333333, 0.7606666666666666, 0.766, 0.7713333333333333, 0.7766666666666666, 0.782, 0.7873333333333333, 0.7926666666666667, 0.798, 0.8033333333333333, 0.8086666666666668, 0.8140000000000001, 0.8193333333333334, 0.8246666666666668, 0.8300000000000001], "correct_counts": [5, 6, 15, 32, 65, 152, 180, 167, 199, 166, 163, 72, 43, 30, 6, 3, 3, 0, 0, 0], "incorrect_counts": [0, 0, 0, 3, 3, 18, 10, 13, 15, 23, 23, 16, 25, 13, 8, 12, 7, 0, 3, 1]}, "dynamics/drop": {"edges": [64.10532930493355, 78.80553761435051, 93.50574592376749, 108.20595423318446, 122.90616254260144, 137.60637085201842, 152.30657916143537, 167.00678747085234, 181.70699578026932, 196.40720408968627, 211.10741239910325, 225.80762070852023, 240.50782901793718, 255.20803732735416, 269.90824563677114, 284.6084539461881, 299.3086622556051, 314.00887056502205, 328.709078874439, 343.409287183856, 358.10949549327296], "correct_counts": [9, 55, 217, 358, 369, 180, 66, 19, 11, 10, 8, 3, 0, 2, 0, 0, 0, 0, 0, 0], "incorrect_counts": [0, 6, 22, 22, 25, 26, 21, 17, 7, 5, 8, 13, 9, 2, 3, 2, 1, 2, 0, 2]}, "dynamics/residual": {"edges": [1.1507504383722942, 1.166901803513368, 1.1830531686544419, 1.1992045337955157, 1.2153558989365896, 1.2315072640776634, 1.2476586292187373, 1.2638099943598111, 1.279961359500885, 1.2961127246419588, 1.3122640897830329, 1.3284154549241065, 1.3445668200651806, 1.3607181852062544, 1.3768695503473283, 1.3930209154884021, 1.409172280629476, 1.4253236457705498, 1.4414750109116237, 1.4576263760526975, 1.4737777411937714], "correct_counts": [6, 9, 15, 37, 63, 94, 128, 188, 171, 147, 140, 124, 85, 52, 27, 10, 6, 3, 1, 1], "incorrect_counts": [0, 3, 0, 7, 6, 17, 20, 18, 24, 18, 27, 20, 17, 7, 3, 2, 2, 1, 1, 0]}}}, "feature_ablation": {"full": 0.024274182941440476, "drop_basin": 0.04808611569032442, "basin_only": 0.05844366219985415, "drop_energy": 0.031331435374528864, "energy_only": 0.040657703926391564, "drop_curv": 0.02375638180042061, "curv_only": 0.1256930708759337, "drop_dynamics": 0.02454171163749289, "dynamics_only": 0.07986208469806944}, "ece_geometry": 0.04318340190787242, "accuracy_ood": 0.21866666666666668, "aurc_ood": {"geometry": 0.7118811950177575, "rho_basin": 0.7176588499789294, "energy_min": 0.7122394640249904, "energy_mean": 0.7126699633509685, "energy_std": 0.7809376529088858, "msp": 0.653860323850022, "temp_msp": 0.6533021641143179, "entropy": 0.6592791559765256, "softmax_learned": 0.648426119528371, "geom_softmax": 0.7027844744493177}, "ood_ltt": {"alpha": 0.1, "lambda_hat": 0.42350354070874574, "selective_risk": 0.7672155688622755, "coverage": 0.8906666666666667, "risk_within_budget": false}, "ood_validity": {"target": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "id_risk": [0.0, 0.03186907838070629, 0.07778608825729244, 0.12750333778371162, 0.12750333778371162, 0.12750333778371162], "ood_risk": [0.0, 0.7385321100917431, 0.7672155688622755, 0.7813333333333333, 0.7813333333333333, 0.7813333333333333], "id_coverage": [0.0, 0.774, 0.8913333333333333, 0.9986666666666667, 0.9986666666666667, 0.9986666666666667], "ood_coverage": [0.0, 0.7266666666666667, 0.8906666666666667, 1.0, 1.0, 1.0]}}, "env": {"python": "3.12.13", "platform": "macOS-26.5.2-arm64-arm-64bit", "jax": "0.11.0", "jax_backend": "cpu"}, "artifact_paths": []} +{"run_id": "70cee2dae378", "timestamp": "2026-07-30T11:18:05.366179+00:00", "git_sha": "67dc7a2", "config_hash": "f5d7b65bb5d0", "config": {"run": {"seed": 2, "task": "graph_planning", "notes": "graph adaptive-halting run"}, "model": {"latent_dim": 32, "hidden_dim": 128, "context_dim": 64}, "inference": {"k_restarts": 12, "steps": 50, "step_size": 0.1, "temperature": 0.05, "init_scale": 1.0, "grad_tol": 0.001, "sampler": "langevin", "anneal_levels": 10, "anneal_steps_per_level": 8, "anneal_step_max": 0.2, "anneal_step_min": 0.01}, "train": {"epochs": 25, "batch_size": 128, "lr": 0.001, "n_train": 8000, "n_neg": 4, "neg_noise": 0.5, "objective": "basin_center", "ired_noise_min": 0.1, "ired_noise_max": 1.5, "ired_ridge": 0.05, "ired_decode_weight": 1.0, "ired_stat_weight": 1.0}, "eval": {"n_eval": 800, "richer_geometry": false}, "conformal": {"alpha": 0.1, "delta": 0.05, "n_calib": 800}, "task": {"graph_planning": {"n_nodes": 7, "ood_n_nodes": 10, "edge_prob": 0.4, "max_len": 4}}}, "task": "graph_planning", "split": "halting", "seed": 2, "metrics": {"k_restarts": 12, "steps": 50, "alpha": 0.1, "n_calib": 800, "n_test": 800, "final_train_loss": 0.178430438041687, "tau_hat": 0.8333, "compute_used": 0.15805000000000002, "compute_saved": 0.84195, "halting_risk": 0.085, "risk_within_budget": true, "full_accuracy": 0.68, "halted_accuracy": 0.66875, "accuracy_drop": 0.011250000000000093, "risk_monotone_in_tau": true, "calib_risk_by_tau": [0.4625, 0.4625, 0.4625, 0.4625, 0.38375, 0.38375, 0.29875, 0.125, 0.125, 0.08875, 0.0125, 0.0125], "calib_compute_by_tau": [0.0, 0.0, 0.0, 0.000625, 0.022874999999999833, 0.022874999999999833, 0.0464749999999998, 0.12292500000000013, 0.12292500000000013, 0.16652500000000037, 0.29684999999999984, 0.29684999999999984], "tau_sweep": {"taus": [0.0833, 0.1667, 0.25, 0.3333, 0.4167, 0.5, 0.5833, 0.6667, 0.75, 0.8333, 0.9167, 1.0], "compute_used": [0.0, 0.0, 0.0, 0.0004, 0.022099999999999998, 0.022099999999999998, 0.04724999999999999, 0.11257500000000001, 0.11257500000000001, 0.15805000000000002, 0.29180000000000006, 0.29180000000000006], "accuracy": [0.4525, 0.4525, 0.4525, 0.45875, 0.51625, 0.51625, 0.565, 0.65, 0.65, 0.66875, 0.68375, 0.68375], "disagreement": [0.47375, 0.47375, 0.47375, 0.46875, 0.3575, 0.3575, 0.28875, 0.135, 0.135, 0.085, 0.01375, 0.01375]}}, "env": {"python": "3.12.13", "platform": "macOS-26.5.2-arm64-arm-64bit", "jax": "0.11.0", "jax_backend": "cpu"}, "artifact_paths": []} +{"run_id": "496414a8f2cd", "timestamp": "2026-07-30T11:18:07.813782+00:00", "git_sha": "67dc7a2", "config_hash": "593dfa402e05", "config": {"run": {"seed": 1, "task": "graph_planning", "notes": "E2 graph selective-prediction"}, "model": {"latent_dim": 32, "hidden_dim": 128, "context_dim": 64}, "inference": {"k_restarts": 12, "steps": 50, "step_size": 0.1, "temperature": 0.05, "init_scale": 1.0, "grad_tol": 0.001, "sampler": "langevin", "anneal_levels": 10, "anneal_steps_per_level": 8, "anneal_step_max": 0.2, "anneal_step_min": 0.01}, "train": {"epochs": 25, "batch_size": 128, "lr": 0.001, "n_train": 8000, "n_neg": 4, "neg_noise": 0.5, "objective": "basin_center", "ired_noise_min": 0.1, "ired_noise_max": 1.5, "ired_ridge": 0.05, "ired_decode_weight": 1.0, "ired_stat_weight": 1.0}, "eval": {"n_eval": 1500, "richer_geometry": false}, "conformal": {"alpha": 0.1, "delta": 0.05, "n_calib": 1500}, "task": {"graph_planning": {"n_nodes": 7, "ood_n_nodes": 10, "edge_prob": 0.4, "max_len": 4}}}, "task": "graph_planning", "split": "selective", "seed": 1, "metrics": {"n_fit": 1500, "n_calib": 1500, "n_test": 1500, "k_restarts": 12, "objective": "basin_center", "sampler": "langevin", "feature_set": "base", "accuracy_id": 0.6913333333333334, "base_error": 0.30866666666666664, "final_train_loss": 0.2080264687538147, "feature_names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual"], "aurc": {"geometry": 0.17444436593752588, "rho_basin": 0.25642611361139067, "energy_min": 0.19964686175564186, "energy_mean": 0.19527316950353119, "energy_std": 0.29427140691389053, "msp": 0.1401480752960069, "temp_msp": 0.1386661861515245, "entropy": 0.13897082240953618, "softmax_learned": 0.13937315202394526, "geom_softmax": 0.14038547133382884}, "temperature": 3.1187914236161625, "best_energy_baseline": "energy_mean", "best_baseline": "temp_msp", "delta_aurc_vs_energy_min": [0.025202495818115983, 0.014181046993774545, 0.03686871923665555], "delta_aurc_vs_best_energy": [0.020828803566005305, 0.010827769204687336, 0.03126253972294629], "delta_aurc_vs_best_baseline": [-0.03577817978600137, -0.047061369377102535, -0.024631857225653937], "delta_aurc_geom_adds": [-0.0010123193098835748, -0.0035237183803231006, 0.0015377190771044024], "geometry_wins": true, "geometry_wins_vs_baseline": false, "geometry_adds_over_softmax": false, "risk_coverage": {"geometry": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.03333333333333333, 0.03333333333333333, 0.05555555555555555, 0.058333333333333334, 0.06666666666666667, 0.0722222222222222, 0.07142857142857144, 0.07916666666666666, 0.0962962962962963, 0.09333333333333335, 0.09696969696969697, 0.1111111111111111, 0.1076923076923077, 0.12380952380952381, 0.12888888888888886, 0.13125, 0.14313725490196078, 0.14629629629629626, 0.15263157894736842, 0.16, 0.16984126984126982, 0.16818181818181818, 0.16956521739130434, 0.17500000000000007, 0.18, 0.18333333333333332, 0.19012345679012346, 0.19285714285714287, 0.1965517241379312, 0.20222222222222222, 0.2075268817204301, 0.2125, 0.21818181818181817, 0.22156862745098038, 0.22666666666666677, 0.23055555555555554, 0.23513513513513515, 0.2394736842105263, 0.24188034188034188, 0.25, 0.2544715447154471, 0.26031746031746045, 0.26666666666666666, 0.27121212121212124, 0.2777777777777778, 0.2797101449275362, 0.2865248226950356, 0.2930555555555556, 0.3013605442176871, 0.30866666666666664]}, "rho_basin": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.25061025223759165, 0.25061025223759126, 0.25061025223759165, 0.25061025223759187, 0.25061025223759203, 0.2506102522375921, 0.2506102522375922, 0.25061025223759226, 0.25061025223759226, 0.2506102522375923, 0.2506102522375923, 0.25061025223759237, 0.25061025223759237, 0.25061025223759237, 0.2506102522375924, 0.2506102522375924, 0.2506102522375924, 0.25061025223759165, 0.2506102522375909, 0.2506102522375903, 0.25061025223758976, 0.2506102522375892, 0.25061025223758876, 0.2506102522375883, 0.2506102522375879, 0.25061025223758754, 0.2506102522375872, 0.25061025223758693, 0.2506102522375866, 0.2506102522375863, 0.2506102522375861, 0.2506102522375858, 0.2506102522375856, 0.2506102522375854, 0.250610252237586, 0.25061025223758654, 0.2506102522375871, 0.2506102522375876, 0.2506102522375881, 0.25061025223758854, 0.25086440519577363, 0.25830140485312647, 0.26539249754967215, 0.2724913065076973, 0.2813721918639928, 0.2896286231884037, 0.2953804868698466, 0.2993618618618604, 0.3036646916831233, 0.30866666666666537]}, "energy_min": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.08333333333333333, 0.08888888888888889, 0.075, 0.09333333333333334, 0.08333333333333331, 0.09523809523809525, 0.1125, 0.1259259259259259, 0.13333333333333322, 0.1484848484848485, 0.16111111111111112, 0.16666666666666666, 0.1761904761904762, 0.18000000000000013, 0.19166666666666668, 0.19607843137254902, 0.19444444444444442, 0.1912280701754386, 0.19, 0.18888888888888886, 0.1893939393939394, 0.19710144927536233, 0.19305555555555565, 0.2, 0.2141025641025641, 0.21975308641975308, 0.22142857142857142, 0.2229885057471264, 0.22777777777777777, 0.22903225806451613, 0.23125, 0.23737373737373738, 0.24215686274509804, 0.24857142857142855, 0.24629629629629626, 0.25225225225225223, 0.26228070175438595, 0.26837606837606837, 0.26916666666666667, 0.2764227642276422, 0.27698412698412694, 0.2837209302325581, 0.28939393939393937, 0.2911111111111111, 0.29347826086956524, 0.2971631205673758, 0.29861111111111116, 0.30272108843537415, 0.30866666666666664]}, "energy_mean": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.1, 0.06666666666666667, 0.044444444444444446, 0.075, 0.08666666666666667, 0.10555555555555554, 0.10000000000000002, 0.10833333333333334, 0.1111111111111111, 0.10333333333333335, 0.10909090909090909, 0.125, 0.13846153846153847, 0.15476190476190477, 0.16222222222222218, 0.17708333333333334, 0.18235294117647058, 0.187037037037037, 0.18421052631578946, 0.18666666666666668, 0.1857142857142857, 0.18484848484848485, 0.19710144927536233, 0.1972222222222223, 0.20933333333333334, 0.21025641025641026, 0.2111111111111111, 0.21428571428571427, 0.2172413793103448, 0.22333333333333333, 0.22258064516129034, 0.22708333333333333, 0.23232323232323232, 0.24019607843137256, 0.23999999999999996, 0.24444444444444455, 0.24864864864864866, 0.2543859649122807, 0.2606837606837607, 0.2633333333333333, 0.265040650406504, 0.269047619047619, 0.26976744186046514, 0.275, 0.28, 0.286231884057971, 0.29290780141843964, 0.2979166666666666, 0.3006802721088435, 0.30866666666666664]}, "energy_std": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.03333333333333333, 0.16666666666666666, 0.2222222222222222, 0.24166666666666667, 0.2733333333333333, 0.29999999999999993, 0.3285714285714286, 0.30833333333333335, 0.3074074074074074, 0.30000000000000004, 0.296969696969697, 0.2972222222222222, 0.29743589743589743, 0.3047619047619048, 0.3044444444444444, 0.30833333333333335, 0.307843137254902, 0.3111111111111112, 0.30526315789473685, 0.3, 0.30476190476190473, 0.31212121212121213, 0.3101449275362319, 0.3097222222222222, 0.30933333333333335, 0.308974358974359, 0.30987654320987656, 0.31547619047619047, 0.31149425287356314, 0.31222222222222223, 0.31290322580645163, 0.31666666666666665, 0.31414141414141417, 0.31470588235294117, 0.31619047619047613, 0.31481481481481477, 0.31261261261261264, 0.3131578947368421, 0.31367521367521367, 0.31333333333333335, 0.3113821138211382, 0.3063492063492063, 0.3077519379844961, 0.30606060606060603, 0.3074074074074074, 0.30579710144927535, 0.30780141843971626, 0.3069444444444444, 0.3054421768707483, 0.30866666666666664]}, "msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.006666666666666667, 0.005555555555555555, 0.009523809523809526, 0.029166666666666667, 0.037037037037037035, 0.04000000000000001, 0.045454545454545456, 0.04722222222222222, 0.05128205128205128, 0.05476190476190476, 0.062222222222222394, 0.08125, 0.08235294117647059, 0.08703703703703702, 0.09298245614035087, 0.1, 0.11428571428571427, 0.11818181818181818, 0.1289855072463768, 0.13472222222222233, 0.14133333333333334, 0.14487179487179488, 0.15185185185185185, 0.15595238095238095, 0.16321839080459769, 0.1711111111111111, 0.17956989247311828, 0.18854166666666666, 0.19696969696969696, 0.2, 0.2066666666666668, 0.212037037037037, 0.21711711711711712, 0.22631578947368422, 0.23504273504273504, 0.24083333333333334, 0.24634146341463425, 0.25555555555555565, 0.26124031007751936, 0.2689393939393939, 0.28074074074074074, 0.28695652173913044, 0.2914893617021276, 0.2965277777777777, 0.3006802721088435, 0.30866666666666664]}, "temp_msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.019047619047619053, 0.025, 0.02962962962962963, 0.030000000000000002, 0.045454545454545456, 0.05, 0.05384615384615385, 0.06428571428571428, 0.06666666666666665, 0.07083333333333333, 0.0784313725490196, 0.08888888888888888, 0.09649122807017543, 0.10333333333333333, 0.11904761904761903, 0.12272727272727273, 0.13478260869565217, 0.13888888888888887, 0.144, 0.14871794871794872, 0.15432098765432098, 0.1630952380952381, 0.1701149425287356, 0.17333333333333334, 0.18064516129032257, 0.1875, 0.1919191919191919, 0.1931372549019608, 0.19999999999999998, 0.21018518518518517, 0.21711711711711712, 0.22631578947368422, 0.23162393162393163, 0.23666666666666666, 0.24065040650406502, 0.2476190476190476, 0.2527131782945736, 0.25984848484848483, 0.2674074074074074, 0.27608695652173915, 0.28439716312056734, 0.29375000000000007, 0.3, 0.30866666666666664]}, "entropy": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.006666666666666667, 0.005555555555555555, 0.009523809523809526, 0.029166666666666667, 0.037037037037037035, 0.04000000000000001, 0.045454545454545456, 0.04722222222222222, 0.05128205128205128, 0.05476190476190476, 0.06444444444444443, 0.07916666666666666, 0.08235294117647059, 0.08703703703703702, 0.09298245614035087, 0.1, 0.11269841269841269, 0.11818181818181818, 0.12753623188405797, 0.13472222222222233, 0.14266666666666666, 0.14615384615384616, 0.15185185185185185, 0.15714285714285714, 0.1597701149425289, 0.1688888888888889, 0.17956989247311828, 0.1875, 0.19595959595959597, 0.1980392156862745, 0.20380952380952394, 0.20833333333333331, 0.2153153153153153, 0.2236842105263158, 0.23247863247863249, 0.23916666666666667, 0.24715447154471543, 0.2523809523809525, 0.2596899224806202, 0.26666666666666666, 0.27185185185185184, 0.2753623188405797, 0.2836879432624115, 0.2923611111111112, 0.3020408163265306, 0.30866666666666664]}, "softmax_learned": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.019047619047619053, 0.025, 0.02962962962962963, 0.030000000000000002, 0.045454545454545456, 0.05, 0.05384615384615385, 0.06428571428571428, 0.06666666666666665, 0.07083333333333333, 0.0803921568627451, 0.09074074074074073, 0.09473684210526316, 0.10666666666666667, 0.11587301587301586, 0.12575757575757576, 0.1318840579710145, 0.13750000000000012, 0.14266666666666666, 0.15, 0.1580246913580247, 0.1619047619047619, 0.16896551724137948, 0.17555555555555555, 0.1860215053763441, 0.190625, 0.19494949494949496, 0.2, 0.20380952380952377, 0.21018518518518517, 0.218018018018018, 0.22543859649122808, 0.23162393162393163, 0.23666666666666666, 0.24227642276422762, 0.24841269841269853, 0.25193798449612403, 0.25984848484848483, 0.2688888888888889, 0.27608695652173915, 0.28581560283687957, 0.29513888888888884, 0.3013605442176871, 0.30866666666666664]}, "geom_softmax": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.008333333333333333, 0.006666666666666667, 0.005555555555555555, 0.009523809523809526, 0.016666666666666666, 0.018518518518518517, 0.036666666666666674, 0.045454545454545456, 0.05, 0.05384615384615385, 0.06428571428571428, 0.07333333333333332, 0.075, 0.08235294117647059, 0.08888888888888888, 0.09473684210526316, 0.10666666666666667, 0.11587301587301586, 0.12121212121212122, 0.13623188405797101, 0.138888888888889, 0.14533333333333334, 0.14743589743589744, 0.15308641975308643, 0.1630952380952381, 0.1701149425287356, 0.17777777777777778, 0.17956989247311828, 0.18645833333333334, 0.19292929292929292, 0.19705882352941176, 0.20190476190476186, 0.21018518518518517, 0.21981981981981982, 0.22719298245614036, 0.2358974358974359, 0.24083333333333334, 0.24715447154471543, 0.25396825396825407, 0.262015503875969, 0.26515151515151514, 0.2733333333333333, 0.27898550724637683, 0.2879432624113475, 0.2944444444444444, 0.3013605442176871, 0.30866666666666664]}}, "ltt": {"alpha": 0.1, "delta": 0.05, "lambda_hat": 0.10002331995594467, "coverage": 0.05533333333333333, "abstain_rate": 0.9446666666666667, "selective_risk": 0.04819277108433735, "selective_accuracy": 0.9518072289156626, "risk_within_budget": true}, "coverage_validity": {"alpha_grid": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "target": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "achieved_risk": [0.0, 0.0, 0.04819277108433735, 0.09177215189873418, 0.14555765595463138, 0.257781324820431], "coverage": [0.0, 0.0, 0.05533333333333333, 0.21066666666666667, 0.3526666666666667, 0.8353333333333334]}, "feature_diagnostics": {"names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual"], "auroc": {"basin/rho": 0.6105604512101905, "basin/entropy": 0.3894051831687602, "basin/dispersion": 0.4935923737479979, "energy/mean": 0.32511127171542764, "energy/min": 0.34090696080861266, "energy/std": 0.49929498407726225, "curv/lmax_mean": 0.460030699954804, "curv/lmax_best": 0.4852384036856608, "curv/trace_mean": 0.45787191412343714, "curv/trace_best": 0.44889311458747716, "dynamics/steps": 0.5, "dynamics/monotonic": 0.6295333981767476, "dynamics/drop": 0.6543776594304471, "dynamics/residual": 0.504791400680231}, "hist": {"basin/rho": {"edges": [0.3333333333333333, 0.36666666666666664, 0.4, 0.43333333333333335, 0.4666666666666667, 0.5, 0.5333333333333333, 0.5666666666666667, 0.6000000000000001, 0.6333333333333333, 0.6666666666666667, 0.7000000000000001, 0.7333333333333334, 0.7666666666666667, 0.8, 0.8333333333333335, 0.8666666666666667, 0.9000000000000001, 0.9333333333333333, 0.9666666666666668, 1.0], "correct_counts": [0, 0, 0, 0, 0, 10, 0, 15, 0, 19, 0, 0, 14, 0, 20, 0, 0, 38, 0, 921], "incorrect_counts": [1, 0, 0, 0, 0, 12, 0, 16, 0, 18, 0, 0, 18, 0, 41, 0, 0, 49, 0, 308]}, "basin/entropy": {"edges": [0.0, 0.05493061443340548, 0.10986122886681096, 0.16479184330021646, 0.21972245773362192, 0.2746530721670274, 0.3295836866004329, 0.3845143010338384, 0.43944491546724385, 0.4943755299006493, 0.5493061443340548, 0.6042367587674603, 0.6591673732008658, 0.7140979876342712, 0.7690286020676768, 0.8239592165010822, 0.8788898309344877, 0.9338204453678932, 0.9887510598012986, 1.0436816742347041, 1.0986122886681096], "correct_counts": [921, 0, 0, 0, 0, 38, 0, 0, 20, 0, 13, 18, 22, 1, 0, 1, 2, 1, 0, 0], "incorrect_counts": [308, 0, 0, 0, 0, 49, 0, 0, 41, 0, 17, 18, 20, 1, 0, 0, 6, 1, 1, 1]}, "basin/dispersion": {"edges": [1.6302329015759547, 1.6519526525410289, 1.6736724035061028, 1.6953921544711767, 1.717111905436251, 1.738831656401325, 1.760551407366399, 1.782271158331473, 1.8039909092965472, 1.8257106602616213, 1.8474304112266953, 1.8691501621917692, 1.8908699131568434, 1.9125896641219176, 1.9343094150869915, 1.9560291660520654, 1.9777489170171396, 1.9994686679822138, 2.021188418947288, 2.0429081699123617, 2.064627920877436], "correct_counts": [3, 7, 19, 24, 43, 69, 85, 117, 140, 123, 109, 96, 71, 49, 39, 21, 8, 11, 2, 1], "incorrect_counts": [3, 2, 5, 13, 20, 26, 34, 70, 48, 47, 61, 42, 37, 21, 14, 12, 3, 2, 2, 1]}, "energy/mean": {"edges": [-0.2394584914048513, -0.18958720415830613, -0.13971591691176097, -0.08984462966521581, -0.03997334241867065, 0.00989794482787451, 0.05976923207441967, 0.10964051932096483, 0.15951180656751, 0.20938309381405515, 0.2592543810606003, 0.3091256683071454, 0.3589969555536906, 0.4088682428002358, 0.4587395300467809, 0.508610817293326, 0.5584821045398712, 0.6083533917864165, 0.6582246790329616, 0.7080959662795067, 0.7579672535260519], "correct_counts": [1, 2, 17, 33, 48, 57, 92, 101, 101, 103, 96, 84, 85, 75, 68, 32, 24, 12, 3, 3], "incorrect_counts": [0, 0, 3, 0, 3, 12, 13, 34, 37, 34, 43, 35, 55, 53, 47, 40, 31, 13, 6, 4]}, "energy/min": {"edges": [-0.5504263639450073, -0.4960724890232086, -0.4417186141014099, -0.38736473917961123, -0.3330108642578125, -0.2786569893360138, -0.22430311441421513, -0.16994923949241642, -0.1155953645706177, -0.06124148964881898, -0.006887614727020264, 0.0474662601947784, 0.10182013511657706, 0.15617401003837583, 0.2105278849601745, 0.26488175988197327, 0.31923563480377193, 0.3735895097255706, 0.42794338464736936, 0.48229725956916814, 0.5366511344909668], "correct_counts": [5, 10, 14, 45, 55, 80, 92, 92, 122, 96, 106, 90, 70, 53, 51, 31, 16, 3, 4, 2], "incorrect_counts": [0, 0, 0, 7, 7, 11, 33, 38, 24, 44, 49, 50, 60, 56, 31, 25, 13, 7, 6, 2]}, "energy/std": {"edges": [0.07470156430161684, 0.08989143155875684, 0.10508129881589684, 0.12027116607303683, 0.13546103333017684, 0.15065090058731684, 0.1658407678444568, 0.1810306351015968, 0.1962205023587368, 0.2114103696158768, 0.2266002368730168, 0.2417901041301568, 0.2569799713872968, 0.2721698386444368, 0.2873597059015768, 0.3025495731587168, 0.3177394404158568, 0.3329293076729968, 0.3481191749301368, 0.3633090421872768, 0.3784989094444168], "correct_counts": [6, 12, 28, 49, 74, 114, 143, 136, 129, 114, 85, 68, 39, 21, 5, 2, 7, 3, 1, 1], "incorrect_counts": [1, 0, 7, 21, 47, 48, 58, 69, 65, 51, 29, 26, 17, 9, 5, 5, 1, 1, 1, 2]}, "curv/lmax_mean": {"edges": [1.000376323858897, 1.0019486089547476, 1.0035208940505982, 1.0050931791464488, 1.0066654642422994, 1.00823774933815, 1.0098100344340006, 1.0113823195298512, 1.0129546046257019, 1.0145268897215525, 1.0160991748174033, 1.017671459913254, 1.0192437450091045, 1.0208160301049551, 1.0223883152008058, 1.0239606002966564, 1.025532885392507, 1.0271051704883576, 1.0286774555842082, 1.0302497406800588, 1.0318220257759094], "correct_counts": [105, 233, 210, 179, 108, 77, 40, 38, 17, 14, 6, 6, 0, 2, 1, 0, 0, 0, 0, 1], "incorrect_counts": [26, 89, 110, 82, 57, 35, 24, 16, 7, 5, 7, 1, 1, 3, 0, 0, 0, 0, 0, 0]}, "curv/lmax_best": {"edges": [1.0, 1.0040147721767425, 1.008029544353485, 1.0120443165302277, 1.0160590887069703, 1.0200738608837128, 1.0240886330604553, 1.0281034052371978, 1.0321181774139405, 1.036132949590683, 1.0401477217674255, 1.044162493944168, 1.0481772661209106, 1.0521920382976533, 1.0562068104743958, 1.0602215826511383, 1.0642363548278808, 1.0682511270046233, 1.072265899181366, 1.0762806713581086, 1.080295443534851], "correct_counts": [618, 188, 73, 67, 29, 18, 13, 8, 9, 5, 3, 3, 0, 0, 1, 1, 0, 0, 0, 1], "incorrect_counts": [273, 79, 45, 17, 16, 9, 11, 4, 2, 5, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0]}, "curv/trace_mean": {"edges": [32.01093864440918, 32.016456858317056, 32.02197507222493, 32.02749328613281, 32.03301150004069, 32.03852971394857, 32.04404792785645, 32.04956614176432, 32.0550843556722, 32.06060256958008, 32.06612078348796, 32.07163899739584, 32.077157211303714, 32.08267542521159, 32.08819363911947, 32.093711853027344, 32.09923006693522, 32.104748280843104, 32.11026649475098, 32.11578470865886, 32.121302922566734], "correct_counts": [4, 3, 28, 47, 70, 105, 117, 148, 132, 120, 77, 60, 45, 29, 22, 13, 9, 3, 3, 2], "incorrect_counts": [0, 0, 4, 15, 26, 42, 52, 66, 59, 61, 50, 30, 20, 11, 10, 7, 4, 3, 2, 1]}, "curv/trace_best": {"edges": [32.004268646240234, 32.017548370361325, 32.03082809448242, 32.044107818603514, 32.05738754272461, 32.0706672668457, 32.083946990966794, 32.09722671508789, 32.11050643920898, 32.12378616333008, 32.13706588745117, 32.15034561157226, 32.16362533569336, 32.17690505981445, 32.19018478393555, 32.20346450805664, 32.21674423217773, 32.23002395629883, 32.24330368041992, 32.25658340454102, 32.26986312866211], "correct_counts": [57, 188, 204, 172, 154, 85, 58, 46, 29, 18, 10, 5, 5, 4, 0, 1, 0, 0, 0, 1], "incorrect_counts": [19, 73, 80, 73, 62, 48, 48, 14, 16, 11, 9, 6, 2, 1, 0, 0, 0, 0, 1, 0]}, "dynamics/steps": {"edges": [1.0, 1.05, 1.1, 1.15, 1.2, 1.25, 1.3, 1.35, 1.4, 1.45, 1.5, 1.55, 1.6, 1.65, 1.7000000000000002, 1.75, 1.8, 1.85, 1.9, 1.9500000000000002, 2.0], "correct_counts": [1037, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [463, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "dynamics/monotonic": {"edges": [0.64, 0.6465833333333334, 0.6531666666666667, 0.6597500000000001, 0.6663333333333333, 0.6729166666666667, 0.6795000000000001, 0.6860833333333334, 0.6926666666666668, 0.69925, 0.7058333333333334, 0.7124166666666668, 0.7190000000000001, 0.7255833333333335, 0.7321666666666667, 0.7387500000000001, 0.7453333333333335, 0.7519166666666668, 0.7585000000000002, 0.7650833333333336, 0.7716666666666668], "correct_counts": [0, 1, 6, 33, 38, 72, 88, 131, 131, 143, 112, 110, 69, 46, 27, 15, 9, 1, 3, 2], "incorrect_counts": [1, 4, 8, 23, 37, 38, 59, 79, 63, 52, 40, 32, 11, 6, 4, 4, 1, 0, 1, 0]}, "dynamics/drop": {"edges": [13.908043444156647, 21.40915985889733, 28.91027627363801, 36.41139268837869, 43.912509103119376, 51.413625517860055, 58.914741932600734, 66.41585834734141, 73.9169747620821, 81.41809117682278, 88.91920759156346, 96.42032400630414, 103.92144042104482, 111.4225568357855, 118.92367325052619, 126.42478966526687, 133.92590608000756, 141.42702249474823, 148.92813890948892, 156.4292553242296, 163.93037173897028], "correct_counts": [208, 321, 228, 142, 66, 39, 18, 7, 1, 2, 2, 1, 0, 0, 1, 0, 0, 0, 0, 1], "incorrect_counts": [171, 167, 74, 36, 8, 5, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "dynamics/residual": {"edges": [1.1329888900121052, 1.1492741659283636, 1.1655594418446222, 1.1818447177608808, 1.1981299936771392, 1.2144152695933976, 1.2307005455096562, 1.2469858214259149, 1.2632710973421732, 1.2795563732584316, 1.2958416491746902, 1.3121269250909489, 1.3284122010072072, 1.3446974769234656, 1.3609827528397243, 1.3772680287559829, 1.3935533046722413, 1.4098385805884996, 1.4261238565047583, 1.4424091324210169, 1.4586944083372753], "correct_counts": [2, 2, 8, 21, 32, 71, 96, 106, 140, 134, 137, 98, 88, 51, 28, 13, 7, 2, 0, 1], "incorrect_counts": [0, 2, 5, 9, 17, 29, 50, 50, 43, 70, 55, 47, 42, 21, 15, 5, 1, 1, 0, 1]}}}, "feature_ablation": {"full": 0.17444436593752588, "drop_basin": 0.19274098660894265, "basin_only": 0.25332924341284363, "drop_energy": 0.18924590060362054, "energy_only": 0.19443145021840322, "drop_curv": 0.17517378968044742, "curv_only": 0.2700948061609721, "drop_dynamics": 0.1729185443228303, "dynamics_only": 0.20883041928582574}, "ece_geometry": 0.026691927047032264, "accuracy_ood": 0.256, "aurc_ood": {"geometry": 0.7974446165188469, "rho_basin": 0.7490819589151309, "energy_min": 0.739655450497817, "energy_mean": 0.7320500499695907, "energy_std": 0.7382724006497628, "msp": 0.8071610870796354, "temp_msp": 0.7964482735224696, "entropy": 0.8055287758926124, "softmax_learned": 0.7951769722870716, "geom_softmax": 0.7840875769994762}, "ood_ltt": {"alpha": 0.1, "lambda_hat": 0.10002331995594467, "selective_risk": 0.8401486988847584, "coverage": 0.17933333333333334, "risk_within_budget": false}, "ood_validity": {"target": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "id_risk": [0.0, 0.0, 0.04819277108433735, 0.09177215189873418, 0.14555765595463138, 0.257781324820431], "ood_risk": [0.0, 0.0, 0.8401486988847584, 0.7979094076655052, 0.7672955974842768, 0.7467778620166793], "id_coverage": [0.0, 0.0, 0.05533333333333333, 0.21066666666666667, 0.3526666666666667, 0.8353333333333334], "ood_coverage": [0.0, 0.0, 0.17933333333333334, 0.38266666666666665, 0.53, 0.8793333333333333]}}, "env": {"python": "3.12.13", "platform": "macOS-26.5.2-arm64-arm-64bit", "jax": "0.11.0", "jax_backend": "cpu"}, "artifact_paths": []} +{"run_id": "5bd7ea239e4f", "timestamp": "2026-07-30T11:18:10.314612+00:00", "git_sha": "67dc7a2", "config_hash": "0c62cba9cde6", "config": {"run": {"seed": 3, "task": "arithmetic", "notes": "E1 selective-prediction main run"}, "model": {"latent_dim": 32, "hidden_dim": 128, "context_dim": 64}, "inference": {"k_restarts": 12, "steps": 50, "step_size": 0.1, "temperature": 0.05, "init_scale": 1.0, "grad_tol": 0.001, "sampler": "langevin", "anneal_levels": 10, "anneal_steps_per_level": 8, "anneal_step_max": 0.2, "anneal_step_min": 0.01}, "train": {"epochs": 7, "batch_size": 128, "lr": 0.001, "n_train": 6000, "n_neg": 4, "neg_noise": 0.5, "objective": "basin_center", "ired_noise_min": 0.1, "ired_noise_max": 1.5, "ired_ridge": 0.05, "ired_decode_weight": 1.0, "ired_stat_weight": 1.0}, "eval": {"n_eval": 1500, "richer_geometry": false}, "conformal": {"alpha": 0.1, "delta": 0.05, "n_calib": 1500}, "task": {"arithmetic": {"modular": false, "n_operands": 6, "ood_n_operands": 6, "max_operand": 7, "ood_max_operand": 9}}}, "task": "arithmetic", "split": "selective", "seed": 3, "metrics": {"n_fit": 1500, "n_calib": 1500, "n_test": 1500, "k_restarts": 12, "objective": "basin_center", "sampler": "langevin", "feature_set": "base", "accuracy_id": 0.8166666666666667, "base_error": 0.18333333333333332, "final_train_loss": 0.943260669708252, "feature_names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual"], "aurc": {"geometry": 0.05870278336778761, "rho_basin": 0.09557280117969337, "energy_min": 0.23308855480058674, "energy_mean": 0.2321489497773461, "energy_std": 0.18295923460701716, "msp": 0.05696229801237003, "temp_msp": 0.05587579661080262, "entropy": 0.07586100765649918, "softmax_learned": 0.05495199413638246, "geom_softmax": 0.050518887585592184}, "temperature": 0.3191283771478964, "best_energy_baseline": "energy_std", "best_baseline": "temp_msp", "delta_aurc_vs_energy_min": [0.17438577143279912, 0.1451627099062947, 0.20605252294076107], "delta_aurc_vs_best_energy": [0.12425645123922954, 0.1000795529758454, 0.14889441517869045], "delta_aurc_vs_best_baseline": [-0.002826986756984985, -0.013129993995088966, 0.00820655055054469], "delta_aurc_geom_adds": [0.004433106550790274, -6.164051281485425e-05, 0.010993126415850296], "geometry_wins": true, "geometry_wins_vs_baseline": false, "geometry_adds_over_softmax": false, "risk_coverage": {"geometry": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.016666666666666666, 0.011111111111111112, 0.025, 0.02, 0.022222222222222334, 0.03333333333333334, 0.0375, 0.03333333333333333, 0.030000000000000002, 0.030303030303030304, 0.030555555555555555, 0.028205128205128206, 0.02857142857142857, 0.026666666666666665, 0.025, 0.023529411764705882, 0.027777777777777773, 0.028070175438596492, 0.028333333333333332, 0.033333333333333326, 0.031818181818181815, 0.030434782608695653, 0.03472222222222222, 0.03333333333333333, 0.03333333333333333, 0.03333333333333333, 0.03214285714285714, 0.03678160919540229, 0.04, 0.043010752688172046, 0.05, 0.05656565656565657, 0.061764705882352944, 0.06857142857142856, 0.07499999999999998, 0.08288288288288288, 0.09210526315789473, 0.09658119658119659, 0.1025, 0.1089430894308943, 0.11746031746031757, 0.12558139534883722, 0.1340909090909091, 0.14222222222222222, 0.1492753623188406, 0.15886524822695047, 0.1659722222222222, 0.1727891156462585, 0.18333333333333332]}, "rho_basin": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.08056394763343407, 0.08056394763343401, 0.08056394763343389, 0.08056394763343398, 0.08056394763343411, 0.08056394763343418, 0.08056394763343414, 0.08056394763343395, 0.08056394763343383, 0.08056394763343372, 0.08056394763343364, 0.08056394763343357, 0.08056394763343351, 0.08056394763343364, 0.08056394763343382, 0.08056394763343397, 0.08056394763343411, 0.08056394763343423, 0.08056394763343434, 0.08056394763343444, 0.08056394763343454, 0.08056394763343462, 0.08056394763343469, 0.08056394763343476, 0.08056394763343482, 0.08056394763343488, 0.08056394763343493, 0.08056394763343497, 0.08056394763343502, 0.08056394763343507, 0.0805639476334351, 0.08056394763343515, 0.08056394763343518, 0.08431372549019736, 0.08825396825396971, 0.09197530864197688, 0.09549549549549718, 0.09883040935672696, 0.10364611747590659, 0.10983156028368969, 0.11571527417401978, 0.12196428571428705, 0.12901162790697826, 0.13573863636363812, 0.14465408805031627, 0.15422477440525198, 0.16295752832803626, 0.17129629629629786, 0.1791748957647589, 0.18333333333333446]}, "energy_min": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.43333333333333335, 0.4666666666666667, 0.45555555555555555, 0.43333333333333335, 0.41333333333333333, 0.3722222222222222, 0.35714285714285704, 0.35, 0.3296296296296296, 0.31000000000000005, 0.2909090909090909, 0.275, 0.258974358974359, 0.25952380952380955, 0.24666666666666662, 0.24166666666666667, 0.23921568627450981, 0.2259259259259259, 0.22280701754385965, 0.22, 0.21428571428571425, 0.2106060606060606, 0.2028985507246377, 0.1958333333333333, 0.18933333333333333, 0.18717948717948718, 0.18271604938271604, 0.17857142857142858, 0.1758620689655174, 0.17555555555555555, 0.17311827956989248, 0.16979166666666667, 0.16767676767676767, 0.16372549019607843, 0.16190476190476188, 0.16018518518518515, 0.15945945945945947, 0.15789473684210525, 0.15726495726495726, 0.155, 0.15609756097560987, 0.15555555555555553, 0.1573643410852713, 0.15984848484848485, 0.16074074074074074, 0.16231884057971013, 0.1617021276595746, 0.16805555555555565, 0.17482993197278912, 0.18333333333333332]}, "energy_mean": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.43333333333333335, 0.43333333333333335, 0.43333333333333335, 0.39166666666666666, 0.38, 0.3722222222222223, 0.37619047619047624, 0.3458333333333333, 0.32592592592592595, 0.3133333333333334, 0.296969696969697, 0.28888888888888886, 0.2743589743589744, 0.2619047619047619, 0.2533333333333333, 0.24583333333333332, 0.2411764705882353, 0.2333333333333333, 0.22456140350877193, 0.22333333333333333, 0.2126984126984128, 0.20606060606060606, 0.19855072463768117, 0.19444444444444453, 0.19066666666666668, 0.18846153846153846, 0.18518518518518517, 0.18095238095238095, 0.17816091954022986, 0.17555555555555555, 0.17419354838709677, 0.16979166666666667, 0.16767676767676767, 0.16470588235294117, 0.16190476190476188, 0.16018518518518515, 0.16036036036036036, 0.16052631578947368, 0.1606837606837607, 0.1575, 0.15772357723577232, 0.15555555555555553, 0.15271317829457365, 0.15378787878787878, 0.15703703703703703, 0.1608695652173913, 0.16241134751773048, 0.16527777777777786, 0.17551020408163265, 0.18333333333333332]}, "energy_std": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.2, 0.2, 0.2, 0.2, 0.18, 0.16666666666666674, 0.16666666666666669, 0.175, 0.17407407407407408, 0.1766666666666667, 0.17575757575757575, 0.17222222222222222, 0.17435897435897435, 0.1738095238095238, 0.17111111111111107, 0.17291666666666666, 0.16666666666666666, 0.17407407407407421, 0.1824561403508772, 0.18, 0.17777777777777776, 0.17575757575757575, 0.1855072463768116, 0.1833333333333333, 0.184, 0.18846153846153846, 0.18641975308641975, 0.1880952380952381, 0.18620689655172412, 0.18222222222222223, 0.1817204301075269, 0.18125, 0.18282828282828284, 0.17941176470588235, 0.17809523809523806, 0.17870370370370367, 0.1810810810810811, 0.18333333333333332, 0.18461538461538463, 0.18416666666666667, 0.18292682926829265, 0.18095238095238092, 0.17984496124031008, 0.18181818181818182, 0.18074074074074073, 0.17971014492753623, 0.18085106382978722, 0.1819444444444444, 0.18231292517006803, 0.18333333333333332]}, "msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.03333333333333333, 0.016666666666666666, 0.011111111111111112, 0.008333333333333333, 0.006666666666666667, 0.005555555555555555, 0.004761904761904763, 0.004166666666666667, 0.003703703703703704, 0.006666666666666668, 0.006060606060606061, 0.005555555555555556, 0.005128205128205128, 0.007142857142857143, 0.008888888888889071, 0.014583333333333334, 0.01764705882352941, 0.02037037037037037, 0.01929824561403509, 0.023333333333333334, 0.025396825396825393, 0.02727272727272727, 0.03333333333333333, 0.0361111111111111, 0.037333333333333336, 0.038461538461538464, 0.043209876543209874, 0.04880952380952381, 0.05057471264367815, 0.05555555555555555, 0.05913978494623656, 0.06354166666666666, 0.06868686868686869, 0.07450980392156863, 0.0790476190476192, 0.0842592592592594, 0.0936936936936937, 0.10263157894736842, 0.10854700854700855, 0.1125, 0.11951219512195134, 0.1238095238095238, 0.12868217054263567, 0.1340909090909091, 0.13925925925925925, 0.14492753623188406, 0.1546099290780143, 0.16388888888888897, 0.17210884353741496, 0.18333333333333332]}, "temp_msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.03333333333333333, 0.016666666666666666, 0.011111111111111112, 0.008333333333333333, 0.006666666666666667, 0.005555555555555555, 0.004761904761904763, 0.004166666666666667, 0.007407407407407408, 0.006666666666666668, 0.00909090909090909, 0.011111111111111112, 0.010256410256410256, 0.011904761904761904, 0.01111111111111111, 0.014583333333333334, 0.01568627450980392, 0.018518518518518514, 0.02456140350877193, 0.02666666666666667, 0.025396825396825393, 0.025757575757575757, 0.02608695652173913, 0.027777777777777773, 0.03333333333333333, 0.035897435897435895, 0.037037037037037035, 0.036904761904761905, 0.04482758620689673, 0.051111111111111114, 0.054838709677419356, 0.059375, 0.06565656565656566, 0.06862745098039216, 0.07238095238095237, 0.07870370370370369, 0.08738738738738738, 0.09736842105263158, 0.10341880341880341, 0.10583333333333333, 0.1113821138211382, 0.11904761904761903, 0.12713178294573643, 0.13257575757575757, 0.13925925925925925, 0.14565217391304347, 0.15460992907801416, 0.16180555555555554, 0.17142857142857143, 0.18333333333333332]}, "entropy": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.011111111111111112, 0.008333333333333333, 0.006666666666666667, 0.005555555555555555, 0.004761904761904763, 0.008333333333333333, 0.014814814814814815, 0.01666666666666667, 0.015151515151515152, 0.030555555555555555, 0.03333333333333333, 0.04047619047619048, 0.04666666666666666, 0.04791666666666667, 0.054901960784313725, 0.06296296296296294, 0.07192982456140351, 0.07166666666666667, 0.073015873015873, 0.07727272727272727, 0.08115942028985507, 0.08472222222222221, 0.08533333333333333, 0.0858974358974359, 0.08888888888888889, 0.0880952380952381, 0.09080459770114942, 0.09444444444444444, 0.09462365591397849, 0.09479166666666666, 0.09696969696969697, 0.09607843137254903, 0.09904761904761904, 0.09999999999999999, 0.1036036036036036, 0.10614035087719298, 0.10854700854700855, 0.11333333333333333, 0.11707317073170731, 0.12301587301587312, 0.12713178294573643, 0.1303030303030303, 0.1385185185185185, 0.1463768115942029, 0.15815602836879444, 0.16666666666666674, 0.17551020408163265, 0.18333333333333332]}, "softmax_learned": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.03333333333333333, 0.016666666666666666, 0.011111111111111112, 0.008333333333333333, 0.006666666666666667, 0.005555555555555555, 0.004761904761904763, 0.004166666666666667, 0.003703703703703704, 0.006666666666666668, 0.006060606060606061, 0.005555555555555556, 0.007692307692307693, 0.007142857142857143, 0.008888888888888887, 0.0125, 0.013725490196078431, 0.014814814814814814, 0.01929824561403509, 0.018333333333333333, 0.023809523809523805, 0.025757575757575757, 0.03188405797101449, 0.033333333333333326, 0.034666666666666665, 0.03717948717948718, 0.04197530864197531, 0.04404761904761905, 0.04712643678160919, 0.04777777777777778, 0.054838709677419356, 0.058333333333333334, 0.0595959595959596, 0.06568627450980392, 0.07142857142857158, 0.07777777777777777, 0.08828828828828829, 0.09824561403508772, 0.10512820512820513, 0.11166666666666666, 0.11544715447154484, 0.1222222222222222, 0.12868217054263567, 0.1340909090909091, 0.13925925925925925, 0.14492753623188406, 0.1496453900709221, 0.16319444444444453, 0.17210884353741496, 0.18333333333333332]}, "geom_softmax": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.004166666666666667, 0.003703703703703704, 0.003333333333333334, 0.0030303030303030303, 0.002777777777777778, 0.002564102564102564, 0.002380952380952381, 0.004444444444444444, 0.00625, 0.01568627450980392, 0.014814814814814814, 0.015789473684210527, 0.018333333333333333, 0.020634920634920763, 0.022727272727272728, 0.02608695652173913, 0.029166666666666664, 0.032, 0.03076923076923077, 0.0345679012345679, 0.04047619047619048, 0.04367816091954022, 0.04666666666666667, 0.05268817204301075, 0.05520833333333333, 0.06060606060606061, 0.06568627450980392, 0.06952380952380968, 0.07870370370370369, 0.08198198198198198, 0.09210526315789473, 0.09658119658119659, 0.10416666666666667, 0.1105691056910569, 0.11666666666666665, 0.12015503875968993, 0.12878787878787878, 0.1385185185185185, 0.1463768115942029, 0.1546099290780143, 0.16388888888888886, 0.17346938775510204, 0.18333333333333332]}}, "ltt": {"alpha": 0.1, "delta": 0.05, "lambda_hat": 0.19510629211616212, "coverage": 0.7346666666666667, "abstain_rate": 0.2653333333333333, "selective_risk": 0.08076225045372051, "selective_accuracy": 0.9192377495462795, "risk_within_budget": true}, "coverage_validity": {"alpha_grid": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "target": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "achieved_risk": [0.0, 0.03294117647058824, 0.08076225045372051, 0.13363705391040243, 0.18333333333333332, 0.18333333333333332], "coverage": [0.0, 0.5666666666666667, 0.7346666666666667, 0.878, 1.0, 1.0]}, "feature_diagnostics": {"names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual"], "auroc": {"basin/rho": 0.753239332096475, "basin/entropy": 0.2466538033395176, "basin/dispersion": 0.48666716141001853, "energy/mean": 0.5139591836734694, "energy/min": 0.512973654916512, "energy/std": 0.4924705009276438, "curv/lmax_mean": 0.4856994434137291, "curv/lmax_best": 0.5174768089053803, "curv/trace_mean": 0.48462486085343226, "curv/trace_best": 0.49396066790352505, "dynamics/steps": 0.5, "dynamics/monotonic": 0.3888029684601113, "dynamics/drop": 0.36415881261595545, "dynamics/residual": 0.48698775510204084}, "hist": {"basin/rho": {"edges": [0.5, 0.525, 0.55, 0.575, 0.6, 0.625, 0.65, 0.675, 0.7, 0.725, 0.75, 0.775, 0.8, 0.825, 0.8500000000000001, 0.875, 0.9, 0.925, 0.95, 0.9750000000000001, 1.0], "correct_counts": [19, 0, 0, 38, 0, 0, 22, 0, 0, 0, 46, 0, 0, 61, 0, 0, 126, 0, 0, 913], "incorrect_counts": [12, 0, 0, 49, 0, 0, 31, 0, 0, 0, 34, 0, 0, 33, 0, 0, 36, 0, 0, 80]}, "basin/entropy": {"edges": [0.0, 0.053756966200268666, 0.10751393240053733, 0.161270898600806, 0.21502786480107466, 0.2687848310013433, 0.322541797201612, 0.3762987634018807, 0.43005572960214933, 0.483812695802418, 0.5375696620026866, 0.5913266282029553, 0.645083594403224, 0.6988405606034926, 0.7525975268037614, 0.80635449300403, 0.8601114592042987, 0.9138684254045674, 0.967625391604836, 1.0213823578051047, 1.0751393240053733], "correct_counts": [913, 0, 0, 0, 0, 126, 0, 0, 58, 0, 44, 21, 52, 5, 0, 1, 2, 3, 0, 0], "incorrect_counts": [80, 0, 0, 0, 0, 36, 0, 0, 30, 0, 33, 29, 55, 4, 0, 2, 2, 2, 0, 2]}, "basin/dispersion": {"edges": [1.6335429597808506, 1.6563980282608786, 1.6792530967409065, 1.7021081652209342, 1.7249632337009622, 1.74781830218099, 1.770673370661018, 1.793528439141046, 1.8163835076210737, 1.8392385761011016, 1.8620936445811296, 1.8849487130611575, 1.9078037815411855, 1.9306588500212132, 1.9535139185012411, 1.976368986981269, 1.999224055461297, 2.022079123941325, 2.0449341924213527, 2.067789260901381, 2.0906443293814085], "correct_counts": [4, 17, 24, 49, 72, 120, 146, 160, 154, 138, 136, 83, 51, 36, 13, 11, 7, 3, 0, 1], "incorrect_counts": [2, 5, 6, 11, 12, 20, 36, 41, 31, 32, 28, 18, 14, 9, 5, 4, 1, 0, 0, 0]}, "energy/mean": {"edges": [0.2628147800763448, 0.32859045515457785, 0.394366130232811, 0.4601418053110441, 0.5259174803892772, 0.5916931554675102, 0.6574688305457433, 0.7232445056239765, 0.7890201807022095, 0.8547958557804427, 0.9205715308586757, 0.9863472059369087, 1.052122881015142, 1.117898556093375, 1.183674231171608, 1.2494499062498412, 1.3152255813280742, 1.3810012564063072, 1.4467769314845405, 1.5125526065627735, 1.5783282816410065], "correct_counts": [4, 11, 34, 36, 54, 72, 130, 190, 248, 209, 112, 45, 20, 17, 12, 10, 5, 10, 4, 2], "incorrect_counts": [4, 9, 24, 18, 26, 14, 19, 22, 26, 26, 13, 16, 8, 4, 5, 7, 9, 16, 7, 2]}, "energy/min": {"edges": [-0.1004953682422638, -0.031788365542888636, 0.03691863715648652, 0.10562563985586168, 0.17433264255523684, 0.24303964525461197, 0.31174664795398715, 0.38045365065336234, 0.44916065335273747, 0.5178676560521126, 0.5865746587514877, 0.655281661450863, 0.7239886641502381, 0.7926956668496132, 0.8614026695489885, 0.9301096722483635, 0.9988166749477387, 1.067523677647114, 1.136230680346489, 1.2049376830458642, 1.2736446857452393], "correct_counts": [2, 5, 20, 30, 59, 66, 94, 170, 219, 219, 154, 79, 37, 25, 15, 4, 8, 10, 6, 3], "incorrect_counts": [2, 5, 15, 26, 20, 21, 11, 28, 20, 22, 22, 22, 10, 2, 12, 3, 13, 10, 9, 2]}, "energy/std": {"edges": [0.0727182563720062, 0.08970085308900838, 0.10668344980601056, 0.12366604652301275, 0.14064864324001491, 0.1576312399570171, 0.1746138366740193, 0.19159643339102148, 0.20857903010802364, 0.22556162682502584, 0.24254422354202804, 0.2595268202590302, 0.2765094169760324, 0.29349201369303457, 0.31047461041003677, 0.3274572071270389, 0.3444398038440411, 0.3614224005610433, 0.3784049972780455, 0.39538759399504764, 0.41237019071204983], "correct_counts": [3, 12, 31, 78, 116, 175, 179, 175, 161, 122, 66, 39, 36, 17, 7, 4, 2, 1, 0, 1], "incorrect_counts": [1, 4, 5, 17, 25, 32, 50, 37, 36, 23, 15, 11, 8, 7, 2, 1, 0, 0, 0, 1]}, "curv/lmax_mean": {"edges": [0.9985763331254324, 0.9991151432196299, 0.9996539533138276, 1.0001927634080252, 1.0007315735022226, 1.0012703835964203, 1.001809193690618, 1.0023480037848156, 1.002886813879013, 1.0034256239732107, 1.0039644340674083, 1.0045032441616057, 1.0050420542558034, 1.005580864350001, 1.0061196744441987, 1.0066584845383961, 1.0071972946325938, 1.0077361047267914, 1.008274914820989, 1.0088137249151865, 1.0093525350093842], "correct_counts": [5, 75, 353, 340, 144, 100, 71, 43, 38, 24, 11, 8, 3, 3, 3, 2, 0, 0, 1, 1], "incorrect_counts": [2, 42, 63, 45, 18, 19, 13, 16, 14, 11, 3, 8, 9, 4, 2, 1, 4, 0, 0, 1]}, "curv/lmax_best": {"edges": [0.9949193000793457, 0.9967200636863709, 0.998520827293396, 1.000321590900421, 1.0021223545074462, 1.0039231181144714, 1.0057238817214966, 1.0075246453285218, 1.0093254089355468, 1.011126172542572, 1.0129269361495972, 1.0147276997566224, 1.0165284633636475, 1.0183292269706725, 1.0201299905776977, 1.021930754184723, 1.023731517791748, 1.0255322813987733, 1.0273330450057983, 1.0291338086128234, 1.0309345722198486], "correct_counts": [2, 7, 798, 299, 53, 28, 20, 8, 3, 2, 2, 3, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [0, 5, 168, 50, 14, 12, 7, 5, 6, 1, 2, 2, 0, 0, 1, 0, 0, 1, 0, 1]}, "curv/trace_mean": {"edges": [31.87547731399536, 31.885582343737283, 31.89568737347921, 31.90579240322113, 31.915897432963053, 31.92600246270498, 31.9361074924469, 31.946212522188823, 31.956317551930745, 31.96642258167267, 31.976527611414593, 31.986632641156515, 31.99673767089844, 32.00684270064036, 32.016947730382284, 32.02705276012421, 32.03715778986613, 32.04726281960806, 32.05736784934998, 32.0674728790919, 32.077577908833824], "correct_counts": [1, 4, 3, 10, 7, 8, 13, 5, 9, 18, 31, 112, 275, 293, 177, 112, 112, 28, 6, 1], "incorrect_counts": [1, 7, 7, 14, 9, 5, 1, 4, 6, 9, 12, 14, 26, 39, 16, 30, 32, 25, 14, 4]}, "curv/trace_best": {"edges": [31.76333999633789, 31.782129287719727, 31.800918579101562, 31.8197078704834, 31.838497161865234, 31.85728645324707, 31.876075744628906, 31.894865036010742, 31.913654327392578, 31.932443618774414, 31.95123291015625, 31.970022201538086, 31.988811492919922, 32.00760078430176, 32.026390075683594, 32.04517936706543, 32.063968658447266, 32.0827579498291, 32.10154724121094, 32.12033653259277, 32.13912582397461], "correct_counts": [0, 0, 1, 1, 2, 4, 6, 5, 8, 24, 28, 67, 429, 398, 160, 69, 15, 6, 2, 0], "incorrect_counts": [1, 0, 0, 1, 3, 4, 8, 7, 8, 13, 11, 20, 49, 61, 38, 23, 19, 7, 1, 1]}, "dynamics/steps": {"edges": [1.0, 1.05, 1.1, 1.15, 1.2, 1.25, 1.3, 1.35, 1.4, 1.45, 1.5, 1.55, 1.6, 1.65, 1.7000000000000002, 1.75, 1.8, 1.85, 1.9, 1.9500000000000002, 2.0], "correct_counts": [1225, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [275, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "dynamics/monotonic": {"edges": [0.7166666666666667, 0.7225, 0.7283333333333334, 0.7341666666666667, 0.74, 0.7458333333333333, 0.7516666666666667, 0.7575000000000001, 0.7633333333333334, 0.7691666666666668, 0.7750000000000001, 0.7808333333333334, 0.7866666666666667, 0.7925000000000001, 0.7983333333333335, 0.8041666666666668, 0.8100000000000002, 0.8158333333333334, 0.8216666666666668, 0.8275000000000001, 0.8333333333333335], "correct_counts": [0, 4, 12, 26, 70, 112, 154, 221, 164, 186, 98, 80, 49, 26, 14, 6, 2, 1, 0, 0], "incorrect_counts": [1, 2, 2, 7, 7, 17, 26, 34, 29, 36, 31, 18, 23, 13, 12, 9, 4, 1, 2, 1]}, "dynamics/drop": {"edges": [61.334716603159904, 75.90191677833597, 90.46911695351203, 105.0363171286881, 119.60351730386415, 134.1707174790402, 148.7379176542163, 163.30511782939234, 177.8723180045684, 192.43951817974448, 207.00671835492054, 221.5739185300966, 236.14111870527267, 250.70831888044873, 265.2755190556248, 279.84271923080087, 294.4099194059769, 308.977119581153, 323.54431975632906, 338.1115199315051, 352.67872010668117], "correct_counts": [9, 83, 206, 341, 278, 152, 75, 26, 21, 10, 7, 4, 3, 4, 1, 4, 0, 0, 0, 1], "incorrect_counts": [4, 13, 31, 49, 51, 30, 23, 11, 10, 10, 9, 9, 6, 2, 5, 3, 3, 2, 3, 1]}, "dynamics/residual": {"edges": [1.1505316942930222, 1.167380820463101, 1.18422994663318, 1.201079072803259, 1.2179281989733377, 1.2347773251434166, 1.2516264513134956, 1.2684755774835745, 1.2853247036536535, 1.3021738298237324, 1.3190229559938111, 1.33587208216389, 1.352721208333969, 1.369570334504048, 1.386419460674127, 1.4032685868442059, 1.4201177130142848, 1.4369668391843635, 1.4538159653544425, 1.4706650915245214, 1.4875142176946003], "correct_counts": [5, 5, 18, 43, 58, 105, 154, 167, 200, 158, 129, 72, 53, 26, 19, 5, 5, 2, 0, 1], "incorrect_counts": [2, 1, 5, 12, 12, 28, 27, 35, 35, 42, 25, 19, 14, 11, 3, 4, 0, 0, 0, 0]}}}, "feature_ablation": {"full": 0.05870278336778761, "drop_basin": 0.1212250149219976, "basin_only": 0.10427090599426178, "drop_energy": 0.06538158737574412, "energy_only": 0.22639015279001842, "drop_curv": 0.06922486202954758, "curv_only": 0.1307398782738768, "drop_dynamics": 0.05576840446247957, "dynamics_only": 0.14639777564824447}, "ece_geometry": 0.06136028513989912, "accuracy_ood": 0.228, "aurc_ood": {"geometry": 0.7893819156935301, "rho_basin": 0.7193987522326036, "energy_min": 0.775468743535553, "energy_mean": 0.7822174952275276, "energy_std": 0.78991149550709, "msp": 0.6564852425996128, "temp_msp": 0.668475764448379, "entropy": 0.645240327250456, "softmax_learned": 0.6585912199386484, "geom_softmax": 0.688751813066032}, "ood_ltt": {"alpha": 0.1, "lambda_hat": 0.19510629211616212, "selective_risk": 0.7557117750439367, "coverage": 0.7586666666666667, "risk_within_budget": false}, "ood_validity": {"target": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "id_risk": [0.0, 0.03294117647058824, 0.08076225045372051, 0.13363705391040243, 0.18333333333333332, 0.18333333333333332], "ood_risk": [0.0, 0.7344497607655502, 0.7557117750439367, 0.7721611721611722, 0.772, 0.772], "id_coverage": [0.0, 0.5666666666666667, 0.7346666666666667, 0.878, 1.0, 1.0], "ood_coverage": [0.0, 0.5573333333333333, 0.7586666666666667, 0.91, 1.0, 1.0]}}, "env": {"python": "3.12.13", "platform": "macOS-26.5.2-arm64-arm-64bit", "jax": "0.11.0", "jax_backend": "cpu"}, "artifact_paths": []} +{"run_id": "470896cedd8b", "timestamp": "2026-07-30T11:18:11.814580+00:00", "git_sha": "67dc7a2", "config_hash": "11f62a585e6b", "config": {"run": {"seed": 3, "task": "graph_planning", "notes": "graph adaptive-halting run"}, "model": {"latent_dim": 32, "hidden_dim": 128, "context_dim": 64}, "inference": {"k_restarts": 12, "steps": 50, "step_size": 0.1, "temperature": 0.05, "init_scale": 1.0, "grad_tol": 0.001, "sampler": "langevin", "anneal_levels": 10, "anneal_steps_per_level": 8, "anneal_step_max": 0.2, "anneal_step_min": 0.01}, "train": {"epochs": 25, "batch_size": 128, "lr": 0.001, "n_train": 8000, "n_neg": 4, "neg_noise": 0.5, "objective": "basin_center", "ired_noise_min": 0.1, "ired_noise_max": 1.5, "ired_ridge": 0.05, "ired_decode_weight": 1.0, "ired_stat_weight": 1.0}, "eval": {"n_eval": 800, "richer_geometry": false}, "conformal": {"alpha": 0.1, "delta": 0.05, "n_calib": 800}, "task": {"graph_planning": {"n_nodes": 7, "ood_n_nodes": 10, "edge_prob": 0.4, "max_len": 4}}}, "task": "graph_planning", "split": "halting", "seed": 3, "metrics": {"k_restarts": 12, "steps": 50, "alpha": 0.1, "n_calib": 800, "n_test": 800, "final_train_loss": 0.16939383745193481, "tau_hat": 0.8333, "compute_used": 0.16787500000000002, "compute_saved": 0.832125, "halting_risk": 0.09, "risk_within_budget": true, "full_accuracy": 0.695, "halted_accuracy": 0.6725, "accuracy_drop": 0.022499999999999964, "risk_monotone_in_tau": true, "calib_risk_by_tau": [0.445, 0.445, 0.445, 0.4425, 0.36625, 0.36625, 0.3, 0.14625, 0.14625, 0.09, 0.01375, 0.01375], "calib_compute_by_tau": [0.0, 0.0, 0.0, 0.00034999999999999994, 0.020199999999999853, 0.020199999999999853, 0.04342499999999971, 0.11570000000000019, 0.11570000000000019, 0.16180000000000028, 0.28835, 0.28835], "tau_sweep": {"taus": [0.0833, 0.1667, 0.25, 0.3333, 0.4167, 0.5, 0.5833, 0.6667, 0.75, 0.8333, 0.9167, 1.0], "compute_used": [0.0, 0.0, 0.0, 0.00025, 0.021575, 0.021575, 0.04630000000000001, 0.11372500000000002, 0.11372500000000002, 0.16787500000000002, 0.3007, 0.3007], "accuracy": [0.4825, 0.4825, 0.4825, 0.4825, 0.5375, 0.5375, 0.57375, 0.65, 0.65, 0.6725, 0.6975, 0.6975], "disagreement": [0.45625, 0.45625, 0.45625, 0.455, 0.36625, 0.36625, 0.29375, 0.15, 0.15, 0.09, 0.00625, 0.00625]}}, "env": {"python": "3.12.13", "platform": "macOS-26.5.2-arm64-arm-64bit", "jax": "0.11.0", "jax_backend": "cpu"}, "artifact_paths": []} +{"run_id": "52b5abb0b7f5", "timestamp": "2026-07-30T11:18:16.252284+00:00", "git_sha": "67dc7a2", "config_hash": "4a05527e96db", "config": {"run": {"seed": 4, "task": "arithmetic", "notes": "E1 selective-prediction main run"}, "model": {"latent_dim": 32, "hidden_dim": 128, "context_dim": 64}, "inference": {"k_restarts": 12, "steps": 50, "step_size": 0.1, "temperature": 0.05, "init_scale": 1.0, "grad_tol": 0.001, "sampler": "langevin", "anneal_levels": 10, "anneal_steps_per_level": 8, "anneal_step_max": 0.2, "anneal_step_min": 0.01}, "train": {"epochs": 7, "batch_size": 128, "lr": 0.001, "n_train": 6000, "n_neg": 4, "neg_noise": 0.5, "objective": "basin_center", "ired_noise_min": 0.1, "ired_noise_max": 1.5, "ired_ridge": 0.05, "ired_decode_weight": 1.0, "ired_stat_weight": 1.0}, "eval": {"n_eval": 1500, "richer_geometry": false}, "conformal": {"alpha": 0.1, "delta": 0.05, "n_calib": 1500}, "task": {"arithmetic": {"modular": false, "n_operands": 6, "ood_n_operands": 6, "max_operand": 7, "ood_max_operand": 9}}}, "task": "arithmetic", "split": "selective", "seed": 4, "metrics": {"n_fit": 1500, "n_calib": 1500, "n_test": 1500, "k_restarts": 12, "objective": "basin_center", "sampler": "langevin", "feature_set": "base", "accuracy_id": 0.8186666666666667, "base_error": 0.18133333333333335, "final_train_loss": 0.8476110100746155, "feature_names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual"], "aurc": {"geometry": 0.06427468985900764, "rho_basin": 0.0938341849888178, "energy_min": 0.16779814857981604, "energy_mean": 0.16802618351251167, "energy_std": 0.18329896021538222, "msp": 0.04726303074136766, "temp_msp": 0.04804395890364349, "entropy": 0.055194322709484635, "softmax_learned": 0.04644594370809505, "geom_softmax": 0.04320914064226967}, "temperature": 0.32684245355705593, "best_energy_baseline": "energy_min", "best_baseline": "msp", "delta_aurc_vs_energy_min": [0.1035234587208084, 0.08007861277177743, 0.12748935525846986], "delta_aurc_vs_best_energy": [0.1035234587208084, 0.08007861277177743, 0.12748935525846986], "delta_aurc_vs_best_baseline": [-0.01701165911763998, -0.029246370318950058, -0.00695285101955189], "delta_aurc_geom_adds": [0.0032368030658253832, 0.0009814347315345426, 0.005577143152198716], "geometry_wins": true, "geometry_wins_vs_baseline": false, "geometry_adds_over_softmax": true, "risk_coverage": {"geometry": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.06666666666666667, 0.03333333333333333, 0.044444444444444446, 0.03333333333333333, 0.04, 0.03888888888888888, 0.038095238095238106, 0.04583333333333333, 0.04814814814814815, 0.04333333333333334, 0.03939393939393939, 0.03888888888888889, 0.041025641025641026, 0.04047619047619048, 0.03777777777777796, 0.041666666666666664, 0.0392156862745098, 0.03888888888888888, 0.03859649122807018, 0.03833333333333333, 0.0365079365079365, 0.03636363636363636, 0.036231884057971016, 0.03888888888888888, 0.03866666666666667, 0.041025641025641026, 0.040740740740740744, 0.039285714285714285, 0.04137931034482758, 0.044444444444444446, 0.04408602150537634, 0.046875, 0.052525252525252523, 0.058823529411764705, 0.06666666666666665, 0.07314814814814813, 0.08018018018018018, 0.08421052631578947, 0.08803418803418804, 0.09583333333333334, 0.1032520325203252, 0.1119047619047619, 0.12170542635658915, 0.1318181818181818, 0.13703703703703704, 0.14130434782608695, 0.14751773049645386, 0.15694444444444455, 0.16870748299319727, 0.18133333333333335]}, "rho_basin": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0789724072312084, 0.07897240723120841, 0.07897240723120832, 0.07897240723120841, 0.07897240723120853, 0.07897240723120862, 0.07897240723120866, 0.07897240723120871, 0.07897240723120874, 0.07897240723120877, 0.0789724072312088, 0.07897240723120881, 0.07897240723120882, 0.07897240723120871, 0.07897240723120849, 0.07897240723120831, 0.07897240723120814, 0.07897240723120799, 0.07897240723120785, 0.07897240723120774, 0.07897240723120763, 0.07897240723120753, 0.07897240723120744, 0.07897240723120737, 0.07897240723120728, 0.07897240723120721, 0.07897240723120715, 0.07897240723120709, 0.07897240723120703, 0.07897240723120698, 0.07897240723120694, 0.07897240723120688, 0.07897240723120684, 0.0789724072312068, 0.07897240723120677, 0.08567985794013051, 0.09224978403060469, 0.09847392453736972, 0.10437887835148014, 0.11010416666666574, 0.11656504065040564, 0.12271825396825312, 0.1285852713178286, 0.13747294372294272, 0.14625081221572334, 0.1526061530638181, 0.15899906329452637, 0.16708595387840589, 0.1745857317285879, 0.18133333333333207]}, "energy_min": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.2, 0.16666666666666666, 0.14444444444444443, 0.14166666666666666, 0.14, 0.13888888888888887, 0.14761904761904746, 0.15416666666666667, 0.15555555555555556, 0.17333333333333337, 0.18787878787878787, 0.18611111111111112, 0.17692307692307693, 0.18095238095238095, 0.18888888888888886, 0.19166666666666668, 0.18823529411764706, 0.187037037037037, 0.18421052631578946, 0.18333333333333332, 0.17936507936507934, 0.17575757575757575, 0.1753623188405797, 0.16944444444444443, 0.17466666666666666, 0.17051282051282052, 0.16790123456790124, 0.16547619047619047, 0.164367816091954, 0.1622222222222222, 0.16129032258064516, 0.16145833333333334, 0.1606060606060606, 0.1588235294117647, 0.15714285714285728, 0.1583333333333333, 0.15855855855855855, 0.1587719298245614, 0.15726495726495726, 0.1575, 0.16016260162601623, 0.1603174603174603, 0.15968992248062017, 0.15984848484848485, 0.16, 0.16376811594202897, 0.16879432624113486, 0.17430555555555552, 0.1761904761904762, 0.18133333333333335]}, "energy_mean": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.23333333333333334, 0.11666666666666667, 0.14444444444444443, 0.125, 0.13333333333333333, 0.1333333333333333, 0.14285714285714288, 0.1625, 0.17037037037037037, 0.1766666666666667, 0.1787878787878788, 0.16944444444444445, 0.16666666666666666, 0.17142857142857143, 0.17777777777777776, 0.18333333333333332, 0.1843137254901961, 0.1907407407407407, 0.18947368421052632, 0.18333333333333332, 0.1857142857142857, 0.18333333333333332, 0.17971014492753623, 0.18055555555555564, 0.18, 0.17435897435897435, 0.16913580246913582, 0.17023809523809524, 0.16666666666666663, 0.16333333333333333, 0.16344086021505377, 0.16354166666666667, 0.1606060606060606, 0.1607843137254902, 0.15809523809523807, 0.15648148148148144, 0.15675675675675677, 0.1570175438596491, 0.1564102564102564, 0.15583333333333332, 0.15528455284552842, 0.1539682539682541, 0.1558139534883721, 0.15833333333333333, 0.1637037037037037, 0.16376811594202897, 0.1652482269503547, 0.17013888888888887, 0.1761904761904762, 0.18133333333333335]}, "energy_std": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.26666666666666666, 0.18333333333333332, 0.17777777777777778, 0.2, 0.17333333333333334, 0.18888888888888886, 0.17619047619047623, 0.17083333333333334, 0.18888888888888888, 0.1866666666666667, 0.18484848484848485, 0.19166666666666668, 0.18974358974358974, 0.19285714285714287, 0.1933333333333333, 0.18958333333333333, 0.1843137254901961, 0.1796296296296296, 0.1824561403508772, 0.17833333333333334, 0.17936507936507934, 0.18484848484848485, 0.1826086956521739, 0.18055555555555552, 0.17866666666666667, 0.18205128205128204, 0.17901234567901234, 0.17857142857142858, 0.17931034482758618, 0.18, 0.17849462365591398, 0.17604166666666668, 0.17575757575757575, 0.17941176470588235, 0.17999999999999997, 0.1851851851851853, 0.1837837837837838, 0.1850877192982456, 0.18461538461538463, 0.1825, 0.18455284552845538, 0.1841269841269841, 0.1813953488372093, 0.18181818181818182, 0.18074074074074073, 0.17898550724637682, 0.17872340425531913, 0.18055555555555564, 0.18095238095238095, 0.18133333333333335]}, "msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.002777777777777778, 0.007692307692307693, 0.007142857142857143, 0.008888888888888887, 0.008333333333333333, 0.00980392156862745, 0.01111111111111111, 0.014035087719298246, 0.013333333333333334, 0.02063492063492063, 0.019696969696969695, 0.020289855072463767, 0.02222222222222222, 0.025333333333333333, 0.029487179487179487, 0.028395061728395062, 0.03214285714285714, 0.03563218390804597, 0.03777777777777778, 0.04086021505376344, 0.04583333333333333, 0.050505050505050504, 0.05392156862745098, 0.060952380952380945, 0.06481481481481495, 0.07387387387387387, 0.08421052631578947, 0.09145299145299145, 0.09916666666666667, 0.10894308943089444, 0.11507936507936506, 0.12558139534883722, 0.13484848484848486, 0.14148148148148149, 0.1463768115942029, 0.15248226950354607, 0.1597222222222223, 0.17006802721088435, 0.18133333333333335]}, "temp_msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.003333333333333334, 0.0030303030303030303, 0.002777777777777778, 0.002564102564102564, 0.002380952380952381, 0.002222222222222222, 0.00625, 0.0058823529411764705, 0.007407407407407407, 0.008771929824561403, 0.013333333333333334, 0.017460317460317457, 0.019696969696969695, 0.01884057971014493, 0.02083333333333333, 0.024, 0.02435897435897436, 0.028395061728395062, 0.02976190476190476, 0.033333333333333326, 0.03666666666666667, 0.03870967741935484, 0.04895833333333333, 0.05454545454545454, 0.05784313725490196, 0.06761904761904777, 0.07314814814814813, 0.07927927927927927, 0.08421052631578947, 0.09487179487179487, 0.10416666666666667, 0.1121951219512195, 0.11904761904761915, 0.1263565891472868, 0.13787878787878788, 0.14444444444444443, 0.1536231884057971, 0.16099290780141842, 0.16666666666666674, 0.1727891156462585, 0.18133333333333335]}, "entropy": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.004166666666666667, 0.003703703703703704, 0.003333333333333334, 0.0030303030303030303, 0.005555555555555556, 0.010256410256410256, 0.009523809523809525, 0.013333333333333332, 0.01875, 0.01764705882352941, 0.02037037037037037, 0.02631578947368421, 0.03333333333333333, 0.0365079365079365, 0.03636363636363636, 0.04057971014492753, 0.04166666666666666, 0.044, 0.047435897435897434, 0.04938271604938271, 0.05119047619047619, 0.0528735632183908, 0.05555555555555555, 0.05806451612903226, 0.059375, 0.06363636363636363, 0.06862745098039216, 0.07428571428571444, 0.08148148148148163, 0.08828828828828829, 0.09385964912280702, 0.1, 0.10416666666666667, 0.11056910569105703, 0.11746031746031757, 0.12170542635658915, 0.12727272727272726, 0.13555555555555557, 0.14420289855072463, 0.15602836879432622, 0.1659722222222222, 0.17482993197278912, 0.18133333333333335]}, "softmax_learned": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.007692307692307693, 0.007142857142857143, 0.008888888888888887, 0.008333333333333333, 0.00784313725490196, 0.009259259259259257, 0.012280701754385965, 0.013333333333333334, 0.014285714285714284, 0.019696969696969695, 0.01884057971014493, 0.01944444444444444, 0.025333333333333333, 0.026923076923076925, 0.027160493827160494, 0.02976190476190476, 0.03448275862068965, 0.03666666666666667, 0.03870967741935484, 0.041666666666666664, 0.047474747474747475, 0.050980392156862744, 0.057142857142857134, 0.06388888888888887, 0.07027027027027027, 0.08333333333333333, 0.0905982905982906, 0.09916666666666667, 0.10731707317073169, 0.11825396825396824, 0.1263565891472868, 0.1340909090909091, 0.14074074074074075, 0.1471014492753623, 0.15248226950354607, 0.16041666666666665, 0.16938775510204082, 0.18133333333333335]}, "geom_softmax": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.002380952380952381, 0.006666666666666666, 0.00625, 0.0058823529411764705, 0.007407407407407407, 0.008771929824561403, 0.008333333333333333, 0.009523809523809523, 0.010606060606060607, 0.013043478260869565, 0.013888888888888886, 0.016, 0.016666666666666666, 0.01728395061728395, 0.02261904761904762, 0.022988505747126433, 0.02666666666666667, 0.03118279569892473, 0.034375, 0.04040404040404041, 0.05, 0.056190476190476187, 0.06944444444444443, 0.07567567567567568, 0.08596491228070176, 0.09145299145299145, 0.1025, 0.10813008130081313, 0.11428571428571427, 0.12093023255813953, 0.12727272727272726, 0.1348148148148148, 0.14057971014492754, 0.14751773049645386, 0.15555555555555567, 0.1673469387755102, 0.18133333333333335]}}, "ltt": {"alpha": 0.1, "delta": 0.05, "lambda_hat": 0.2286224610513153, "coverage": 0.7453333333333333, "abstain_rate": 0.25466666666666665, "selective_risk": 0.08050089445438283, "selective_accuracy": 0.9194991055456172, "risk_within_budget": true}, "coverage_validity": {"alpha_grid": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "target": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "achieved_risk": [0.0, 0.0, 0.08050089445438283, 0.1198443579766537, 0.180787191460974, 0.180787191460974], "coverage": [0.0, 0.0, 0.7453333333333333, 0.8566666666666667, 0.9993333333333333, 0.9993333333333333]}, "feature_diagnostics": {"names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual"], "auroc": {"basin/rho": 0.7546809134891742, "basin/entropy": 0.24459157645142748, "basin/dispersion": 0.5018112904771029, "energy/mean": 0.44983773232420005, "energy/min": 0.4503197451619084, "energy/std": 0.5016077074152137, "curv/lmax_mean": 0.44767017148879096, "curv/lmax_best": 0.4978010035447404, "curv/trace_mean": 0.5762478444146388, "curv/trace_best": 0.5414815457942135, "dynamics/steps": 0.5, "dynamics/monotonic": 0.4247101935236635, "dynamics/drop": 0.38611324008430736, "dynamics/residual": 0.4858749281471546}, "hist": {"basin/rho": {"edges": [0.4166666666666667, 0.44583333333333336, 0.47500000000000003, 0.5041666666666667, 0.5333333333333333, 0.5625, 0.5916666666666667, 0.6208333333333333, 0.65, 0.6791666666666667, 0.7083333333333333, 0.7375, 0.7666666666666666, 0.7958333333333334, 0.825, 0.8541666666666666, 0.8833333333333333, 0.9125, 0.9416666666666667, 0.9708333333333332, 1.0], "correct_counts": [1, 0, 19, 0, 0, 24, 0, 0, 32, 0, 0, 26, 0, 0, 60, 0, 0, 98, 0, 968], "incorrect_counts": [1, 0, 20, 0, 0, 29, 0, 0, 25, 0, 0, 30, 0, 0, 36, 0, 0, 48, 0, 83]}, "basin/entropy": {"edges": [0.0, 0.05387781635334003, 0.10775563270668007, 0.16163344906002008, 0.21551126541336013, 0.2693890817667002, 0.32326689812004017, 0.3771447144733802, 0.43102253082672026, 0.4849003471800603, 0.5387781635334004, 0.5926559798867403, 0.6465337962400803, 0.7004116125934204, 0.7542894289467604, 0.8081672453001005, 0.8620450616534405, 0.9159228780067805, 0.9698006943601206, 1.0236785107134607, 1.0775563270668007], "correct_counts": [968, 0, 0, 0, 0, 98, 0, 0, 56, 0, 27, 28, 38, 3, 0, 2, 3, 3, 1, 1], "incorrect_counts": [83, 0, 0, 0, 0, 48, 0, 0, 32, 0, 30, 20, 40, 4, 0, 4, 1, 3, 3, 4]}, "basin/dispersion": {"edges": [1.5954158897257813, 1.6171290768517932, 1.6388422639778053, 1.6605554511038172, 1.6822686382298293, 1.7039818253558412, 1.7256950124818533, 1.7474081996078652, 1.7691213867338773, 1.7908345738598892, 1.8125477609859013, 1.8342609481119132, 1.855974135237925, 1.8776873223639372, 1.899400509489949, 1.9211136966159612, 1.942826883741973, 1.9645400708679852, 1.986253257993997, 2.007966445120009, 2.029679632246021], "correct_counts": [2, 3, 4, 17, 21, 42, 76, 99, 132, 162, 172, 143, 128, 92, 50, 34, 32, 5, 11, 3], "incorrect_counts": [1, 2, 2, 2, 7, 11, 18, 14, 40, 28, 30, 37, 29, 17, 9, 11, 9, 2, 2, 1]}, "energy/mean": {"edges": [0.7072046200434366, 0.7702707141637801, 0.8333368082841237, 0.8964029024044672, 0.9594689965248107, 1.0225350906451542, 1.0856011847654976, 1.1486672788858412, 1.2117333730061848, 1.2747994671265284, 1.337865561246872, 1.4009316553672153, 1.4639977494875587, 1.5270638436079023, 1.5901299377282458, 1.6531960318485894, 1.716262125968933, 1.7793282200892766, 1.8423943142096197, 1.9054604083299633, 1.9685265024503071], "correct_counts": [3, 7, 28, 55, 67, 56, 50, 51, 47, 44, 61, 68, 152, 167, 172, 90, 63, 27, 14, 6], "incorrect_counts": [2, 3, 2, 7, 14, 15, 16, 6, 14, 15, 14, 13, 20, 24, 23, 26, 21, 20, 12, 5]}, "energy/min": {"edges": [0.34151938557624817, 0.41106958836317065, 0.48061979115009307, 0.5501699939370155, 0.619720196723938, 0.6892703995108604, 0.7588206022977828, 0.8283708050847053, 0.8979210078716278, 0.9674712106585502, 1.0370214134454727, 1.106571616232395, 1.1761218190193174, 1.24567202180624, 1.3152222245931624, 1.384772427380085, 1.4543226301670074, 1.5238728329539297, 1.5934230357408523, 1.6629732385277747, 1.7325234413146973], "correct_counts": [4, 7, 14, 36, 72, 61, 65, 60, 49, 64, 85, 112, 140, 168, 120, 87, 57, 20, 5, 2], "incorrect_counts": [2, 2, 2, 6, 9, 11, 25, 11, 17, 15, 13, 18, 17, 28, 25, 23, 29, 10, 7, 2]}, "energy/std": {"edges": [0.05760100396148682, 0.07262729619130567, 0.08765358842112453, 0.1026798806509434, 0.11770617288076225, 0.13273246511058112, 0.14775875734039998, 0.16278504957021883, 0.1778113418000377, 0.19283763402985654, 0.2078639262596754, 0.22289021848949425, 0.23791651071931313, 0.25294280294913196, 0.26796909517895084, 0.2829953874087697, 0.29802167963858855, 0.31304797186840744, 0.32807426409822626, 0.34310055632804515, 0.358126848557864], "correct_counts": [1, 1, 5, 16, 47, 70, 130, 148, 166, 165, 142, 111, 89, 55, 34, 20, 9, 9, 7, 3], "incorrect_counts": [0, 1, 0, 7, 7, 19, 28, 32, 36, 34, 38, 25, 16, 6, 12, 3, 5, 2, 1, 0]}, "curv/lmax_mean": {"edges": [0.9986266444126765, 0.9987258836627007, 0.9988251229127247, 0.9989243621627489, 0.9990236014127731, 0.9991228406627972, 0.9992220799128214, 0.9993213191628456, 0.9994205584128697, 0.9995197976628939, 0.9996190369129181, 0.9997182761629422, 0.9998175154129664, 0.9999167546629905, 1.0000159939130147, 1.0001152331630387, 1.000214472413063, 1.0003137116630871, 1.0004129509131112, 1.0005121901631355, 1.0006114294131596], "correct_counts": [4, 4, 12, 17, 40, 61, 83, 170, 202, 224, 175, 145, 55, 20, 5, 7, 1, 0, 1, 2], "incorrect_counts": [0, 0, 2, 4, 6, 9, 19, 27, 41, 47, 55, 42, 16, 2, 1, 0, 0, 1, 0, 0]}, "curv/lmax_best": {"edges": [0.9928669929504395, 0.9932882905006408, 0.9937095880508423, 0.9941308856010437, 0.9945521831512452, 0.9949734807014465, 0.9953947782516479, 0.9958160758018494, 0.9962373733520508, 0.9966586709022522, 0.9970799684524536, 0.997501266002655, 0.9979225635528565, 0.9983438611030578, 0.9987651586532593, 0.9991864562034607, 0.9996077537536621, 1.0000290513038634, 1.000450348854065, 1.0008716464042664, 1.0012929439544678], "correct_counts": [1, 0, 0, 0, 3, 0, 1, 3, 3, 8, 12, 17, 24, 46, 87, 216, 787, 16, 1, 3], "incorrect_counts": [0, 0, 0, 0, 1, 1, 0, 1, 0, 2, 1, 3, 7, 7, 19, 52, 177, 1, 0, 0]}, "curv/trace_mean": {"edges": [31.90089813868205, 31.906382902463278, 31.911867666244508, 31.917352430025737, 31.922837193806966, 31.928321957588196, 31.933806721369425, 31.939291485150655, 31.944776248931884, 31.950261012713113, 31.955745776494346, 31.961230540275576, 31.966715304056805, 31.972200067838035, 31.977684831619264, 31.983169595400494, 31.988654359181723, 31.994139122962952, 31.999623886744182, 32.00510865052541, 32.01059341430664], "correct_counts": [1, 0, 3, 9, 12, 18, 28, 35, 60, 119, 205, 224, 185, 111, 75, 58, 48, 26, 7, 4], "incorrect_counts": [2, 3, 6, 9, 10, 17, 11, 14, 24, 20, 26, 26, 24, 24, 16, 20, 11, 8, 0, 1]}, "curv/trace_best": {"edges": [31.815509796142578, 31.82562713623047, 31.83574447631836, 31.84586181640625, 31.85597915649414, 31.86609649658203, 31.876213836669923, 31.886331176757814, 31.8964485168457, 31.906565856933593, 31.916683197021484, 31.926800537109376, 31.936917877197267, 31.947035217285155, 31.957152557373046, 31.967269897460938, 31.97738723754883, 31.98750457763672, 31.997621917724608, 32.0077392578125, 32.01785659790039], "correct_counts": [2, 0, 0, 0, 2, 2, 5, 7, 10, 27, 39, 61, 89, 141, 208, 232, 221, 134, 40, 8], "incorrect_counts": [1, 0, 0, 0, 1, 6, 4, 6, 3, 3, 18, 17, 23, 23, 46, 36, 38, 37, 8, 2]}, "dynamics/steps": {"edges": [1.0, 1.05, 1.1, 1.15, 1.2, 1.25, 1.3, 1.35, 1.4, 1.45, 1.5, 1.55, 1.6, 1.65, 1.7000000000000002, 1.75, 1.8, 1.85, 1.9, 1.9500000000000002, 2.0], "correct_counts": [1228, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [272, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "dynamics/monotonic": {"edges": [0.7133333333333334, 0.7192500000000001, 0.7251666666666667, 0.7310833333333334, 0.737, 0.7429166666666667, 0.7488333333333334, 0.75475, 0.7606666666666667, 0.7665833333333334, 0.7725, 0.7784166666666666, 0.7843333333333333, 0.79025, 0.7961666666666667, 0.8020833333333334, 0.808, 0.8139166666666666, 0.8198333333333333, 0.82575, 0.8316666666666667], "correct_counts": [1, 0, 1, 16, 23, 81, 100, 182, 139, 225, 208, 103, 97, 28, 14, 7, 0, 2, 1, 0], "incorrect_counts": [0, 0, 1, 5, 12, 15, 21, 26, 25, 34, 30, 27, 29, 18, 14, 4, 6, 1, 2, 2]}, "dynamics/drop": {"edges": [67.27662276228268, 83.98704353620607, 100.69746431012948, 117.40788508405288, 134.11830585797628, 150.82872663189966, 167.53914740582306, 184.24956817974646, 200.95998895366984, 217.67040972759327, 234.38083050151664, 251.09125127544002, 267.80167204936345, 284.5120928232868, 301.2225135972102, 317.9329343711336, 334.643355145057, 351.3537759189804, 368.0641966929038, 384.7746174668272, 401.4850382407506], "correct_counts": [32, 125, 299, 388, 250, 73, 29, 10, 8, 4, 4, 1, 3, 2, 0, 0, 0, 0, 0, 0], "incorrect_counts": [7, 27, 53, 44, 42, 26, 18, 14, 10, 9, 6, 5, 5, 2, 1, 0, 0, 1, 1, 1]}, "dynamics/residual": {"edges": [1.1324152251084645, 1.1493771066268283, 1.1663389881451924, 1.1833008696635563, 1.2002627511819204, 1.2172246327002842, 1.2341865142186481, 1.2511483957370122, 1.268110277255376, 1.2850721587737401, 1.302034040292104, 1.318995921810468, 1.335957803328832, 1.3529196848471958, 1.36988156636556, 1.3868434478839238, 1.4038053294022879, 1.4207672109206517, 1.4377290924390156, 1.4546909739573797, 1.4716528554757435], "correct_counts": [0, 2, 3, 12, 27, 70, 95, 129, 174, 178, 180, 122, 103, 65, 37, 16, 11, 3, 0, 1], "incorrect_counts": [1, 0, 3, 3, 4, 7, 21, 40, 32, 37, 33, 34, 27, 15, 8, 3, 4, 0, 0, 0]}}}, "feature_ablation": {"full": 0.06427468985900764, "drop_basin": 0.14673850433647762, "basin_only": 0.09760604064778008, "drop_energy": 0.06786327554908943, "energy_only": 0.16911957473150377, "drop_curv": 0.06430618294867559, "curv_only": 0.16700293935048874, "drop_dynamics": 0.08141400025100816, "dynamics_only": 0.15248745127459712}, "ece_geometry": 0.04329479322788476, "accuracy_ood": 0.20533333333333334, "aurc_ood": {"geometry": 0.8356043455050088, "rho_basin": 0.7448845820671743, "energy_min": 0.8118924901529568, "energy_mean": 0.8124673916344924, "energy_std": 0.7923987846566023, "msp": 0.6753445881260701, "temp_msp": 0.6744023348611631, "entropy": 0.680160911605275, "softmax_learned": 0.6733183754314852, "geom_softmax": 0.7053885590469356}, "ood_ltt": {"alpha": 0.1, "lambda_hat": 0.2286224610513153, "selective_risk": 0.7856517935258093, "coverage": 0.762, "risk_within_budget": false}, "ood_validity": {"target": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "id_risk": [0.0, 0.0, 0.08050089445438283, 0.1198443579766537, 0.180787191460974, 0.180787191460974], "ood_risk": [0.0, 0.0, 0.7856517935258093, 0.7973568281938326, 0.7945296864576384, 0.7945296864576384], "id_coverage": [0.0, 0.0, 0.7453333333333333, 0.8566666666666667, 0.9993333333333333, 0.9993333333333333], "ood_coverage": [0.0, 0.0, 0.762, 0.908, 0.9993333333333333, 0.9993333333333333]}}, "env": {"python": "3.12.13", "platform": "macOS-26.5.2-arm64-arm-64bit", "jax": "0.11.0", "jax_backend": "cpu"}, "artifact_paths": []} +{"run_id": "9202e8561acd", "timestamp": "2026-07-30T11:18:17.733920+00:00", "git_sha": "67dc7a2", "config_hash": "882cd7965583", "config": {"run": {"seed": 2, "task": "graph_planning", "notes": "E2 graph selective-prediction"}, "model": {"latent_dim": 32, "hidden_dim": 128, "context_dim": 64}, "inference": {"k_restarts": 12, "steps": 50, "step_size": 0.1, "temperature": 0.05, "init_scale": 1.0, "grad_tol": 0.001, "sampler": "langevin", "anneal_levels": 10, "anneal_steps_per_level": 8, "anneal_step_max": 0.2, "anneal_step_min": 0.01}, "train": {"epochs": 25, "batch_size": 128, "lr": 0.001, "n_train": 8000, "n_neg": 4, "neg_noise": 0.5, "objective": "basin_center", "ired_noise_min": 0.1, "ired_noise_max": 1.5, "ired_ridge": 0.05, "ired_decode_weight": 1.0, "ired_stat_weight": 1.0}, "eval": {"n_eval": 1500, "richer_geometry": false}, "conformal": {"alpha": 0.1, "delta": 0.05, "n_calib": 1500}, "task": {"graph_planning": {"n_nodes": 7, "ood_n_nodes": 10, "edge_prob": 0.4, "max_len": 4}}}, "task": "graph_planning", "split": "selective", "seed": 2, "metrics": {"n_fit": 1500, "n_calib": 1500, "n_test": 1500, "k_restarts": 12, "objective": "basin_center", "sampler": "langevin", "feature_set": "base", "accuracy_id": 0.676, "base_error": 0.324, "final_train_loss": 0.178430438041687, "feature_names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual"], "aurc": {"geometry": 0.171055380110865, "rho_basin": 0.27572696927818435, "energy_min": 0.36405839474901913, "energy_mean": 0.36247574254442705, "energy_std": 0.31446505154630827, "msp": 0.1492232493426811, "temp_msp": 0.1400399776291398, "entropy": 0.14691654031820944, "softmax_learned": 0.13872939213414487, "geom_softmax": 0.13922822696499798}, "temperature": 3.0741721331379974, "best_energy_baseline": "energy_std", "best_baseline": "temp_msp", "delta_aurc_vs_energy_min": [0.19300301463815414, 0.16374700571598869, 0.2228419667503917], "delta_aurc_vs_best_energy": [0.14340967143544328, 0.11666052233185831, 0.17035195020051339], "delta_aurc_vs_best_baseline": [-0.031015402481725185, -0.041904924234315785, -0.02069570943951534], "delta_aurc_geom_adds": [-0.0004988348308531121, -0.004262736925938849, 0.002823621624244659], "geometry_wins": true, "geometry_wins_vs_baseline": false, "geometry_adds_over_softmax": false, "risk_coverage": {"geometry": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.03333333333333333, 0.03333333333333333, 0.05, 0.06, 0.0611111111111111, 0.052380952380952396, 0.058333333333333334, 0.05925925925925926, 0.07333333333333335, 0.07575757575757576, 0.08055555555555556, 0.08717948717948718, 0.1, 0.11777777777777775, 0.12083333333333333, 0.13333333333333333, 0.13888888888888903, 0.15263157894736842, 0.15666666666666668, 0.16825396825396824, 0.16515151515151516, 0.1753623188405797, 0.18055555555555552, 0.18133333333333335, 0.18461538461538463, 0.18518518518518517, 0.19166666666666668, 0.19885057471264383, 0.20555555555555555, 0.2129032258064516, 0.2125, 0.21616161616161617, 0.22254901960784312, 0.22761904761904772, 0.2296296296296296, 0.23603603603603604, 0.24473684210526317, 0.2504273504273504, 0.25583333333333336, 0.26178861788617896, 0.2682539682539682, 0.27674418604651163, 0.2825757575757576, 0.2911111111111111, 0.2920289855072464, 0.298581560283688, 0.30555555555555564, 0.3149659863945578, 0.324]}, "rho_basin": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.2713365539452495, 0.27133655394524997, 0.2713365539452501, 0.2713365539452501, 0.2713365539452494, 0.27133655394524897, 0.2713365539452487, 0.27133655394524847, 0.2713365539452483, 0.27133655394524814, 0.271336553945248, 0.2713365539452479, 0.2713365539452478, 0.27133655394524775, 0.2713365539452477, 0.27133655394524764, 0.2713365539452476, 0.2713365539452475, 0.2713365539452475, 0.27133655394524747, 0.2713365539452474, 0.2713365539452474, 0.27133655394524736, 0.27133655394524736, 0.27133655394524736, 0.2713365539452473, 0.2713365539452473, 0.2713365539452473, 0.27133655394524725, 0.27133655394524725, 0.27133655394524725, 0.2713365539452477, 0.2713365539452485, 0.2713365539452493, 0.2713365539452501, 0.27133655394525075, 0.2713365539452514, 0.271336553945252, 0.27133655394525263, 0.2713365539452532, 0.27133655394525374, 0.27440476190476587, 0.27932816537468025, 0.2850505050505078, 0.2945185185185211, 0.30036231884058223, 0.30460992907801665, 0.30988247863248064, 0.3159183673469404, 0.32400000000000173]}, "energy_min": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.3, 0.38333333333333336, 0.3888888888888889, 0.35, 0.36666666666666664, 0.3722222222222223, 0.3809523809523808, 0.38333333333333336, 0.3888888888888889, 0.3999999999999999, 0.4, 0.39166666666666666, 0.382051282051282, 0.38571428571428573, 0.38222222222222235, 0.38125, 0.3843137254901961, 0.3833333333333333, 0.3824561403508772, 0.38166666666666665, 0.3841269841269841, 0.38181818181818183, 0.3739130434782609, 0.36805555555555564, 0.37066666666666664, 0.36923076923076925, 0.36419753086419754, 0.35714285714285715, 0.3551724137931034, 0.3511111111111111, 0.3473118279569892, 0.34479166666666666, 0.34545454545454546, 0.34215686274509804, 0.3447619047619049, 0.3462962962962962, 0.3441441441441441, 0.34385964912280703, 0.33931623931623933, 0.3458333333333333, 0.3430894308943089, 0.33968253968253964, 0.3372093023255814, 0.33636363636363636, 0.337037037037037, 0.33695652173913043, 0.33333333333333326, 0.32986111111111105, 0.32789115646258504, 0.324]}, "energy_mean": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.36666666666666664, 0.4, 0.3888888888888889, 0.4, 0.38, 0.3944444444444445, 0.3857142857142858, 0.3875, 0.3962962962962963, 0.4033333333333334, 0.4121212121212121, 0.4, 0.3871794871794872, 0.38571428571428573, 0.38888888888888884, 0.38333333333333336, 0.3764705882352941, 0.36851851851851847, 0.36666666666666664, 0.37, 0.37301587301587297, 0.36515151515151517, 0.36231884057971014, 0.3555555555555555, 0.36, 0.3641025641025641, 0.36419753086419754, 0.3595238095238095, 0.3597701149425287, 0.3566666666666667, 0.3602150537634409, 0.3572916666666667, 0.3525252525252525, 0.34901960784313724, 0.3419047619047618, 0.3398148148148149, 0.3396396396396396, 0.33596491228070174, 0.33589743589743587, 0.3325, 0.33333333333333326, 0.3333333333333334, 0.33565891472868215, 0.3340909090909091, 0.3340740740740741, 0.33260869565217394, 0.3319148936170212, 0.32986111111111105, 0.3272108843537415, 0.324]}, "energy_std": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.23333333333333334, 0.38333333333333336, 0.32222222222222224, 0.31666666666666665, 0.3333333333333333, 0.3222222222222222, 0.31428571428571433, 0.325, 0.32222222222222224, 0.33000000000000007, 0.3242424242424242, 0.3138888888888889, 0.3128205128205128, 0.3119047619047619, 0.3088888888888888, 0.31875, 0.3196078431372549, 0.31666666666666676, 0.32105263157894737, 0.32, 0.32222222222222235, 0.32575757575757575, 0.32608695652173914, 0.31944444444444436, 0.316, 0.3128205128205128, 0.3135802469135803, 0.31547619047619047, 0.32183908045977005, 0.3211111111111111, 0.3150537634408602, 0.315625, 0.3151515151515151, 0.3196078431372549, 0.3133333333333333, 0.31481481481481494, 0.3162162162162162, 0.31666666666666665, 0.3162393162393162, 0.315, 0.31219512195121946, 0.311904761904762, 0.31317829457364343, 0.3151515151515151, 0.3148148148148148, 0.31956521739130433, 0.3205673758865248, 0.3208333333333333, 0.3224489795918367, 0.324]}, "msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.011111111111111112, 0.025, 0.02666666666666667, 0.027777777777777887, 0.038095238095238106, 0.04583333333333333, 0.044444444444444446, 0.04333333333333334, 0.048484848484848485, 0.05277777777777778, 0.05641025641025641, 0.05952380952380952, 0.06888888888888887, 0.075, 0.0784313725490196, 0.08333333333333347, 0.09298245614035087, 0.09833333333333333, 0.10793650793650805, 0.11212121212121212, 0.12318840579710146, 0.12916666666666665, 0.13866666666666666, 0.14871794871794872, 0.16049382716049382, 0.16666666666666666, 0.17931034482758637, 0.18444444444444444, 0.19462365591397848, 0.20416666666666666, 0.2111111111111111, 0.21568627450980393, 0.2257142857142858, 0.2296296296296296, 0.23603603603603604, 0.24298245614035088, 0.24871794871794872, 0.25166666666666665, 0.2560975609756097, 0.26269841269841265, 0.27364341085271315, 0.2825757575757576, 0.29185185185185186, 0.29927536231884055, 0.30283687943262405, 0.3131944444444445, 0.31700680272108844, 0.324]}, "temp_msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.02666666666666667, 0.027777777777777773, 0.028571428571428577, 0.03333333333333333, 0.044444444444444446, 0.046666666666666676, 0.048484848484848485, 0.05, 0.05128205128205128, 0.05238095238095238, 0.05777777777777776, 0.06666666666666667, 0.07254901960784314, 0.07962962962962962, 0.0824561403508772, 0.09, 0.09523809523809536, 0.10909090909090909, 0.11159420289855072, 0.12083333333333332, 0.128, 0.1358974358974359, 0.14320987654320988, 0.15119047619047618, 0.16206896551724134, 0.1688888888888889, 0.17526881720430107, 0.184375, 0.19090909090909092, 0.2019607843137255, 0.21047619047619062, 0.21666666666666679, 0.22342342342342342, 0.23333333333333334, 0.24017094017094018, 0.24666666666666667, 0.24959349593495947, 0.25555555555555565, 0.26434108527131783, 0.271969696969697, 0.28074074074074074, 0.2898550724637681, 0.29787234042553196, 0.3083333333333334, 0.3142857142857143, 0.324]}, "entropy": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.011111111111111112, 0.025, 0.02666666666666667, 0.033333333333333326, 0.038095238095238106, 0.04583333333333333, 0.044444444444444446, 0.04333333333333334, 0.045454545454545456, 0.05277777777777778, 0.05641025641025641, 0.05952380952380952, 0.06888888888888887, 0.07291666666666667, 0.0784313725490196, 0.08333333333333331, 0.09298245614035087, 0.09833333333333333, 0.10476190476190489, 0.11060606060606061, 0.12028985507246377, 0.12777777777777774, 0.13733333333333334, 0.14743589743589744, 0.15679012345679014, 0.16547619047619047, 0.17241379310344845, 0.18333333333333332, 0.1935483870967742, 0.2, 0.2080808080808081, 0.21666666666666667, 0.21999999999999997, 0.22500000000000012, 0.23333333333333334, 0.23684210526315788, 0.2452991452991453, 0.2475, 0.25447154471544725, 0.26349206349206344, 0.2689922480620155, 0.2765151515151515, 0.2822222222222222, 0.2898550724637681, 0.2971631205673759, 0.30763888888888896, 0.3149659863945578, 0.324]}, "softmax_learned": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.02666666666666667, 0.027777777777777773, 0.028571428571428577, 0.03333333333333333, 0.044444444444444446, 0.046666666666666676, 0.048484848484848485, 0.05, 0.04871794871794872, 0.05238095238095238, 0.05777777777777776, 0.0625, 0.07254901960784314, 0.08148148148148147, 0.0824561403508772, 0.09166666666666666, 0.09999999999999999, 0.10303030303030303, 0.11159420289855072, 0.1222222222222222, 0.12533333333333332, 0.13076923076923078, 0.13703703703703704, 0.14523809523809525, 0.15632183908045974, 0.16444444444444445, 0.17096774193548386, 0.17916666666666667, 0.19292929292929292, 0.2019607843137255, 0.21047619047619062, 0.21111111111111108, 0.21711711711711712, 0.2236842105263158, 0.23162393162393163, 0.24, 0.24878048780487802, 0.25714285714285723, 0.26744186046511625, 0.2765151515151515, 0.2822222222222222, 0.28840579710144926, 0.30000000000000004, 0.30555555555555564, 0.31564625850340133, 0.324]}, "geom_softmax": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.03333333333333333, 0.016666666666666666, 0.011111111111111112, 0.016666666666666666, 0.02, 0.02222222222222222, 0.03333333333333334, 0.0375, 0.04814814814814815, 0.05000000000000001, 0.048484848484848485, 0.04722222222222222, 0.05128205128205128, 0.05714285714285714, 0.059999999999999984, 0.06875, 0.07058823529411765, 0.07037037037037051, 0.07894736842105263, 0.08666666666666667, 0.09206349206349206, 0.10151515151515152, 0.10579710144927536, 0.11805555555555565, 0.12933333333333333, 0.1371794871794872, 0.14814814814814814, 0.15, 0.1586206896551726, 0.1688888888888889, 0.16989247311827957, 0.178125, 0.18282828282828284, 0.19215686274509805, 0.20380952380952377, 0.20925925925925937, 0.21621621621621623, 0.22105263157894736, 0.22735042735042735, 0.24, 0.2512195121951219, 0.25555555555555565, 0.2651162790697674, 0.2727272727272727, 0.2814814814814815, 0.28840579710144926, 0.29574468085106387, 0.30555555555555564, 0.3142857142857143, 0.324]}}, "ltt": {"alpha": 0.1, "delta": 0.05, "lambda_hat": null, "coverage": 0.0, "abstain_rate": 1.0, "selective_risk": 0.0, "selective_accuracy": null, "risk_within_budget": true}, "coverage_validity": {"alpha_grid": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "target": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "achieved_risk": [0.0, 0.0, 0.0, 0.08443271767810026, 0.17467248908296942, 0.29086892488954347], "coverage": [0.0, 0.0, 0.0, 0.25266666666666665, 0.458, 0.9053333333333333]}, "feature_diagnostics": {"names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual"], "auroc": {"basin/rho": 0.6013891932695351, "basin/entropy": 0.3986422593972451, "basin/dispersion": 0.5095047929805765, "energy/mean": 0.5601111192279283, "energy/min": 0.5643785358885074, "energy/std": 0.4847708216654086, "curv/lmax_mean": 0.41285074796470805, "curv/lmax_best": 0.4264839571107377, "curv/trace_mean": 0.4132007857079082, "curv/trace_best": 0.4264727964870415, "dynamics/steps": 0.5, "dynamics/monotonic": 0.6341983831300071, "dynamics/drop": 0.6880605676901973, "dynamics/residual": 0.5087742794295501}, "hist": {"basin/rho": {"edges": [0.4166666666666667, 0.44583333333333336, 0.47500000000000003, 0.5041666666666667, 0.5333333333333333, 0.5625, 0.5916666666666667, 0.6208333333333333, 0.65, 0.6791666666666667, 0.7083333333333333, 0.7375, 0.7666666666666666, 0.7958333333333334, 0.825, 0.8541666666666666, 0.8833333333333333, 0.9125, 0.9416666666666667, 0.9708333333333332, 1.0], "correct_counts": [0, 0, 4, 0, 0, 11, 0, 0, 17, 0, 0, 27, 0, 0, 13, 0, 0, 37, 0, 905], "incorrect_counts": [3, 0, 11, 0, 0, 19, 0, 0, 22, 0, 0, 27, 0, 0, 32, 0, 0, 35, 0, 337]}, "basin/entropy": {"edges": [0.0, 0.05387781635334005, 0.1077556327066801, 0.16163344906002014, 0.2155112654133602, 0.26938908176670023, 0.3232668981200403, 0.3771447144733803, 0.4310225308267204, 0.4849003471800604, 0.5387781635334005, 0.5926559798867406, 0.6465337962400806, 0.7004116125934206, 0.7542894289467607, 0.8081672453001008, 0.8620450616534407, 0.9159228780067807, 0.9698006943601208, 1.023678510713461, 1.077556327066801], "correct_counts": [905, 0, 0, 0, 0, 37, 0, 0, 13, 0, 26, 16, 13, 1, 0, 1, 1, 0, 1, 0], "incorrect_counts": [337, 0, 0, 0, 0, 35, 0, 0, 30, 0, 29, 22, 23, 0, 0, 0, 0, 6, 1, 3]}, "basin/dispersion": {"edges": [1.5494269388744626, 1.5753932117990324, 1.6013594847236021, 1.6273257576481717, 1.6532920305727414, 1.6792583034973112, 1.705224576421881, 1.7311908493464507, 1.7571571222710203, 1.78312339519559, 1.8090896681201598, 1.8350559410447296, 1.8610222139692993, 1.8869884868938689, 1.9129547598184387, 1.9389210327430084, 1.9648873056675782, 1.990853578592148, 2.0168198515167175, 2.0427861244412875, 2.068752397365857], "correct_counts": [1, 0, 1, 5, 9, 20, 41, 60, 100, 132, 118, 143, 138, 96, 78, 44, 16, 7, 3, 2], "incorrect_counts": [0, 0, 0, 3, 4, 10, 19, 30, 52, 59, 67, 66, 66, 53, 25, 11, 12, 6, 2, 1]}, "energy/mean": {"edges": [0.35141998281081516, 0.40477818859120207, 0.45813639437158904, 0.511494600151976, 0.5648528059323629, 0.6182110117127497, 0.6715692174931367, 0.7249274232735237, 0.7782856290539106, 0.8316438348342975, 0.8850020406146843, 0.9383602463950713, 0.9917184521754583, 1.0450766579558453, 1.0984348637362322, 1.151793069516619, 1.205151275297006, 1.258509481077393, 1.3118676868577799, 1.3652258926381666, 1.4185840984185536], "correct_counts": [1, 5, 7, 15, 31, 47, 60, 110, 134, 136, 153, 124, 76, 67, 24, 14, 6, 3, 1, 0], "incorrect_counts": [1, 2, 3, 12, 22, 29, 42, 64, 65, 66, 54, 51, 41, 23, 7, 2, 1, 0, 0, 1]}, "energy/min": {"edges": [0.005636274814605713, 0.06192324459552765, 0.11821021437644959, 0.17449718415737153, 0.23078415393829346, 0.2870711237192154, 0.34335809350013735, 0.39964506328105925, 0.4559320330619812, 0.5122190028429031, 0.5685059726238251, 0.624792942404747, 0.681079912185669, 0.737366881966591, 0.7936538517475128, 0.8499408215284348, 0.9062277913093567, 0.9625147610902787, 1.0188017308712005, 1.0750887006521226, 1.1313756704330444], "correct_counts": [0, 4, 12, 17, 26, 45, 70, 99, 99, 133, 142, 135, 100, 66, 30, 18, 10, 4, 3, 1], "incorrect_counts": [3, 1, 4, 14, 15, 27, 50, 56, 59, 65, 48, 70, 33, 28, 5, 5, 2, 0, 0, 1]}, "energy/std": {"edges": [0.08805419068698145, 0.1022881234272592, 0.11652205616753694, 0.1307559889078147, 0.14498992164809243, 0.15922385438837017, 0.17345778712864793, 0.1876917198689257, 0.20192565260920342, 0.21615958534948115, 0.23039351808975891, 0.24462745083003667, 0.2588613835703144, 0.27309531631059214, 0.2873292490508699, 0.30156318179114766, 0.3157971145314254, 0.3300310472717031, 0.34426498001198086, 0.35849891275225865, 0.3727328454925364], "correct_counts": [3, 20, 34, 71, 87, 101, 112, 135, 119, 103, 97, 48, 35, 18, 14, 7, 3, 3, 2, 2], "incorrect_counts": [1, 6, 20, 34, 42, 43, 58, 54, 56, 47, 43, 25, 24, 14, 7, 5, 4, 1, 1, 1]}, "curv/lmax_mean": {"edges": [0.9987722436587015, 0.9993831271926562, 0.9999940107266108, 1.0006048942605654, 1.00121577779452, 1.0018266613284748, 1.0024375448624292, 1.0030484283963839, 1.0036593119303385, 1.0042701954642932, 1.0048810789982479, 1.0054919625322025, 1.006102846066157, 1.0067137296001116, 1.0073246131340663, 1.007935496668021, 1.0085463802019756, 1.0091572637359303, 1.0097681472698847, 1.0103790308038394, 1.010989914337794], "correct_counts": [25, 409, 306, 100, 76, 38, 17, 22, 9, 4, 2, 2, 2, 0, 1, 0, 0, 1, 0, 0], "incorrect_counts": [7, 163, 122, 58, 52, 28, 18, 13, 1, 6, 5, 4, 3, 2, 1, 0, 0, 0, 1, 2]}, "curv/lmax_best": {"edges": [0.994626522064209, 0.996795529127121, 0.998964536190033, 1.001133543252945, 1.003302550315857, 1.005471557378769, 1.0076405644416808, 1.009809571504593, 1.0119785785675048, 1.014147585630417, 1.0163165926933289, 1.0184855997562408, 1.0206546068191529, 1.0228236138820648, 1.024992620944977, 1.0271616280078888, 1.0293306350708007, 1.0314996421337128, 1.0336686491966247, 1.0358376562595368, 1.0380066633224487], "correct_counts": [2, 32, 858, 74, 12, 17, 9, 6, 1, 2, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0], "incorrect_counts": [0, 8, 380, 49, 20, 10, 8, 4, 4, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 1]}, "curv/trace_mean": {"edges": [31.937888781229656, 31.94588975906372, 31.953890736897787, 31.961891714731852, 31.96989269256592, 31.977893670399986, 31.98589464823405, 31.993895626068117, 32.00189660390218, 32.00989758173625, 32.01789855957031, 32.02589953740438, 32.03390051523844, 32.04190149307251, 32.04990247090658, 32.057903448740646, 32.06590442657471, 32.07390540440878, 32.08190638224284, 32.08990736007691, 32.09790833791097], "correct_counts": [2, 2, 7, 13, 44, 105, 162, 171, 165, 124, 86, 62, 42, 14, 7, 6, 0, 1, 1, 0], "incorrect_counts": [1, 0, 0, 2, 19, 37, 62, 71, 62, 58, 56, 53, 25, 21, 6, 8, 2, 2, 0, 1]}, "curv/trace_best": {"edges": [31.915437698364258, 31.92645502090454, 31.937472343444824, 31.948489665985107, 31.95950698852539, 31.970524311065674, 31.981541633605957, 31.99255895614624, 32.00357627868652, 32.01459360122681, 32.02561092376709, 32.03662824630737, 32.047645568847656, 32.05866289138794, 32.06968021392822, 32.080697536468506, 32.09171485900879, 32.10273218154907, 32.113749504089355, 32.12476682662964, 32.13578414916992], "correct_counts": [1, 3, 4, 15, 24, 72, 187, 272, 219, 83, 64, 32, 11, 14, 5, 3, 1, 2, 2, 0], "incorrect_counts": [0, 0, 1, 4, 8, 28, 72, 113, 98, 67, 31, 25, 20, 7, 2, 2, 4, 0, 0, 4]}, "dynamics/steps": {"edges": [1.0, 1.05, 1.1, 1.15, 1.2, 1.25, 1.3, 1.35, 1.4, 1.45, 1.5, 1.55, 1.6, 1.65, 1.7000000000000002, 1.75, 1.8, 1.85, 1.9, 1.9500000000000002, 2.0], "correct_counts": [1014, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [486, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "dynamics/monotonic": {"edges": [0.6266666666666666, 0.6335833333333333, 0.6405, 0.6474166666666666, 0.6543333333333333, 0.6612499999999999, 0.6681666666666666, 0.6750833333333333, 0.6819999999999999, 0.6889166666666666, 0.6958333333333333, 0.70275, 0.7096666666666667, 0.7165833333333333, 0.7235, 0.7304166666666667, 0.7373333333333333, 0.74425, 0.7511666666666666, 0.7580833333333333, 0.765], "correct_counts": [0, 1, 0, 2, 11, 27, 70, 66, 104, 139, 106, 151, 103, 104, 56, 41, 17, 8, 5, 3], "incorrect_counts": [1, 0, 2, 1, 9, 19, 51, 53, 70, 86, 70, 49, 38, 21, 6, 7, 3, 0, 0, 0]}, "dynamics/drop": {"edges": [13.154254739483198, 18.10311585428814, 23.051976969093083, 28.000838083898028, 32.94969919870297, 37.898560313507915, 42.847421428312856, 47.796282543117805, 52.745143657922746, 57.69400477272769, 62.64286588753263, 67.59172700233758, 72.54058811714252, 77.48944923194746, 82.4383103467524, 87.38717146155734, 92.33603257636229, 97.28489369116723, 102.23375480597217, 107.18261592077711, 112.13147703558207], "correct_counts": [30, 195, 219, 177, 127, 90, 75, 41, 23, 14, 11, 6, 2, 0, 0, 1, 2, 0, 0, 1], "incorrect_counts": [47, 166, 126, 86, 28, 17, 9, 4, 2, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "dynamics/residual": {"edges": [1.1386153648296993, 1.1545183179279168, 1.1704212710261346, 1.1863242241243521, 1.2022271772225699, 1.2181301303207874, 1.2340330834190052, 1.2499360365172227, 1.2658389896154405, 1.281741942713658, 1.2976448958118758, 1.3135478489100934, 1.329450802008311, 1.3453537551065287, 1.3612567082047462, 1.377159661302964, 1.3930626144011815, 1.4089655674993993, 1.4248685205976168, 1.4407714736958344, 1.4566744267940521], "correct_counts": [3, 6, 12, 16, 39, 62, 84, 127, 128, 130, 105, 107, 82, 46, 33, 17, 8, 3, 4, 2], "incorrect_counts": [0, 5, 7, 9, 16, 35, 39, 59, 65, 55, 57, 52, 34, 24, 17, 5, 5, 2, 0, 0]}}}, "feature_ablation": {"full": 0.171055380110865, "drop_basin": 0.1829945994753527, "basin_only": 0.26530562690580933, "drop_energy": 0.17457640922091583, "energy_only": 0.2742022787633486, "drop_curv": 0.17754820658338663, "curv_only": 0.2744174989631136, "drop_dynamics": 0.23819230827142496, "dynamics_only": 0.19387207186931513}, "ece_geometry": 0.02548324869690796, "accuracy_ood": 0.3373333333333333, "aurc_ood": {"geometry": 0.6989345686143652, "rho_basin": 0.6359528685037144, "energy_min": 0.7674459887692104, "energy_mean": 0.7665803824715032, "energy_std": 0.6744965485414989, "msp": 0.6434757978241759, "temp_msp": 0.6232780631954414, "entropy": 0.6410680306163067, "softmax_learned": 0.6172112155416478, "geom_softmax": 0.6429259653235739}, "ood_ltt": {"alpha": 0.1, "lambda_hat": null, "selective_risk": 0.0, "coverage": 0.0, "risk_within_budget": true}, "ood_validity": {"target": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "id_risk": [0.0, 0.0, 0.0, 0.08443271767810026, 0.17467248908296942, 0.29086892488954347], "ood_risk": [0.0, 0.0, 0.0, 0.7046332046332047, 0.6484284051222352, 0.6544011544011544], "id_coverage": [0.0, 0.0, 0.0, 0.25266666666666665, 0.458, 0.9053333333333333], "ood_coverage": [0.0, 0.0, 0.0, 0.3453333333333333, 0.5726666666666667, 0.924]}}, "env": {"python": "3.12.13", "platform": "macOS-26.5.2-arm64-arm-64bit", "jax": "0.11.0", "jax_backend": "cpu"}, "artifact_paths": []} +{"run_id": "ed919980f35e", "timestamp": "2026-07-30T11:18:18.349654+00:00", "git_sha": "67dc7a2", "config_hash": "9345e2bfd6b6", "config": {"run": {"seed": 4, "task": "graph_planning", "notes": "graph adaptive-halting run"}, "model": {"latent_dim": 32, "hidden_dim": 128, "context_dim": 64}, "inference": {"k_restarts": 12, "steps": 50, "step_size": 0.1, "temperature": 0.05, "init_scale": 1.0, "grad_tol": 0.001, "sampler": "langevin", "anneal_levels": 10, "anneal_steps_per_level": 8, "anneal_step_max": 0.2, "anneal_step_min": 0.01}, "train": {"epochs": 25, "batch_size": 128, "lr": 0.001, "n_train": 8000, "n_neg": 4, "neg_noise": 0.5, "objective": "basin_center", "ired_noise_min": 0.1, "ired_noise_max": 1.5, "ired_ridge": 0.05, "ired_decode_weight": 1.0, "ired_stat_weight": 1.0}, "eval": {"n_eval": 800, "richer_geometry": false}, "conformal": {"alpha": 0.1, "delta": 0.05, "n_calib": 800}, "task": {"graph_planning": {"n_nodes": 7, "ood_n_nodes": 10, "edge_prob": 0.4, "max_len": 4}}}, "task": "graph_planning", "split": "halting", "seed": 4, "metrics": {"k_restarts": 12, "steps": 50, "alpha": 0.1, "n_calib": 800, "n_test": 800, "final_train_loss": 0.21295076608657837, "tau_hat": 0.8333, "compute_used": 0.16995000000000002, "compute_saved": 0.83005, "halting_risk": 0.0625, "risk_within_budget": true, "full_accuracy": 0.7125, "halted_accuracy": 0.69875, "accuracy_drop": 0.01375000000000004, "risk_monotone_in_tau": true, "calib_risk_by_tau": [0.47375, 0.47375, 0.47375, 0.47375, 0.37875, 0.37875, 0.305, 0.1525, 0.1525, 0.0925, 0.00375, 0.00375], "calib_compute_by_tau": [0.0, 0.0, 0.0, 0.0002, 0.02342499999999982, 0.02342499999999982, 0.04669999999999976, 0.12322499999999997, 0.12322499999999997, 0.16990000000000008, 0.3011249999999997, 0.3011249999999997], "tau_sweep": {"taus": [0.0833, 0.1667, 0.25, 0.3333, 0.4167, 0.5, 0.5833, 0.6667, 0.75, 0.8333, 0.9167, 1.0], "compute_used": [0.0, 0.0, 0.0, 0.0006000000000000001, 0.02255, 0.02255, 0.04832500000000001, 0.12350000000000001, 0.12350000000000001, 0.16995000000000002, 0.28452500000000003, 0.28452500000000003], "accuracy": [0.50875, 0.50875, 0.50875, 0.50875, 0.575, 0.575, 0.60875, 0.68125, 0.68125, 0.69875, 0.715, 0.715], "disagreement": [0.46, 0.46, 0.46, 0.4575, 0.365, 0.365, 0.29625, 0.12625, 0.12625, 0.0625, 0.0075, 0.0075]}}, "env": {"python": "3.12.13", "platform": "macOS-26.5.2-arm64-arm-64bit", "jax": "0.11.0", "jax_backend": "cpu"}, "artifact_paths": []} +{"run_id": "8bc3880fccd3", "timestamp": "2026-07-30T11:18:25.395740+00:00", "git_sha": "67dc7a2", "config_hash": "e71a6b480861", "config": {"run": {"seed": 3, "task": "graph_planning", "notes": "E2 graph selective-prediction"}, "model": {"latent_dim": 32, "hidden_dim": 128, "context_dim": 64}, "inference": {"k_restarts": 12, "steps": 50, "step_size": 0.1, "temperature": 0.05, "init_scale": 1.0, "grad_tol": 0.001, "sampler": "langevin", "anneal_levels": 10, "anneal_steps_per_level": 8, "anneal_step_max": 0.2, "anneal_step_min": 0.01}, "train": {"epochs": 25, "batch_size": 128, "lr": 0.001, "n_train": 8000, "n_neg": 4, "neg_noise": 0.5, "objective": "basin_center", "ired_noise_min": 0.1, "ired_noise_max": 1.5, "ired_ridge": 0.05, "ired_decode_weight": 1.0, "ired_stat_weight": 1.0}, "eval": {"n_eval": 1500, "richer_geometry": false}, "conformal": {"alpha": 0.1, "delta": 0.05, "n_calib": 1500}, "task": {"graph_planning": {"n_nodes": 7, "ood_n_nodes": 10, "edge_prob": 0.4, "max_len": 4}}}, "task": "graph_planning", "split": "selective", "seed": 3, "metrics": {"n_fit": 1500, "n_calib": 1500, "n_test": 1500, "k_restarts": 12, "objective": "basin_center", "sampler": "langevin", "feature_set": "base", "accuracy_id": 0.6993333333333334, "base_error": 0.3006666666666667, "final_train_loss": 0.16939383745193481, "feature_names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual"], "aurc": {"geometry": 0.1692573776986855, "rho_basin": 0.24609437563326478, "energy_min": 0.3844670869106948, "energy_mean": 0.38671775365041733, "energy_std": 0.3004012009214803, "msp": 0.13953947812056736, "temp_msp": 0.1394850202783934, "entropy": 0.1389181572354234, "softmax_learned": 0.14027532142085386, "geom_softmax": 0.14356994986612653}, "temperature": 3.4489855894245665, "best_energy_baseline": "energy_std", "best_baseline": "entropy", "delta_aurc_vs_energy_min": [0.2152097092120093, 0.17912220172475968, 0.24781429220048076], "delta_aurc_vs_best_energy": [0.1311438232227948, 0.10332126037643946, 0.1586253487976973], "delta_aurc_vs_best_baseline": [-0.030339220463262123, -0.04263629548652781, -0.01865656419709403], "delta_aurc_geom_adds": [-0.0032946284452726737, -0.007710391733187864, 0.00034702386219870264], "geometry_wins": true, "geometry_wins_vs_baseline": false, "geometry_adds_over_softmax": false, "risk_coverage": {"geometry": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.03333333333333333, 0.05, 0.044444444444444446, 0.041666666666666664, 0.04, 0.055555555555555546, 0.06666666666666649, 0.08333333333333333, 0.09259259259259259, 0.09666666666666668, 0.10606060606060606, 0.1111111111111111, 0.12051282051282051, 0.12142857142857143, 0.12888888888888886, 0.13125, 0.13529411764705881, 0.1444444444444444, 0.15087719298245614, 0.15333333333333332, 0.15555555555555567, 0.1621212121212121, 0.16521739130434782, 0.1708333333333333, 0.172, 0.1782051282051282, 0.18271604938271604, 0.18452380952380953, 0.18965517241379307, 0.19333333333333333, 0.1989247311827957, 0.20416666666666666, 0.2111111111111111, 0.21568627450980393, 0.2200000000000001, 0.22129629629629627, 0.22612612612612612, 0.2307017543859649, 0.23675213675213674, 0.24083333333333334, 0.24471544715447152, 0.2452380952380952, 0.25193798449612403, 0.26212121212121214, 0.26814814814814814, 0.27608695652173915, 0.28297872340425545, 0.2881944444444445, 0.2925170068027211, 0.3006666666666667]}, "rho_basin": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.24052718286655675, 0.24052718286655714, 0.2405271828665573, 0.24052718286655736, 0.24052718286655742, 0.24052718286655744, 0.24052718286655747, 0.2405271828665575, 0.2405271828665575, 0.24052718286655753, 0.24052718286655753, 0.24052718286655753, 0.24052718286655753, 0.24052718286655755, 0.24052718286655755, 0.24052718286655755, 0.24052718286655755, 0.24052718286655736, 0.2405271828665566, 0.24052718286655594, 0.24052718286655536, 0.2405271828665548, 0.2405271828665543, 0.24052718286655386, 0.24052718286655345, 0.24052718286655306, 0.2405271828665527, 0.24052718286655236, 0.24052718286655206, 0.24052718286655178, 0.2405271828665515, 0.24052718286655125, 0.240527182866551, 0.24052718286655078, 0.24052718286655056, 0.24052718286655075, 0.24052718286655134, 0.24052718286655186, 0.2405271828665524, 0.24052718286655286, 0.2436781609195365, 0.24937055281882495, 0.254798182304193, 0.2613517992424205, 0.26839120370370007, 0.27427536231883703, 0.2794175343292557, 0.28733747044916935, 0.2942050894431817, 0.3006666666666636]}, "energy_min": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.5666666666666667, 0.55, 0.5222222222222223, 0.525, 0.49333333333333335, 0.4555555555555556, 0.4619047619047618, 0.45416666666666666, 0.4444444444444444, 0.43000000000000005, 0.41515151515151516, 0.4222222222222222, 0.4153846153846154, 0.4023809523809524, 0.4044444444444444, 0.38958333333333334, 0.4, 0.39999999999999997, 0.39473684210526316, 0.39166666666666666, 0.38888888888888884, 0.3893939393939394, 0.3826086956521739, 0.37638888888888883, 0.376, 0.3769230769230769, 0.37777777777777777, 0.37142857142857144, 0.36896551724137927, 0.36666666666666664, 0.36451612903225805, 0.359375, 0.3565656565656566, 0.35, 0.3447619047619049, 0.3444444444444444, 0.3441441441441441, 0.3368421052631579, 0.3299145299145299, 0.32416666666666666, 0.32032520325203245, 0.31746031746031755, 0.3147286821705426, 0.31136363636363634, 0.30666666666666664, 0.30434782608695654, 0.3056737588652483, 0.3055555555555555, 0.30272108843537415, 0.3006666666666667]}, "energy_mean": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.6, 0.5, 0.5111111111111111, 0.5416666666666666, 0.5, 0.4833333333333333, 0.46190476190476204, 0.4375, 0.4148148148148148, 0.4133333333333334, 0.4121212121212121, 0.40555555555555556, 0.4230769230769231, 0.42857142857142855, 0.4288888888888888, 0.4166666666666667, 0.40784313725490196, 0.4074074074074075, 0.39824561403508774, 0.39, 0.3857142857142857, 0.3803030303030303, 0.37681159420289856, 0.3777777777777778, 0.37466666666666665, 0.36923076923076925, 0.36666666666666664, 0.36666666666666664, 0.36436781609195396, 0.3622222222222222, 0.36236559139784946, 0.36041666666666666, 0.35454545454545455, 0.35, 0.3428571428571428, 0.33703703703703697, 0.3324324324324324, 0.3298245614035088, 0.32735042735042735, 0.32666666666666666, 0.32520325203252026, 0.31904761904761897, 0.31705426356589145, 0.31363636363636366, 0.3148148148148148, 0.3101449275362319, 0.3085106382978723, 0.3055555555555555, 0.3034013605442177, 0.3006666666666667]}, "energy_std": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.3, 0.25, 0.26666666666666666, 0.30833333333333335, 0.32, 0.3277777777777777, 0.31428571428571417, 0.3125, 0.3074074074074074, 0.29666666666666675, 0.29393939393939394, 0.2833333333333333, 0.28717948717948716, 0.2976190476190476, 0.30222222222222217, 0.29791666666666666, 0.2980392156862745, 0.29444444444444456, 0.3017543859649123, 0.295, 0.29841269841269835, 0.3015151515151515, 0.3072463768115942, 0.3083333333333334, 0.30133333333333334, 0.2987179487179487, 0.29506172839506173, 0.2976190476190476, 0.29540229885057484, 0.29444444444444445, 0.2913978494623656, 0.296875, 0.29494949494949496, 0.296078431372549, 0.298095238095238, 0.2981481481481481, 0.3, 0.3017543859649123, 0.30085470085470084, 0.3, 0.30162601626016255, 0.3007936507936507, 0.3023255813953488, 0.29924242424242425, 0.3, 0.30144927536231886, 0.30212765957446813, 0.30347222222222214, 0.3034013605442177, 0.3006666666666667]}, "msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.03333333333333333, 0.016666666666666666, 0.022222222222222223, 0.025, 0.02, 0.033333333333333444, 0.042857142857142864, 0.0375, 0.037037037037037035, 0.04333333333333334, 0.051515151515151514, 0.05, 0.05384615384615385, 0.06190476190476191, 0.06444444444444443, 0.075, 0.07647058823529412, 0.08518518518518532, 0.08771929824561403, 0.09666666666666666, 0.10000000000000013, 0.10757575757575757, 0.12173913043478261, 0.1236111111111111, 0.128, 0.13974358974358975, 0.15185185185185185, 0.1595238095238095, 0.16781609195402297, 0.17444444444444446, 0.17956989247311828, 0.184375, 0.1898989898989899, 0.19509803921568628, 0.20190476190476203, 0.20740740740740737, 0.21081081081081082, 0.21929824561403508, 0.22393162393162394, 0.23166666666666666, 0.23739837398373995, 0.24365079365079376, 0.2527131782945736, 0.25757575757575757, 0.26666666666666666, 0.2746376811594203, 0.28014184397163117, 0.28750000000000003, 0.2945578231292517, 0.3006666666666667]}, "temp_msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.03333333333333333, 0.016666666666666666, 0.03333333333333333, 0.025, 0.02, 0.03888888888888888, 0.042857142857142864, 0.0375, 0.044444444444444446, 0.05000000000000001, 0.051515151515151514, 0.05277777777777778, 0.05897435897435897, 0.06666666666666667, 0.0711111111111111, 0.07916666666666666, 0.0803921568627451, 0.08888888888888888, 0.0912280701754386, 0.1, 0.10793650793650793, 0.11666666666666667, 0.12753623188405797, 0.13611111111111107, 0.14266666666666666, 0.14871794871794872, 0.1506172839506173, 0.14761904761904762, 0.15172413793103465, 0.16, 0.16774193548387098, 0.17395833333333333, 0.18181818181818182, 0.19019607843137254, 0.1933333333333333, 0.19814814814814827, 0.2036036036036036, 0.21403508771929824, 0.2230769230769231, 0.23, 0.23577235772357735, 0.24365079365079376, 0.2503875968992248, 0.25681818181818183, 0.26666666666666666, 0.27318840579710146, 0.2794326241134753, 0.28750000000000003, 0.2945578231292517, 0.3006666666666667]}, "entropy": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.03333333333333333, 0.016666666666666666, 0.022222222222222223, 0.025, 0.02, 0.033333333333333444, 0.042857142857142864, 0.0375, 0.037037037037037035, 0.04333333333333334, 0.051515151515151514, 0.05, 0.05384615384615385, 0.06190476190476191, 0.06444444444444443, 0.075, 0.0784313725490196, 0.08518518518518517, 0.08947368421052632, 0.09666666666666666, 0.10158730158730157, 0.11060606060606061, 0.12173913043478261, 0.1236111111111111, 0.13066666666666665, 0.1371794871794872, 0.15185185185185185, 0.1595238095238095, 0.1701149425287356, 0.17333333333333334, 0.17849462365591398, 0.18125, 0.18686868686868688, 0.19607843137254902, 0.19904761904761903, 0.2046296296296296, 0.21081081081081082, 0.2149122807017544, 0.2222222222222222, 0.22833333333333333, 0.23495934959349604, 0.24285714285714283, 0.25116279069767444, 0.25681818181818183, 0.26222222222222225, 0.27246376811594203, 0.2780141843971631, 0.2881944444444445, 0.2945578231292517, 0.3006666666666667]}, "softmax_learned": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.03333333333333333, 0.016666666666666666, 0.03333333333333333, 0.025, 0.02, 0.03888888888888888, 0.042857142857142864, 0.0375, 0.044444444444444446, 0.05000000000000001, 0.051515151515151514, 0.05277777777777778, 0.05641025641025641, 0.06666666666666667, 0.0711111111111111, 0.07916666666666666, 0.0803921568627451, 0.09259259259259257, 0.0912280701754386, 0.1, 0.11111111111111122, 0.11666666666666667, 0.1289855072463768, 0.13750000000000012, 0.13866666666666666, 0.14871794871794872, 0.14814814814814814, 0.1523809523809524, 0.15632183908045993, 0.16, 0.16344086021505377, 0.17291666666666666, 0.17777777777777778, 0.18529411764705883, 0.19428571428571442, 0.20092592592592604, 0.21171171171171171, 0.21842105263157896, 0.2264957264957265, 0.23166666666666666, 0.2373983739837398, 0.24920634920634932, 0.2558139534883721, 0.2628787878787879, 0.2688888888888889, 0.27318840579710146, 0.28085106382978736, 0.2895833333333334, 0.2965986394557823, 0.3006666666666667]}, "geom_softmax": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.03333333333333333, 0.03333333333333333, 0.044444444444444446, 0.041666666666666664, 0.04666666666666667, 0.04444444444444444, 0.04761904761904762, 0.04583333333333333, 0.05555555555555555, 0.05000000000000001, 0.051515151515151514, 0.05555555555555555, 0.0641025641025641, 0.06904761904761905, 0.0711111111111111, 0.07708333333333334, 0.07647058823529412, 0.08333333333333347, 0.09649122807017543, 0.10333333333333333, 0.10952380952380951, 0.11818181818181818, 0.12753623188405797, 0.13888888888888887, 0.14933333333333335, 0.15, 0.15679012345679014, 0.15714285714285714, 0.1597701149425287, 0.16111111111111112, 0.17419354838709677, 0.18229166666666666, 0.18686868686868688, 0.19215686274509805, 0.19809523809523824, 0.20833333333333331, 0.2126126126126126, 0.21666666666666667, 0.2205128205128205, 0.2275, 0.2373983739837398, 0.24285714285714297, 0.2503875968992248, 0.2606060606060606, 0.2688888888888889, 0.2746376811594203, 0.28297872340425545, 0.29027777777777786, 0.29523809523809524, 0.3006666666666667]}}, "ltt": {"alpha": 0.1, "delta": 0.05, "lambda_hat": null, "coverage": 0.0, "abstain_rate": 1.0, "selective_risk": 0.0, "selective_accuracy": null, "risk_within_budget": true}, "coverage_validity": {"alpha_grid": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "target": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "achieved_risk": [0.0, 0.0, 0.0, 0.04, 0.1474820143884892, 0.24862313139260425], "coverage": [0.0, 0.0, 0.0, 0.08333333333333333, 0.37066666666666664, 0.8473333333333334]}, "feature_diagnostics": {"names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual"], "auroc": {"basin/rho": 0.6182722855047252, "basin/entropy": 0.3819771337500185, "basin/dispersion": 0.5171306639836483, "energy/mean": 0.6203141414376272, "energy/min": 0.6201915455327532, "energy/std": 0.4974624761413573, "curv/lmax_mean": 0.40259966730007885, "curv/lmax_best": 0.42140334264075807, "curv/trace_mean": 0.3792493748665713, "curv/trace_best": 0.3965438523438012, "dynamics/steps": 0.5, "dynamics/monotonic": 0.6171382733846404, "dynamics/drop": 0.658876894688004, "dynamics/residual": 0.4726389191268635}, "hist": {"basin/rho": {"edges": [0.4166666666666667, 0.44583333333333336, 0.47500000000000003, 0.5041666666666667, 0.5333333333333333, 0.5625, 0.5916666666666667, 0.6208333333333333, 0.65, 0.6791666666666667, 0.7083333333333333, 0.7375, 0.7666666666666666, 0.7958333333333334, 0.825, 0.8541666666666666, 0.8833333333333333, 0.9125, 0.9416666666666667, 0.9708333333333332, 1.0], "correct_counts": [0, 0, 7, 0, 0, 11, 0, 0, 16, 0, 0, 21, 0, 0, 27, 0, 0, 45, 0, 922], "incorrect_counts": [2, 0, 10, 0, 0, 16, 0, 0, 31, 0, 0, 21, 0, 0, 37, 0, 0, 42, 0, 292]}, "basin/entropy": {"edges": [0.0, 0.05387781635334005, 0.1077556327066801, 0.16163344906002014, 0.2155112654133602, 0.26938908176670023, 0.3232668981200403, 0.3771447144733803, 0.4310225308267204, 0.4849003471800604, 0.5387781635334005, 0.5926559798867406, 0.6465337962400806, 0.7004116125934206, 0.7542894289467607, 0.8081672453001008, 0.8620450616534407, 0.9159228780067807, 0.9698006943601208, 1.023678510713461, 1.077556327066801], "correct_counts": [922, 0, 0, 0, 0, 45, 0, 0, 26, 0, 20, 14, 18, 2, 0, 2, 0, 0, 0, 0], "incorrect_counts": [292, 0, 0, 0, 0, 42, 0, 0, 37, 0, 21, 28, 21, 0, 0, 3, 0, 3, 2, 2]}, "basin/dispersion": {"edges": [1.6360123866978022, 1.6586197833827276, 1.6812271800676528, 1.7038345767525782, 1.7264419734375036, 1.749049370122429, 1.7716567668073542, 1.7942641634922796, 1.816871560177205, 1.8394789568621301, 1.8620863535470555, 1.884693750231981, 1.907301146916906, 1.9299085436018315, 1.952515940286757, 1.975123336971682, 1.9977307336566075, 2.020338130341533, 2.042945527026458, 2.0655529237113837, 2.088160320396309], "correct_counts": [6, 20, 22, 37, 64, 93, 127, 135, 130, 132, 92, 72, 54, 31, 12, 12, 8, 0, 1, 1], "incorrect_counts": [6, 2, 14, 24, 28, 35, 62, 54, 59, 52, 38, 36, 17, 9, 10, 3, 2, 0, 0, 0]}, "energy/mean": {"edges": [0.3838568776845932, 0.4383304126560688, 0.4928039476275444, 0.54727748259902, 0.6017510175704956, 0.6562245525419712, 0.7106980875134468, 0.7651716224849223, 0.819645157456398, 0.8741186924278737, 0.9285922273993492, 0.9830657623708248, 1.0375392973423003, 1.092012832313776, 1.1464863672852514, 1.2009599022567272, 1.2554334372282028, 1.3099069721996783, 1.3643805071711541, 1.4188540421426297, 1.4733275771141052], "correct_counts": [2, 0, 1, 4, 18, 35, 82, 85, 88, 112, 96, 98, 124, 85, 91, 75, 28, 14, 8, 3], "incorrect_counts": [0, 1, 1, 10, 16, 40, 40, 61, 47, 41, 46, 43, 25, 29, 23, 15, 8, 4, 1, 0]}, "energy/min": {"edges": [0.027186661958694458, 0.08629510253667831, 0.14540354311466216, 0.204511983692646, 0.26362042427062987, 0.32272886484861374, 0.38183730542659755, 0.4409457460045814, 0.5000541865825653, 0.5591626271605491, 0.618271067738533, 0.6773795083165168, 0.7364879488945006, 0.7955963894724846, 0.8547048300504684, 0.9138132706284523, 0.9729217112064361, 1.03203015178442, 1.0911385923624037, 1.1502470329403878, 1.2093554735183716], "correct_counts": [2, 1, 3, 4, 10, 29, 65, 87, 108, 97, 108, 132, 100, 108, 80, 51, 38, 20, 5, 1], "incorrect_counts": [0, 2, 3, 3, 16, 34, 39, 45, 64, 55, 50, 43, 30, 13, 17, 19, 12, 4, 2, 0]}, "energy/std": {"edges": [0.07382017921196296, 0.09055781346466538, 0.10729544771736779, 0.12403308197007021, 0.14077071622277262, 0.15750835047547504, 0.17424598472817746, 0.19098361898087987, 0.2077212532335823, 0.2244588874862847, 0.24119652173898712, 0.25793415599168956, 0.2746717902443919, 0.2914094244970944, 0.30814705874979675, 0.3248846930024992, 0.3416223272552016, 0.35835996150790406, 0.3750975957606064, 0.3918352300133089, 0.40857286426601125], "correct_counts": [1, 15, 27, 55, 114, 137, 142, 168, 125, 101, 66, 37, 24, 18, 9, 6, 1, 1, 0, 2], "incorrect_counts": [1, 7, 6, 31, 44, 60, 67, 55, 63, 45, 29, 22, 12, 5, 3, 1, 0, 0, 0, 0]}, "curv/lmax_mean": {"edges": [0.9988120545943578, 0.9993241287767888, 0.9998362029592196, 1.0003482771416505, 1.0008603513240815, 1.0013724255065124, 1.0018844996889433, 1.0023965738713743, 1.002908648053805, 1.003420722236236, 1.0039327964186668, 1.0044448706010978, 1.0049569447835287, 1.0054690189659596, 1.0059810931483906, 1.0064931673308215, 1.0070052415132524, 1.0075173156956831, 1.008029389878114, 1.008541464060545, 1.009053538242976], "correct_counts": [74, 546, 194, 97, 68, 34, 12, 9, 8, 3, 1, 0, 1, 1, 0, 0, 0, 0, 0, 1], "incorrect_counts": [20, 160, 119, 73, 36, 18, 8, 8, 3, 3, 1, 0, 0, 1, 0, 1, 0, 0, 0, 0]}, "curv/lmax_best": {"edges": [0.9947617650032043, 0.9963587850332261, 0.9979558050632477, 0.9995528250932694, 1.001149845123291, 1.0027468651533127, 1.0043438851833344, 1.005940905213356, 1.0075379252433776, 1.0091349452733993, 1.010731965303421, 1.0123289853334427, 1.0139260053634644, 1.015523025393486, 1.0171200454235076, 1.0187170654535294, 1.020314085483551, 1.0219111055135728, 1.0235081255435943, 1.025105145573616, 1.0267021656036377], "correct_counts": [5, 26, 176, 761, 49, 14, 9, 6, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1], "incorrect_counts": [0, 6, 66, 325, 30, 11, 6, 4, 0, 0, 1, 1, 0, 1, 0, 0, 0, 0, 0, 0]}, "curv/trace_mean": {"edges": [31.904047807057697, 31.912987621625263, 31.92192743619283, 31.930867250760397, 31.93980706532796, 31.948746879895527, 31.957686694463092, 31.96662650903066, 31.975566323598226, 31.98450613816579, 31.99344595273336, 32.00238576730092, 32.01132558186849, 32.02026539643606, 32.02920521100362, 32.03814502557119, 32.04708484013875, 32.05602465470632, 32.06496446927389, 32.07390428384145, 32.08284409840902], "correct_counts": [3, 7, 33, 45, 64, 85, 106, 128, 85, 99, 93, 93, 107, 63, 22, 15, 0, 0, 0, 1], "incorrect_counts": [0, 1, 4, 9, 13, 25, 22, 40, 37, 54, 55, 54, 65, 44, 16, 10, 1, 1, 0, 0]}, "curv/trace_best": {"edges": [31.818470001220703, 31.8322847366333, 31.8460994720459, 31.859914207458495, 31.873728942871093, 31.88754367828369, 31.90135841369629, 31.915173149108888, 31.928987884521483, 31.94280261993408, 31.95661735534668, 31.970432090759278, 31.984246826171876, 31.99806156158447, 32.01187629699707, 32.02569103240967, 32.03950576782226, 32.053320503234865, 32.06713523864746, 32.08094997406006, 32.094764709472656], "correct_counts": [1, 0, 0, 4, 3, 11, 18, 37, 62, 91, 120, 152, 183, 189, 104, 40, 23, 5, 4, 2], "incorrect_counts": [0, 0, 0, 0, 1, 2, 2, 11, 13, 22, 34, 64, 79, 101, 73, 30, 12, 5, 0, 2]}, "dynamics/steps": {"edges": [1.0, 1.05, 1.1, 1.15, 1.2, 1.25, 1.3, 1.35, 1.4, 1.45, 1.5, 1.55, 1.6, 1.65, 1.7000000000000002, 1.75, 1.8, 1.85, 1.9, 1.9500000000000002, 2.0], "correct_counts": [1049, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [451, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "dynamics/monotonic": {"edges": [0.6383333333333333, 0.64425, 0.6501666666666667, 0.6560833333333334, 0.662, 0.6679166666666667, 0.6738333333333333, 0.67975, 0.6856666666666666, 0.6915833333333333, 0.6975, 0.7034166666666667, 0.7093333333333334, 0.71525, 0.7211666666666667, 0.7270833333333333, 0.733, 0.7389166666666667, 0.7448333333333333, 0.75075, 0.7566666666666667], "correct_counts": [1, 3, 7, 17, 31, 48, 58, 112, 80, 141, 139, 94, 114, 59, 54, 36, 29, 17, 5, 4], "incorrect_counts": [1, 4, 5, 13, 15, 41, 39, 54, 53, 60, 62, 37, 27, 20, 7, 7, 4, 0, 2, 0]}, "dynamics/drop": {"edges": [13.959155107537905, 18.96183481787642, 23.96451452821493, 28.967194238553446, 33.96987394889196, 38.97255365923047, 43.97523336956898, 48.977913079907495, 53.98059279024601, 58.98327250058452, 63.98595221092303, 68.98863192126154, 73.99131163160007, 78.99399134193858, 83.9966710522771, 88.9993507626156, 94.00203047295412, 99.00471018329263, 104.00738989363114, 109.01006960396965, 114.01274931430817], "correct_counts": [79, 187, 237, 186, 131, 96, 46, 37, 21, 12, 6, 5, 1, 4, 0, 0, 0, 0, 0, 1], "incorrect_counts": [55, 154, 109, 67, 37, 17, 6, 3, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "dynamics/residual": {"edges": [1.157560517390569, 1.1732419848442077, 1.1889234522978465, 1.2046049197514852, 1.220286387205124, 1.2359678546587625, 1.2516493221124012, 1.26733078956604, 1.2830122570196787, 1.2986937244733174, 1.3143751919269562, 1.330056659380595, 1.3457381268342337, 1.3614195942878724, 1.3771010617415111, 1.3927825291951499, 1.4084639966487886, 1.4241454641024274, 1.439826931556066, 1.4555083990097046, 1.4711898664633434], "correct_counts": [3, 8, 18, 25, 60, 82, 109, 133, 138, 137, 115, 81, 48, 43, 25, 11, 6, 6, 0, 1], "incorrect_counts": [1, 3, 6, 13, 16, 30, 48, 44, 73, 63, 52, 36, 28, 22, 6, 4, 2, 3, 0, 1]}}}, "feature_ablation": {"full": 0.1692573776986855, "drop_basin": 0.18552230550261628, "basin_only": 0.2468147564151358, "drop_energy": 0.16719687851591933, "energy_only": 0.2326850496812418, "drop_curv": 0.16787121777273564, "curv_only": 0.21755660344088715, "drop_dynamics": 0.20149229331418345, "dynamics_only": 0.1919227889302135}, "ece_geometry": 0.03336447508959324, "accuracy_ood": 0.3913333333333333, "aurc_ood": {"geometry": 0.5517668728716215, "rho_basin": 0.5982679701853475, "energy_min": 0.7015110277968405, "energy_mean": 0.6965205335334852, "energy_std": 0.5928168612920839, "msp": 0.5847105053531803, "temp_msp": 0.5671211036886489, "entropy": 0.5831236345427478, "softmax_learned": 0.5653040354208368, "geom_softmax": 0.5762981866215301}, "ood_ltt": {"alpha": 0.1, "lambda_hat": null, "selective_risk": 0.0, "coverage": 0.0, "risk_within_budget": true}, "ood_validity": {"target": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "id_risk": [0.0, 0.0, 0.0, 0.04, 0.1474820143884892, 0.24862313139260425], "ood_risk": [0.0, 0.0, 0.0, 0.508130081300813, 0.5472636815920398, 0.5952569169960474], "id_coverage": [0.0, 0.0, 0.0, 0.08333333333333333, 0.37066666666666664, 0.8473333333333334], "ood_coverage": [0.0, 0.0, 0.0, 0.164, 0.402, 0.8433333333333334]}}, "env": {"python": "3.12.13", "platform": "macOS-26.5.2-arm64-arm-64bit", "jax": "0.11.0", "jax_backend": "cpu"}, "artifact_paths": []} +{"run_id": "fadafc6f2be1", "timestamp": "2026-07-30T11:18:32.832325+00:00", "git_sha": "67dc7a2", "config_hash": "68757964fe31", "config": {"run": {"seed": 4, "task": "graph_planning", "notes": "E2 graph selective-prediction"}, "model": {"latent_dim": 32, "hidden_dim": 128, "context_dim": 64}, "inference": {"k_restarts": 12, "steps": 50, "step_size": 0.1, "temperature": 0.05, "init_scale": 1.0, "grad_tol": 0.001, "sampler": "langevin", "anneal_levels": 10, "anneal_steps_per_level": 8, "anneal_step_max": 0.2, "anneal_step_min": 0.01}, "train": {"epochs": 25, "batch_size": 128, "lr": 0.001, "n_train": 8000, "n_neg": 4, "neg_noise": 0.5, "objective": "basin_center", "ired_noise_min": 0.1, "ired_noise_max": 1.5, "ired_ridge": 0.05, "ired_decode_weight": 1.0, "ired_stat_weight": 1.0}, "eval": {"n_eval": 1500, "richer_geometry": false}, "conformal": {"alpha": 0.1, "delta": 0.05, "n_calib": 1500}, "task": {"graph_planning": {"n_nodes": 7, "ood_n_nodes": 10, "edge_prob": 0.4, "max_len": 4}}}, "task": "graph_planning", "split": "selective", "seed": 4, "metrics": {"n_fit": 1500, "n_calib": 1500, "n_test": 1500, "k_restarts": 12, "objective": "basin_center", "sampler": "langevin", "feature_set": "base", "accuracy_id": 0.7233333333333334, "base_error": 0.27666666666666667, "final_train_loss": 0.21295076608657837, "feature_names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual"], "aurc": {"geometry": 0.13283935829490207, "rho_basin": 0.2203375977159469, "energy_min": 0.22303149688717805, "energy_mean": 0.2229615383788109, "energy_std": 0.2828103710565153, "msp": 0.11262886470219956, "temp_msp": 0.10870154503937372, "entropy": 0.11070403088349459, "softmax_learned": 0.10852657318026214, "geom_softmax": 0.11090419152467276}, "temperature": 2.696299147690236, "best_energy_baseline": "energy_mean", "best_baseline": "temp_msp", "delta_aurc_vs_energy_min": [0.09019213859227598, 0.06793479403590773, 0.11328673363081866], "delta_aurc_vs_best_energy": [0.09012218008390882, 0.06776140718139022, 0.11351940825686613], "delta_aurc_vs_best_baseline": [-0.024137813255528343, -0.03391303611106098, -0.015078370648686666], "delta_aurc_geom_adds": [-0.002377618344410623, -0.00461672765694182, -0.00014286112737259382], "geometry_wins": true, "geometry_wins_vs_baseline": false, "geometry_adds_over_softmax": false, "risk_coverage": {"geometry": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.011111111111111112, 0.03333333333333333, 0.04666666666666667, 0.0611111111111111, 0.06190476190476192, 0.058333333333333334, 0.05925925925925926, 0.06333333333333335, 0.06060606060606061, 0.06388888888888888, 0.0641025641025641, 0.07142857142857142, 0.07333333333333332, 0.08333333333333333, 0.08627450980392157, 0.09259259259259257, 0.09473684210526316, 0.1, 0.09999999999999999, 0.10606060606060606, 0.1144927536231884, 0.12499999999999999, 0.132, 0.1371794871794872, 0.13950617283950617, 0.14285714285714285, 0.14942528735632182, 0.15777777777777777, 0.16451612903225807, 0.16770833333333332, 0.1696969696969697, 0.17745098039215687, 0.182857142857143, 0.18333333333333346, 0.1873873873873874, 0.19210526315789472, 0.19743589743589743, 0.20833333333333334, 0.21382113821138207, 0.22142857142857153, 0.22713178294573644, 0.23712121212121212, 0.2437037037037037, 0.2492753623188406, 0.2553191489361703, 0.26111111111111107, 0.2673469387755102, 0.27666666666666667]}, "rho_basin": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.21457489878542518, 0.21457489878542532, 0.21457489878542535, 0.21457489878542538, 0.21457489878542538, 0.2145748987854248, 0.21457489878542438, 0.21457489878542407, 0.21457489878542382, 0.21457489878542366, 0.21457489878542416, 0.21457489878542454, 0.2145748987854249, 0.21457489878542518, 0.21457489878542543, 0.21457489878542566, 0.21457489878542588, 0.21457489878542604, 0.21457489878542618, 0.21457489878542632, 0.21457489878542643, 0.21457489878542657, 0.21457489878542665, 0.21457489878542677, 0.21457489878542685, 0.21457489878542693, 0.21457489878542702, 0.2145748987854271, 0.21457489878542715, 0.2145748987854272, 0.21457489878542726, 0.21457489878542732, 0.21457489878542738, 0.21457489878542743, 0.21457489878542746, 0.21457489878542751, 0.21457489878542754, 0.2145748987854276, 0.21457489878542763, 0.21457489878542765, 0.21457489878542768, 0.22157872157872383, 0.22962497381102207, 0.2366573902288203, 0.24228269085412094, 0.2488657844990567, 0.2561229314420826, 0.26513888888889064, 0.27217068645640213, 0.2766666666666679]}, "energy_min": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.16666666666666666, 0.15, 0.15555555555555556, 0.15833333333333333, 0.18666666666666668, 0.2000000000000001, 0.19999999999999984, 0.20416666666666666, 0.21481481481481482, 0.21000000000000002, 0.20606060606060606, 0.19722222222222222, 0.2, 0.1976190476190476, 0.19777777777777775, 0.20208333333333334, 0.20588235294117646, 0.20185185185185184, 0.19649122807017544, 0.20666666666666667, 0.20793650793650792, 0.21212121212121213, 0.2217391304347826, 0.2208333333333333, 0.21733333333333332, 0.21666666666666667, 0.21728395061728395, 0.21785714285714286, 0.21494252873563216, 0.21444444444444444, 0.21505376344086022, 0.21979166666666666, 0.22525252525252526, 0.23039215686274508, 0.23523809523809536, 0.2425925925925927, 0.24684684684684685, 0.24912280701754386, 0.24957264957264957, 0.25916666666666666, 0.26260162601626025, 0.2666666666666668, 0.27054263565891473, 0.271969696969697, 0.2785185185185185, 0.2811594202898551, 0.28014184397163117, 0.27986111111111106, 0.2789115646258503, 0.27666666666666667]}, "energy_mean": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.23333333333333334, 0.3, 0.23333333333333334, 0.21666666666666667, 0.19333333333333333, 0.17777777777777776, 0.18095238095238098, 0.17916666666666667, 0.17777777777777778, 0.17333333333333337, 0.16363636363636364, 0.17777777777777778, 0.17692307692307693, 0.1880952380952381, 0.19777777777777775, 0.2, 0.20588235294117646, 0.20185185185185184, 0.2, 0.205, 0.20793650793650792, 0.20909090909090908, 0.21304347826086956, 0.21111111111111108, 0.21333333333333335, 0.2153846153846154, 0.21481481481481482, 0.21785714285714286, 0.21954022988505742, 0.22111111111111112, 0.221505376344086, 0.22395833333333334, 0.22626262626262628, 0.22941176470588234, 0.23428571428571426, 0.23703703703703702, 0.24234234234234234, 0.24736842105263157, 0.25213675213675213, 0.2575, 0.26178861788617896, 0.2682539682539682, 0.2689922480620155, 0.27424242424242423, 0.2777777777777778, 0.27753623188405796, 0.2773049645390072, 0.2770833333333333, 0.2755102040816326, 0.27666666666666667]}, "energy_std": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.16666666666666666, 0.25, 0.28888888888888886, 0.2916666666666667, 0.29333333333333333, 0.30000000000000004, 0.2952380952380951, 0.2916666666666667, 0.2962962962962963, 0.2933333333333332, 0.296969696969697, 0.2972222222222222, 0.2923076923076923, 0.29523809523809524, 0.2955555555555555, 0.28958333333333336, 0.28627450980392155, 0.28703703703703715, 0.28421052631578947, 0.285, 0.28095238095238106, 0.2772727272727273, 0.2768115942028985, 0.2791666666666666, 0.2773333333333333, 0.2794871794871795, 0.2814814814814815, 0.2785714285714286, 0.2770114942528737, 0.2788888888888889, 0.2849462365591398, 0.28125, 0.2797979797979798, 0.28431372549019607, 0.2828571428571428, 0.2814814814814816, 0.2828828828828829, 0.2850877192982456, 0.28034188034188035, 0.2833333333333333, 0.2821138211382113, 0.28333333333333344, 0.2837209302325581, 0.2803030303030303, 0.2837037037037037, 0.2826086956521739, 0.2808510638297872, 0.27847222222222234, 0.2782312925170068, 0.27666666666666667]}, "msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.013333333333333334, 0.016666666666666663, 0.028571428571428577, 0.03333333333333333, 0.03333333333333333, 0.04000000000000001, 0.03939393939393939, 0.03888888888888889, 0.038461538461538464, 0.04523809523809524, 0.048888888888888885, 0.05, 0.049019607843137254, 0.0537037037037037, 0.06140350877192982, 0.06833333333333333, 0.07619047619047618, 0.0803030303030303, 0.08695652173913043, 0.09583333333333333, 0.10133333333333333, 0.1064102564102564, 0.11481481481481481, 0.12142857142857143, 0.1287356321839082, 0.1288888888888889, 0.13225806451612904, 0.13854166666666667, 0.14242424242424243, 0.14901960784313725, 0.15904761904761902, 0.1666666666666668, 0.17477477477477477, 0.18070175438596492, 0.18888888888888888, 0.19833333333333333, 0.20569105691056908, 0.21428571428571438, 0.2248062015503876, 0.23484848484848486, 0.24074074074074073, 0.24710144927536232, 0.2553191489361703, 0.2625000000000001, 0.2687074829931973, 0.27666666666666667]}, "temp_msp": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.013333333333333334, 0.01666666666666678, 0.028571428571428577, 0.03333333333333333, 0.02962962962962963, 0.03333333333333334, 0.030303030303030304, 0.03333333333333333, 0.03333333333333333, 0.0380952380952381, 0.04444444444444444, 0.04375, 0.054901960784313725, 0.0611111111111111, 0.06140350877192982, 0.065, 0.07142857142857141, 0.07575757575757576, 0.08550724637681159, 0.09305555555555554, 0.10133333333333333, 0.1064102564102564, 0.10987654320987654, 0.1130952380952381, 0.11954022988505764, 0.12111111111111111, 0.12580645161290321, 0.13125, 0.13535353535353536, 0.14313725490196078, 0.15238095238095237, 0.15925925925925924, 0.17117117117117117, 0.17982456140350878, 0.18974358974358974, 0.1925, 0.19756097560975622, 0.20555555555555566, 0.21627906976744185, 0.22272727272727272, 0.22814814814814816, 0.23840579710144927, 0.2482269503546099, 0.26111111111111124, 0.2693877551020408, 0.27666666666666667]}, "entropy": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.013333333333333334, 0.016666666666666663, 0.028571428571428577, 0.03333333333333333, 0.03333333333333333, 0.04000000000000001, 0.03939393939393939, 0.03888888888888889, 0.038461538461538464, 0.04523809523809524, 0.048888888888888885, 0.04791666666666667, 0.050980392156862744, 0.05185185185185184, 0.05964912280701754, 0.07, 0.07301587301587315, 0.0803030303030303, 0.08550724637681159, 0.09444444444444455, 0.10266666666666667, 0.1064102564102564, 0.11358024691358025, 0.12142857142857143, 0.12413793103448273, 0.12777777777777777, 0.12795698924731183, 0.13229166666666667, 0.1414141414141414, 0.14705882352941177, 0.15523809523809537, 0.16388888888888903, 0.17117117117117117, 0.17982456140350878, 0.18632478632478633, 0.19583333333333333, 0.20325203252032517, 0.21111111111111108, 0.21627906976744185, 0.225, 0.23037037037037036, 0.2391304347826087, 0.24822695035461, 0.25972222222222235, 0.2687074829931973, 0.27666666666666667]}, "softmax_learned": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.0, 0.013333333333333334, 0.01666666666666678, 0.028571428571428577, 0.03333333333333333, 0.02962962962962963, 0.03333333333333334, 0.030303030303030304, 0.03333333333333333, 0.03333333333333333, 0.0380952380952381, 0.04444444444444444, 0.04583333333333333, 0.052941176470588235, 0.0611111111111111, 0.06315789473684211, 0.065, 0.07142857142857141, 0.07575757575757576, 0.08695652173913043, 0.09444444444444443, 0.1, 0.10384615384615385, 0.1111111111111111, 0.11071428571428571, 0.11379310344827584, 0.12111111111111111, 0.12903225806451613, 0.13333333333333333, 0.13737373737373737, 0.14411764705882352, 0.15428571428571441, 0.16388888888888903, 0.17027027027027028, 0.17543859649122806, 0.1811965811965812, 0.19166666666666668, 0.20000000000000012, 0.20555555555555566, 0.2131782945736434, 0.22045454545454546, 0.22666666666666666, 0.23768115942028986, 0.2503546099290781, 0.25972222222222235, 0.2673469387755102, 0.27666666666666667]}, "geom_softmax": {"coverage": [0.02, 0.04, 0.06, 0.08, 0.1, 0.12000000000000001, 0.13999999999999999, 0.16, 0.18, 0.19999999999999998, 0.22, 0.24, 0.26, 0.28, 0.30000000000000004, 0.32, 0.34, 0.36000000000000004, 0.38, 0.4, 0.42000000000000004, 0.44, 0.46, 0.48000000000000004, 0.5, 0.52, 0.54, 0.56, 0.5800000000000001, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7000000000000001, 0.7200000000000001, 0.74, 0.76, 0.78, 0.8, 0.8200000000000001, 0.8400000000000001, 0.86, 0.88, 0.9, 0.92, 0.9400000000000001, 0.9600000000000001, 0.98, 1.0], "risk": [0.0, 0.0, 0.0, 0.008333333333333333, 0.02, 0.02222222222222222, 0.038095238095238106, 0.03333333333333333, 0.02962962962962963, 0.026666666666666672, 0.030303030303030304, 0.03333333333333333, 0.03333333333333333, 0.0380952380952381, 0.04222222222222222, 0.05, 0.056862745098039215, 0.0611111111111111, 0.06315789473684211, 0.06333333333333334, 0.07777777777777777, 0.08181818181818182, 0.08405797101449275, 0.09444444444444443, 0.09733333333333333, 0.10384615384615385, 0.10740740740740741, 0.11666666666666667, 0.11839080459770113, 0.12, 0.12688172043010754, 0.134375, 0.14343434343434344, 0.15196078431372548, 0.15619047619047632, 0.16851851851851865, 0.17747747747747747, 0.18421052631578946, 0.18803418803418803, 0.195, 0.2032520325203253, 0.21428571428571438, 0.22170542635658916, 0.22424242424242424, 0.23037037037037036, 0.24130434782608695, 0.2503546099290781, 0.2590277777777779, 0.2687074829931973, 0.27666666666666667]}}, "ltt": {"alpha": 0.1, "delta": 0.05, "lambda_hat": null, "coverage": 0.0, "abstain_rate": 1.0, "selective_risk": 0.0, "selective_accuracy": null, "risk_within_budget": true}, "coverage_validity": {"alpha_grid": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "target": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "achieved_risk": [0.0, 0.0, 0.0, 0.09854604200323101, 0.1673511293634497, 0.2629043358568479], "coverage": [0.0, 0.0, 0.0, 0.4126666666666667, 0.6493333333333333, 0.9686666666666667]}, "feature_diagnostics": {"names": ["basin/rho", "basin/entropy", "basin/dispersion", "energy/mean", "energy/min", "energy/std", "curv/lmax_mean", "curv/lmax_best", "curv/trace_mean", "curv/trace_best", "dynamics/steps", "dynamics/monotonic", "dynamics/drop", "dynamics/residual"], "auroc": {"basin/rho": 0.6281605685414469, "basin/entropy": 0.3710199322636167, "basin/dispersion": 0.518520903892066, "energy/mean": 0.40323580034423406, "energy/min": 0.4087679751263117, "energy/std": 0.5148586974626617, "curv/lmax_mean": 0.565753150852257, "curv/lmax_best": 0.5177969018932874, "curv/trace_mean": 0.6267170062739437, "curv/trace_best": 0.6067603131419688, "dynamics/steps": 0.5, "dynamics/monotonic": 0.6530875576036866, "dynamics/drop": 0.6858031203153628, "dynamics/residual": 0.515407251124313}, "hist": {"basin/rho": {"edges": [0.3333333333333333, 0.36666666666666664, 0.4, 0.43333333333333335, 0.4666666666666667, 0.5, 0.5333333333333333, 0.5666666666666667, 0.6000000000000001, 0.6333333333333333, 0.6666666666666667, 0.7000000000000001, 0.7333333333333334, 0.7666666666666667, 0.8, 0.8333333333333335, 0.8666666666666667, 0.9000000000000001, 0.9333333333333333, 0.9666666666666668, 1.0], "correct_counts": [1, 0, 1, 0, 0, 8, 0, 14, 0, 14, 0, 0, 20, 0, 25, 0, 0, 32, 0, 970], "incorrect_counts": [0, 0, 0, 0, 0, 8, 0, 19, 0, 31, 0, 0, 26, 0, 24, 0, 0, 42, 0, 265]}, "basin/entropy": {"edges": [0.0, 0.06430286684371896, 0.1286057336874379, 0.19290860053115688, 0.2572114673748758, 0.32151433421859477, 0.38581720106231376, 0.4501200679060327, 0.5144229347497516, 0.5787258015934706, 0.6430286684371895, 0.7073315352809085, 0.7716344021246275, 0.8359372689683464, 0.9002401358120654, 0.9645430026557843, 1.0288458694995033, 1.0931487363432222, 1.1574516031869413, 1.2217544700306602, 1.286057336874379], "correct_counts": [970, 0, 0, 0, 32, 0, 0, 23, 22, 14, 22, 0, 0, 0, 0, 0, 1, 0, 0, 1], "incorrect_counts": [265, 0, 0, 0, 42, 0, 0, 22, 23, 28, 22, 5, 2, 2, 2, 1, 1, 0, 0, 0]}, "basin/dispersion": {"edges": [1.596239525312261, 1.6174650761676754, 1.63869062702309, 1.6599161778785045, 1.681141728733919, 1.7023672795893336, 1.723592830444748, 1.7448183813001625, 1.766043932155577, 1.7872694830109916, 1.808495033866406, 1.8297205847218205, 1.8509461355772352, 1.8721716864326496, 1.893397237288064, 1.9146227881434785, 1.9358483389988932, 1.9570738898543076, 1.978299440709722, 1.9995249915651367, 2.020750542420551], "correct_counts": [2, 3, 9, 15, 23, 41, 58, 97, 122, 144, 143, 151, 93, 75, 34, 31, 25, 11, 4, 4], "incorrect_counts": [3, 1, 4, 6, 10, 23, 30, 38, 50, 46, 53, 37, 32, 28, 23, 17, 7, 2, 2, 3]}, "energy/mean": {"edges": [0.7065413296222687, 0.7633304566144943, 0.82011958360672, 0.8769087105989456, 0.9336978375911713, 0.9904869645833969, 1.0472760915756225, 1.1040652185678481, 1.160854345560074, 1.2176434725522993, 1.2744325995445251, 1.3312217265367507, 1.3880108535289764, 1.4447999805212022, 1.5015891075134276, 1.5583782345056534, 1.615167361497879, 1.6719564884901046, 1.7287456154823302, 1.785534742474556, 1.8423238694667816], "correct_counts": [20, 74, 153, 170, 154, 92, 69, 54, 36, 38, 27, 31, 32, 27, 39, 26, 25, 10, 4, 4], "incorrect_counts": [3, 23, 26, 55, 45, 32, 28, 22, 20, 30, 24, 27, 21, 22, 11, 10, 8, 3, 5, 0]}, "energy/min": {"edges": [0.33243313431739807, 0.39178149551153185, 0.4511298567056656, 0.5104782178997993, 0.5698265790939331, 0.6291749402880669, 0.6885233014822006, 0.7478716626763344, 0.8072200238704681, 0.8665683850646019, 0.9259167462587357, 0.9852651074528694, 1.0446134686470032, 1.103961829841137, 1.1633101910352708, 1.2226585522294044, 1.2820069134235381, 1.341355274617672, 1.4007036358118057, 1.4600519970059396, 1.5194003582000732], "correct_counts": [10, 39, 87, 128, 161, 124, 119, 69, 56, 36, 42, 32, 28, 35, 24, 38, 20, 16, 16, 5], "incorrect_counts": [4, 3, 27, 35, 38, 48, 28, 24, 32, 32, 19, 30, 27, 19, 21, 10, 9, 5, 2, 2]}, "energy/std": {"edges": [0.0686106895390655, 0.08331282795302085, 0.09801496636697621, 0.11271710478093155, 0.1274192431948869, 0.14212138160884225, 0.1568235200227976, 0.17152565843675294, 0.18622779685070828, 0.20092993526466363, 0.21563207367861897, 0.23033421209257432, 0.24503635050652972, 0.25973848892048507, 0.2744406273344404, 0.28914276574839576, 0.3038449041623511, 0.31854704257630645, 0.3332491809902618, 0.34795131940421714, 0.3626534578181725], "correct_counts": [2, 1, 21, 20, 52, 85, 126, 148, 145, 127, 113, 81, 52, 50, 24, 11, 8, 12, 6, 1], "incorrect_counts": [1, 0, 4, 10, 22, 40, 50, 52, 51, 58, 40, 36, 24, 9, 6, 5, 6, 1, 0, 0]}, "curv/lmax_mean": {"edges": [0.9988646656274796, 0.9989987172186374, 0.9991327688097954, 0.9992668204009533, 0.9994008719921113, 0.9995349235832691, 0.999668975174427, 0.999803026765585, 0.9999370783567428, 1.0000711299479008, 1.0002051815390587, 1.0003392331302166, 1.0004732847213744, 1.0006073363125325, 1.0007413879036904, 1.0008754394948483, 1.0010094910860061, 1.001143542677164, 1.001277594268322, 1.00141164585948, 1.0015456974506378], "correct_counts": [2, 6, 14, 38, 85, 156, 260, 277, 129, 58, 33, 11, 10, 4, 1, 0, 0, 0, 0, 1], "incorrect_counts": [1, 3, 5, 14, 39, 80, 117, 100, 27, 11, 10, 2, 2, 3, 0, 0, 0, 0, 0, 1]}, "curv/lmax_best": {"edges": [0.9917203187942505, 0.9926185190677643, 0.993516719341278, 0.9944149196147919, 0.9953131198883056, 0.9962113201618195, 0.9971095204353333, 0.998007720708847, 0.9989059209823609, 0.9998041212558746, 1.0007023215293884, 1.0016005218029023, 1.002498722076416, 1.0033969223499297, 1.0042951226234436, 1.0051933228969574, 1.0060915231704712, 1.006989723443985, 1.0078879237174987, 1.0087861239910125, 1.0096843242645264], "correct_counts": [2, 0, 0, 1, 3, 5, 6, 44, 279, 726, 15, 2, 0, 0, 1, 0, 0, 0, 0, 1], "incorrect_counts": [0, 0, 0, 0, 1, 2, 1, 17, 112, 273, 6, 0, 3, 0, 0, 0, 0, 0, 0, 0]}, "curv/trace_mean": {"edges": [31.799927234649658, 31.81066793600718, 31.821408637364705, 31.83214933872223, 31.84289004007975, 31.85363074143728, 31.864371442794802, 31.875112144152325, 31.88585284550985, 31.896593546867372, 31.907334248224895, 31.91807494958242, 31.928815650939942, 31.939556352297465, 31.95029705365499, 31.961037755012512, 31.97177845637004, 31.982519157727562, 31.993259859085086, 32.004000560442606, 32.01474126180013], "correct_counts": [1, 0, 2, 4, 24, 15, 36, 40, 48, 41, 49, 36, 35, 35, 34, 51, 95, 234, 256, 49], "incorrect_counts": [1, 0, 0, 6, 6, 16, 18, 29, 34, 23, 23, 30, 11, 18, 23, 24, 37, 53, 54, 9]}, "curv/trace_best": {"edges": [31.509716033935547, 31.5363130569458, 31.562910079956055, 31.58950710296631, 31.616104125976562, 31.642701148986816, 31.66929817199707, 31.695895195007324, 31.722492218017578, 31.749089241027832, 31.775686264038086, 31.80228328704834, 31.828880310058594, 31.855477333068848, 31.8820743560791, 31.908671379089355, 31.93526840209961, 31.961865425109863, 31.988462448120117, 32.01505947113037, 32.041656494140625], "correct_counts": [0, 0, 0, 0, 0, 1, 1, 3, 6, 12, 14, 11, 30, 29, 60, 81, 113, 245, 460, 19], "incorrect_counts": [1, 0, 0, 0, 0, 0, 0, 0, 1, 5, 4, 11, 18, 29, 31, 44, 73, 90, 100, 8]}, "dynamics/steps": {"edges": [1.0, 1.05, 1.1, 1.15, 1.2, 1.25, 1.3, 1.35, 1.4, 1.45, 1.5, 1.55, 1.6, 1.65, 1.7000000000000002, 1.75, 1.8, 1.85, 1.9, 1.9500000000000002, 2.0], "correct_counts": [1085, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], "incorrect_counts": [415, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "dynamics/monotonic": {"edges": [0.6349999999999999, 0.6424166666666666, 0.6498333333333333, 0.6572499999999999, 0.6646666666666666, 0.6720833333333334, 0.6795, 0.6869166666666666, 0.6943333333333334, 0.7017500000000001, 0.7091666666666667, 0.7165833333333333, 0.7240000000000001, 0.7314166666666668, 0.7388333333333335, 0.7462500000000001, 0.7536666666666668, 0.7610833333333336, 0.7685000000000002, 0.7759166666666668, 0.7833333333333335], "correct_counts": [2, 1, 10, 26, 66, 85, 136, 130, 185, 113, 121, 110, 51, 23, 14, 4, 5, 2, 0, 1], "incorrect_counts": [0, 3, 13, 28, 34, 39, 87, 60, 73, 38, 14, 18, 7, 1, 0, 0, 0, 0, 0, 0]}, "dynamics/drop": {"edges": [14.541723385453224, 20.501742911587158, 26.461762437721095, 32.42178196385503, 38.381801489988966, 44.3418210161229, 50.30184054225684, 56.261860068390774, 62.22187959452471, 68.18189912065864, 74.14191864679258, 80.10193817292651, 86.06195769906046, 92.02197722519439, 97.98199675132832, 103.94201627746226, 109.90203580359619, 115.86205532973013, 121.82207485586406, 127.782094381998, 133.74211390813193], "correct_counts": [123, 310, 256, 163, 111, 52, 33, 12, 7, 6, 6, 2, 2, 1, 0, 0, 0, 0, 0, 1], "incorrect_counts": [107, 178, 75, 29, 18, 7, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}, "dynamics/residual": {"edges": [1.1456641207138698, 1.162844986965259, 1.180025853216648, 1.1972067194680374, 1.2143875857194266, 1.2315684519708157, 1.248749318222205, 1.2659301844735942, 1.2831110507249832, 1.3002919169763725, 1.3174727832277617, 1.3346536494791508, 1.35183451573054, 1.3690153819819293, 1.3861962482333183, 1.4033771144847076, 1.4205579807360968, 1.4377388469874859, 1.4549197132388751, 1.4721005794902644, 1.4892814457416534], "correct_counts": [1, 2, 5, 8, 34, 72, 93, 117, 158, 165, 146, 97, 86, 51, 28, 12, 9, 0, 0, 1], "incorrect_counts": [1, 0, 5, 5, 17, 30, 33, 51, 49, 66, 50, 46, 31, 16, 9, 2, 1, 3, 0, 0]}}}, "feature_ablation": {"full": 0.13283935829490207, "drop_basin": 0.14997742438206202, "basin_only": 0.23296177845321314, "drop_energy": 0.13721472001078186, "energy_only": 0.22409660098635795, "drop_curv": 0.13097753105946705, "curv_only": 0.2233216171978382, "drop_dynamics": 0.18117722135638095, "dynamics_only": 0.1620572670509188}, "ece_geometry": 0.0349920061619123, "accuracy_ood": 0.30666666666666664, "aurc_ood": {"geometry": 0.7404444300750076, "rho_basin": 0.6925842472273304, "energy_min": 0.6068299457403026, "energy_mean": 0.6018766592700887, "energy_std": 0.6989784553631774, "msp": 0.7052802742194463, "temp_msp": 0.697350347427352, "entropy": 0.7029716683577881, "softmax_learned": 0.6957988290761644, "geom_softmax": 0.7041596483562519}, "ood_ltt": {"alpha": 0.1, "lambda_hat": null, "selective_risk": 0.0, "coverage": 0.0, "risk_within_budget": true}, "ood_validity": {"target": [0.02, 0.05, 0.1, 0.15, 0.2, 0.3], "id_risk": [0.0, 0.0, 0.0, 0.09854604200323101, 0.1673511293634497, 0.2629043358568479], "ood_risk": [0.0, 0.0, 0.0, 0.7001114827201784, 0.6923751095530236, 0.6935593220338983], "id_coverage": [0.0, 0.0, 0.0, 0.4126666666666667, 0.6493333333333333, 0.9686666666666667], "ood_coverage": [0.0, 0.0, 0.0, 0.598, 0.7606666666666667, 0.9833333333333333]}}, "env": {"python": "3.12.13", "platform": "macOS-26.5.2-arm64-arm-64bit", "jax": "0.11.0", "jax_backend": "cpu"}, "artifact_paths": []} diff --git a/src/edc/plotting.py b/src/edc/plotting.py index 516dbe2..2a91003 100644 --- a/src/edc/plotting.py +++ b/src/edc/plotting.py @@ -45,26 +45,34 @@ def use_style() -> None: def _selective_row(rows: list[dict[str, Any]]) -> dict[str, Any]: - """Metrics of the headline ``split == 'selective'`` run for F2/F3 (raises if none). + """Metrics of the headline ``split == 'selective'`` run for F2/F3/F5/F6 (raises if none). - Prefers the full-fold run — largest ``n_test`` (E1), ties broken by ledger order — so the - single-run figures reflect the authoritative experiment, not whatever reduced-fold sweep cell - happened to land last in the ledger. + These are the arithmetic E1 figures by design (their captions name arithmetic), so we prefer + the arithmetic full-fold run — largest ``n_test``, ties broken by ledger order. Preferring + arithmetic matters since Phase 4l added full-fold (``n_test=1500``) graph rows that would + otherwise tie and flip the figures to graph. Falls back to the global run if no arithmetic row. """ sel = [r for r in rows if r.get("split") == "selective"] if not sel: raise ValueError("no split='selective' ledger row; run experiments/run_experiment.py first") + arith = [r for r in sel if r.get("task") == "arithmetic"] + pool = arith or sel # largest n_test wins; ties broken toward the latest ledger row (freshest full-fold E1). - headline = max(enumerate(sel), key=lambda iv: (iv[1]["metrics"].get("n_test", 0), iv[0]))[1] + headline = max(enumerate(pool), key=lambda iv: (iv[1]["metrics"].get("n_test", 0), iv[0]))[1] return headline["metrics"] def _halting_row(rows: list[dict[str, Any]]) -> dict[str, Any]: - """Metrics of the latest ``split == 'halting'`` run for F4 (raises if none).""" + """Metrics of the headline ``split == 'halting'`` run for F4 (raises if none). + + F4 is the arithmetic halting Pareto by design, so prefer the latest arithmetic halting row; + Phase 4l added graph halting rows (appended later) that would otherwise flip F4 to graph. + """ sel = [r for r in rows if r.get("split") == "halting"] if not sel: raise ValueError("no split='halting' ledger row; run experiments/run_halting.py first") - return sel[-1]["metrics"] + arith = [r for r in sel if r.get("task") == "arithmetic"] + return (arith or sel)[-1]["metrics"] def risk_coverage_curve(rows: list[dict[str, Any]], out_path: str) -> str: # F2 diff --git a/tests/test_aggregate.py b/tests/test_aggregate.py index 8f3e6c7..054d703 100644 --- a/tests/test_aggregate.py +++ b/tests/test_aggregate.py @@ -7,7 +7,7 @@ def _row(k, seed, delta, lo, geo=0.08, best=0.14, task="arithmetic", baseline_delta=None, - geom_adds=None, feature_set=None): + geom_adds=None, feature_set=None, ood=None): m = { "k_restarts": k, "accuracy_id": 0.8, "base_error": 0.2, "best_energy_baseline": "energy_min", @@ -26,6 +26,12 @@ def _row(k, seed, delta, lo, geo=0.08, best=0.14, task="arithmetic", baseline_de if geom_adds is not None: # Phase 4j field ga, glo = geom_adds m["delta_aurc_geom_adds"] = [ga, glo, ga + 0.02] + if ood is not None: # Phase 4l fields (include_ood rows only) + acc_ood, risk_ood, cov_ood, within = ood + m["accuracy_ood"] = acc_ood + m["ltt"]["selective_risk"] = 0.07 # ID selective risk (already present) + m["ood_ltt"] = {"selective_risk": risk_ood, "coverage": cov_ood, + "risk_within_budget": within} fs_tag = "" if feature_set is not None: # Phase 4k field (older rows omit it) m["feature_set"] = feature_set @@ -116,6 +122,66 @@ def test_feature_set_filter_and_fallback(): assert hr["delta_aurc_geom_adds"][0] > hb["delta_aurc_geom_adds"][0] +def test_ood_aggregation_and_fallback(): + # Phase 4l: rows run with include_ood carry accuracy_ood + ood_ltt; aggregate_cell summarises + # them and counts seeds whose OOD risk stays within budget. + rows = [_row(12, s, 0.09, 0.07, baseline_delta=(0.05, 0.03), + ood=(0.4, 0.5, 0.8, False)) for s in range(5)] + a = aggregate.aggregate_cell(rows) + assert a["has_ood"] is True + assert a["accuracy_ood"][0] == 0.4 + assert a["ood_selective_risk"][0] == 0.5 # OOD risk >> alpha => guarantee breaks + assert a["ltt_selective_risk"][0] == 0.07 # ID risk stays low + assert a["seeds_ood_within_budget"] == 0 + + # one seed happens to stay within budget on OOD + mixed = [_row(12, s, 0.09, 0.07, ood=(0.4, 0.09 if s == 0 else 0.5, 0.8, s == 0)) + for s in range(5)] + assert aggregate.aggregate_cell(mixed)["seeds_ood_within_budget"] == 1 + + # pre-4l rows (no OOD fields) fall back cleanly: has_ood False, NaNs, no crash + old = aggregate.aggregate_cell([_row(12, s, 0.09, 0.07) for s in range(3)]) + assert old["has_ood"] is False + import math + assert math.isnan(old["accuracy_ood"][0]) + + +def _hrow(seed, compute_saved, halting_risk, within=True, monotone=True, tau=0.9, + task="arithmetic"): + m = { + "k_restarts": 12, "alpha": 0.1, "tau_hat": tau, + "compute_saved": compute_saved, "compute_used": 1.0 - compute_saved, + "halting_risk": halting_risk, "risk_within_budget": within, + "risk_monotone_in_tau": monotone, "full_accuracy": 0.8, "accuracy_drop": 0.0, + } + return { + "run_id": f"h{seed}", "seed": seed, "split": "halting", "task": task, + "config_hash": f"{task}_halt_{seed}", "git_sha": "abc1234", + "env": {"python": "3.12", "jax": "0.11", "jax_backend": "cpu"}, + "config": {"inference": {"k_restarts": 12}}, "metrics": m, + } + + +def test_halting_aggregation(): + # split="halting" rows are excluded from selective_rows and aggregated separately (Phase 4l). + hrows = [_hrow(s, 0.5 + 0.01 * s, 0.02) for s in range(5)] + selrows = [_row(12, s, 0.09, 0.07) for s in range(3)] + mixed = hrows + selrows + assert aggregate.halting_rows(mixed) == hrows # halting split filter + assert len(aggregate.selective_rows(mixed)) == 3 # halting rows excluded + + a = aggregate.aggregate_halting_cell(hrows) + assert a["n_seeds"] == 5 and a["task"] == "arithmetic" + assert abs(a["compute_saved"][0] - 0.52) < 1e-9 + assert a["halting_risk"][0] == 0.02 + assert a["seeds_within_budget"] == 5 and a["seeds_risk_monotone"] == 5 + + # a seed over budget is counted honestly; a None tau_hat (infeasible) doesn't crash the mean + over = [_hrow(0, 0.3, 0.2, within=False, tau=None), *[_hrow(s, 0.5, 0.02) for s in range(1, 4)]] + ao = aggregate.aggregate_halting_cell(over) + assert ao["seeds_within_budget"] == 3 and ao["n_tau_feasible"] == 3 + + def test_feature_ablation_aggregates(): rows = [_row(12, s, 0.09, 0.07, geo=0.08, baseline_delta=(0.05, 0.03)) for s in range(4)] fa = aggregate.aggregate_cell(rows)["feature_ablation"] diff --git a/tests/test_make_tables.py b/tests/test_make_tables.py index 9b364e3..f606d52 100644 --- a/tests/test_make_tables.py +++ b/tests/test_make_tables.py @@ -6,7 +6,7 @@ import aggregate import make_tables -from test_aggregate import _row +from test_aggregate import _hrow, _row def _rows(): @@ -47,5 +47,25 @@ def test_tables_well_formed(): assert t3.count(r"\\") >= len(rows) +def test_t6_halting_and_t7_ood(): + # T6 from split="halting" rows; T7 from include_ood selective rows (Phase 4l). + hrows = [_hrow(s, 0.55, 0.02, task="arithmetic") for s in range(5)] + t6 = make_tables._t6(hrows) + assert r"\label{tab:halting}" in t6 + assert t6.split(r"\midrule")[1].split(r"\bottomrule")[0].count(r"\\") == 1 # one task row + assert "arithmetic" in t6 + + ood_rows = [_row(12, s, 0.09, 0.07, baseline_delta=(0.05, 0.03), + ood=(0.4, 0.5, 0.8, False)) for s in range(5)] + t7 = make_tables._t7(ood_rows) + assert r"\label{tab:oodstress}" in t7 + assert "arithmetic" in t7 and r"\toprule" in t7 + + # both are gated: absent data -> empty body + assert make_tables._t6([]).split(r"\midrule")[1].split(r"\bottomrule")[0].strip() == "" + no_ood = [_row(12, s, 0.09, 0.07, baseline_delta=(0.05, 0.03)) for s in range(3)] + assert make_tables._t7(no_ood).split(r"\midrule")[1].split(r"\bottomrule")[0].strip() == "" + + def test_delta_pm_formatting(): assert make_tables._pm((0.0736, 0.012)) == r"$0.074 \pm 0.012$" diff --git a/tests/test_run_halting_sweep.py b/tests/test_run_halting_sweep.py new file mode 100644 index 0000000..2937266 --- /dev/null +++ b/tests/test_run_halting_sweep.py @@ -0,0 +1,41 @@ +"""run_halting_sweep expands a seed grid and appends split='halting' rows via evaluate_halting. + +Offline + CPU-only and FAST: ``evaluate_halting`` and the ledger ``append`` are monkeypatched, so +no reasoner is trained and the real ledger is never written — we only check the runner's plumbing +(grid expansion, one halting row per cell, correct split/seed). +""" + +import run_halting_sweep + + +def _fake_metrics(): + return {"tau_hat": 0.9, "compute_saved": 0.5, "halting_risk": 0.02, + "risk_within_budget": True, "alpha": 0.1} + + +def test_run_halting_sweep_appends_one_halting_row_per_cell(tmp_path, monkeypatch): + sweep = tmp_path / "smoke_halting.toml" + sweep.write_text( + '[sweep]\nbase = "configs/smoke.toml"\n[sweep.grid]\n"run.seed" = [0, 1, 2]\n') + + calls = {"eval": 0} + captured = [] + + def fake_eval(cfg, task): + calls["eval"] += 1 + return _fake_metrics() + + def fake_append(record): + row = record.to_row() if hasattr(record, "to_row") else record + captured.append(row) + return row + + monkeypatch.setattr(run_halting_sweep, "evaluate_halting", fake_eval) + monkeypatch.setattr(run_halting_sweep, "append", fake_append) + + rc = run_halting_sweep.main([str(sweep)]) + assert rc == 0 + assert calls["eval"] == 3 # one evaluate_halting per seed cell + assert len(captured) == 3 + assert all(r["split"] == "halting" for r in captured) + assert sorted(r["seed"] for r in captured) == [0, 1, 2] diff --git a/tests/test_run_sweep.py b/tests/test_run_sweep.py index 36f198d..276ae6e 100644 --- a/tests/test_run_sweep.py +++ b/tests/test_run_sweep.py @@ -35,7 +35,17 @@ def test_cell_config_applies_override_and_cell(): def test_load_sweep_reads_real_config(): - base, override, cells = run_sweep.load_sweep("configs/sweeps/k_restarts.toml") + base, override, cells, opts = run_sweep.load_sweep("configs/sweeps/k_restarts.toml") assert len(cells) == 25 # 5 K x 5 seeds assert override["eval.n_eval"] == 600 assert base["task"]["arithmetic"]["n_operands"] == 6 # base experiment config resolved + assert opts["include_ood"] is False # default: OOD skipped in sweeps + + +def test_load_sweep_include_ood_flag(): + # Phase 4l: a sweep can opt the OOD fold back on for multi-seed OOD / certification runs. + base, override, cells, opts = run_sweep.load_sweep("configs/sweeps/graph_cert_seeds.toml") + assert opts["include_ood"] is True + assert len(cells) == 5 # 5 seeds + assert base["conformal"]["n_calib"] == 1500 # full folds (no reduced-fold override) + assert "conformal.n_calib" not in override

    • ^vPzsE_Kqy+q8GHx{#K#n3oumqW{RdhIroaa6+;BM~^R!RrZb=X0> z_}G0&^C2kI_lNk-#EYrk8v%~lZA>+NTe>bx#~7E~Di{MOP_p`abVBVkj2Rvfm?bu% z3`Lq4O6rkM7gw+mEh{rkqodZfqjWiFecz%EA|x?;!i6y-OaJ>mF8|-0#pNCXr91(k z*sYk3QVIpn9)SFG2dC;LYUp8;U%LMnT&3BN{V!O>ZZ&cd1cvf>PzNS!VTzgw(OHMq zY45R|;%D@HM!o+<{()4veS=npv7vRk$h0Y})dnDCvKRpjZ1<6xcsbK73sSF@NozuQIo`dR-FZR=ra{dq*x@UFfC1MUI4*u$OVuqpPFr{Wej@7Is{ z?-I#++P<4%>1&m1UYoG**9(|M)rACrwa|@IHk;s9a)>M?6I*Vtr9#BPfd-Jvw-yZ5 zhwU8%f((ps@WR^#S8f|d*b{#e~dB3>h%0?m$9W*y=0hzufrP@mQlO<^>Mdz*zKx)7rG=@s*4vwEP;zFE{fP8(yND!5 zZMa>#rl|JU?D<56m6Xrzeiwc$46)H{wz#?Fu;!SOVE)t5deq@25Z0V}@hiFgC@2D1 zt7lA`O<&o(FJ-;&`~TrP(XNmimvDkM!QVy{sOK&5{+%WAF%qk{no=_-&zA_C#+8$w z;GASz#lo&SbofvJY1RldRcY!9WYEh5F-2G0&CoSZ6@w%$uc$yxYmD`2vrvg{xjY=_ z&-+tj{Y=Fi=R^O}2QUJ4e)Jbfq;mx@h&013Fv|c(YRzm#X5aue=OiLoM;t_$I1DG0hFu3igtm^v>-pX)rmRCKx&phMH2S9rGko@7m?TlP5 zA@~AGX0DBA4LB|4JBqmx$=(Vgp8XAN0x}%Fzk%{@a(tWCspXc+M2)Y)6b&9%a!TBs8@ z;jOj#nOW`^lRn269)08-lP?)tRL9MRxE<*#<3xG)JzqO>#KH9pi3q`mwc1?2f#W7U znH*w*3j6_Ga-uV&eQBVr-eML=og}|hE^SHC9Zd+^NCiLjDiI>w^>?Li$40m zfklO!(zzTwy*S*2BR>#5+2TaQW&H{?lcOV}CXdN~V7doN>3RQ=h2EmJr-QXt5SM1> zwm9vTR%F@NzJPnXNSBgvFHu}XZh*G+8?VBRoFIIJU6pTe3Qx)4oPD6;EtNwXy_3*8 z#NoYNc;=2jOopXu*z}KSNR6+9Mc}9C6nyv=*Ut(VfIkm?!e!o#sGVq|u!Kqo7A}t~ zuJQQ_=$^nFK2synM;lbbsBr#1JHnyb>Tj9k^lZz4@Zmc65penv@vw#DTqcjtgMm_3 z^+b8-$t5f*jDwHv--^KOG9+}};EBOaRl>`u$t$feM)K~b#=5q3pZp7c#HaEv*7Ya> zw~LNtS)*sc1Ox)qk-25im`8_wQjE$>WyY3;r5hNk6{AtGr)UttRw}9sq@|)ehb?n%O^q z9Omq1GNo>o&E2~-Yg^Za#<(cfK*~`+pKOEs%zqAGGblWWELFkjqWB_I$F$K~j%MeO zGhotsYSg* zsugg4iLJy}F1o#vms3Uf8Wrc7((dQaHP5x(&+;0j zzFUMSm#=Tx#UYz@m6M`zdHV+KF1421C>&jiUbZ4uiJ1}dyZ9;U=h^=AO**()JDO4E z#^W_9t3Q1vDrHBSk?)t1J2JwMn#cZi|7vqs#^oxXWzT{2Ru|tHvcl}6|DU@IMBLW+ zfist6c7X47RM;_m9VdCn%C^%4xJq8W-ZsKDT=2O>Foa6EmD=vH-rnn>FJm_DWf%u+ z<%&ER>72*Vgy|da6bOXNsA=a%ztmR@Ev9Q@9*Z5#7KN}dQR&yq?Nw)AJ5QR)QBw_LD= z_8LP+Ep#w)Evix@$^^BPh3{7?C*YFz2 z5c)Rw=#5xZBW1EK-G)7zFJ~d4xe<|!pHP&dj@ke%NNVh?$7i7XBv+kS4ZY@*d5uw> zP46(2t-aU^dc!em6dk?}L^*|P5niOx7&->Yt9f}u=*#skkaf~!L0uRPAphd{iQY%H zmtQ6*9QK_W*+|H~2h4~jhznSn|9D;AeUR%ROP?zHm}5%5+mEuNblNsTn#5gsaPKw~uAS?0k<@!1Aso;Y5b5)=wxIh*y5 zQ91wDxH_KfNAVEimD5sdy)}lt)=4=K?)teo?ofBhsl2ihX!39HpPx~yTAXusZn+|| zLi8Gj92sHw?eXw0sON7Y*rmz$`90i2egD#!HdY zPEbCenXypf0UYmt^$j* zV~mfOb^26*-;cl8i9s>Y?${JVjdLr*XY(Tq$g&12yFHxtWNG`ALl-M(d|Z)SF6 zDqJ%{d@{}gsg8^yw)Bc$*oD$a2I5Bm2QJ5q!Y=`mJ{GK1<1l7+s`5VdCQVsloD?cT z2uo|vwWVFoHv(U*fO`V-d%C*>kJOoP(+7P8mrCZX+7XQ)Y;cnGM#-wwqz#GS`~r2x zVfqs<xy;SB&enT6B+x}9{=K|<5$k5A= z;)WT?qYQ*f9_uM`PEW1VHd4L42bu}F43;lS*66Y!qEQ4zK@!$;gk%~mS59eD^fPK6 zPUFyOhFE*rJ0XIMj!Kf!<*yyU)5%HAEH2roR|$!v+-3F)68p}qKe8GwsC0!#T(>YG zEBcU@T0Zq`<~xQ%%F{Qze1PnAOj^GF{$u3oBmGZwsYSTFiu*2kY1GqXlN!^lUlCHu z%4g560=|3ut)4AJGOg<`E)MSG0d`-YCSKPM=hujCIx#e42t4?`+!CiyX6Xi@%=me>ZDnnmRr<%lQC-u^2T!=c5D zwSi4$jSn?gB{P|<4Azke7TbPma9`sQ*&m;r`WUU%)Bq=no#f1+uS}Xo*IFsIQz1$X zin-CYxD#))h3V*WnS|$@czA2;KgzG0h#PIpITtCcVYug^sX$C&-t9v&59U6qOATk* z{$51;F3XYu%GY#dM}}D7{-*n{?Ly!l6biXm0&sN;?xSJ!Rz4Jkbb; zu@GaBINfos$|)`DzfV?gXxbcbGR~1Fvgv7q^lI8GYoEmwADMiSp3_WF78Og9a8w=l zR0iWI0_7?7tZWgh;?G{x-xENeb5}Dr_R3g<;wn?>u0zJ>3@v7?Z}ugjuKq^Tg-++I zV>zrHyI*J0s7HT8RU`v^)_I!B+f3vftxyi6>je3M7JDhD3uPL+w#BBih@5&eFby$4 z%iHZ}GCB^L#^(<(?K?ry5FCD-V$LSxC3SAYY|0|D>3`vJQ_iL`rmv9!rmy{})OTL3 zSuxk$Hp2OAtxDwjADTv++qvqq{p>?#{H>A&d#-d~n$-}riv#V2i z*=V{{Rhz^ioCa$lAz`CE%*b^_W2K=vOcupb)z$a@QH~*& zKl(9B$dP_U5U`cs$mGbLYh$=&oX9U-!|6nP_@B`KsQuX_c{nre?E?>sk`2MErV=4g z_E@hYt~i}-u!;>&oEe%!lSnEFwE<*%$RKML#3DwUKSX8+-JL^baW}>2dAp2?FVEi_ zClC4^eJ%lzbnVmC+b|nO=)Gwuz<-7hf|DpDAzi!9QeT zbU*-D-B;(GQUb|LO?Xk&PnUun-yj1W530N?_G-HPF$<{8uM?_oDG6e@11n2$O&3Y% zL3?5chvJFdF2-X21{CdV$=Y)U!V}@RZI-b>0#5e;JDk06=u*g<8@RHL%;!TrXiT~g zs2p0S0znOKceoz~b%Qb?^IqQBdFXYLcXd4AZwkHsmuz(5Bdaw{^g{tQwqKXp$m69y z^2BQ_c3E$D>!Qh3mFNra@dv8@bqy(xpl#pq(!cR$QW_brc#GA?Nhe^!bg`Z&^ z(8^dfB05C|z>rio+yx{739_k1!+6gSRSvzlA{1?;Xdx}D`hd)!;`zU4EB~T)kY#Ks zPUSqWH$O>|dZ5G6MS|bU^S*+I1+41hB){bW)qlIqJbSqyJ|UFL$jaNI>-VllW*-CK zsAJnG6>I5*=(4a`Zekux6XGU}(}7uwnj^`l-3gAoDP&bY^Sv^uMkH~SsPoccdmb)A zmZEc67Wz|XU#dwH1n|^J!m+^~aUqKmp9mHCHo%^_aT!~w92T5-`q$F0^eel)%R4u= za?gS_!Z`5dFF&hLKaNttNDcxq#St+NNzyZEqlY|x^Fk$RToB>W4K4Jf;pny!KrjC` zMPnsI(L5!hSbXw~x#(P5n@OhBHn`so+-)Zc7|Bi*R5+E3cfI=vPFCSGl~`#+>-vb}0wt?${@6 z4%$jF>(!$Y0Akyh117=(Yi_=*uj%k6!PqQ@s5pk6+5_H9Nj|_KzU@f{1y) zG~ZG2FZY`nxO_0~NRFu-{(iXDnA+!a0Q$2rndWrVP$nTA_ji6MUt}d3t2{- z{}mNd^qE#RL7(Bb4MK^6S`>9_hR*CSRdYEwmAF%6aK^(2!>eNf0)JgVkF}YMn&&J? zucOC?Z(bsxINs7{Y%UwG4@k(N)|XxFFOC(r<=A%IKeSRYE1lm#dY{#&{w z$DMy!74DnATP1@ROJ3=va+G13dP0j!GgJVMNR%EJq{L@Zft6bwux!sw;QxRa^*wkIhS%=q4%pXPMyS zB|CB!Q!#;_vieVF_ezkyu#mk=$jLGkfaq46wweED`3S4b8vFS(r^NzK8iEJy5+ zJY`bc*6+5eKEkrhs;-$O%}=$(VuKI2 z7pG>D`vq}os$|k!wfDDot%Y6tmx6UXpmHg2Jr`bFew`E9DeCvc6Y2L}*_esbuk+y1 zkxLg3&uVUULf*%5oH?xr*GCd1)*G~UM2S^tigVzPFwUX50P{HDC*u`f19kLclTdgI zldL?W^qgx=Ca*nC1`QHaa$Ck)c zw>H|AGdsb9n|i#w>0yt@z`mhQ(2d1@mHCo0F6~K@0T6cur@gcDvHh7qSyLKo51nWB zjSqxgaR->`vhNSm6K$>mlu2QFR&z6GgVVQ^BFvaU*W#5ct}B8zUM11md4pswa7 z4XYzH7Jb_alM&NPuE#ZZ9mD?-q{S>7dGO<#x3@dHv?O2nT#UFV%6q88s4wUk&0OK% zUDXL)UwQ<{F!J+ST$`dkZnPrVj!5OGGnEri3!+053AXo91QJyH{(BOWoYv}f*wMhJ zpL<`x+aJWE`OBb?4XZ{pNhdH>MN!96JBBC1`ZkB*|pj78!Jj@1Qq$9$ZVYomBJf5K7DFhO?Au&~pXwPFF!4maUaz9@C zfmJ5eErwTgJ5KdU5IG9T6T&ma%Y?o}rt|)v!%b`^mA*y1ik{j5d8mvb_|P0HB2VuY z!lZr1ap2tMb_4IjV(8qSv5^ozvE&buJmEYL2|B@@SfAR^=gPq*g%5fXAJmyc%&>)- zru60YO)JT@PMM{nHJW6~r04pSvv%`A2_$;is^ILS%(6FINB1C9O&sX&P0(y6c|*b^ zRCN+zR>QF$gShobT{OCcQzNT`Hogt*k~%RK4c@>Ixyo#WNU(3SNH7b> zhoQl0tbYiv=yl|{=)`U1D*+QqpxmTI%Ko%wpuHkKaAY3il;4Ov0@qduPE3?Bct0^Cz@B03+MsOz$J+n3$ zgMqcY%&`ilAY)sz6l~}|2mpssh!{X(N>q=iXzD$TDI>r+JP>#|_5%eo3C_SD^hGoM z5{XO43HlE5P5~ew8QntG@P3X^0hH|>9!3OGVy7NnUML~V47rp+R0m*biOSgYyBIL` zf8K#$oV43DDlY4KZE;O={OcB+bR7(|PaxQOn zK@T3^Q5s8}f8M?MpHW?*idEX@}%!qrqobQav;M0>IRWT@FzPbR1 zY@hF8xT0-$b}XmlI1n_}q?NTMaNnez{W*ADP=ndAx4bAq@CO35Q|E2dIJU6C_R&eT zMkwd`IHZrUL;yyJLQ(;_>ZZ~>CD3)uhm+POmVQ4&o=(f3`4&%{pa@tJLbTg;5YIXg z6(5HrWJ51LB$i?b8$+%&-32k49pNH^H9x)X9vyxSA;i&g&(W0BSXA0d=qe~R`&JIr#_}B zK7L&=zyq^ii*lkV64Gq+F&f!fF2%O5k z_rCo_sZGJu4|xc(R`ap^tqTCP4D0SmGGRX2!5}QH_SJTU+-PMT%qb+=} z$;xwWZ7=?KPOV^2J8;ove5jnU%c^gf7^s&>sBQ%7E)O)Nuf7z9P!1dJbNjOh+PMAW zBi4+^fjZ7ugv9dXOm)?ZWG1L=P}`v?OG&2f9u%C=2wce&D1F@=mG0%S3ttK>SO1N0 z@kE|j$&Vl3C$V?xR$w&;3b5SG$8LYPg*yJ?ku+^elFAo6)M+1><-bkrNTYt&d8Izo zo>&ylhLCW!32z7l4s3e|{ePRjuBz2Z{1|fIh{BrSzvC)5~jteBt`Xw0s)e|7)pj zWUXa;p|H^yT+Mw3i^ZVAKmzcS`!+R?z_#zlxfY&=enr;wFiwE-2;#L)F0!f(>TKYd zO_@|$O^$y}uzQ`)Afpo?J35;O{ieXu#leYb=?OSveb)qpOsk!Rsez+!6Z zEr-{EANqmsjydmfqBRga*_B6vSbYB?&+egBQ8f|(`;sE(7{VBV)&m8`|#P5ZzHd-?LySumwV z={jAyy7PNe+O(G4(}9&9V|_w8l$&~R2ns`R78Cw3#m11MW>|RGUoa5OLh!*2bxa~J z4p`H0dTf)Puf#?|f~vkwKZ&RLLe(XnK zNyNUeEXMlDeo#=wf0iJIA@Mn#(M}ci$iX>xsfR*j8GBKWq;CI1p~G~N;)7y}(Ho_- z7TRG_cvjx0eL%JMNSbE9oQ)5LWgu-CRF`k1_@BtJGvt`#czc1^^vPJ%E7Etk+5Dt} zG5fPm#WL~zSpUBMu)`morwMNF^`FpInYDegzvt}!4Jr|3X2Sc{9O^#WBo1NRX8kKc zUkW`la(Cjm(a`P>Gtfm_)ZQeEk9rfU(lH;$lrahJpI(w@3BR(v%JiSthFk9ZEWRVp zv~e6d9qiNvLDdI^jG8Y!U24|#dO!w{1Nw*AQ={WAjK4VPd??obx8gG-Fuovu^{>H>w%a#;V)#RKD(LDjVK!K$J~^< z!@IrD|B+Hb*Tx<3H&4;!#P>$n{5f&(*{zB`cUOm#Zc9yLk^2NfzWRbX(5Ji4%$2iW zpR>^<{XQr?$6mdctF-Bax3b#XWY4VQI}jeh*d6`p1U* zW?iG(xBMmOo44Ft>(#ww6_~WhpaNG7tNo^R4)?c_&m{i~m0@M!_<>V9bc2M?`7iR#hlaP$NtmE%pJE#LbaJzm~#uJAwW` z5lO8`&Cmf_F;!R*ZPX2X3elf?3mf_i7g32~%}P9wcoWmG{9Z27=RW^WToD=d)p?A5JXR`Csb^pEsQt{J6P2TvPu_48%6IY;dzNv06mBb)b5-$ieZYTThoD%?s=3GNcBc!dPv_7vUxWNYHg$4fVqUino zWEnuC@CNm{s^8cs&5ZnhR>Y|p-j#@IO&8eYa3^U^nD$IR@-^YrRnE3)5#*!c*2?Q12i@)Gse=eAGl8PD?eW3jIr`cnC) zuYiuH;NusdGiD_w0AI|R&4v6eNl42y6y~=(Nc3MEZI^a$7ePrP5##>ozVRcQSQzL> z=B?-+V|GH@HvKB|%X%NWNPivK4E@o4lv;C}3Q%iy!sW657MGeazBkZ8h2IZDr<5;? z4)jk6@*_;+x*(!S%B8&(c_FJQJUWh~cubD83XM&ZsIqCZKA6S)%5jX<68iqZSjg&J z$OSZbUWp-kudfrj#H6inO6@gr8Byf8l%AE$KmPwAa`2s%*BHRSIg(F`{za1m}{pD@rA=!ueSn9Dc#O zxnfch{eQpbDv?6~6IT4grl0Xq{bQdHc^>U-WmqZ=|fZ-uP3b{j{b-og}F=!x$l7 z(OcSAnzVe{&v3KyGO_+t!+@WATqT?L$Lm3DH8`nw?nwn9j53Y#av9S}<^dMfvfvhU z(LSjam;NYlb#9EBzg<$AS1*eYqDoTERUKmZ?);jFv`M^j&U50Oce6ZgLmEoZr3CG+ zSX!Jn-AHwWme`@8uws4hFS+&+Z|oJU_y=(~ioDN0aj$gWZAs4Z2{~J3y-u>%E%FM8 z+)^$pR(c;6Hj#)Uc}S#saV{%XeVJ)&qz|?_n-px#M^l#Pg9qRVb|i{S`2dPUM(Bi{ z;rS!?r4Ph{4a4renex1z;FY#*ZUg9m$Di-V_&D|+Ff#kl#XFC{!Q<^2`UZIx%0#7EmRFwwZN=N% zW4m%1?!w`FfTKobCq8{tUzZ?-)`$)K)t+rgH5;8cBV_ruLx0!u%{XIfJ~sAU@m#r5 zxFKW!;?MofF!LUbeCN^|sBeNaq$=J(FL~iZmr}jI~Xe(ZnW zht52A?Bx|vIFn)kN8v-pX$Qn&hAGZ)vQZb&7t)X zXqhgpYeIsMBvGSsBQNi_Ss`rSy-t#AD{-kil%Qm3kpTzajLUOqdQ$1wWz{*bBVKn~ z=!s#X)zuj-eILn;cj&S;v%X{9bL&wBzAi0ljSemH!=H>_1Xy^L$dtXTQLJK-eWJ;}W1SPP;0J z446drIy*PwgYNpaFWANY`DDuqdKdQACWesv0U)Kv)sjFp`0ocEECG>p!Np%ym%1(o z?>?W;;4JhRHb_Ro zDr9<1q?~^G&Ce18TJl07bLVz)yk|_}OQI#MS%N2S^Q5z+E}gzVWDTZRUpHsjLERqF z2n2ou0VjJj!T7U4@g1rTcd1#Il21P}+g;PlXDloZmsUTQOV7s{rUK|88h{Oxes+_78b$+$WLEC*ZP*t7`B9BY(rqQ1cX;#e z6pg(;Hpl|%kObnpa>bLDZ)*ETs()7{F&5z@2%fsllg!uGgikL#4xgW425%+?^g!`y zdVH3n_mlU!#@E(AVF9U@XV?7RUkZvmBo*k+1uLb%EpCP{C-g1jD2u9$N1-b-$GR#U z&LO$N0oOvSOTe@fk+a!aXkVQyL^eI3^|9m@D}!QN`~1t-Uz)`>S4aS{)BqM*b*Rkz zRht+`S-#!oU$V@KdPZHu*;#K(o5kskP3C}Hzb2v?i@v19=8!!SfN#~;X@pD8O29n_ zF9o?Tmi&;k{BM7dH~`80{m-|}o9JyvoDCy0XAnh#l_5~0Fgjz%ZYL1-mrqhi zBST+P_X7To4|JS=0cFcNJ(CQ&cKAhSm(31XQ;4!?mDl-O)!x z20p=#|2i2{{9+DCfhJ(`c%aLB0Nvfe#4pmQIRzYR`~$pRF66$$ZD0YjLJi#Yq%#7D zT_NJqK6oi94MpHh7OXl3KjYq7K2(S<_4YgTF|a&O;Mls62rc(DotTqW*%FV6w;s0u z)!Fdii9F(ixCdfklJ%EmZFXh!C0Qsj5& z6kSUrz!4@2XZJQHyG-uU>gp&vIUTVr4^En-+e*|3Kq;8niEC>EHrq91-onbb2kF@0 zLOSNLWM$`nFkX^awad)457HRLSuf+6QfTR34a3MZBrKWa)no>=TL;lM@FDn6eDmoy z(0x@sdH{d?ykK1NDH7YinPu$5)7pzK^k%j-oQX3hiOm1zD-NtI0{xP)s~y4b3xbug zp0N|1-RMvV3rC&fMlCMqMgwJ1=hdO$s_h$k?b``*E29mgsh{>3fZcPt$luP?enz0g zX*SA{L&#cX=|=b!FEKhhnI@M680QsK;7$xwCVQ|czVz;ogiXR&>1|~t1_*X$18`&)xsqFym_hLsLmG~Y7aI5gbvQ46ike#yU>Ly_RkW8e-D(^PARK6U+vpRw^ z8f+~yTAV|*=OLHX28Pe;{sq`D3o=Wg8*H>(y&!LO2g#%3zsy580b|60air;SAC%si zlt@3!FyyQ;`0W74;4}T z%7OB)L6>N4wY7_h9$e{4|yZ#E-M<*0k68F z2!~CE(}i{%XDIfKYU%Y4%+Bl3rp`B^syTcP#MyxL?8gt|>8N1CX7JTNmLg@XGrDr@ z)>tem-~ASQg84eYRhU<=LCaj{Ntd{Ow!EYc2kib zLf)?v5h^+^m0CoGDiVA%U#h7sX2s!9q@?_wu3gBWwg{HS8)9msYcXrD8L`k%NO>Q{ z$2>Os+@oi$+V$2>weRYFy>RWrtwD3sj0!YKl-KZ85Bpmx){MRp;K`;(KYN_XoFsEX z*y2$#^AsgEKq>YYjyz4Xl}vg@Gu-QZ#^^D7virDTtztN?E~oH-j+{SHre&LAki(Xe z@&3Y+>W&C^-i%iXZUQSsY@a(@e0Y}BKy32#;h#W&9G#1%nI(}mM%n0#<)=`p&kl+y z+3$|`mRK40qb1#S|I;PvB_OE^$P$y+{-h$z^izKyIF=96fU=jc{%BlN@kCrb3~#Ty z>J0XhPdMK*$2zI@EZa3=`65L2cy{jPansg?{O@_}j^U?1foJpQ8ZsxJpK6;zbCc(h zOPggrj`ZtDR0JRfC>5t{Td;Ln&)n=4qJ=^lYN_7fmX?7XEuf;jx+*Jr|CzXtmKneH zl%9RkukojCf0pdadFR@T!5mCVwxLx<1Td?#`v#%i`UzlLt(mxv9XZP}*5#nE|Dq=M z9wqzo@N@reY4cS>Gd2*dkeld6A+&>Ue_o}9>kY#CUQT;x#XyB-7*FT4^3gWfOw zmc>cj0V@NzF?+TVI}OAoMeRG$F_w?cxc>FN1>*>Bt{CAt`r89?Q5z}(izcKa@ossd z3pSJYWM9dvkcrP=PgJd*RllRtAnrMqwc(T`mA&-x9$lhZm!!nqTeJdBDh_s%*W7h; z7i=yVep#A&U@#1789&!q!v54fNgu3 zvCcm4HXhek?usTi$nHQl;;F*{vA)N?8H*kj9zMFU^;>}3IT}_s+DSweN@T)ZY(woF zHN5dgpj}$Sbw;RQE_jT}z8_;T|E`QK+zMkeY$)Q~YBR0WdsRpVw4l8z5((N_CsxMP zQ-O_Dw}47s!DByKpk;9gBNrNahQL1MJ}U|_4EXl<=RtP-5$OA#e@^vF%OZl6p=aKK zuRRcEn$=Dxb|GUlv=c>TayarcpU5eQMtr<9M_TrM zN?Mk>DAUpDVM;ki9-Np56T=$y%}s62Y+8l@kdG7jnBY4J#jzWsGAJq-e2^S&z@xu? z;zKbePk3>NNpxo7YyIjHHXZC9Pl?l3JNIDK#b7tbTeU*)%Y_=M4HqF6(6gq_x*Z=B zUvfpgK2RHz4~lMY2LMRhTu-{u)<}&llC?dd3KrQ)L1b4ML9--oqVu72u|q(du9;#NHy(2u_;k z5gd2W(D7RD293v840XQ5|1&fFcg9Kn<$q0nZ9w EH>#L^4FCWD diff --git a/analysis/figures/F4_halting_pareto.png b/analysis/figures/F4_halting_pareto.png index 68613ea073a83a833043e8401c73b96a89016072..23fec14e814427e57fd1851c325c1048a801b4d9 100644 GIT binary patch literal 44333 zcmb@ubyQVt_$?};pa>#}AdNIgBOoo(AxL+#>Fx%L5Tv`MySovU*o1V0zy_pc)4A`v z{C?+-^T)kooIeg@@CB^B*1O(#o;l|;A41&hFTW^eA9c>+4ZLLhG++ofxRu1;uY@D2IoGet9uC9(Q z0_^N||L+xS4$c!e;4b$aWwc#x-6Fz=zR@f)N5Qj7lVv5u)I8I67ci3^s}o`T zu1UX3n6T)x-*)=&GxeL4toKwQ(|9e@ukh|lb}^c(H2z~p})!NM4TE|AdRFrtSpsDA(7N28x$C;hussZr)X3{3CM7YF^4h*dRVe2KfoGh}-yrM! z6eT2>?*0LTPToXp)Z5NbJ2^QOu_cNI_%rD@yB5RAg}v3-^qX^U46|OLF{u~Fo#kK? zbM=0}#5Mcz`N?QCOkVcEUp#{r_p*zCiDGp+E-tQ#UDWAPK%U)Peaz%StCuRO+wAgW zSEa`MO{7Ri-u2q>Wm?E6uFnu`lUz4xtY~kc1)j`pp`yR{gUB#0^f7~u{(67>bCnK% zzb60d=4MgPzeeXbm;3tlWoG&+Dlso?CKav&j<+U8YpqoGxcF8@7^nj}U6Oa7aum3n zBiCwX-`mP>bKN#$)yl6eb6Wk0Pf0mEerwMKOeZ?}y6avaL#9IA-I)}x!}X#L`Zp;u zYHD$5!akFEbJFE?zUR(+D!9-+k$%`oea4pRQrpg2T1g5iFJxzHtzykj=moU(+kLg! z^ctzq(9l#G9gX|a1setX|gl^;EvTzFPw~ znyVW6^T)C;i8E)`eZG0LFP35AqN)PS5QCVQn0<1#}d-&nmf9= zA;#pM6$Gq0V@`((gmwsHWZIhQeQ~4=L>8mOF2{FR-0O1qi-=;B`sANi%0H1 z77|V9x|*%CnLC>K!5hbJC^E6(zUH+&S*E+?_eIZNZ$e@z>DRBfPIsjj`Kqd_oL9cz zZ}QmBn5Fy?Mx@!(n{miZ0aiu*I7L2=F?!`d#;sH2JC|ph^Yf*d)B3gzE~U(k zF*4VSoTCDnXzD*ZMLq{Vh{z43_ZQnH1VpZnRCC-OQV3~nupVvRsweWhnAfoe z1%ORFIhZauwYA#duvA{6S`aa^T3McG=&$W@d19vz)+eRjg%%IDjGI>DOOcy1 zQ~jH(b5)^s`-SFaVs7(Bgwu4#tZ%Cl>qtV^P2~ykHQ4WNpB@lRD#2su!Zc6%M)Tw+ zj~JoKqONjW zKIZJFf<3t1?7EqaQm_mF2W0HaQ_k1B466A({#R!Q)?VE`iHypxQ9AgKA5SzR5q+^H zEfM5`>!XJqHT!lx!>~Pp;Y$#hHsfm04rA!#$5)@L6=sR{<$ekzp8H^T4+m%B39F7; zdahhNbKZmd_eTOMd%}rLPpWLD$|pVr2X7}n*DO!N%BW*+55ghcnlfcOb3AIoQA&D~ z$~zGtt6rkh=S!FM@F9wz z%Y5*xi|erK_rJUUh7GG2Z4UT+o4;}XA>uJ%qub=1=g+@m(W=pEK|@V#>piTP%q58; zyFA@**i34)Udj*+FlZC4U8@|tcU2{d7ij%10TRe*D*ZJiuW%)JHHIj`nZ&BiCNowkMxJRM9Zf2RfN-qB2o@Uvi- zUgEnaRRe?P2k7<1&`nfzMXK28gb}j3f>=x#DZIW)SOK1*(+pxn;Z#}M4Gt;q9|5n6L&ulC;dd8=JGlq;nVu{6{T{?_Qi_Bm z_%Ez}h{Wyv`Gr)qbSxE(TZjD7K+Z(>BORGa{aGS%#pa-2%|TNi&tHDspKDMO1t-&I zEBMoLg_TQJaL%xu=eO~832fc4zn3Sf))_nNA8m`*ba0jwlon^ zOGk#){X0H8Hr*RM7j!ctVp+7$tbQSbwaSu=Der{a zti}EQ*(V|cN3Xv@lSFpp>jt7ad9@nPYBh*;T0N_7s_qdC%+y#!w$ou4TGn&T^I6{K zBzm;7(4yHf0Cu@+xd2-Rb`rUlou%DqyM$BG4B2kjFz+-f@j9*6EYIPba6~h1TUpAMO{sv@;m3k8#cz2MAdmd3A+CIebmfm z(8?(2zVmi2{2}gH_cqT|)AD`?N~7VaR<`X-wfWTPPf~}*wI-b7-8s!_^T8Ao0RdI) zryQ@Y0}fVxtXTgU&uQBOF*P>{1s1dToqe@gU(Cw%K!&jH!eoi263B2C?$>`DgJRq( zT{l&N>+vd<^m@o#^Tm5;U;16t>PCdM@FQmvShTb4p)ks%U7b5~+lM3qQnvRy1c5vshtIk*gYO^3yWdL=)k*SW! z>UttqW}5YU_<6HNzs@7bZy} ze$CChvvnmOgM&$hAEKcVFssQd&FPZ98vjc;L0YT)jQm~3pM>9VV&VD&*ha+aihV>Y z@2;?Jfg*WWWE}xnr)q~KkveSV(5%0|<@M}*vi4b=F~-(OxLtfrj`yqABwan+g9LFy_N z_2}#)D4SdNht1*tTCLIv!|1pEI)QCSp>v<+4h0AUu^&I)$>~{<$0$wz_!b|V!Q1zK z6S4_M+sD7#Yudo$&HzaH@0iDG{ja_K0c2g1R!PMgZaR~A))PwAdjW=KNI`%>a*x8ad~(GtNN{_W;dIb z`X3{HX7F=;n= z3D*}N9Djgg{9CE&G_t@8tm(((03|pu?S_2X@bE$&9q@e5<~$$y}O9!0{E4g*3%)xO4r) zvK}(1xT%;H`PT?GZZ+mg&Tcy%32zVjh4tLpLe7Gt;q{sCBorfYr^-GTW{FmlkUf$k zjhniy)3VhXxQ|uKW}3P_oLTXHmsPjEa38bw<0|ePzyFnIYQujoz2Mk3d+3?MH;gK} zuh_lerDP1&Ta2UB?;w9trOK%5_AzE-tvZNekvlL}3oKTmbseG3QJIg81i|w9mUva) zXN4>>=8pIW{LY_>h$`yo_-BJG%7_PF=Zv5%M4 zX~Ic1>~F@8EVRZg)6hDv#rcMN)f`6e-HYZmB1DENCvZ2oT*|&D=sX#_5E&7=+^InV z0@(-h<;0pxtLV-72IcjM=S8cLr~fOqwgjbsv%ki~oF*T!8V}IXml*2vAA((uMZlOf zq%0jlR&}w8Uxu1RF8=O4UsB!Q{bqG0w9mA2B+c8!t|On0UMIS%$?orGJ+@<|@1EeU zQt8^Xcd|Q27|b+@j3r{#F{_e%C?43d+Jy6Ug09Z%&tH@2+wbM~W{DDN%0dC?&>&ZE zl*Un<#&0ov%a~XGH6`AQbK*8nIEY^T4nuECJop3~g3yBKvM-oZru7_k#CAo8JKTlz_0ZR* zPBK=_*&`6>7cD6i-DhRTW+qki>LW!Na9V}^EQKi=@j^B>C^qlMzL!39MsM^KPjs1w;o2kgp*!=cP0OOx!(bgV^Y-{ZQl9gh4$oishj8c$p_(U zb+QHNLXofJ8vw@SYb`noFbluDMDJFPNJFLjIm~&p`7q&RZU}#3B1Z5vd+e9%JmR&R z&DEj4INmN`(|>kPO5~aIgApuyHdLAS@zz%E1)RO37l1Bx=KE~I%V8#fRWolqVz5Zl z8*rO-TyMfLCN6z&9>VApya9n)g(MC!VfOmR+q|FTQyQ;6%fntqf9>JZdFc4Qra|TQ zZBF0i722{jVe78sJSwz!jlVhQ{9%lJM+4l$g=D;OOW$BZ2-fUcLLC8K`}sdr?6hE` zxPTcWe}`|}nD#}{V88vcoLM!9P>PIOFQx}z44qk@$$6NIZHER1eDR3M;)Jyw0k{AS zPzyBX0yloRs}-xsx&?I${KEG^B4l$GPa;GfR*=8L@IT5veiF~}=#~^fwEYicaDEBI z@ppsSt)+XC*#bUZoOL&G!{=m&soZEEOfg+0tKDp^G7du{G9*DZSCXo-`=Ez5PXbeeuhy!{eDK+OJUOm3V%6#H_BJTa6Og(45EBdhG`QluC8XPMZ?7$w+ zmPaqn*MChH+Wg}Pyo4XN;`1#MGXMpKiKY$ePr@7uKFa{)C5jN5(H5yORZS}4#sal1W;K( z9j*_mis$LfVxGZMcuIb}Galfa=Sz^vxD)Oi1mkc{f4=m1@+7h5a(-qi|BKONyPr=@ z1r9m~gkD6yB{g*grV(!r&-j&qgdX0|B#O?9aAxc_aW|Tdx2FVaStE96Y(=Hm#xa~Yue<-at&42bllkak z;3fkk;O~g9nTbmGa+q^rbD8p5)U5o(pZ={v>gsc?EB|1!TQ*S{-|s|Tf8(6Rjrk^* zi8Z=$3`u>3s}(6l9vv@7oALx^{{jVnQz3sab zq5!}yA``7u`3k$$6l$3#W|N(V_BhHVSf{{~OyGN#+kMu%`AuBkS?>Gl12^z=$fq?%2VV_ zdI47l!QLIKNKUqqoDX<%5nr#8N+^P-g2x}%#$-RiC#OGnx?|{?BtDy5&;CnbqqhpV zE;5n@Q^~<`U&cDgAo|rGMmPk6ClsMpeFHUAS`#?~?;*3vcUM<>MZTTnbW@+)rwC4f z>AmZ6>N&}xTb!T14JKbD0>Ha zH|yMrIKu5{370hBwwYvka*xXKXJS?kx>VGDXBLyph}Z^)UFChb7SEega*#YE5IMij z`&2+A{?mQWF3>oPn57`{6dfsFr>G-=5K!o&qDCq=(taLNkU@LZ#HNQYX~hz^^)2g3 zAz@(Lt?*IVoBN4NPQ;Y3$ww@PK=X2oaB(#grRItRs$_KJ>mfR1f^;N>ox2Eki1_)J2=e15iC zRk2O*n{%b@ZMgCIoUVsc06yM1F?1huorK{CGWO<=5-lfR*hIvJ!pSNqlq4i>K0wPT zJH$20v4fLv92u$Pd3J^{+tZ|`7KH@q-dh4^GbusF_bCEAru_f69~T=pep4qZ3l0fs z{kIdu(qZZXLwEOV=W-^&9JfP196=es#m^Cd5B=JtUEXrr(*-x}1ULPI;&`9=obJa@ zrzJ2xG-!OY_c?_I@-C?vO)>tvEc!-BAsAkJ|GxRg-8*+=9RKEsn^^H%B)(BlP@7xs zgFU0A{k{!1@u^yfii@M|(~t|u`vm*&{yptqR5$Bf4NF2d@}?U4=LuWb>FveG^(>cE z&y)!>o9SSB6ETk-Q9a|YDQL#YxR^gAJ9spV*M8WfWV3P4fferZ@v!RMa}J}EN+726 zj~A&94gs+%s@7&oF(BaLXp>M>0tjfa3>luU0|J1$gHL2}gLEj@Ysxk4iMY6gm`;nw zeu_RX(B4PuY&5U0&kr}?K+PIpO9fKq=1fh&&CSY>FgkkrpF7pWTxL-seiwiCQ2;?W zZ;sN{z*pqk9hTZJE}v_ZM!9SZh0>v8VE;lvYB&CqCorcOfOHH{&vZKc+sFjXy#hl4 z+?4n9Y`j3N52mjD`YKMCxI2;)47Pb83fxB@Ktv0t8KB?Gob1d@pgi^$SzFw-DTw!q z3Nm6q@Ge17zI1$5Z@*yeGy|@c2dd-*3YSuZg_O^}w$XO3-Y(B^`P&u>jK$jTI=t`{NTzL|<5qO7zEx^)fl%TzP3XyB7U`k-o5`dcv&s7S-WN*Ez9q8N0f`!U6?he{Qor(PPXr#-NPfP#;Nik~GCu8H@=})|JXgiLU`B_xJfBy)8 z$Im&%@X%`0X7^?E7D>Jvfhd8? zX>>vb&hehdrg^D?mY2jDlc)?w=JtH6ugDX|DaHiuiQ~1tXT)HolC8@;noJH=5_JOgvwlMlcvZLz*fBg<JJN};L7xFZJ* z9cL-Czh}Jo;B}y2J*t+~dhrWJ+_00i6W(m@p3JiRY_vhvYICYK&J0eleqcvSoqg1^ zfxdor=BeH8+XUn*CBGSsQZ41#db{Vw#u0weQqt0JkiZ#0tx9?Oxk_#bRB8hylx?IP zLw0aXtXs7&>TY~a>;61WO$(iF{#)W-rQt$zjk#v`ao5Vhpd$}91J6Me0zxALzxPq$ z$PtKj@K&$GTW-Am^?0+uWdg0P7Z@tOwF(KOO>RMJRl=j2?3 znRfL?Yq7TzVuiH^j`-Y5#a_>tgN)<;^vQAmF`MTSp*GFT_ zN!N$NqH7>v$AH?+leZha(cSsSewVuq$n{i*ViX^UQ6TOxfP!7M-5jrWoppRlMBsV4 z$D)&GD*p@K5%EyjzG|Zl?Q{S0XZ%iU5i<)hd|Z0!t$&m&8b1bK969(guQyp-a zsm|Lv^4Oc_Zvy2M1SqUbxUFzn3QNgJ=SY1CU%v)#y;Tp&HEh05ByRToeX|GVJyqyS z1(-GMiBEBU#SNuO2GugLD`~LPitW(6l8~^9Qo!NG~omJBN% z3S>~qsAV#e>i=w+E>?)d!>p@LD<(0oaFgEJ~%K9t4;@y@b= z-glEi=ZIJ?NOc#m8&v20^t(+W6rZ>3Lod<7=P!vx+YxZ3lE}V6twSaSbu5d-so8|Y zgeQk3lzAYLTBU=Rs80w**X@Y%*qQ$79ABZ=lwE8F_L_8_uPsQjgseKWAYaA-0$7q@ zLCxiUy@!5oHOQIVe9|}{>N8E*skXbqzcv8^xYLXfc3fykZ^4;aXQ>;q#p^Kt2dvgB zi2Qch-KjR8Q;jZUr05Of2Mipp6j8qm*UhnlNF9u(v97L3k2z5E+qL$hN2-rB?89?W z5&cKU40OCABRgC;@*B~bV9?%lE}U;cG|fcB!q$K)Y(wxO72^k}B3Y9#8&chqdSh-A zK`F`xQlSPD5ube$C9Kq_>%H^xx4ZEh;Ujm*yQ-ZL>}(6+@%Nuc{>y1%uVZFlDe=Pv zD(uz-@XddbDduSLTX=;3JsrH(8*u1Gfs&VVE}zJz-o+(#CC{ogi6J(klX7s`rAH!d zVv;>lEwMS$WOIrLei95Dnq&-B@=ZE1D}J+f8)==2RB3W_k942}n_ckC78E<$yo=^= zMX%nL8?+T?Mi8!?F8T~x65}u%7#YQYau4YncGdjw;?j3Uo=1O1m@*!8!*cQj7#U?i zIea5uQzFhi{c}5w!_Yka7m;geb6cTgYOS9R0}RDaV%GL+qcbo0%qoF@!v6<@Di;+m zuLd}i)C`JgZ_9d>gpcE_iKQ@w#|_v3;z3iB-@Mg$2x;7p=@lG3QyFz_u^_ETo+mDNid*oMMa7%>y^y%2!c z%5Gj*YVR#|bckfpBRw#Om!kF)$YbbyhQpdN$GhI4jpivTdo`(hF{M0*vAP0+6tGgm zIzV!#iSay0`Qoq2BNjgUd9C=5+@EIqdmVsjq4H>FW^CsoJ0}OW(Cp@XBXs+2IsxO` zS7q!#u2yt-uA9VEHs`XQ0;MMnvPMgHe@n-#qe4wR18Z#!EVE(6zfFL|$iz+ZyANVn zNqu^Fv=hjAv)zeE>qKI_k;#;-5g`u1$v52T)B9=gqU1==3vgrI|6QoUz zlup&BhmzS`VJ^0%ffPt)zu6e5vt2{cxT8BTfAr+QZX&pycZbmX@bKu#Z3v}<^=i)3KJ20!HcI| zkDA9xbszrXb1?q}ynB(+18^%@8M%bv_5EmoBxY-cip{}oqTPU{=SpH=_a*U5w^&NvKN?*R+Xo|yG?^R5pC`l1<|F}#z)k8_K&;+OJIe_~ zUb9P9UU^QuS1FKS3+qD05s}EedKDN9y=JdR)u-fM=^qUAeGCD+(N^0^Pv78z?T9O= z5%2#-uIMsV>3gr?oYJ=Rg*XX@P69Cta3jqCBomGMx!upu;L0o3R@806(mP_nnCq*9 zUy&hmgFnfQgJ&jS);j=1_4WX)!Hn*|NDERxfsqi^k%t(NB&ZCk^ zmHIIQ_EgqZvE9n_)udcYx<)R*4dw%(2YE(C(NYf1(Q)%-aHUX?J*Dav)Qq(tjW<)| z0&|oKmQ-$)ZK;_Z-G7#`MbVHuC4-wKBUEC|Wd)yHcKqVE0kb_HMZgE27s zH}p&7hcD<_>+QhKI(h^&1$O>`>x_QQ#elt_k)zh=WoE-B?-4Q~UMO+hf_^{kx$_x2 zrCGatiDyW&Twka_VMo)@*}}LTOs}#IK#slL=l!B(BJ`Q;^Ur1J&-;gY*qA8YO8cR4 zWWRtm1Lk3T@X##V1NStZ6beEzMw*t1xPooM)hZClK3KVfa1{OP%%K1<>g!I zZ{liM;V;5L_~O)p<{lkja(10snFXkL$+QkYxB_q-5V30{4yHGe-2i#M$n6M}`d*8R zr_Y)rDP{ZmxctwJbF3a;OeRjoV|m(vwT0%9s|pFo3O+eGtFf`M3MkMM9F_odpj~6G zo_?m7QCp@_&L0{YiUg3S1Vyv+pcoeSTytn|$S?mK zmWE%OVKtGGE9FEk6^Z}!>F+uy9Z;Vh^bxL4PiwiQa+*qo!7UH>7X8MtEGvLJsz}%U zghm{GVi=W^XurLm*h~H_Qk}Io9c(*9fp7IEepYic2A6sR`#kf2d}BWuNS{jf|RHqYZ(rf@!%*P^#wuTuOX_I&vlk*tiPk zfR$A6h)gwKK{@>z6s2Hwbg83>6Txio8yCK#(+3;9wB*)=2V1-#@J9Qbn{|c@8L5>v z0;do2r8zS_z0?{TO=1%uGVT?WWn-!Zis~o}3sEjxeTs&Gv#l=qvVK{aC?)@JXicQv zwWq|q`&!yniBW@5mxwoS^g9M@^|`hiV1nxh7O#oFvlAtn9(}yA`*W^BI_^^XV@t!N z#BBODM92S);1Y}go4x%kyZL&jW;i^k#GuCfi|^5hv=rQOeJK4oWIn5#NJ_u%B^#Sx z6Jza{(~{|DW`*X?(6@SwGK+>n>Rh+d=v}SVm=qQfTd4}iP#S;Ex~UE@Mtz}h0e;7s z7LN)EDJdO2cDar+ojSQS)>8nmb}K&kc1Ngc!d9!K`49PNzU5k6uNW+V9ZQYG&a^Ye z2wy)P65KBolZ%3f8-bwk+92TC?Hp7#$w-zRfC-N|9UYB-Qr@iG&Fj@67e2*SO$Q$c zEWHoW1SEn*9;P2X;#keg%9|N|GzqX{EYJ#zTGxR)FGYX->|oV-e^I!OLi)L_R(5=R zjhcJJX#Vv^xqOTA^P|%mbL{xC%-X~OVu}PlS}-8gI?)@mKX)wQ_q-XU3*vKYzK{O0i|F454C+#ah_n~(o!l0QBwdePeK+tUH5z8>&awIdz6GsqWda{~Gy1^zo zR|o_L#PpwNXckjJdeC-czKCF%^#UhYTqX)g7b%xvAmpcm9B0;r`~0DNwOHn~1qg0w zc*KnzVC6?4+1fm+fkaBq72E zdi?+_ihBs6ESJ;yq+fE|dEAXZ&~sYHtI19Qs8JRGdAavHLR}gmW>S6Qzo8?U`!+&h z9;o0Fn}Cs1qfhO9?>qX!aD0k>Tf)FzKjRXNgOMC$C9~fJJUs}1-2(5A-)OSs2!YY93(T=FijUxSYTIWFIoRfJR4U(ajW za<*hZAKG;2*hx(KsV%#;dd*&j0UZD%jApsoSJ*@L^E_7Y%LO?Fxdcj$TwDEzoqzqo zeIy_SYP9TA%TRxE$?4JQbAs@=^(RO#Hz-~*?t;bJRB^6W{K+mDm%YcOn+R=ANZUwE z*b!!DD4!o$Vvpwwt1Z#{RWb+N&gwDvr3WyY`&x`P3+1};-^VuF10V#?i^?d2&mn`B zj;_hOPJ*pw>o*<0>vH~$e{@>%THn^EreNsS+-NR|Z2Gw%Y;ld^WO1md_U!vOv4>}cNrF)%Sp64pRS>Rav(hV@20bKaTe z+?*&$gw$@S@Tbob*$v${=pSwRHOx1$T({y4GgEnqCZC>ioChj93T&aU-5A_mW zmUucVQH#q)ghHl>zCSP|^Sl08hW#bsvFv!+oAy>lKo4vV1;$k)J;VTOpM?N&BmX${~Iw?K?{e-jG2k0s| zQGM0U5*a6}{WLLgSA||p($_07g7PnM??Bs~&uFpHDd*0eJB7;E17aJZ152Q-Kt7Ym zEp2sWMXy;gT&nltSNrD57p>$P9=QNfTE@2+b&T5+|A4>}0vYHvki@w~_#rwe=Z*9( z%;GCGg#a-y!D?`s^-?V_FRK)*G1cM7%E~@0I|RDqKe2LiG>?vjMPbck6=mpq@M*r) zYi!ewemLo92`^MhVRC6-rKJ$?w$K)c*|lbNw*ryWEo`=b-Es_QUdu-=ZAYW>7Y{!a z?uI99KW0$;cdh_NQd!&GaXQ)2wgTEAnW(?sg0!~Ko#jSU7bc}lhL9HmGO~KN^tGAT z_w)^XDL6VsN55_6SmHdt3!NQaog{Wc|Cd)Ba`5C>V4`;mi@r�zUpHESQ*>@L^FD z3osR`2HcCm6BZ)bV{2SFfP$ zr?!Gmp>OP2Y;Fuxks8eg5I;hE6XEFV08JOq#WEjn}2 ze0#!CfLj0A$Gj8LIxq>*H#vDU7vSj|-_4bc9T=KYVceQ5+y3i9B}y#DN~Og+QDG?R z24+tuRq!^oQ4o+oOVVJ-Hyx$IY^M~oTl=CKd0BBoi4*??&UV&ca7qShE-9(8$zLgOzsgKT9$#I-m) zQ7?SkwUe3m;)yd(uE~6JW>&{iwTXYx+bFLMXzejD*&sI@z;n`D<3%~Q5unFKR#8#I z>M4k~>WArHC#0k{MC9M%Km3KCPU1>9jx5gf+2beXv5o^8jjnp{tL;NTjPd|J=K@ZM zg8I~Xh7+-D1p_-@9%C6Bjh_IFXTQ+WQ2a2CarJeLu*(9JYt?&%G+$9Y0G7ErwJ3#T zm5R174KoOwVnqa8?()^`{<-z^$&)+*!W+PMm=_RGr;uAZCN9MbLGrQz1Vx8~0MpdI z$knsEU-}WzBtIzJU(rF!Jtl6NGb#~5#;-@nf{u(`22(`n5E{Ii+pG8&Ga0J~a=mY~VDrrb5PqOWCI`-tLr7DnG}^uj2!I5K zBO|zFKkj)Cx?#QeM7RntUNK4&@Go2M;Q9>EW6^}*iQ2wu;*l$#_m1;cE%8(=(&ZxZ z;DqAPZd^E6r+%nwBa1RYg`qN><`>Bo=(8wEFdR3%0kkW8J(849aiQL7>}Mg2bne~P zijl1^fpd>ASjC>;#%0?qwYQB3)w1j_br>4PCXYsZ%r!eV{;^C@2A_R`4#u1Pf`}t0 z))3_89TM0uoPoAF%CdS*U`5A`4i4}8+EBm% zUa#>4Ap00bv&7I-r-Bnq?~H?dQ2bdMG|J@yeue~Pls-@apNLwlu^4u@V&mm!VB**Q z$!@U?VU(sscuoh{kc6TlB*AzT4Z@kg8n!4C{Vu#Wlz&| z_O_!KEZ0HxsaKd1W3JdZ%%;F1D3xb!yU?tTqxW9%Lt*#TtWAY>*`-3p>G0m6GXTjq z{>i@Dd2V^Pg$SU$tb!&$gbRaupwy?H)+M^E_2&)KNs^mip?affRLb@Et>zo6`iT^k zYO9?S`iG)-L5Xg33l4KW8fCq%rOYT{Ah#x1w+9$MyEZ-MUFTB{xs2uC)U&^-AY9O} zUab3Q#`?#*FsUx*D!t@9zqh61I1U4#cj4c`rcE}{v)T-YOilig@|_32JG zrNLVcAE`~Pl6xNXU_#HOI7U$I_uk#lkT>U%zh@hft=6P7>!? z@aEeIiSkPBt>RKqZ^-~r>2;1x?T$MhY|ywhI_E+3K9)vU=jU7;J7`F81vw!k!8{0Ab^Q5T$>kkP+0aMh843|EibERiSBaJEN@H)u+p^UPdvoXkG9!{UDbIOGUAa zyz}e+V|p2I1Yp3tq0;PXWi^uhiGWE(;`U(pH(u`}OHEPX^RN4wWQ6H^BYwF6d=jh= zLV!2hmG)KeN^iNI@vFc=c=K7&SS!{O=EAuTm|ue3=;4Yz^YK43S)&oC4$Zf(?cof& zbQTgAd%VzF3=*z`w`lT?kBhjjT&V=5NduhE+>jnaz4R~%IAHd+)KvuBmM?^p0p+-X zDhRF9t)>5E;!(;(j|&>-`1>UhGOtytZ*IoiKFi+N5;uK-R-i#QDUyue2PMH%D)u_W>Am6SNp2+w+q%p1pYN%)llV z(D5ekHZUj9o)SR2OG!@3`u+R&0)62hGEHX!lv+3GZR7nZ$=1*jCJFjPYVn%)v!FjCv-nqMr{k3vduJ<8-T3 zmUdkYGMME07tPO`Lf25D9fAPFy3PhkLlZ{z7LJdRhKC+$RaQ+M^_k}XI+Sl7b95XI z2?B(CUS#7z(dY^ZpVS?U&!uw#L|z|V12Y~n{|=4Y>rGFJq^(Z5YFW1Z5+QU^kGUL(8K)yy#hyxY{YSh+I1B7mMTioqx4(m2StC8a| zSTt>~-`NsN5kNsor~9DBl%~JNLW6_O3NP(|=G&i$!$#b|0-i9qa|I-@k~u&w8=(OK z-ah*0h9SMA@A1R~AG?2SOeiKMn%hFr%?Fw5#F0CauZva46Q7i{WN(%56qM3!4a*Rmk1hEHDh98%>MlP zicd_;^sf!6F>*6$DO>V^rYO&ONWe*2Ea~aW@u{v_FT-6GCU!{XK=7sC7<~`HQz_Ho zIR{n>sFNGgNP)`CjI&zSuowe2e#_B0;MQiK`zd>DGZ%D6sbovsZ<%^GzXcwwo0bWB zi80RExcsi~=Ujr~pzB2OKSWMr6jZL?QrrP-3=p3;8%$Au|3UxKm6MZm6vS9<(C#*J z1i&O5Z8{gH`wmVbI^SoQ_gck??E51MZ7wc9xtPV|uOXX_2Em9WL0Jh=z{=kqzx&flhE1O7`3s=t*!3#jYUObBY{ zc~GHJ(t;l*cO>VU9vNGRf&(EJiFOFuLdNYt2iV}yRK5*b1&w0b%V&LkeZ_j|O;>+_ zt}ro*4UE2ZHVa_jD(RQaZrk~Ch_(KBc?i=1km}~?BWor-uFmg#kl?Qw*Ow@1o#^?J znHIz}<>r24ppR5fe60Y;s@$GkukrU^7r;PO7j(82LN@r_w&LcSj~QiZBmwoy)oXGt zmOKJY4$c=xRy+dB8NxnbWANY;5+-a3Fsl_BRk^0q(CuG!IczKCGtvc*LwGRTb?RFh zIB|ZCmJz%QDzEGw)Uu%$n`pg{!ikOm)> zVH^fbh_b+(`6d7BU4kNFor!QDHci*Esr}(NO*GI=L`{>>9faZrAYS$8e4?Wk`u-oK zL)86$pX1U!Tw<>8%NvY3N5&7nV`KZ_Tgyxn|IJZb^Xi@>{TJFcImJmprhjUZ zPR~Hz=ux-M@CfI9(2D^iv?7w;=;-KT7YchP$}!HRWaY!Cmhe|kw_h7nF^$8d0yBU; ziV2tiVk=}2^G(5F*FIG1 zgW4FGnSzT+!AVF26il#nQ0EvM!Cin!RK67J-Hn2<%nxerp#F&A%sA|4xPoQ?7vi1Y zZovIQj*o9rca@qGWnX1ELd_G$Q>^=2>BDbA*7=~WYM#PEWsv2yf$R2yTRmTaIGM*L zLAA@-`{;6A^}>jl1Dp55Kkmk3kDIh`zuOiB^f-SYcV!l*L%p?|i)}6Kt?6GuFOqzP zL7Q4SY%GhSto>f|2$BrYO`UW=tQl}fUPJ+zVG}+cZHFHd13SrAl^XNrF%rOl#z@hS zXU3x6tTw?V;&W0uBS$#PwJGZ@*LUV~%1?X-r@y}b2|MiF|B&kaUsf5Fc)%^N@U=zl=E)4pRbczO_isQ{H8n;KMsYs8*mSnAiwmX6qbyH-pUSp|Db!>%lZTx zitzJ8Ysv$_VK&Yx*FZd~k_R)a_*2d&B>$r)0l)!eN7U08Yz< z-Mia&$61SEw=hUJu}k|puId+GlIQSApFoM9aswGi?A2dsF-+kWCh(I3)=v-{XiQT1OD3ooxJkXT%;8z?#ByTAdIAM~99*cZ))(zHRe zFV6=tc6>tVjloH(6;l%i9W4h(=YP36jvjp-{JD)jkRD(w-$S?jzmV{ZKbiy_QI!?y z71GSx0A%H=r)=6&yK>cd7U$z#%>FUAu>S=1kWn*j{&8(K?NQGtthhvO7Eb9Ki@o|W z{KTkSfI^SI#jvU-{fDP}pJJI`Jbd9QJJ1ivzoQjk-;Ush4k?%?Q)T9Yr@6gyf zMwy5E)YH5_#T2qD#w1-iPP_ck%BeKNdT_7>87%Oc@X^hy=aPX3V4-l$dCusb?VzFH zP=9B_TCU8{z{g4Prp^>UKdUzl)SkgFS15 z{X$HW%Z42A?CD*PBhWyVzS4>x+W@M?AaLhZ=uCmo!~Bwh*EWd-)TvEOOvbOF$^@t= zsy*jE&%b#iFP~reTTM6dl80%n>YvJsYzpKDrndEgWExn1Tr_aS$pU#!{TTbs=+4f@ z+--F9JYX2n_yCpI%Mzf?jl{GkZKMT0HD;wI#s7z|_YTMM|Nn=xWvk50q=d_ikQpID zWMvE4o3dviWHc_avqzE{l994AvS;=tdyo5h>ixZc_whaM41*aw&gVG&k6L^e{q&0poEKpSN#DGNpYoZCCbg;#PZotp)T!9@GJdxfWu^X!{VA`>$U$rf%;Ro0Ggzz$yN34SVq!jg?x*ROUnAlFty5HiJ12 zNgxZ>+}!$=$gW{m!oL_d;u)1)^C$a00}x^>XUB*y7ilb%Nhtuhdcp$&l3DoKY61d$ zjB>*P+2u4HKQkRym_P@&FXN*BD)oiSa14I316A}Vk#itwyf><%pr8;Fh~Co;u4w?m z6ro#Tsp$bc;ChI}A`G8lkmhhBvkz5ORTt6pCth-(MH6;70p<;Kw!l}E|F^%GW6e?- zxJivx?+jTZmAwY})S!8CZ^s{?F)Hs;5N{BFj6izFh4-zkE!(T3x+HW$OdyDi0!WwU zrbX02H$n`J?kx7UvUWV1Y9QX9ij*+@Q|-4?#Wcps6}oRe(U_yLylCMn>vI!VZXb3N zdYxTpLNpeeK+!$%dVDMQF== z04cOLTisxIGTmF7Kzf?e@NX;x4Gj$~pefQ&6u@+-3fi)H)$A~!)XI+piKTA+oWbB< zk%5=Vi^ssb4fdjSoq91i1P#Z9H;mio z&CY6`3^^`)W7$!&J!m`A^}jwnSb#$Np)aDrS^4ru0bo~p5(QYad3L3B58AqBnPUbi z2e^8oxFE2@Ypo8C#5ShpkqO+=?hVr5+t-m!M@5f;hf&z+$8zmeotG5|@8U<}i|CBz zg-g_5PEVUcSugFtVp~rM{0SQBjuwD=5#lVtIL9VIQ?s3F+>NHskc5ON>LA1EzpCZv z3ad2X<6_h1R6A*D>E0FiXFd67cK`7nRVc*|^qwd_&>=yBN)`{0wkp}@AP0#n0dQk_ z?dkrT?AWhPJ45v*xB_8wZ$Xpc`Uyd=6u9+3W$g#gw?21@E~r6vbOwYw8mKg{{f*p9 zdRF=MIZ{K|I`0uJdcZ71K~evb``477f)(B5;S!$}mtmEgYLfdbO`1Ek6dA(H?gJ&P zv0@&6b^fnVB~mFw$LqvKXAyy~TJ;OmpAWtu;BCx4APOH>O)iPaL3&eOodTed`!{A` zLwb;5$|TtVRAcT^cgvVD+cGr4w8#3DTRf4^0Hs+x0lJY@f0jGOr)ui;w*5miKhD`O+!~Xxey<5U<_&U)hHuy6$M2c+{n6kUl ziE}>32RS2;>*G2sMiI8@SD(eTL%CgRwTQGsVJISNRLt33 z0aqusZpg~^Q3K=t9fgK>y1N+-cwr#vplG>vT0J`S-(_lP@PUb$EtOdOj%BMzw@Y}@ z{YHdE+xj)7x&%&y-h9-|Skq=7b@(b?o8j5E(#|T}28~0BCG8@80jL3w8Ci$eZ0{*3 z3tGXZ&a7>9iJ#$8(Aieuc&}Cz92oJLmZ20J*9^V4-*F_x=~BV%p@Oteq7*#+`MUk1DpJY= zPx)kVe#Kog+fs1eBa?pjeCnHC~P-Ni9OHpky8ID%FD8{qBZDj@gd zp(LJ$jk5?kY2+uH4WbKh0!10Dz+!gAEx1kR-hgRVc3O=MSkLubdukqy)7S*_Rc zfvanLRFu(EA1|*eao4~q^N#mRP8>TEg?>S18}jmGjfDsa!#A^mpM%(1`_v*fzP}*s zwFP}9)b6HunI^~+b-!V>3cj04I2f|;#|tX``5iuN=zFZ8bW1Oy$^>=$2!4iq@Tz21 z_ri}C45DsKD&qU;;E&fEB_$z213IDzNC^-Qxt4%kXrelejTiEs|Bu)O?^`83Vo}(* zO?Y+xG`L!375~xUgvo!ctV{Rza3(hhfjg@=^eT#wh6I|$Uev0I$ZXaZBViPxUIM}; zb7s@S^$cx3DkCXo1w3{Gii$i5myNFomI~@lP`r+=yO?GP1as4j+N%jN?lQx5M_S!o zW8az)zq`F_9n-`*@xktN6nV#|i$^Zp@CK}g516KRp~kPm4llnZ?xP0$Ex7({MLN`m ze<+YRm}l8OW!K1+&roD4BvR!q>bl9+ z?xt*zBJiRAErv&$TmH}V@s~OH&}InjI8!p0lA!JKVj>>U=;d2 zpPAEkZu8Ft>a(5d%Eq1Aj~dkj5S*5$rNl)?a+`-qk`~=u?|t^K7W~`+<3@H-_qoJp zYWc6FlAM9O(oj==m?1CNjDDGhLe3EPs_s8f{$)UN9N)85J*3Jku^Cizn&Fe+!Byyp ze%8|b-x?gop4ZbO8cb3)UC1YK{-6Fx>|_MI5^Rb)3H^Tz_hbokRKr<#{(bOSdo+0u z_@}oJc4h!uTNdUTY?LjKTyqtG*$4zisR#t6U=HE4p0BFRH~J>#yjZNuqEAxShSKDdDhh}Gl?gnW+2JobD4jz z?K{E%ix}~f!4+3qoj*bPP4Gtv7I0s^Dg${oTu@@hKzd#iKr?0_qC@hL?m;dUr$Lo= zrg9>CwZ-;?qU>(`+t#I0TV!Fr-b#A?0f6{W0zNqBSl8Fr4@i8dxbz<6T!c;C#p+m_ zv2PQY(bjC$w5@K4w{%7VVrd=8T5nny=_)P)1*Dd)hnSgiwF4>9_}R4sayD)RCq4yx>OJ1m zUM)k7FR<;s`|(32zRwaRA?&#myk9Nsd|+_B;_@y-(;0H4IQQ!c;yXDLACqS+`MJa9 z_WO&yA);)_|CX8Lz4OTRw)5=$;N&S_k(@SXmClor#=yErk2M5bkz%^x=g*v*#ByS8 zYmyK?!`*v8M@!3oe6W4du=u2jg11f0{NM2xgY_g`4%r;E481cpUMhhOT4nT;&UW_2 zxNlsBKCzGeN=Ot@^%g$`_4+j7A|9fe0qK5-;SSG1!~7fU((|`hGH_epAKrG~NOv2e zeRE{?-|i%BWiR-#~bgtZW z3B5;~{^_IGMs|W8xM_O0COj=!kr&2?f`>%|v!v^|6Vpb!f`sCF2~&@%u`y||i>&z_ z>xD~xnvXq2cSoGrJ7`_scS$);e1E}?cL+y`mf1{_*G98?KT>w*qk29$P5(w!I4w;? z5@zf2#XJwL%ul|KY-WZ|?g~MDQtc)0G$BfH?$?r=C4c_rXoh0L0&Wll30jnV%W4dX-K;NSP=N%z)8b#REfL#gl=dD)pyh;CC@En$Z5 z?Ly3RNWm<3@mZ4)Z8DE$JuS%+B0H{EYNfFNeiecyUbmM8+9qT?csRK7X?*Tg zm55GPrkY*STzm6Z@CN~E?w>mCe6lBBJ$FA5jF4Z|4d`XW8DK$p5%XotHVQ5f}9UlWtdk^f1{~=3;*q)wP zO^%S5E=V7VCMVpbQL@()Bz?w#tnD!PeX9StIn6fL=Pzyg=y=~>jA>_yrjhR#D|J(ZfFPDW_`-t*9J>qNJv7x-yk0I|4}Z8mMNY6a41GGX2b;M z7eq_a^atWAr21Dy9Tv_u0#h2SCyXEzXv=6to>8OY%l%Gw@m-6p zD@|L$|4||sA|FTn)847(;KGGV7QAQ5gNU88+82|q{6{bz7cZ0_peeG{xKX1}FWJf$ z+<@F)Ct}<-ajU$0uV$#?&}e4cmjA>AM-&hFj;k>~e4jy8$ z`}XhsK0&@*r5lT8yTRA28pC`GlJkV;|L2i%z(ZjQXbD2^O_x{=#;EW-McH2Un` zwj0+jo0KJm3sxG&Y$INPkP+m9p6wChljS^0*hlbYnbjYDgTW)!jNEwKBRzk(V>{IShyPR= z(>~&(-;c)Yu-qZJU;TkD><~y?>ogG4V_us|6%drFzjOfkGa!QG0w|5Lv>GH%h%&jy zqXzaBIF`{Rn!XhzOaIy82qERr_9-+6itx6AY*=>G9vUZvImX0#c&Zcy zS^CV(%m;@SkJ^*OzCwDu=H%pLPYI;?Ai#dYkOB6D!`ir1FvFjT;de!V@}8Wma*N1+ z;-pC!I$KQ}-ITLH8{`@Zn>2(M*$nbguqBIo>p&(B3SK*_A@Ho(>QG^>0RP-nIo^r48Tc$!H3L};%VN&@zBSNIoI(3C4C8=y|9E+^cO{t7 zhYk9WJ{52t=XgPS#p3CkYf43y`iM~wN|D~ji9l)}yI@ZtJ^4+GPRY^TC$&`^i*aDE zsidwwIxEyW@V9!EeE=`lI+bhvoP+rEHK%~BZ346_!Kd13rX@$${Psl=ldw960VG%l zRu(_t-{MeR{Ag^<6E&IUioerXgNZ|sol67K3`Y?mLrCEus3q1d`ZhXI!1b@Ka(}LZ ze)rYq*qE!J^tFAqgCk@L9T!F1p@@J&#C_rg6(TQr>blBLHJ`Q^^bN_;m|KubJSG5= z0;kBl$n$Kof2*t$+f_*v^bU33a`S;lyDHoDThO26=zlFcsQI#d+f$n1>eYq6#b*e5 z^xU2)kfl>nHZ~uBB4G@Pd&|7wR-^&OrYBcJQDX$3l=RDfA-bjEk(ImhFUszNR#s=> z5(Z>bRJe|6gW6R6i+qK=f=ADWm3&@fxZV@bZ-=X2*|iaJ`lDWj8iZ2(j4z9G*~`=A zTL~A3ycn73zi@ZjlO<{v2ZCL~Ah13owg#BV3aF-X8zE%gF}^=+9Fl%d5O1Ucz)Yra z=37lm*efU~qgS#RaM~)a@W~hIMU=)Ni&$0Ud@LX#BnLg`A7&>K0biyifEJ(*!LHi^ z;7Uytx9%C(#Ln0&(8J`YV)C19_%v1`5}MI+nKTz7^<~7pc%@^lneIs~z87__M_m9; zE+xNC5V<^5q6K`OJ2n(-h@?w462Ya+I#Z;6r{B8>I>1+|XMIH8hNDKVdi0rqQ!WNF zLT^Bhf*ai9|DG(m1Gf#jzt@mdeD`2yg#$@8@ukIXMVJCmau1h$f<>BRM6PdVb)k~+_RDWa`#f7h!y2T|KoE+z$Iq&vN&3)tv!rA)>D^oN3MCYUmhf zuf8KN+x)NSmm`Yp&Mxj>ZJs%6mML3rFiVN*A=_V413u>~1y$dQt7oJ}NR>ZEae>Po zfTbBy)LYlNxY}8lLct;n$oW?rL@uWVn<^Il$(=x~46Ep6gHeVn&JQ-}q&vTUM{G?% zVeu0*{K9F-L{6%TcokB;$pR$2gJ0|K;$JOPg8p}D_v|K*>s?C!J57Ny8sO_Z<|*$k z#Gr#mLKg+TS;z(;Bzyx*n1}6GzzvU$FQIX5Z@c%-i*Lp>$Ne9DVM*84V0>D-LCy%` zoWAg$w}<66?~NPxh?VTwVR%9*wKaylyu5s_B^5yRuM-kDK_Ao`+4QmolmNMNT}I#b z>Of(Ln>TD!DK>!nQ35OtIOG2Qv)tvz-hFD4-|OoidW(ya`jk_jxYaF5i$HYD`&R~p zndU8T?@181jyn$62YseX>K`J#qSzX{%Fjtf*H(Z9Mxof((YUr^#&mmt1uBwl#S)?g zz`w9c*LksS8bXOOB>gej&v6)5SABd`++gt@BvkTaRqnlvm_HzsrUiORPP_iEf2%p5 zU3vpM<{~3fsm3ZbOX|o23Z*uRi~YjQMn>j2ffGRNED-pnu?sMX>H0QAk>0BzhTUAV zbxJdS2ijpBuVsX4gr&B)4~sto=nE&v{}r%@MQEiDX=?r4f3j6s1RzAWYaM9l%9B#8 z9Ytt;dzNHUP7xaltl(Ekw@S@Qyai30$eKasNzef3z*#Wco(@2y{DL>q*GYW+F)(t| z=`u*^8$YVu5H1T8bt6B=po#Su=7TOw2Uc)nOU}NeJUSSrQ;Z9VC{9wIh-d13$M^B3+efT<-G_K%u2d zd(z`!Ml@vUk9ENp^I^3B48aO-**;*J+0xo|T`~M(r{7C^3LP$Nb^6n#{RR(Fjc|Ot z1qpZdaR?(}8E0Bu^>-KS71&naAzdTmWzdAWG>bD|*|-iHORO!xsN%R%NyZVJYNY2- znQ?F|_O`qnCz7JmQO&ylK2^t0tdJOn$EeojS>`!_Az5$ZD9G~i-V%hq6UcWtt^-!2 zSwN4Gvk-}o-I=FVpowuSz+|K`*|;{m&jGs#v91@Sfq_xBPT1hK8awR9*4=!x={81} z1@Py)JnT?nFM#-aT$|9h#2wSN0?18=w1Q0Ca@naFk)qfy%+Cp^rTWd`Rx~@Eytm{h zel?O#z%)tfuAclQ$gvIV`_S>ZIHlq+LOc>(_5tz{=aQ{NoYHq;=*`B9XhuEInfA8i zOF4J}HZlmz3qb_r7oZ?P5e?UxTT78c169XO53Y1&YTUUdDcLZ0C=m5qV?9^OTwY}t zVD!^v(j#W#NkDva`B_BpujreCESEiBrqxi^3g=|ZeUAQJnk|VMpPH_0;D|*@rFqB6 z8Xef1=zB3&aw~_B_66S5b5>khUgrK@g1!Ymu3x(h18W9rWCh;8B!U3w1tDvmnk1@V z&BlxqP{g-@tpW)-V;95frRiK~r*(NQO@MXMX=SYvWz~)tKL2>{CL&?MT!z)0FhnN? z0aZV_rep~Eo*Zoe<&}pvw(gaKAv|M(m#fgg+I zQITnvigX5{s3CxtTlvHIsS*&}_Flv_m*6wEK~;9&V5ZXLcVmxR`DovEOK%e`8>%n5 z0Sw$5c}cw!DhU&KStVPYzN)s)>>OhkZ}QS_7ScuEz?RQ0wz*-^x5)F=dhG&sy#={| zlCLDaalPUpl8g*CCxYlVVSV>q6!<^hvJqiApKG2%bXn~JnUGaR@BNr!*UO!gp^_K; z`2!xLDNf9WE>=6DuYxh z7wMf|L`#P!Jn4~avWOFK{^kgF;zZ_SyV36ziskxFk} zDAPmO09`QXtFg3fzF_A)Z{7B{KmF}LS$zlt6d-17zNMzw>+%MfbsIJIU-}}|5v_zw z(obby%i2VtsR@7SK&)P#pYlj6dnSIk^4GXF0oO(hN{16X6^-a>C58Wb*?Dqs4KLA+M7F+Ja&RF_qYHEspKn(&k z9>X`uO8whZ`ZcmT)|#^8hSc)7ocLB?0USHX3@_w#xt}P~JKN|3Sk_`8dNqC%7^uAH z{gDL-R8Y%NV|8630lG}R&RFLbp#RNKyWg}Q(}V4u8LEXcP|l18k|Gi;q%6K+P*pyZ z45GHO@*~J-&49!UtBM|doQzB|oB#n`{X%i4~ch5J!s}K*i z$L=c1k_K}CjwB@XeFpT~NW!Ws#Rq`va>RE)#)F2Ess*U242H|}^!Wi3B=jQ4Nv*~i ziS}A1#fI1%OA6$-J^4LRpxm5*C`&wk44z z?D|a1%rl?YckJo`RMrS{vKlroPblpnwz*jL>*igYU2?a_!|U2d`V~$;CpSOb)zWSM zZQTn{bWYzOd|#29e>&lQWoQ>Z-uSJ^opT5~Ni?Eaf%DM#rdp0LB1Na|Edlm%(Ff*s z!7w?QZ|LL0-7M{sdQu7X{H`HqfX40;#eQq!@fC*17j~+aRnUSJ^@1ve38}hHLGs`UD7Sc>i8yDUpu%ayaXF1^c(7ja)dt~#@n|;Zh!C7%T{2@^GOnbLPfn1zr%V@ zS~}j`0#M*Z?=1j)5%jBf`vSQ;#3683y6^*D5 zd`_S{PA+YCws>R9R zJjgATPYoOdD`gL3=5V($d44>8kYbFF_HLC$SLy~Bx^w-0J0N0>+!2!!YOFGx*6qK4 z{; zPgjR zeu9GWVM(YKY&1We@W(+xZ*h};fjX#jBXpchgSqpH4Zc6W!JBUwn*arwfzg$HmrBA1 zMuMe_bXn#QH+5G-0)v7Q?{&D)alBAVJh3AO$yI^bN1J`R_Bo7>pqF|Bg0nov3wx?+ zNFB--$w)eXHIz|EnULm4rPzPv;V5e;FfDBRNRLr<6BgzS{o zF2yhZa!Nwh-o{AY@cusIpy*pyS7-OAPovYZz2pwghOa&jsx)XxPMJ%G3$|*@enBSs zbHb$%c6uwW8XA*Jt<2+hxACN*)w)?3Noj-~7v$bJ#HgcQ`q~EnvF$#Fe-Dkz0~ujK zl(wvO{={bLCUeroL@cEVVYPVs^J;=o${%nF{q#(a*epP^;6ux97pqAi)k36GN6Z{^ zIq@8C#~NPr1Slt`daA&L3mI0aH-k}a21Qdn8eIzvqd@QNhan@3sw}5hQ-5RrCg@ym zdq*Y{0ksauetRp)Kkh8`cdXjhituO?;LgKK|6=i6c7_l29Xl5Q5PO_pW7P6V!9SPM;>F$0h&ibT zFE8O@uf>6xMERR$l@IoeMYC)!mQEJ0Sn2+33JdGvjWiKFrysnJ8AawkbYID}rN-?l zrww>ZjxxiE^xmI|(I9k!7i|W1!dz$89r_(&@@VdjebC*zvMD{Mq&s4~ZmakG2UQMC zAPN{NvUv5Ks?`K0+@X_BZ{Tq@uWZaQIo(qCll-MG4QECoQECt)~_*e(qzT66UW4 z`)LH$lD8@W?FY76l~Ot<2VQ;aW1RF0DSX%gmg+&0tzu58syoVgBL20PQH}WDyY7TZ z_NiqMrSmWi&;bf)^}gYt^b#?#?D#MYwpkq6MSA7&%l?IwL|KA==&`O3$?83WeuNYDCtaXuCifeS*F%#>|R}?rfT|)T0 zFIBs&y$I%en3?FfmhnU-=SoRb$~IzplHK>&Q*-HkP^Dw!|NDW8PqEt5#OloCFXl3M z%{1%tioqejNsinEmF)Ky^9s}n1#}*Cy#}-5NvpYTY}W*?+3QYTR#5sZUO?%*0~$R$ zz;JqynVYMWVU!lT5Ck`0tjO;Wl6W+jn4+GQgC*mz-Jt&ucl7OmIzCm}AZH96^RDlq znrcL2{(R2!wGB{=Hk|?p7p|Cuk9Isy6(d>7r`WV>4MAqKM(1M49EH0n<-baPenp!_ zt&|7*s2yQ!S5ZRjQMPf7ARz1+mOAF*J&m%Z`0CS=EqJo00JELW0Zt8-O)tu6nJ8HI zt};E>n^foXEtdyrdC~$yHF29avCwc*sOkh)S%C0H^aOd? z+J5LYR<_pITy4BgGdkyY#&lc8Lr_{;C=M9eEG#S^Qj^s>JW{#yQbZBeL1q6V)EHAX z@%29;ib5;kzQ9bESKiGxbvf4t^J2S>Ad*|)El0YbQ*73#iu`Z9HpIZYQ1$3r)!k|C z3-+Vvy;$qlUZ?)pLBaPdKG9uB=q(eOM!u&x27iz(VwwU4s0_H3lTs-c0a}@XsNh^p zE3n;NMcw23SfeMJg^kPS`<^aKQ5k*v_HJvbuVzc)if2^d*8d(r;TKr{^YDV#s;cq@ z>7lIB4smc`a3vGPegOX;5+&0r$LM_(eq)bTE20BH=ExnG5g}y;noFM@biim}EJDA4 zI9+&Z?1PEZYW&L!zRhoy{5a9ed)}_h%F;Y_q2DIAsU#0}BJF{yk)T zFxY;64Fe#wJoZiZm@!dL_4YTYcMw~9@zNdPCMdLXOE`|!emd-eC|bq!G8`tXNY8r& z%`c+KiTPFr#^q)>$gctx`3hSH#N6?WVH;?nT}hxl5L4(4w3J~O0CLR`IB*q$RMcRK zUGv-Kokd{~!do$#)UsOh%(p;cp01KCeqi!us0OEI0=G2IcSC`i^prMZB|s>}|?16E#O zRmipkkg-qyn@8_FFQS{;j04;7V{Z`3R6-}Pc*?H%F&n+olk>)B2yonlp;o5t5MfiEXy=SAuvNLeD^ry#&oE7f3}a z5eRT!`x#89umgRs=9eib@D>vsX4`1=;Yqu~k^8K9Nf$(YU{b%+pR2*Keef9tLh!PM z!h-69TrS|nDv`9vevtK-CIL!DjS5`hg!t!A0OVGJ=ynOxn^RjseYgr`YN7k(!c|BN z8p-Sr*c<}9$VTTc;JiH>Ae$1j598R0M@4!=$M^{-6)QuSC#-Bc2ks4*lOJ8)7E9HL z>WB-BXO&Qe-7fJ43$oSdo~U)Q?a?b5{)P`7-(kLnE6BMQJ#WD)Qwa(%nqP+S5_|#> zv$DKgNZ%uHa~P^px&zLGTqug?t4|KB^>rqSG5C=lRSRC=ZGV9y4p{hh(4O8vf1%`= zGJRah&|L!-+Hy`Ku7gE7eLb%O_o+8aBVuknt44!777bJSGoTZyge9;5PNr-`+UFw@ zFb;3-A~HPDyOxfU^2yc!NTD8je}YTQTD=a97`K|25Bc-U;7&>;`2LV^b?pTKa0)8x zW=kQ!;4Og`Jr@~uzST|knQ1@*v=zOLlZ|(vF|h+3-hm}xK>1vt1h0fgw3IaweBy_J zwIy|KhdzJ=o6mw;{uitb7m$9J-OwI^@I#$vK%BKFVifZq{_1O-1jph|7j%I73_Md} zZtIhG6Tb)*ldcTjLYj+DKy)eNv_Wy}&}=>d?b?Gm0&T=r>F+nr2#+)i{wCHAJYJPs z7N9!!*-z_?!F-aeY$elsUJug%!ak58a8|L)ASr8w{vsXxw>Qv@B{1jXy;5WC%)Pea zon3&>VRXsG*RN;ZpI@4JVpG-HcW;5koW_zGM&L9uFeb09tv5Gw4JIKrhtbKRkH&Xt z^{RbgpsU8vTUOX&G`}G1(idc!3zEhLEf_o4uy}PSV^Y_JzhFm7;f*5>E@;w`8SBwd zHs+k#kCb;!^n(wz$H{l<)vc}WB;OPFZJjKo_$S+EXKm%%0mz;?F2TUU%`PcK>^Bnm z(8wdH^ydi)yZ_!!uwJXup$!v|mqM@z)j4y$g9d9k4Dsl%5rAVKvJj z((Nrff!!D(72?&feNOhx5wKKuCq6%t7!+lo*NUgFY9WUdup+xr5%;D1FFl8($EGE_ zPpm-RmgzLtUtn0E@}l;nF+UAIE`(CfNbcoj481IYYRE?24>=Xx2Fe1eI!n7O7LYc+|CLQIwie8jhP|M}xd?!@-+^?Hcf^ME#IytezTao_q6+Y+f8x zjEmd?O@u#^Gq$~;7M;8YG2@D-$$~8j%u2#;=LfOrk2q&Jhw&q8zFODF^n9G`@o<o>tVthqsWjn3n)pf;< z_sFR%9p2^x+wWBL8->KqsA#-US6xDCjE>+rL9q@d?VmeVZ8LKdt&}AzpRqCUM8u~Q z8*chZ*FI-t-g_*~U<5IE)3CS!{Z1@I&3F47eT5ySd2 zESWH9-!9r`fG=N2x`q>}XNXPaX6vTHQb2Y1lUCjId(mF1oboEHe~pxW;BL0PH@*{c zXU~Wpj};>UN9tpun4e*s`S1tV@#7$)i^+e$e3H&KXSHOGG3>v~HclzQ0I)O)h`2}S zX7Z)`OLen)!fh&cQY_#EONh}9Mk5@v33x5Tm+UmrIbhBqJ93nnW8V&aQ1#`!IP8@b+*j4W=6K@w&6Gj;|W zD{%f=ORjl<$8mcKM~XWMPoiP1N=Vs_i`amp?-0%?Mc|fghQyb@!j_U}`@OIncK0ga zY^*N-zJTP`(+AA{7icmzc4-46J7|JH>-5orOKtcYT9KJYi{<77`Fs}9=e#OXHA)?) z8Wvj#JNKLVT;Ip>7zdo33N46l_dMKykh}Exlyq+l?S};hV&0treHXFgGCM7KODTo1 zmG@rP#)OxNmrNT=82xz^K%%q9C)JqQ)AD$-^`gWjvh>)%aZ(IF-28h%Fk6^o`>t*I zDKaw@%ABPe)hQrUe6$ZV+jgMz`>8e`=h?!t?}wBUw0)oS^Hy2*q&fqb29w%Nk*%7* zhpY{ZRqCaw%uR556{u!8tJLf+zt~-bJ4Z(lOWrfNNNR9P$C#Yw6{Ca!j^vIHC7Jv; zF!QlXO6T#k6Yue!KrUjgAITw>7Dng6V2S0!JbHh}Uc$&aPX>vNbErNYzX(lz&a~sE zu1|eTA<`UW6QnMp2Dg*P@z_OfT1j2~M8y2Lem$AbOd-@e#OM+#b3mmWGl-FG?|!P0 zfBu>oC&UezGbUrLc;zFy7DQOE*dp$; zJifnHwYO|BhgH(YlJXab%3s65CWSydy53b9{)^-twE8mL4g(T?t$(MRD7HCMq&3$i zfpBE7OfJ9Uy~wX2vwQWehkYq`@uaYHE;{5~&i7=I4Sx_5S-L$~$t7oTgLbF3EH014 zpBrywTQ>ll5VLoAg1Eh^3`$-z8}WSO%#&IRqg2Fp*?X)^T)58tSr^>7acR?t#JgOYuc^|J=R^XB$FAs$fFzBXOG>Z!$SZYyxO5J&^?R6p=rt9 zRUHmq(HETlp5}ksi}pL65=eW&pruCEWPRi507D6nwHA%dpg0}jk+#6d96r2OS*!ru z`$4#mo9D$QJfr{bxVJ3KCtl5yMeNF%Ws-S;#+`oLbS2m<&q(s02w7Ukn2gp%w@zBr zi8xL^AJiQ0X&L!IfgfWV!_kNA`&2r9{t)DE6Yo~@wt*=AjQYrE>6*o_8iO=+g zo6B^Okj|@4TEP0v-v+)!zX!Ttpd_+WwRKG(JVMaL&?@MTp5hx3Gndxg?A&Sa>|k&^)vi=9=C(cdp1+@2=?%d4-sddvyLmb^K8x3PT-zy~IB=}yd-k)ok0ed# z`ef+ag+oGAdEbF|`p%%-7%#ex^zzDPH8K;Lrfy#$Q1c1)kWkXz2-7(^GTV-wiEri} zoDSQKTg5Z3V_e)o;bOR zPk;*!s*|*rkZg9PHL&7B=1|UT+Vgh{9uFEW4o9f@yVgNPeG}_)VlLwOsM69@K;p(vb5(evDM|v=z?yAIvTWpamlM7*B)aasTTJ^ zZMf2DE)}Te%Ep_MNLU8`@s>`@L)r{#Cmf3(fR)QdK;jeWhlCr+l{oo4e;Q^Q~I#mu450?{QPn|0D>pySS zd-M$z%B|Q)Q?sN()ttMW#B6iD#YgA$ka-PE!l?wN>@sSz=b=iD83am&Tg^?;Mez9x zJ56^Y*y2-r#vO8oFx$@Y&EPo&MYEZK@5^RL__YV?!>7wkd=zmU!L+MGtd=ww7#Ci5 z@3-og(}lg3u`rXgrsFc7a^ldr*cmTUsw~rF`(tJpi3onx`I5wQUONblH_e}V_dKe7 zzQXoX?;xEhLyXnB_H@e!*A~_JrspIaO|3sN?v~J{rcvnx66>mV1_bnJP48K0aChDb zXgw%zi)gK4Ykl4)S90aDCV6d!stq?g$2ecnM22F?I25fTRvfeu7&`toIhRZcP|_Cn z?oyDK1bq0(NKYJXZAEZMeuHQr%VI(%DJ$bGjyY1q<5V?v61=!cDcN|XiedSpbz0zH z+!f=2ik1XqP`lW5ot)?|FHM3ig&Hh`@1C%4+@hao6E89U!$iukK$02rBbp^V!ZRpi z{QksdsOH^i%(Qo8sCO0=pxs1s*mS@n@3XABnzs^&Xz?TE3lqr~O?LIq9^ zorS>96R9DpB&z&9`OVIaTt^&?{DJJ({_2=-2l7YNitA=--0i#)=KaDutmFFcP=`;MoEDwT7yUUfZR)qKU{|>ji4aeo_O+Z$ zh(##S4VceQzdR03(9eG=im7%KE4j}{cO82HK<|6MhUV~dL}P;(Zw7tN*P#)R9$?%n zeAz)YcD-p`U^zQvMP`NsJ3Ct?e3uyWes?uC{!Ly}p|bNiZ1RbnYWh*K&&MwoU>s7l zro@F92g^Oo@!lysjShX$AI290b*YxbgS8z4)n2Dw6*?w1i6@><>HD`wU0!`&N?}CJrJ?pGz z3WX}cBKn#Uzi~^(CKPcbE_I4Xn{r_D%$L7l%%1nlld`^|(?rhpL7Hih6t#kXWqn_! z^Iml%s*M3Hp({%>T3p7?(rEO4nL;(~_48ThPfu z$kE8-eGw5wgnr3jOjwL(S@CuV-&`i5{MmeLg2%Jx|7#ZK=F5-~?EMk>qbJ++g{iFJc6kB3+Sk-Nrq8{E)myv|IzIAibXwo~Kx?VrN3TBxb z372r5rD+y1{u|9XZB;pKS^-U~-m!TGY;oDsdXj#*LD8G*Xm)d9zZe`4+|zyk6fYxA zB*l}fKmPv?$RuoumFj);RAWI1(Y=PlkcMPQhqiObP?l|&AqBr$#uNvJ5+Hv#U{nL5 z5_tqefe>9t69G(NK>ld{^HyIe3V<1x@F$vrZlNO({sjWDg8X5kIg1tkpMn-5g#e|b z)8K)S!%MtTNal74lBRlNCIqVL z6>(I>nMoD}e>+jQZ75oH^0KklS7UAC8Oj-nFoBqRqqA(u|Jyb98J}rNiUMGu>Hq)5 zpT(j4-#&}n#`apqD#%WXp8YkrcmMudAX$;YEx&sXns^~!rpS3txx!&00H%H*FfJ>U zJS6!Xa4!%qBxqAK8%sg`p!VSo&UvEgzZ&)-FJDH(0#MY`OE~ZXUb%7EzdpUS)WSC?^5( z`O(mz@bUsHB8ykfC!L!F)k_5!n=WzbX)DBVp&%oTYuKRZFMrryxuY5oh$}PV{ZJ7q z1NcsJ?p%zIf$MY>!hRq*5&GcGB>Ad(XbrMX@Y=R4EwMDBKq00D@|d)#D2Uk^g+LSA zeh-k3Wg~h*E&A^YQjp>Umxwxrkbr%P|{d5Oy#l=xG7?F_bEi;Bg^FB!(2x>dr$T^XP_1(*TG=UBo=LIsr&* zofk^D{Y-f-8w9mWEdcP8RKqZWQ5e#x_f(ID=XN&4Z7%v@b1re9|lsui%TmU%0A`ss%0Hc<+sS}o4-*IGF50~X5 z#hWa~7I;A1vPESGE!NXY06(Ziq6!e$HFKsT+3?x^hFNg`H*m+7f)ix}Y)+rO4|j4H zO9oEy9oFCHKPt{V;oku%(BQXrEd%$vr*2iZ)(nu?H7-c%utLD?kqw@%e0YNqH_@d)TY7b*SyjeNrjqm;T z!Ax@C><)9LdL04u{B%-2xw%Y06idH`1Ox>}fn8=4^d?)~h~AGwzp^B57vvQlw%*g_ z03T%^zcQ=|%9lqvrZ@%$hEHE%XsFT>f;q1O)TFp>f@fHUaJVOgV_lRyMFn@It(j%4 z(`Nw+JYbXopP-XP%;e$&fZ8<3PCzLHz9}ATaUmZ>mslB^0`lyjYH|<*Tof#@2W9^O z%tdFBoreU|txd4nsJCvyv8s0>!e$os;pq`@Pr1oHOMNJP_~5NMOI>E_1`~uHfRz2X zt|}y_&ez9xZ5)I2Mf@+ABybCiFZ$rpuPitocfD!U5NRq}jTE2|UA&mp0YV>vsBuz> zkI;AxV(5GhR20fYb3RJ7fA?8{%xN2T0eH=|4rx(`vpQSkWX3083QrvSxyPICedBg_ z(Y-QcXva`LtTq8p4UEmsI}=SYtO3)XKl8ln;|0|Y>=m!uP*lnW@EL0 zJf^Bh+^G#T>OE%(3r{c2k5m+QIkijTrKY4XZ3nlXpEelc<=>?N)os!=ytE*a$q)%2 zFo9Kvj2GKqt0&dCuD`*VaJy?Q&?FYxv@D5YxGgdD!K8Voh);xTOvkC$oRZ$E>(MUo zkwvacn(=*F#ZQ!6fyUUmv|s{bn32oNQbNYB?z&v=aDL`234 zS|)T<6j^U^pE^Vcy@f#+$ER?^84MYPm4|E49!A#P`i^wqJ}2E~wudEX+VZw}VLRFk zB*j`9Be z&}(ym1R{Fk1Z{KMzaiRDFHo)Vxh|@25B_wgCVmLyh~oqYlkef!lSjh0N0BVkNv-UR zA|ng2n|gh3VBfSn)EfS?DnbHLI%l(IN;$!&Ka}IDy4dMN^x?w?%GPjUCmT5!>H>_U zDvEBePOg%f@cF zyo{%udX99RzY!Jh_CzAAwUb4RipbmICgnDf+@s9WLGpdEpJ;30)k=ra z=ZnE)I*ieoFFcm_3Ws6jaE#tqMo(lOM)Oc8fB&=yqT4AacI{1R{ zO_bBegQRo8McI(~ajngD)SoR-FoBai6lXx~=AYhwmy;ky(j;XMmTOvVV>t10QY0u?a`)`oQV&DllmPcYqv637FtUKI%RfZ3-RASA4My_7ZQx3=ccYu-dd z6ua`gUCP2$(k}k#e*4=#;R2A>An({GU}+aA^W{H@lUUSQKvq`EqhyF+sAj>eU=(pk zbtUuX(jiF{D=X*oL$m4K*h4zrNX8XX7EQ~_UbOUQLAw#5<|yXCuETjm zigx-3^yMBB(Vj-9CfMTeNFiV+Zx}o&>E&L3vQ!vdDKOIfxF9C3Isdxv{+J-_M9TPf zyvrzy`f*xhgU_|_H{>rPdbPB;9X(ZI(1YkI2NiEM`3A!7JOE?VV-D8{-MUimZv42Y zvc}^c`?RgiyMgCOSdFvuz)!ECC#9FXQz{#1eOX^1KD0JFU%xoqbGo+c`(&M_`}d86 z{kF$!`YTm7pZK|a+$=561zBQESDRR`@`qJv5WOQt5(ixMg_2pQ1x150OBO78nErSs z+l*B`5)iFi^c?2QBJ9qP2D4X2?vfVNLi)fGc2&WX&vU2NXd_!BY>yLA+y$>prK$I{ z2X(g0a~C2m;VOq;zvyE0H%Z!rc$V^5)&_d&@Z`8ba@=#HP;LBSEFy`t4b#Jm4bX;&Ui^}e@{207K) zqG&LkQc5L5WTu0t%p57MC}pOw&24ldrxZ=6b_z*kXp4j>Lv`3rhOO9UI>-<egghW!`>IZ7{3w#<KfP{-sL5;jTyx|1UYa8VyF0 zt`0=$3=p4I`#8c@UDRzt&U$q6QjQ zvVpsNr$GO7R{E6V|4KgmQsH}^RqJ**OiY~pTafkk_YBAHvduPaK1_m{GeB^T`eW;TKf@l2JJx98>oz zZ>FE_`{_bGU8Lp=Sh&)S1win)IpsfFsjPUgSx_?jqzyTc6vjlBI;}Xt=PpyJwD5L? zkZxw&GhsP(++`BW3R0OoVK7-$P|vuP5US+U2S<2-qv@#wC_dhSzRoAsRBKkx?-3U|O)H@|sFn}n7`9?BWjwi+L}(eT_`=r@|AzF&I=-yV zn*rd?mZ1y*jV_uaA6sB7lvU{xfKi<#6Ku7TO|<+L@M z6kQ%@<)*zJUCHx%oZYkU?b;UYKX-2YPy4lh?^plT;EBE!{y!SqN(v#qA$_Y7UWuH} zGbCvsmW#sHqPnd#tn|5ilL;x*ZlqmMTBitTNS%sC zDMQri4bH(39}=bAh>CVOzNv!6a~E@V+)X;YQnN>gXm@an6WLk9!*VYe)$Ba#g}w~R zU~{UPU7UNmF^A?2zD{08r`J`gkg4d}(8hDRZp_bx{RuZ=5}01LadhptDR~wef8$QN zgR(BD@2;~&wR-T^v?30n(06Ye?UCOcVsrOPizZkgPi*ootJj*s&}@rOE(!TUsh)!C z6V!R65Fr8RsTZdG)sENo$LQBIVN06gh4Q9atpD)_Im@ERD#nY5BN$f;rTwchK8@1@ zAqDqN&cb|?I5nn#u_*{Y8<1^opd+&NFt1aqA|^J(CsXYJ+y)JCF7#gD*Q&GcBrGru zjP!3nN&FXGK6zz^|7fSDAC257BuS^E$>*f3eTm2O1@zYJ(|Q_DRaXh`Hk-ibr~Sgv z`3)|jk^9P!dJ(H#|1>tG#Z1*XgO~!5u(|t&>w{j}Ww!oxU$Qcbex9h_K=v3tbU5 z)~`~L%A}EWHtyvnGgP|V;IqRyG2aK+0>Z+ndnBianoMEO+uPt=);JxRXA%|@>YvRb z-i+zLrr#)vWp0)u-lFjaWmBHnz4W7wRU!59iAo`?wfR0xLCh9bJD6>c&|^xAJ2H|V zFy=IY}sEwRWKLuFg_cP_RI=u9Q18gifMmQV2jo0A!imaWF~E>xsFf zzXXv&)JR~cTT}{1s_6Np^M`YRmDpiL%j{{x{MZgyhcWYYr)LD(fg#{0r@qXh&~<%I zG030BP167p*qKlgQ1_gIot`#5-w&Rhkf4C~r;1rx*LgBp72C+>du~^_N;-M;ayLqmY+V5P` zIics)PV4mU+pCQ0n}8g4u{q9(t-<#dk!@y1vJadl7Po!GG4es|{_5y;(hp5M0Mj($ z(}MvZFm;e}`jQu{l#8&q4wP!7A_v4+<|8O;9azTfJ4(M@L(6>Jn0W{_a))$@i<~T> zk`WO?{|jIUo{eQ9VB(I+i`-5fe>y-*4g(_l`s zm$TN=^kUiF%d2Zc{S};RpNPOZx?5%nS7&F&c#lmlu$YLHTCWqfv!Sl_FEikWu%sxZ$8PH%Kh{=UPx3 z8xBDa<5Se!-2Ceu&<-w#2KjEU#yB%=7X81%{DBF`T#k8oKS61>M_T$Xmrf$hdBqeD zsUDt#ZF|B=+RQj}{ic$cKX75xcu!ozf4&PC>9 zaAx*-79_X*ZB*ugoox4gC#Qe;Gl+Se)lMAc6;Xo3zE;fxja6E_XFw}Y$zc(lHQR@z zG_rdumF1(bC|8T~RPNO?@pkhcZcdUFTPJ$|8Fdz3qzxP_zZ$I)>(sm{v6ZrwcOD&sPq_dVUH#O zliP_ToZSAnD^n8;tMY(pJ<sXHqFFWszEIuim z)99+5h>=WG<;dH_6mA}t@MXKS(-w6oN>JX`CowHsE3hHsaX+WoQWtZ|$tLtn(uqW# zV(LBhLk}nZ@E25yFfj;^F}U4V&SXBON6B6U^OD{{;Y_}mIK@~lJ@cF1&*9n&-@aLA zeyzIO=5L*Tgb_zN6!2ts;GtRBw>C3Dg1CX+OT;ruvMrNS+Ax)?S+|jo<(aiNU1m+h zv7O3MCjLZzOjt_npo*y;r8wK#_pI>b_Af23sQ`rhoJ87Zzn?Vh{vQJp14L)5U773; zj@8Bi)3b(cQq+WP-`k$m5_IgYRzdBjCKDhS+Wo<}UcZNAjSMM-R;&3(O&9$h7`xYg03jWtyn zua1jMp8;+an=xV_1TDxx&wqknhT ztxdWcrO%mny>PQ{#HLyzJ_BXF#dE%gN-OI*Rw;*t{Gxw&$A0ctGTCVR0my{a$l;fV z&r`2J9jMOTc#uQ8ep_3;v-M)b_Z8FPIT5tef(~{_+Fs!>tbJXapxC#B(~0V zYM-0QF2c*rs2P%y=;my7+hV%+$(nFPx2%_KM5TKu0JRzW!Xi{&@`D}`;te}$QYJ$` z0#YA5MnhpmYuQB&Cf)=sK1O}#AnzFT2IcyoJj>$oM#hA(3xX-iufF4H-baYHYec-`kS<;ELqj8u$uX9g;TE1}x_V z*TIbOaLxGR!yWp!dUW}x(O3_mO@16)lKkc^pw!IKnx}pmzNbQd3K@1U9?e@RWwSy) zM4g+Y_FG=<{&c^&+^}u2xmq;bm7I@#PbcE>u;-Z3wv@e zxCsZg!+AX`Hxeh(!N=Mmc#LyTRDBY%E7tal8EE+qJb=A_Qfk0Zqj3Ns*KN=c9+I2h zU@-0KHcH#i)cuuHwl67p%Pb6}ZtY3>bOg}xa7mer3~@-_3>Jj^*}6kQavt(cQM>Tw%-P%Uurou96M9 zGH{u4oBTy!k$8v;>-g3Q{t6196>~j9g0B@wk4w%IQJZo=_?qnC2w{0&drI2*Ii!iE zeK1b@Bbpo4Yrc}-FZI2JD}PA7KyT*RD|MG6WwnNh>s5Mx#dx2}Gx9a9Sm%4cG@-%B zi+8!2r|iJ{+-u9v`=gV0k$gsGh-I4uSfFQmiW?v}&gH>j`~77vak38c#{&^-;Wb#| z6!P)j$jlSf;DqcyE%B1IeK#!pc;Tu`tSg|ljHat|9f8$w&$!<$KQ&@$8&i*dBP-Q0 z8=V8s471I3v)(PW)j`oGoW(fTbUO1evamY2=ItZNkMv5hz=S~h^pz9c*X^$+y{P$P zin?v}W4lKO1u&^IqpdWmYQG-l@EDH7{DII!SjB$=wl!wJe|^;gAOs=}{tJNbuU3Lo zg?24%x%Z<~K&4B%6Ic&C5CP*_2eoi0vZvPl0mO zf?2Sd>4PgL0rBrXlE41|Wi_RDq_G#3aHiU4zV^Su(8!DneEuAL?RU|e3AGr@n{T1y zRmO^`OW_{0d|C<;KZy+;GGqx9F8QA!=zlR4{eK6G|K6znufgikDeWt6^S{lz)1-dD O-=PC0dRaOSzx@-&`J(Iq literal 44200 zcmb?@byQZ}x2{Mj5`qX40+P~QuY^cbY5v$L=>QJ6bB+dIKn zS#AFN3>G^_Ggikb zJJOUKCwpzTGM1KwYDKT*9asA{sWyPg+Vzr*Ad8iaXDPTPWYN{^jiINy zd9Io(7xnowh41ZkUA673b~6(8g3H!;!L3G_zMkW9cjWHH?fF{TaH-yleb+j>`SDw} z$yV>CQSkqw^YgP_@i!)7i)3y zdtX#13wn;{Y)O{ZIIhUp4r4<1T)uq57Tx)?Fm!w+IQZTdm`h=`&GficEiFt-uhmP7 zMYoflNz{JM(ec>!}>1tt25rC^OlTR1E3c$d_!&^KV{J>b(P=g&7eq}Q*R7EtTJob2+ zfc}jpnh+z6OiX#h`#k2dCePE-?U@GE>o)0VT7tTbsT!;PRY&b94Rh+3GQ_ENGxZLE zc8{8cS33^DLe^~j`VLbA?=^esx&mJqNO`g86sGH=J0>=t@aK=c1FEDD@BaS2S5t?Hwx5fkF4{dfa$2HOcXVhoT`SM) zbLFo0?k;MmzCoq2^lUTyaJaeF^K{o@4)G_@`;C;&t?=!Ys<6%e;&)4JbMyS&-dIM0 zI(Ze9C>DcOFUvXBwkwYQL$|%Tv6kDLON((qur)WU9QPN+i>8HIY~FhxcH#)H3!6BKDv7T$z(r9wE2@+`r_Dy(`>-GXRF-1W^JFpMl1<^x^s&C0_^woPmdXM)8d$Qcp|K$A2Ytr-+1=y zS>ppj7EN8#SFik8+;(Ry=j__9UzN2-=)O2)3CM$z!!u=jk zU%2+=tofP=PoJbBW?WKFf&W2{sX{tDSV4?!F{&443-`E-wuc^fz&->8ZOOzu&rnb? z>5Ul*2ng7~*}ngXHo<$R(PhgyGBWbwoIx!=-QR2{q|9OI$Hhv*10hmxyUs_ZhpYX^ z1}fZ^qd&)rwMN#D=ei=>Q=Dk)y1Blz znW8J5s$CmQ(@4Xy5h~`q>uarulCDj$_-Uen!%*T*u;*3SSdoUtWwZC?3E3oVK|w*| z`noyy^@yTy&6!(&BDZXRG{2MSY|0Ba=UV%d=5nVsh3j+tG}CP3Eq^#TsGIi2F<&%-PxN5Uc40e zGVX+^*ZN2tU0`zud_?LX{fK~ zJMR4YXQH~{cjwS4kARV-0cBu`xOj$wdE*kfNUU{u+ep=5a+8x zno!EBo%q(-uZp)j+T@aTxZ&u`_TpRR3#SD7w>@-e>!&-PpNSUlUbcE)7QR)idfSEV zy(LJj@uwx&+n<~?P?+H%%G=aI;oF&iCQ_SfAkBw~P7&!4JulH?Cl%;vx98VbV zHp!XYH6?ZCGvAk;efo3Jba!}Rv_SRPz&MwVI&8esnEuTEs0B+gp+3}X>TU+=&J@jW z4So3cl@1bZgKCq6%gZ5#oEnUbn%Mo^= z>pZUv!zK)0Pz3Z$of6-5=p-)-T*sz4pTr(x08O{&lo7!NV_2h8Q+HELCLVV-C zz#bzXm#d#BI;L7Zn;R#u6beRi)_Zaodngg-kCm;%`P&A-F@4&VR-L*gyL|lO0{A}49~NyETf{n~e9 zIA_sveKhNO7I;;;!z+yPgMGaX2qKT(XHUN zdEG8j=UmY)YCfmcP@R*ICjNsuW`kDUS>TP9R2rQvsu13X{k$cudyfTEZPGSXF$ z$3;m#Fi;IX7!)`RE&O>freV+ntZh^vpV}YeY3*wt_LX&>;qb>JGPHq8Mp;skl!e2X zg+2n;MafZ3=7TfOnMbm57 z^1-FpC_T5ef_~K_?&$u_@@#gk8wv<_k}}~$KUB7 z3B*E=NvTBf&C3}qu1cP)eIAdBih*CgCHT}kh0y}(U7hu#cJ&6wmEHo1l?emiSO~tz zciC^>n{tTF1VAGyaxDNv9Yn?3uH|XFHg&4ci)iNN7PE~t7{AFqx7nM?>=O)Jk0zwkL^qM)M}8O&+RON!QWq4SAY z4K78cx=K$Hko%lvH1gjqnLEpTwbKopsPyUX?8xYa*X2WBvw?ohkLlv$#q3gwuC z8I^3Hx$c?Op50lppi72LKpZk=Q5Y*vHK~O`sJYfxX)+hF8jHYc)_bE1VucDGOk;Mg z=o&kOcG=fY!+lHz8*KK8U+U?5@?A8BE`GGh?4a>vBC#OSZZFO;KLVz^+BjYcWBuK9 zw*;Y?7w~|v(9jbR$Gmxo#Jo^%BuFQpXnfen*BEBa-$ufs;h7e?+H>gZ@v7xaAWa#d zbNPpuRtC)*HEK-S)l#!};%gZ+N|T@IXd%A5LIgAr;FNRWJ!AR4phBoE##C{V%(-WC zLAKiJg*i`7$HFZ9b1c>O27dfKSXaa=av?;tY5P+fLidIY&A7c*)ZS0FcJ7<{<#&<= z&JYn~`>`%Ot3kNqG-dZ!zxz`;jyn7ou+G=Q$%wuhJRM8*qZ6CVjgh>F=F>UPO1C|I zmvhoIfWOW}+k}X`_1Of3DdP1JdBR?2<)Hw4q0XGWd$;#UM>U9aRsy{B>$Q9;$#BhU zhmYgFHa?2>%U5a2X*+N9y;29CZy!nWbg*j6rXbfOofHwa3DsDfRI|L0eL=6;s0+Za zl|#ZCW?^_w6b*sSG8;<$i@E8C+IqvD*iZE_bza){=v{twtmUq-nkX%Nf1f2U!>Om% zX8NeXJnpU^h!^X15lz~hPge@B-}J^#&nRmvP@O!Ih~-3Q_MEG?(W$jA5T4sU=Wsi6 z38Z3r<1FA*9fM>rO`8Y|dh5`3Xb7hPaqq6xXueS`a$VElgE+&ZULsD@bxFaQMwg=g zG8bnOiw4B{Oo`ro?tr!j^xQx5m8teoBnNH>d}?h3<7@U854|Le+U0meT4Tl$)5G$Y zQ7?U}iah3~MXesQ5x)E#g4`0nA7C=@fVDYYc)<(gZWgJMMcsr3UJD=ud`K2`ddII_ zTD`&AdX_ol#5ZX{CvT=B>oECx%R8Kay5#ZwrwEY8TAgqMfZhmu_m*VgpeW}zRyh_M zsf~og7;rSbHoI~~Y2%*+T(^I54eBJ?-YapLs<7PiaWI|Ej3!oj%vQ{Lmqk`2G>&&; zd+YP-^c9b+-UxcyHVpNl7kQOAHFH+%&L95y-@~5 z%j5et!Z<`6)cse1Bgz{cljKjr9ha~6w#}x;ohG;buCiPYMcYu8B+i(Bxy1`X0SoiXO>mi~ebV}h# zhTDlIpd(%hQOe9LT)!^SrirFge7zTP{H;=cA&QozZBd0t!FYW?EDkyO_jXB)-%Ct` zdOm3@c!ajSDZSLooX1z|F9#@PI0P@Zs~Cc?R_g4DaNr@&O9JA;K@w~Fi!iF_B!s_o zW;OifM>gF73A(a{&Y&U=6LX_)LY*{lFVp>;9BdLT0X@Q%vRBwAowV`cp^svpqt;x-~Nj^=>;Uzhi#y11YSZ zuo(W%N7LV`X&!gY0w}Eh(;soz;&F_pQGVL&_X$rk9_-}8z(=}pF$DE6<-=bKUY`a& zKZQA$!=D1X0&~Y2b{#qmvY)X=s-NF#zgGAU_(<86*I359;B&lj`))QJ=cHxCDl};J zM3K;+cAk`YdWu7kdzc)f_boM}EOLe6azWjkhwL>uGA!Zi=X`+Tc`mgv13rDdnxeqZ zegBg0a4s-l3j_+!>vfw|w5SGvTqfI|BDXxj&MBTPSF|P)>r`@dtR%*g)j>lOuXk9I zXq@&1yn|)Fn(XwQy6wp-9nvR1NIPxbS}aSqKQ~K2rz8k_n9oTJ`PlQ#3cgtyc)8P+jb-Aa`gyAH|UlPsn8$B-~Oq~KqYR8t%2p0 zdf0WweHIPEE+7@IR>?5A)9}$eN5m>K2kFsgry%V-_ouLL=`Qz~QF9w&)|LE^dHXy$ z1L9D1Q+1o&D?CD9-?cilYINSn#F(P8BPhW%xA{4aRPsn$zp3QC&>d~BKg8wZ2=e&A zp&n@x{I9@;Pn|frCcU_V62CswagU};$)``ftaCnS^ig$G=3~64wEM)?2F?khQ(tPC zf5O(fTMw3bhSE(`vanniFiuhlfA)qD47A-`fUUEJTphJo6^bPGnQ+IP{NJoUC8&`9 z1FL3Q!oWOvF^zxsoq++_Z3BTqyQ@ksr!AH3LxQ*!Wjz^itLKfYlLplAPw+@2mZ z!FP!^H!P9(hJ~+oISAl=#adNr6n^bQ2L9pLPefXg-{(?@qo;IMchGiw?qBV___Yxc z+%1`L7@b45wf-{*cj`_j0jDN*=pgceY>k4p*fI<_9?Apddg@ODa#Bby8(4Jl#4VU& zHovAn&L{ATxf41peRDs4-hqhRWc(o$;6=1t!kwIrg(*2A05cgG@oIqPAg(Kn#OwsV zEKFI3+rf37=X-N58Md6)5sRXh@#mYlz1R)FW|2XYQyIbC_%Qr8&K%CK#l6)=TgB;w zo!dn|vd2hp@g5MK<6z$LKSbYrlodm=$pR>^*^we~p2hQdxAGBkh$W4D3}PX~j{ZOK zrvon)h0e1Z|6aIf_;eN1`Cz#x)awOAt2!dp{{JIh|C_tT4wDyKrtEQhL$g1HCEb!n z(GiJ@m&#lB!F2G)j;z@762$iMi^L}~xVi1dllT@#HBFxeuRBMiVq!$)!^UV%AC=G) zzJ7ckV^v&`lG5NCAcea{9D>4o8unGG_lMOFIoF*X-KK3_9ts8oC2xdR^h z_ScMv*s|XX9oOE1g)C-t1T7o+YQ$_~$aGS0cNfGZJEohQZzo7+%F5Jkxio&G{z$|r z-I5v{?1c_d(J!3hA_I-=%8Pxy-wbz7#^`_X|)|o7A*o?+=h5X^gh6UM%m6b_LS=VNHGW0Atm62q5(*rtZ4>7 zxHp0Q?efXaw8RK56Anm`0&&Q9qmF4Q=S!-JWx5*LXKg%34>a?@OkJK^x08~SvjXlf z_7RFE@c zwKho*Njm4bC)ktBU)6lRnq+%}2=3M~Qo_rEAqIYKP;XqwLz5!ZZic)HsMG?(2DJi} zO2C?fyG1016D&%E5%++QXmWM7f3V+0K3(0M=6jFU`jh zW+;&;3CPCDwTJ0JfkIO>ajQ0-1?;H9P( zOVc+VY)-&AyhK2MPrqOldT=IJZ>b;pdrjt}+n%(hqhlcHseq)wH>VGKq^GU!;r3n4 zL@qOzhy0Z3$g2jv*Xpwo#N0CEOrn5rGC5iw8g6hh`@-*(+mFPlU5rFemfPsIS9*69 z@IHAHBpYM^!)8sceIj6H0BQ4t@XboR0TY=RU=d>=Qi*&JP+VjJ?uM>_uRbi?$@ucZ zt#qczZE(B!mrvld;lkGd$E|VZszX!2i+dbz=J+-{Z>Ti5?j%maIZb-00g2Llvk#_- zKAuI7;{_{rYP`OO^z9j_V#GPE4Sb{lTwmeME#OZh0V$`GeEOZxV3b^JYoaXvZ5LDg zsmE@^8Z`%p%C#@pG>+>(pM@L(LO?Y$GjsG7B8nY1nc){3c{G5b-Mnpb+f#^EFr%D_+FhkXw})Ke20Uqm3e4r zX!F)=AURUKh-vK)kJY$b8ivcw6SJTn?d?~WZ{7BU%wfU7GxSnsC4$qkJCjwCmZSO4pd6({GtQ?aV+FIs zI^Quem4w$ep8%HaM!{zLGyNS#_fTe?$Fb$H6@CYow$;;saW3oK7SGc}@6>n|G9#HE zqH;ZdW>&_t;&;1eO7cwh2b}Oo((>=F;Ol+%a<)hs4QZns=+(|^#P`0H*SFEVbz`s-V05kqXB;47dx|PY$gy?3rH`r!cp&5cf?bG~lWwBT3+8 zS2x}ts5+R=Rjs63&0VOvxV?@nsh7RG+Wd4nL66t%>3_`Wlr#;P)1RZ}-&lu%9vyD= zf`@3%*iIhR@4t_u9lb}#DLld+HIRr61-y5m2}EGUF==HwW_?b~Sdqew^lj(A+rjR? zOw%+VqwBvyu2K# z5wB&nl6YHK`^?h3U!x-OEBAMs_Hqll+0B1ajCEk8Gz?pqXrSielbTYaXA&e4ldZxyr=jN4T$DM9SM8L1!6o3T+butf3}VMH~JF+Mx56fBN= z^W1fd1&TnxjE*1W=BDp8M+qD-1h`t2FPMiOvKe{aLjSu82-Kv zk7;>T9*QTofr?0vl^4BZ%c41+%9$J=?A3MLgkwB6&dH!8aX%qx=|OIu#T6E^dYCaWCJM+e6$)voquN;<1+N^{j|xP|;Cy zc~gUXt;7ib7L0oVmn*VLXJOhLBp%0M0o=R{ndH_%wZ%y2=EYq!CJ|X~e^a%23d+6s zs6>XX2z}*-9>!C>&&z5E!jay`eJm}+60-5I?=)s{_mbwRWWQIc*k158)LFIZ-oA|` z-kmb1ZcnH9Sg50SATXr8!o-R7KDe9r+*l^?_4QqP?B|RcpmxSxe^2|pk{{{VbGBc| z!*rO8I&xn)#4k?w%Fr~}iv5uw#u2zn{p&VB*%gkE#|4QzS9+~L()|W-;MirHsvN;H zdy6IRU#eEB`z9S_2Ijr^C?fg4bc(L@Xp5^`IfgAZ6xY&DGn^|F6Ys)kZr*(`9-uaY zJ36deQBn&3u43*pSC+71A7*FMCXrYiF;wF~8K{kWHv&6(7KWUK{HB_32J?7*NECHO zY}q?XMaAU!T}UA}d}jd>WtsCdUAscJkh?*(KWm{yX6u5b!PCH-dw!Rr0md)7KgLo$KdDo z5JoGrrxQ{p*a?oD=ShotzebL2W2Ycx|GBzUO-jAdT))DPxG)jO3R0jOOX@_|r8Snq z{xDB_E4ZvzWxJ;`-GaR<-IFG`qcjFU^6pnKQjVE0U>Y~IL89tu^>5($oA#LMdXGib zC1u}poKy4tI`0q%x`XZS(5YgUg$g{Ro_uP@2A3@P9TH6RIuz5lIO=+@98GGx^DO~x zc7xOhwk{j4&3&_u+NO3w%xktkw31ml1jJH{wd8X2Mj5_L(}=1AQA5LB04yCFmSfb; z*M8EdZ@f9~kbNCgSgOsHcXV}|QCKA&=wt`1Zvrfk7>m=%h2WQ#zZ+D$bf&@1p*3I5 zXt=?ZePgUVMvbL6qwr&JcN069q&#ZsobKLhLN(iii!6=e0Ucgggr)qhQr7dJLa^v* z(E5-%UO5)`O*yNq+C@Yro#HVK#B<8lI~VD^^59{jF~P8hUgkg$NGiHbPaj3Aw-{H` zX8BZd%lEEm#IleTsc2Gur=Cp-S^ZXhej_W!HmYo zyn$CbBJgk{4WM~3ooz{@2Fsj3zI?rAt2HCjAtEwds3J0!IFtEC zhX54WrosUA^qG2VA8d7bF)^ffNSw-daMW8i&eXZu zupHhSld&ZYo^1i%_%0?I{uf|3eN$fAZ%XH@_zZH(AF8twHT1Y-9{!ye3{7y_74_+p zlJ5`PzIBH+-H&>~@Cr7;!h(fZR()qHV_8@%=&a@PTgT;3W_GqVSz)6}BO;LeE0oU=o77Vk>6eBjz^07uq zkNT%oV=E3|I7r~Vh2|=)fqUPEIf>3P6{vI6VMM(ce@^>_${j{QNy!m3|5;23)8a{- zKd@3`Y>kvtrHxG8j%`~VUa(7j!5-)idS*8HI;?j!S|3YRdKyewkPtXji3N5gc_6k= zEgHfJt=9hX;vtx6s&FIO<<3g+z)3CxV3ISfEWmzlK7ph)q7ly8VHqSOs<#>QpOYD( z;la+N>dI~-M<$qX1Trm@Xw$f!r|^o7106y)IeXBT?k?SWBD&c&Ek-kI%(%R^;b50o z$a%Cw150jo)$owhW+b%zP9lF>Z{<>;#Fux~dL8PMEu3wXZbC$L4JEr!ws;;40F%_9 z-FWZVFZ2tG;DEMuElCfq;MWoC_&ngG1_J1H%5P`nH;#8;4UgmV(+KcLsSSh;3}9mn zt3xhSV4u5z02)iF&->dW8tMG z6sH9$U--D!RB$mHv}bKeh>77p2ZMnZjG?pi7Tr!izu!}EVWGtt&^uVh9qYmL2qfh( zM$j1%rU zWn7M}uvL`|CG3CaP`m->i2K}})`uq4A>hIk;5hMl0QqGAjsfuI76 z0O)zN+8_0S%k}9fP_a{~dM- zk}Q!gUI$A>VKi2>)51+)V9WYKPubc)-IL?(56t6RKp)B~DSg%j>Ib2#KcbUOZe@Ky zBeQA()FN_k-oEV@a$EX=1=R)xKpB%o#|=aGV-Nkok3k*dbFw9J17-$-7zNr9;h=oXgPbIHj&Zpv? z*oOdj>shdT-Gg^cms!fRswA%46w`z(QhgT!2#-2;K~XftqpUzRuMn!-!)iz|tlbtL zPFx$>$Kh3L z;d5LLR*mKTjGjWTN(%>-)f8PT3^m?eq0eTtC6Rh&wol}JYs9sPMq1)f(1n{}ufXce z_0V||>b3dk%@RIWYF`w%moS9%@aYNZ@!_B_0^|qTP>GMzOd-iYqd;@Zf4Q@6pfz&S*($@PoPg1EeCwOCfbuSn93dEZ;TSE8cF zk9mWcM2ICkIyk-wNv%qNB{~ohB4PtX@zc@k5BDDp3!DP_q0lq5j}Y%5zjkM~$!+W< zHORSZ8@JPZE~Hna-L`)zAFK`T9k|fl`X`&{0PFlPJrMu~t~&@Qn`fed`G#k8^FGm_ z{Ho`=H^=OGvKw5d zB6`XAD(n&7m(3vxZ3))7_}xHwse^^B0N4=Z_}QW{(nPNq8O;i|NzZfleYfN9Gh*jV zH+uc9$~&V91maNn_t12Yb|2CpX|(x6{A-+b5>@D{y&f^<-ZoGANlJ8~u$ za+n|JiNvt^d^r__bSme~r-Bv6{ypIt2raI!geTq1_CB9|lcYN3ejcsp!5z9<3oWiF z@YLJYhMu9tf>-9%-NEG#iKeS+3vU6e5xWTj;rD6jyf+0E;Wkg=$-q)gfNtJP#ZlkxsEg#Gv2GjP|J|75V zDG%(xU56`D-tw3xNHgG)!Da1Onk55D1MsIu7fM4>qYJMt%sW%4Sk_*7rjOb`MN6 zClm%!=4)kT3KfJ)RQuly=gJt_vQkoJ-suFd9=c%_-gDzrcNF1dj2$15R@z}B4u;|f zhVq5ad#5y1$EV$Vs=Jh%T2YPGmJzxPaLL_%fV4HLa+P|e^2tZL3IjFUDrIp-0U9a< zb>3+uDfZ+pX&gsz1_kx530i{C@~{s0Z)HCal8Jg#xetU?o&yMwQ>@pCyPKDXZmCPY z@`ncl;Drj4K*s-L`<;GD z$mSvgkvLI{9e{yU@h%=kgsk>bioQ_WUxz`*PQij8St#~lo%`%KWLcyg=QDqydL>v3 z&CW7NxK!JvonLa&^kX`L(ct5Sa{2G`xWmCYX5J8{-Qpps%PsC;Zx%uh< z5e4M24tbqHI0&GrGnsGoF7(n|>5a{?n{QPC_$nyA0e)!uh1)8Q6n12PZv|pm(}cXU zL1~;t*U2&7hOVqYLwE2sH^l|H_<#Vg0=RvSO9B6NGC>c%=2lSbRyjq- zA#-*oGM;^(jc;)|Y2bLUsdrs%MiLt}=rOvQ3ott6o>Wf)r&&0Vt9Eyh`xa}*(WYxqd1JQeuV(l4Dl$6uuNt{#v<^opso>_5 zDXU7abH|ziK$*t_$zL}hVs&e8mdOlmUAmuZ4x|aH%xx{7jiEs+8ehjo^6EE`+`Tym zoZ>Q|;+01%CrX*^T8`r|96P^$CDrG-t6fbge7+pLDJZAYur`oV4`8K|SZ6iC5^EP4 z^Le_?PS)Q30&61$<^q-^E&;1uF950Pk~%NDH+o5<2M{htBe3!34ABN0H`c(<;KFUR z_)*IEomOKY6@3@zWcZ9rC6SNqZ|(r7&2qRFs6ME4IJ&#nj}~dL_8$UGveE_xfNQa! zfRxkJJUXhT1!(4rTcAQQoR}A%Tez(03iV$h{nEM8rZ{o*&x!cjdrsa#$pli65=f7f zW1h|p2b?v7fjd|=22F`n&eS9>izs=7fZJ|zBdbbwa4#> z!%fJluU0Y#DP>}X{GD-{LF)iK8x*e&dKt^4KzU=wQkusFz*c(o!Wel3#dJ!8tKyGq zqvSH7N?DhlB<4rFRM0z{wTH&d8u*-xqxNcB@MKFt^(=Cb=0ljylbsvXl4V2&{DDz%i?zK$_2RVm=Vo!dYNp&3X53|;DTqezoZo7#50MF1|tQ}49 zPphuYQTHMLI*q&vN_q%@+{^Z6%tGi@bAKI^NiDrB`qQ7ptDforvcRy^HlORpO|{_H zKcH$QKHC~1UJ>AXC2mEOLV)F85q>84@jsYxyS8&>ui zh=BaRic!-Q#eaoXJ~99H_GH@5BDITaxt5Zb7~WI+(~>Ps(37X%Ni%MI=%EGiG8}Q5 z94~yDIw6ex;C`G#%ce{)77Yr3XTjSGRYRzO<$=ms5*z5^NTGeF9q8oww@JAkS2ve? zmcQ|x=qbP!(>Q&S@sPQP*qK-io5wK$*DLE28hm``ObqTjr;ue~C-7RN_i=T}r@wO` ziEN5V?XPyP1oam_>~`TCBp&?7LhCU>NB2>xK@XA2%GBq=cc%m?s%BaG6w(bopN9qXC9rJ@?EXlI&Wahl%2xZK-)`KuS+_+jYUT9r4u*1R% zpLCyo$3hNC`0CuJ3R**wBo?YrCGLy$qTJwzk|xkcpCd6UHL`#bHtMf&81kEC;l7|@ z_j{>B&wf7UAeUKkx`>g*?$KxkQm|tr11BpXU%(#%{UZ(#IKW?4iGH8KOXxWkpS@LzDit?q3 zZZ%NOz^TuCYE5ucDlRza{Q%_1&gI%?(nXM6gk3h(in!xJZEFkTP>mR3kLV_d)_5A(Nw~L8 zzBWMojNGMQAPmx1v=Xra?Y0W2w6oCD(_6Y*V;)5c`}(*izZ@ zJuElOqcL~653l3jbua9D)`zl8c4rza?^&R9=lh2CK9cH85cX~LNPR8{5hhTYL7TUk zCkU8w=VA*1{J-c+F*cSnHYO$;4G)qEdLUWv}STp_h@E{*-Dyd{80qetM@A* zsV>xfe8D7u*9-O%Hk6=i1JPZZfLVKaQ#?AYj=>J*qYkm)2-h@DkKK}!SQ?Nu{NtGJ zCLvTgfK_MvuK@p4Kw#PrWD&;eKQr)v!v}UB4Xm;0qNGQ1He4|Va0;=z(c^$%X17eo zlV1&1+8FZdjDcQ)9JLQf+0{-#A6Y!$T}Z<2oBf*vxZQ_#BP^7k5yKxUtLpC3D05Z@8= z45^CXVVbb7{w&lRIGd%722`Jj=jFNWu1mK5vTJ__ap6bQ&|7XcNaivRBASb^_l0w8 z9J|dF*Ks{?45dC+2OhSL$O-g5?~5blQxJC5PXT4}TJ()#GM{p#X+HxH$f}J5 zVG=EZ4NpGW;OdaHx}~zV5c>c^F=;pnAhc3;MUZR6vZVub2iocIpq$*&C_Fs;bIF@Z4=&s7(THN9=)5dz3$z8uKgmi(0W#=sfQk$>Ig|wnKrE;(nxG+| z36&Os5k}m@#ly?IYl_Vk`8r1?cJvz14SEed-cyChgew6BTgMTeyr1F*JxBkgBDf9Q zLHu@eK@GVT6&0oWT<00wfCLk?)z3)p8i$CC=c0F4!4Ay71U*{y9z)AP<7BPHx4t0J zm5S1vf z1l)6*yfzT*WPyaj1PC@%YOS>_K=%3YE#IN-B>l&2+>9j3+{$JlxL0P;S8wQN8jf*1>-7LNHGLzxsV zC^OlU!u^VYuxDfFZ$i=g*{=a~n3n2OpJzPHE0y(nQ+eY6a#L!)g!j?lDav=8FQ_WN z6GvZ}p!{u(MH>3*1**bHYoI-&xoAGs>)TZ4kK_2HSTj(90jVq=n?}A;$0{7TWQ1JM z`Fy||O3Zwm)R+@h55!CsS3zc(*?uoUSz;g=NTPc(M1!dL`L&2aTL^%(u3(c#XqpW- z&NNd>o`UV0c?TmwqeR8sOZ0X(pR$QZ=S*E78XKeA58ju+PT&*|lsF}pY}1_e z_s{p%efk+4orR&}?SAKg%yv1Z?7=ZqlkB`&0hwfh21D7uSl&&AR7H2U;(w^bej8q1 za4sGr;#up?yguU5B)IK-05PaL7F0O8uz-gKZ1n9Jw0rLS*$7?%X&nQ(2WXa&7%BQc zv|-J~K<3r+XQ8tM`@BE=6N_Az;lDwTf|j-xNGMYW09YQJY=j-h7PQtlpLs7XE#-iV zpte=0Vb^(cRJw+NRliwNeaZa4m0<|e2}crc%k1$XwJ#hdr1}~h8Vcbir~sTifB6rZ zTmnTfV*4;5+z`I_?V|!}Uo@>e$Q=NeYYaLUqUaRKyz@j+T7q!MH3B%yh0#m%<7CM&;vAf~ zX8;&;cpRA@-+~PaJ0@EuP72qH&5lvKUNU`*Aq7YpURG%7O&?V8EOX!0Z~|a7XO~sC z{`VhGpu|$l(VE1$p$iU0wQu5pj8?Xh}amvYpFx=82K?fs99L^Nb-Z8*uui^;6;!5 z2Wk^JOg=RdZT`%7FPA2yhk8h&fI#r`{E0a zO=7>O`3upa&d~M=0;v|L^UJb^r`o6!ZDXdP3XDU{h06fYL{bodn$aT|q-MwSoj7p93kVB5L=#dwgPI z;V|3kU|Lz(n}vMPR_fvldXLADeyY8BlWEC}rh!tR&B6miiQ2)7QZZhtkep^xRQsUD zuRPcp>SZZo& zsIrw~D6)p$&F68H-^j~kAh|K^v)MSKC^sZ*tKI5{f?kzO3bxugK^iIR+dNT#Rj zx-tQ~|3FX#rsMe?PK45vFW~?bb@ZU9iHeJ+F^Ad^nlCenF53dE0&->n`A16F^Twd%H za0*qeTpLjUwe09T4G7S6SE+NbZF9qV7ivL^8rR@k*lI`rmy8N!U-W`s)-x7U-*iu(Z&t`2uU?g z*RvcQL75`k^#1hlTks;mS3-v4GG$*7s65HG^y#$ob*J0%4Uf|DeP{BmeePp9c#05! zkyF;$!|_L;%p@YCqF*Ej<0Uu7f#Sb-b@Ocn>Cg`e`3UI{G(b=b3F!;|5F}FsyqO>w z$mT#SZ3ZzsK9uZV3_ItDNAF|Ssf~d;OQya91B--ebD+y$Opxp?EET9GW1(x>OX^W3 zA33Ss)pwiaGrN8SsH9K`?!_`g4bnE z@TaEopv*bSOiwQz)*CPA=^Bp{A>aBryyN08hMS%t>j(@8m=stfJLicx(sw%>@}t2u zdZto0+F3{sh4a==HLcwy*app18gIzg$$6irXi#>-)gi9JD``6O3vdA4z}!|qxz;kl zLiX@8dEQ%lK5J<2BeUQn3<0h-FD(wJw;BMPTNx+whzUnx1{vHxm<6|y`H?LhtyZoJ zzQ{24wRp1HA{z54OwHflKii7->#?R1()HJ4KDh{O|Kbv`H_ATzH|5z3LHg%VzNTo2cCA!8-Ua?!9tUE@TrGUL1g;* zKrT$#j!Lw_Tf5nP%)b;kW!k-j$UIC?-BUgdBj#2)1yra8vJOPn3QB0BotGN$rdC_D z8t^3BU;YKN=W6XUd5YwJ6KfaHt8U0mee&(&BXyrckU@R#jsd7xiAZefyxn}UmsABOp(OgmiYv$z+0<(VAmw}xS51tI&J*BLiIIvESpR` ziwzgn$_l1zU*z7_Y<^6rc7jh={~9rgF9Bq4)IW6+=-G~J^m%5--wND<`Ww;TDumcB zVC_jSf1{uy(nQIlK~N3p;7INId*{}^6?!6kK8%&V;=8NMqd{-y4MP*7;FY~j@-jvL zm5Ts=F1ix{&ClpP5oojP;u2Y9>J_AmCNRh@FfcfdF~eh!TA zQV&u1HYoG|vol{ucsuap=nXDgabs|(GYvFyIUeh%qD|Q!eHwbZ$_iOy8#v2rbCe!@ zprnYJhUSa6&vdN~Bv6O+_?nf$ zkQO=)fG-LtX)ozpZsH;1ROm9-4L^$1)zGjB)K;=_}6)2S{bdu;q)D8tGncF zy<=o@6sIgooiH>k<~;3$nk!=~&c)_SHudp=7*Fh=0Cq z-rpJaeop74^~uCi%I|MFm)SD03=Lw)3POmh6m^q`PKS!QFD z3hF6a;~%`f{&=f#e2asH9p#a8=;UB_g!b-aTC-!mWYAuD=wJr zKd5Zb^@8fQs^@=k_TKSa_h0;QwrrJTuO#xZBH5%6DSOZCy@l+G%9d3TvdJFVBNZZh zhU~rf=6Bw@zW04U9`}F0{<^LPAMbHq=RDU*ji=s1F3L@k^0Y71g8K}00`LMIT7XPH z=Imt6C-TDwHVkZn&bh5fyL2P*l%g+n$b@&=0mlhg?ruT4hVoZHmlp%S+YS~a3%jk0 zi~JmChMzYY+P%tUFT20hS%Rb>Ss6BIt@b&q)v4GlQ00T)l@hyvPFS|`b1Dpd7MWQz zzqt&9vhwv!10^1WjCT+-fZV&`tgnC&9dA*y)YQ~wNRCK@*g`r-w74tG+R=?)vXptkjPdaQo9_>d>#KpuaM90E zL!Fc$5j0qxAY>RA(9;7eLC@L_q#9&ReC3|LgDu+^HkmBDi=18TF1k}7ya`4`N^S+0PqXEP@F^njI8eR4E=V~|F&scI22mwe?tMI zksF14ott3wFEVIA=Mx6^B|5x6`Eq~^_J!QPLrD-wNcjHzYNLId0bhtmY5Y^fX}Wkp z15#GJC|}*eMjZfH-FH&8R(lKa@OzU)XMFOI72O}lNO50&WujN*Di46y?DjbODLtPm z%SVFESVZfW{lLlNoGpCqdB=bQObnHY{C1xz&tMWz@sLtciSqC~+_rF8dlV>!If}&Q zoUK9W{xGNSHi6|*2fcZpLkBHTPLTjup4P}WWP4N)@wUbt~i?VHXsE&0<0G=USu;95EK;ry(NRwK*+od zfhte894}wKe9j=k;YcCqoY}ZEw>LBWi^~UT?a%{TTYrh{jW{`u*Q52xKHphXMT8uc z7TyOll$PEh^ezF z3|F77AXQPIUmvuqpX(QwWt&z95TQF)ExtneZ~?)etW2?2z#XhX3``89L`F6~9HM{lL1!1 z1Fb-GBsU-}UeWlQdQ@SQbDaDJ|8TLJFjF8FMI?Ab0*4?_55rS!EQ&`qE73P`tMJK1r=U10p;o=!6 z#2v@CGOlYIm;X1DifP65yo=nl&zu<_C-AWOr6GU7liD*j`}<=> znispz`$j`3G^8d5^*VLlUeD^V1T0b{1hMVw5!o~&IaV7E5w-p73^GGy=sjtdlYJPS z%aXr9Hy85jx(UAVR5~)$J3gq_fkV?wul-M(o`GGPJfg@Ev(>nhh>3;uPdW4Y_=bV) za9Qrq0wOs8nlV(pYu<3j4 zL@O2$0H>|qvvK^Ilu75L-#FvwO-s?%w#&&BV+D^~;fQ8A4LlLAM<}>Bk-PQPvhUaO zmd&3M#Xab95MC(3HKCRPMB!3>-2cAE^sMgzT*O6DR@Ro~tv7Gp9QG1!vJ(X&!oF!g z-wG(Di+`0uHz$7t^?8BPQ5X5Hgwi26>=qg@elQ@nu?k}rVyN^+Avoi{H>gCsD|&AfApzvWt#8d!@dPNr*ka(GXY*4!2K{5&+u4uU(eDi%|`k> z5N7-X9hy9+<#%G=hn{1Yn}TgIpagHGva3$Eod*JD7UtN8z?cdB4UJ!35A-{sjB06J zL0*U`MYLiD%og{}X@|Yxa2ngca^WxL?3U=%+2SP~O$~#Pj$JVTyA4h##J#F2O0;lo z796PUx^_?0ta;$zPd7T}TI}nEEqiZL@Y}}A{wR>jc;34Z%+r43-SG^NP7Yh=PpL|M zaB{vF=)iZ*Ec>z_g7Rpvp`+DWcsKKv09XJQqo1DFLd5`ptRWR-uQ7n|U5DsVTVD_h zi*6+iCrdua6!Y|iV~u}b`MCnbHGF!L8AwaZT?KM<8R)Jc?O4A91QfYXLCpoEsqVRw zxG(4i`M)w75r|=$yZHms$#iPW9WM!LQkHQ^zr>cyjbuMSn2V@nfwmOsuozI!{z4i@ z5dUOQ7JN@2q7!8Y#xEKH#C!_?zwZ8Or_Yz2RHJ}x2VohG04N|WEoFNZm~6X2y^;l~ zQ~Q;1OAd|`YSpI>EAzsUtnA0^%h`i@-o z9(7X+`29E8NBD~*8O31PXX$@roB_rxeQ>cgliA-xfD70zJSZ5+)#i7Ha4sN@HN4?@ zJpczl-};CQzC=(l3bF0Y`T!cA3f!j>w1|)f_14ni6x!cjVW3%5*Jg46ETiLZozykk z2_rwtAOrs%QkDdnjUcNulC*Yz5}%A6QDd=%_a!|L0@O-gQnLzrEzo0uUtU&4MdiZ+ z1R^RnK4&65K{@J4s^q4iGiorjRHFHxj%EBCTvK z@82#zbAWGMw07HD)fIr8IpsBI={WeV;Va2^9(1S|ocFNitFJY`kHy=AkG^br|xb)g(r0wEk3#uRHi4voGs<9;aL~)v7-R z0vDhGZDRnelz?#wjLAn0XuW;8s{`PcZmU#yHAFO~j8{R5q$}3w-y6()s&^9IK-|-I ziDL$TX8m7Rh-EGN6YA`ATCAQ@MPs7p=EDzjir7R7E#DLk4GkA*6OP!sK1u!w5P`@+ zHqfi7`!s?61u+0D) zfDJ0zpf3<)S_nV={+hQVrSz&=zTV3E^XXsFa$}U%DCY0>dfLgR&MSXP`#e7L9Yez# z0boQcNDF?ua&2S%J9#%h`?>>#dl;kT6s}DGXtL|RNI0#Z+U!SuyTz^dgotBdMwThz zUXJ}lFtmsZLKhAhdyZ$M{`==T^`1w&HZJ)gt7d}XZ`39Ub2~%t6^x`lyu9OMqd+Pd z7ZCcYIimdoU(Q`Q{8hl@e++=eqq+axz;+n^^y{d3<IRsB@$dA2Cc(Sd%|PsMKCO z2~M30Ml$=((+Hpqp*wM&rtn#Bz*#)Yui4f5fOieLTp!P(&w zC*sEFGYUqZw{lIaU?q;EJ-KV4awAbUg&QHWu>_&jgWgAd59nRhz4QhOsWyzheC*w&$-ls~SV5+4-1u*QE-^FrYBwJtwLtUbjfAaV-BI|!xYT1XJ8*CvdU_y>$ zk|wek$V;FMpvt;_O`vlCN~*cDn7PWGx_^v(GdZ-a-cup`y(d6O&H!L2l{OBcJGaQ< z7+?3#I{gG1emddVU&WN|isbKWPHnl*KyoVpdVE!NPs-$rcfGX2|jGm zmPD~S0$O1WtUoZgUDaP(9|-?fbzLuZGYk#=liE%&a@!6za+QI!lDwcv9CdQ@I)9M}B@ zwD9M^&+BRygZ?@R^Gtx+hDYrr?@V*@-+UJ~j5Hg~Jqf4HLL+aUs7OAwLafy8#qT>- zkm-i8^%~bv7K$bBVYY$|u>`HzNrBZMD?-T>)-4VmTFI)lrqgaPNqPBz1 zSh!LY`D(Sn@ZAK!>;O)g9rgzlj1IjzeKvu~f0GYlSZ7NFh1=dob{}o4>5HjxbcJkO4a#(iKF&Nw&-a z+yn{~06Sq+QSB@Y4I?$kM{R$4e}?71evKCXdD$tC&@u9!BGz58aFCQ~Ki26HYrjm! zFoS2h_un{-4?hlRD*C5HqME*Q{+UnU24ubxdTRY)ZEwZw zpB)nk5h@O=bl`7D+d}?jggbmqmDNn#*mmSH-=BrMsF+|y{?($E!$r2-OBXpck0A_^ zuCL&cC@WyUeeTefv4Fzhcfq;zWK$aX3gDkpVg1l(Ss;88EoF&#L>V1*GB4*d6W{Zi z$z{n~qQk*?!<(<5uxZkf+RfEj)OJqrH6!1?cfo3J`dhAB%GDW)laK%5TF0cyJf6LE z(i~JGwn}-v|CNYD_;~MpD=ZHFHx&B(^c})sPD{;BF60v^eknJ@DDa;;1dOM}#D6Y3 zNrDRw1_dRKI?R-T^Q%=s4L{EITEs1>u#Zt@{|?kDE`wc(CHbw`R_^nms#zBp` zq*ycnT{56a>(*6+BsIvz$&SMwPAil0?91xiDCmf`pIES0K?=rCdz=Qe2@|ZJr=C0D^RWU#d749FkVN z2Wa~U>Gw`Dq1ZR@Ot-yPmI|ZIJr#4Bee^NrmG42t+$Xuq?2^!dqyb#Dcm~=hvNcMg zkN*n9>g+qTEtpYMg6!bA6yErL+2WHO96hr6;-Oo8Ibl-Yom9YyS=o^|R*bgj@$Rry zUi05wU-5)IYBuc6xAMddRB! zr~l@M1{UqO6iDegflNL2_LM&5{@GF-*?yOW?gxNGbGk#=t_jF%TxqGQsW3saY0%!WjvH+#5(4=` z3{A%WBsa6a7yngUyer@=L}r*@v7DTc+g$`>2WS8aB3{14$kIFde&@U6kr z0!pGs%eObMp(MWhJ`LKN8#8#BcnrwTJnfBm6gxhN^qmCKd7JkZa)QTFD)^cqZx}=l zi?HD$-F-JZaUU<`9qXG2B?DQN_e8-Kv2pdk7K=KOhLTEx$I&0c*EwP6RHxbQToeSL zUo`}s3p&Rw*O|ilh3+~ZpT4)j?VbS#?dFBGFg{S)R>4zbI(E6ku(vLlma^(U z#e=5}FZ&wRRut%SC)PcX<2)kq*&#-+KEg9B#)Hg=n?lI-i(=?)bhAc-Axg|P+!#ZE zFc(eRK_j}XQosBf#PLfnOft22SzaJ#>t}-LVzlD^r$JZnoIBy=PGn~E$(`Bl33kk#^~n0&4E`&7?{ zWduJ`zNWa<@Iu-`vb{chcW-U%*$~2~B1f6r%3UvA+UsdR=d;EpHX6Pga9=4MNQHNn zL$lxKbc+tzLZQfI%fWRg0bHTYqn?#SPV_UJ{D!AIY08QwU^p(amku5&|40w&)>Nkz zNJ4ugns}1D15@QMKMWjW7{8@@6pcin&!p7`YagC4{E)qeDIhAGfm3-FIZJED?4o`* z|OVxy7xBegwfZP~OU7Q}s>^C@o{S9mvh>7m_G!V8H$Dm)Lpm@J}G~u=0 z_+z~Z|5P-!*!cH{zInTbZ73b{8HTjg?ao{`+5-1VcV9Gj3#ShgY`qHDT=IS0d@0@7 zEm_;)V$s?N>SW3ht61)IIRdWj=3`usHmp|4M@?I+)QPwa)8N)f1+59&g&d8-&tG(_ z`prK9RF9kry@Y5_K?vAuzZ=tUwZA^GiTI_#Qm~VM1jE{IPzs<@LE68Nhv9#(`VJeL zELH`aH1B|(%WeW|aMAuAfKd1oxb=M4t#2p_YgDOPIaFT$#o#?fz37#ZAkFo99uCgKM)zb_b z{MoMM11e-aQroy%o<4oLjf#UuOO-$Ohbw5q zyGyOqTR!sVSMN^-S;Po{%>njo-2NCjF)xT9wn)f#3k=9l2htv zK6B%8HBGp|Y#Y5{zF zB>3GjFH}MKT--tM7%19Vbo59B+eo#$GLTK_e~m!0dZ#1BrWnBUY&v{%T_51gr{A!A z3P&RMJ_%FPLr@Eec3e$oW#i9#~(YCvp$V6FK{q@gk$f&sHjuc1q7z^ss{J!^X>U$e`)ms6PnuNO*!|a+^?ySN5CeZ48E34n0qRwPd1n? zNR+xtQevTdW9A@2&!v}C==zwgzVu{En~*V{(**jg7ytvZ#5F$#JTGYm3|znTrw=UM z*hFT3Ai7jr9ROQ32J5z^&}5T8YRkMJmy1RM)FMMBun}Xi=_hnOIz{r(KwgAe@`6(s z#I5U{23cUEgsRg_iRm67P$VOAp(M@{!N;@&NaGn^OWs=CUg!a)M(Rr=AY3ZzmVelE zG9J^!R@WK+!M5!X*^9@M>2lb3Z}v&iM+jXBLM?HV;Dkv*x5aF73;t+WAKY(`S%hqx z@R0_dFvQ>Kcts{_5Kkt<Xt+9O^BU)_Rvdx$TrqHZ*X=p9g5uOs4OHSKsR>9%4qf@%&Lsls$4 zR^J5+eSN>xVnXD@CCnirJjXw;#qRZR`b_&IMcmg=phrW_s`3HmqFHvoJvtubBD=3pB3GR02 zA3wmReYDh&b_y(IE2Ml)BjRHbj6RCag?38kmy&oGw&ta!EHZp%kPUX8>;aOl8TiXZ z_7nFGmWw--Ow{_HR$*+S5D8KS)MRMrCPXvp}&X$>``DbY4y@j)t} zEr@~jgjc1q{zx@j<>6Rm2>l_bV`HKHRKyxYoPe#?zzoRR6`Ifz{#r=(z%JutbAjhf z`b0(oix7M;Fh>L-&PU)+emzWfwVh!kcH(~1CgJ;1urOrY)%QT;H!!~+c>?X zcdAS3fb{Rv3Qi&ss4)M1KZCc?qcI&LGRXt`R{;YvnqP{V)^BG~?Ey4Y9`1G0yG*zb zD!+rRXQ#CQD<*LKD%wp!O3Ij*w_ojv=kN$pF_K|eP-yIdC$-gF6thU<4E(D)VuV=~ z={*;s#R^ann|{Yu)*Z;6nlu9gFY5%H@VVd^>uC*y+Kf(+Y?4HovoL;^_IF-o@067) ziD4`eeqTJ*>*8+t&GKZ5i0|#Ms8~wJTS}Go*X%?c)-IRU(h?ILavrqo8d0w8PHACFO#k0&l}Rpv(Mm27Nsc3u>*dORLAGQUu8>O zp@vl?^)|*%7VH2lP$=^V$48PnoqQ^xt5SNp#*+44Oxrl_SYVr*^%IMgR30LwBBe0M z5F`G|u8kLi*0UjeZq8xsfWcfsrm^4(yVd&fS-{XG*YG2#H7?#i<05pn&^k zs!o;b@8p4}`lICzjW%l%zzM*GQn0r#X!A~~_%T;1*cKPotQ|PL+LsfHV6t#Z+(0rz zf9Hc!`$?Y_O5_gmwpbdAJ`?mlE{C|(2cuPu(Bsa)Lp-?x@@7r|qap#3l?4g?G%z>C z0-0X0ID;9=F(UST1OA<1C#Z)UUc?X%_+6&0kz=&hZu>3|`~A;bCedM)DzO-eB0nVL znsaQ1iXK4HC^P73l`%GvEdEb=%=vp`uv`S$c(GxMh=ZW!`~r-Q8T@_iiDLPAjN7ZD zLZbtxYd&Z1xQ%``LzrHk9wO^tg=GSWOJH(xQfHGENRJS-;y@b_hram@eoo$SA2GLk zhh;aLVoWwIaqW@aE2m-NxE2J^N5(>AQ1%>w2qH;JjEkEu+6x2kUF_z&$ZoZW!e=xf zxX;sy>d_W!Jou7kYmiW2o6gGS25;n2=)j~a$P$9q0{hi$MCE*PZZI=HNGlns^%_o| zBLoWjA(Bi2`i4r6c&Ho)NN6E^;5(FAy@;1uI>ZuI9ck}?a`PTG5q0Ki3}Q2z7J;sq z62p{u*kiM@YpODB)6H7<_3{SHCBJ~R=sQD>U0d86q!((AR;#48XhHj~8}o!ydgv>Gc`~CY79!w!-Mr&MMcF$yW0fugk!}> zZYlu&rCufC!ZmZhHS4@~UH z?j}XXofcI_PCwt~Zo-7r5$^mN`=OH{&RT`D0WAptf7w>cfSAYvK>^;k&SZkzSQ&?U z6R+nC;Ukblv{>+Z0FiN|U5dNcItPZk8Gv*xYY{O!gt&IM1qv@itn_`;wzy=icZH$Q z$c2k~9cIpv8jEeU-BwDGB0RZ6CG1x4W9K=90qkrstWluyk+@aMgg*K-v?D`E-OVN= z|2m-eT%N8uW3O5O9JBAh{Z1t>iEj^)OMOoL$8%z55#iw^rK>VS4X%_t#;?z<+ZxY) zJSQPFDd{Bek?tOLKMY7sT?95f`d4eNri=x#q8kkkXbHDoT>%}tQ3dZmFO`(Dz6K@xPcWBFa|HgLH| zA>HkW{_#GNS+ICNBkd0TlOuSj+4z3MZ-4rVDW*UmG;bkB9sWFI?#sC{G^`x(v4`Ld z=p(*(d@~ES_gTIl@Q%N`^}pxPM#PYs>$cB(-hhVBthpwtM!A#z3)Pvj?y=-l_%S4OI2+`kSoa{Kld4 zOv;Z>o{16q_xS|cXg49I#(qBW{Tb$kno92pQs62j5mzLFj?Y!sfz58?G%5GuEb!DZJ$q-*E!LDfDSL{re zr-J5Zge&)oe-uTDKh-@BenGTJf#+iogi2NLYDjYT=SbA|Y3Z8grC8Hbo`XO*<8!Jl zrdL1Xp?BkTZnQ;{+y@)+^E^1>sGUX<#}tUQaOcy9 zL!^(MTB}K@v4453_ULg9Ke~)csrVAxBB7DsW@uLnVHS<~D=CA}Cd3VeR?C{|kd%$3 zWlCBy<%RG$sw)_1)w7`0Rr9%Z>Ww*L-XbN7sg=Pk0)Xg@3k{bb1G=K~qU0@h|5}ivXFhhs)!@Tjh1&k zUH&SUj_($n|Hu;IL^bR;CC&Y?WF>b_T@$_tIclDb1P~{NUvj6LmAw@4Od{utg-0ul z@_AaBUH)%ljDd32w>I4M1)s{0_ac}u4vM8^H-t5X3)$`)V+0ok_G^}Se~!L_VU^TI zEm** z+21Z-f=mnv;^#U64{DTizeH92YUwiygvADd)#f9BBdzZ{YYIeZPy%kYi2G8Uq5&mH zlzjP3jf+0_cs-oT*@9OLDa<9eQscx10q6ucxmr|+wypM2#q%2eI{j-9&<<9e)$ z1!I)sy$)w>Lwa8%)cNch)0<@9S|N}i-t-ZWp0#Ivmzbi3^p5Vg5^W5T95B2X(!tL< z0IPlNB}_|J@V=Oqs$ls`4Y!;dkJ9mqV)ies7G|^S*E4>t>`VK()hM|$-;Mi+w*LJE zd-H7U(YBSO(Y$keMv&qiV8|EWbQK&SvB>HgJ&@N%M^13;5egom^>8CSX0>)f{_MY$ zJvxD=bdqg!sg+7lQM+A+U?%ug%RH!V_4(<%W-!lH8k&SzcLLCdz~)&>ss4|MW$a8$ z?tZ2UTYMu}l24f+T=;w9UDVz#o#LJwBI8E;v#@r!6PPdyvLyotTFkcj3vLir*kYb8 ziA9Bd$f`-J`aP2-K^h6``Liqi@ERKpqbChCf zwv0l-nNR<^qz9a#o6vXUI9&i~O+9wozS2vGTMXKQ@0oc7LYCKG@4*d~{y`&nWA5jZ zi0^XhzdMq8T1IWE`cTydqO6CHeOi=~gNN_j&6{0@Rjj&05)Xea_T}U}u0Xao4%IyW zdMI{Cm4s;weQPQxe@tUtpjjM(eMfA2uP=qc7#9Ju8VsE#6 z-72-{9P94XH&-}x24@F83A)H4a?#Tykq#9n;=qI+LQNKI)woJOj05m3Z0hsE2c|UQ&UqW0E0OKVILi8Cit=< z-^r9J{i}_A+vM%N>32^Ak9C95JReQO!AWNGVF9?|MPXlVBKt>1;`B~N^wxl*!Jv)e zgLwP%<$0;V@hE{e5G<1nQDG(PK zR%rQ$EdYM1zAzGdUoc2;K<9*B6C9|c=>zbQG!m8zX^ZIx)*; zFY)YH8VeuS*}4z(Mx}J!FC0I{k^4oVV69O4taM#O@Q7_$^;Huw?nB1*oSV#VceZD- zFG=66B(xd5vrqcgNa0ArEFbI&*(R_mZf$X}-xgfIk8gAKW1#W$)6p!$UaVxR06_Q6hVbv^Gz0>$n(^|v9V z((B1?*NV;9@ak~wW6sQhs!feH02K2;VpIvjxRY&2G#Z80-v!a`dmx`i7v$w#R&&2U zW`q{Pc8AQSyD{3OmT5tuq5U)hj^EAy!nS91b-1}XOFUNn?4y8s^qs+|lTQV;u*YW% zB!n*DRT-#(q3aOL?eaa)RV>1e2-BTXFduYcVPP%Yu^voH+=U3oMTlqkT;+LC zu+bSo7?EA4hc+Fc6yITPrW8fI^R8 z*tV?pst-oG5fw5U(%xN^s3zy=UCgVC7+l52E5M&1Ts$MH1PhV#{jtEqn%&} zS_{isBr{SBHe9+`{sD;wWJb9zXSS0Sc%OLa0?Wq1z#t(K8WhwM&912gqDMAedaZK% z4T0Ox@*ANk*2I7}~lAuKjdRfGDr(9}OCXWgQ{>?DJlE z{Th!_=$lmH)zU{bSiwceu*^&$!$zsX%&u}aR_>6}=tI{n2!XGvIr+1o2v_l0 zg1QiIyF{#OW1DEh#LOgKk)gls1|>* ziPP`;C*}{%;M&e0LYKrkcs!QGK9-eypIZC{>!$jQO2&%H3~f7=_qc2M(d#dFA8oj~X{#Y=u<}OkVq5=zpLviXTP@nGG^tu8$_D z{WPTH5SlsqP;6HuCMd@>fJ3i3^Hi#uh*hhsz;!t#|Jq+1{&;=$e7)Sfo8}}&6BW?4 zWN5C!VZ9UGwyY@@|J5Pyn<~GZ^~1d(<9uY6Aq;9Q+b^IxglD!-gk#~FH@AV=lyM>i z#^#r&7DMLs!qJJ;d!d2~$pgPDZk-Js{oqqvKYU)bD+dd39;0wrPGPXl5}c1wN`&wT z^NpQ9v1r|>3oi&Knz;Opp2xoOG@EuquZ;_z^rg0}-w5T@!sHC_%VvhhxhMZ}?Y#ZN z5w5js?nxSDrM{%=-fni1%~U@HE-vM8$0i1HYIpWGUcwv$hoD?c2HtDUm3C!*FAm1rlwyN;^`<2-?tCd&2SB5X> zugmyJT=1(SOthHZd0Lekd+6)F@gVHjIhBf`a|tYa=Nm{Nq%8A*=djo;3{2r^dK)w95i7f-eG`_c!wL_^%MAg+~ivm^Bf5j3^kI^Qn_7*|Hd^mVUsLMI-vbo}me!}P_vGu|8fLugT7ntr@Fd*rY#;al83oy10qvvv z6JdLeN0WnH?KrQCNjgCTF*>``SSvn8q@&nVCUlrreu(1;ZK*kTgTxzRQrGf{$4xkw zS)5ns3$R4j?~cn)kki`ws|SRJ z<=a<$n}n8ov78{NzXq|FwKo0PwAPJbD!D@V+(7rY?ny(eqzBE!I%sLDaeuDA3acX* zn0*K|haKdy>S9FSWEGgyq+*ul0%{%9Gi~*m0P3^V)OD zpUt^C&6bwkYvrm9F09#j6(UZS8*W}&jH8tvq4m<>A%1A*+WaEC@FkZM-DXAh3c7ydU9?0SOrP>UCfB&8_WQ=2 z(Dx@Y(-avpo1z(=XzaSOb62Fl+&bmJjM&9&>re6{K`Xr)d=E*e+-(ZGbZq`?d$6gt z;q-y5%*ff--AQTm)37=UF>iZjEWhfQr-3^X_{FZ#4KZ-P>cdnUZ9;u1Bx?9KhI}d3 z<#1w2)`Rf%>+jwqm9ffgPA|aN$Z65J#r^o+LMeNBPTmqErnXL#YQA$@32;a-F`IXe zbUsN}M!Q{W?4U`7J~Jq3~<5b7)7#@JZc3~U~=pLEKs#RIsgpLJEFn%rL$ z^D-`_JuTl$JZ4-`>x(f7qqgnU{c7e})v#p?ueMY0ZO0E#BlYgy%Tb4}r*AChB(4~< zl?&w+U=w&%p^@_}_!M41v@c8oDYzd>3^b2uf4Cka=FYH-D1)C4VT)Jq(vk?Hq4QzB zce5+Ov?3X|l;HR!$C!&QCf=DH8Y2^LXTRD&hF*fJjhAiD(dB>F$$mydj=0M_i|CK4 zNl1(-nwEf6qs1TR-056~m8Pf(qn;TdjMjU9LKld6Fc)S;#mZ=p$|nO?0=?dLxWPo! zH%GI6@UymwnEY7${&hC!|&g(>#O#;eeNdS|XXZPHS$1i$Ksib6USTi5Nx@k0!Q z$i|yQIi#mHxj8n5Js$UH#guY#DKG4Vh%=!%su$XX|Mlw^E-7cI=O4kr9Vfm2`3G?p zZ+|y;JlqvqCvAqi_d=O%=kZ_6Ff~hF2zeZ0E^@z>;uo{l@F}hV;RIW%MbgR}5Dl## z6go$L)6-D*6eVaKWZ1z`5(jxCs&G$`$FnMdM#+fNklhwC=xinG^kH3?#_H>YI+2`&Y$*c`Z|K*-- z2=iQV(+2~9pcH`|+UIBrzdP+VcWp_0nO>KfovaVt9wje7H1e=^kI!gz#=O*6L|unNe1-M^fXpqBxe9hp4p5YzT(i~3hq^2_&WMUGkzZ$;ggyoidD zyEl{x**&SO&s`U~nRSaze_Ly1KsN3CU-vnhgj?}23ZHnL9gRW08fe^ee+eZoM`JGq zU5WxKr{{^>qer}KbU0@R9X`%5b@Q+_;DD%1X&8rja@he4&#INSfa2bXexVJHwGZB~ zac*aSUUBNu3y9!R>qvPSz2HbheK)C}CJlY#YlT(S(4m%O9>=XfM+=~;j4DkYLayBB zTM9|Ot8n@r7I`r+Fl1AIC|;UuyHoOP7{mZSGd&a3u67k%|D5J>pW?nGl25nx)$}*z zNeT44`q9eYUHNp@fFW4?1mcn6R;et(sJPyCETr>j?JVPy<=$Ny=$Nd8EmT@7IA2nd zA!NYi3}QbPVO5db+E+c98U6HW5QTTpP55P6?Le zeY*W&4%vIR6)$=+Uo{0^d67Ca@;xFe^8!7+Yd&2Wsa}vv=lFi(b9ZG;fkZpQT^GV% z)9nxT^Es%gFoYCgN7q9A;CNgj$z7KMK}7zTdp{ zV@8``(NC^yz92Bv#N1t6lu0Y9b29kT@T^wC?}jCZtgHDC^5WBy7j68!3Fiy^s~_q3 z-D_nK^ho9#EFi-|-8A=l5Sp*HgO0jh{w1t+xx6i?=#jbC{aZzbWN12+KMp?^p>r@% z=~dpx_XgoWpBnEnMhSW{)}Wp_#XUK;L$f3@0@Ufexj9D~;V>yH!C<=i#QjehX;D~a zWgm0%u{?iv;=SU%D&E9+EHZebUkay~@4>eMr)UA7sLMQ_lqfOTT|Oe}DF|~QLp#U` z$j#ohH>LVhY!i+X=8V%LH`z=o)g;nb^g7+X`m-u&MZG!meK}^WA6-0OX7Dg%A2IS> zTngcFm1z|J%jRfU+8q_pDHzgmqn+@N83B#9;v(_V%QK_1_c0u|MZVHUW8|g}ipm^) z<9@B?S)I|}bU~e@^KA$lg$#53;rn#UyBIH64XLJ%U%jPvqnmzcWgjY@!``0Iy!9Z^ zfLc49%(as-XbZax75XHVY(gin4kL}JP(FP1l@mI)?Y$u3c2uU$WOHFy&PL{@|4ee4 zJ3vEhGvBW01$mRtUwPAS*{EYuU~;m~&oGL8gr4Owh`sNb8ha_a%d?26SPZMRN8jRm zq@(fWX)5VHcMZ1Rnu;ztLeX{BnIP_6#HY!^x!W~i{F>7wm6kq{RH3q&i%;UkJLX@a z7O`|2Km1N{_b0!Zh(&#BWY+L39CCa})%lRzJp{Mpcb{%4xnLswg>QnBYmy2tOE4w( zBYZh^8^>{bNDLV5rA6sI{-!90_gl!kc>LH6wMr;0*GQ=9(fxu`)U)E+Y|>;9aVxKU zg7|pof|eQcakGTM%X%#Lm)CX(Mfys6$$Mt#-I~T;aw>QpR2d=fdOi8wvdgs^h16e! zXfwW!m0o$XZHp_FD-r(MrJfiUGhu(?b?6tFoCcF;xytD&@k0S+wp1!_>#YctpOBa{ z6%=1($2i1iAd#wN4drLYNx&)X&ZET4xf`>Afu3k#P5$NbFP0CjebPu2)O?G9%>BWVf5>7O9c>;=1mwRk@N-QNM zoT;P?``^cWod{-@ant%0P`hbnN->a=$HvrT@P3(0HRtV1a_2l#e!axfbbykkwMjb1i}LR< zQby!<%4VtNaK8!-hs_hThWJj{-HxoKr0uV%NuSNf#Z!v@NuDTt@vZT?N89trH2-Hr z6i@C?6Nh(%Y*QS^$>(Pi9AIO&QM^XGT_L;NuY2jXY2^GRIwBk!s{$R*eeo8xj&fUe zz|}RI5=HKd;!4VZRNO0b>!JcHC`Hu(k9#XJt9s4R}|;hy}Gzv85u3 zUX_rk<8rQQyCVrFKDdnpUh49{@x115pi)(jI^Um1$+Z8swPA>1|oCz*WtZBiu$pn0Q@f3hDii5Ig)ANY>k^IxZ zqkbQUu|t+s^)uC$#<9{i!*jb$5$v%*c!&rx*z|x917DD3XYlL)-NLL9h=^=oH{<=^ zZ&iMfS<1rdg@g@){9P7={!zEoAA{U^O0|`#DK$O)7Z86lK$9b8tSs%XdB*MWL&+x0 z%*-udc$0`Z7P8~{dz;gan0gIK+@lV&Vac@ z;8WcqouZJ(8OldHpyNV>Sm1pKh2<*(UbKUNUm(&lh!;j7h(7!Ur&!|xL@x=LCK6%> zNjc0Q^K-Ryze^4)1+)N{V}Y;v4$wwlmXN$cp3n#w&7mr|T?3YFA z@$M;AX(AAq_K6%V7gqveL31a?75?b`DoA2z^=SocLEc&kxo#9Ifr$JTv3yp=IDrdx z)Tu)xqZLA7x{BYn1`;V5;4<}U%K?PgmbY!H%AB*rai4|{kb2#-=|20 zIJOABvq9k3TL6uBdH)7?*lJ3cXm?tAx?l-KhWy9$TN(uhA5121N5#c8>%K#S1SMy< zbp5Yhy@F2x*J}~_a-+RXMCM`zpv#9X2jCoC071>-5D*wb`QT6-fkV>;gc7|ozdsI` ziy>EUE)xi*B?Q|lsfP5KQNVF@%JeC?gyuo#qbzitmse$=$V6^*DoiY|6%2$QfJw@& zfbP%^|=5Szkk*QgdmzV*epk2?43uXt`NQzt$5)C zbi+>SCM2k&A7R~y3trH9KB*NFr9Y@;Dlm-Bw8qlt?ZV5_>j*1?I(!9Qz=wNslJsL& z+Be9k3_!3$qh)}vi=V%T=CZqvHj&U%?ax=c{8ms7esG1R^6N3m_zS2YET7{5nhk3H z%;%deEUNZ7SQWdxHxR7xWW0+~+@9m-CV`K6%}&4e2o^;Wq*xrDXn`_>Z13)6E;rYtHFfHayn5duGypD?g&4q!QVFqa3mN(VVXH$D@6HGaP(;dv9d!Va z>)rjCuivPaim2)2!-9i*!GGJ5Zi0AYE5Q+%dQkX+bC~6m+r2ZaHFNxHLj#|aOG!TS zDLFxpd_w#g9uuLGhYl{8NOln}tlk^|3j`OY=%5c|JgBIX@={i|*$;(bAHWVfpVWb~FAvnYhg&7NBm`4@NKPoC z6`QXB9g32V_Y)bJ@jUxt@Fy>NpX{iQ_T{Me8Li;tr=Z>TzCN`BY-T^arccC>P`7M9 zO#b5nS6FI?W8Of-F%hQ)_T^gDZ~fxUeNYAvHqOSXWcj02m$dlHIar7>zkTCp7{iL( z(EAwqB6-m5^p(+qTxKW-&y5ImCDc=sQ6}{DYXu7$$dNLrkM}z_KxWyiHjfZ%mhc*m zg;EQ2&zekHlIU##h+E}m(?a{F@THom(}JxVQ4oXBkB`a20IAwQ6$iq_} zTAwz(3daYd>N#riNY<)9dlYX^64oZ<6TAb(a~qb`lbEDwCao~t3OjI)A!56j2qx+C zwe&!;^oj+ZgouAoe{;M?Z9mse;4b+hU`#8;<6jR;~_HqwLu$VQS4g8HL`z1k+T(U7Uw1EhK0xXC-D z7FwB2p>$PHW$asKbY*?&DjmlrWtHKfp8-FBD&NjWKCARy=DZv9WcXqHwj<4>lNoat z5+2WOlffz_*X~W^{jhSNP1W#4J=oH%fr{dB^lj+GmwFpnKH*0OW~JP>CCWcn)JrrB zNhr+rP>-u+eu%d(L#)l@WGV9#+TuT+%@)YGh-fEE?=zw>Ug*-h{WQZBSeh2aH7|`3 zS6|UyE{35}7G~7GTF{Q>vg=l==w>6aXfr^fBUyfWEjv9<;&4PR-kv6aPj+vPzC@7z zt9+HGne0-$`se#^2;dCN1CPKWZLUn-Ut;qH^O5qxUz+-$g{fsiub>mO|Lj8I$Du`( zF#5He$2aDt4yy(TBn~%HM!B9w_q!8>4LzOFZu)YE&c776MOhw`U&h-4lw#i1hiBkZ z?3;Ttf-y|Tv|iyjb_(zPqBl^>793-suG(5;LzhC+1%uQ(t0$+MH8$YIOduI~I;h`jobBhf-eO`&i$KN>ch z4do{A3B7_}B1#|2En@D1=}fW+VQyVEaLg|Xlo(HN*T%P;G_i+rqY}QoH@YghU^!>v zl;E(@C?|)(^@xgoZ$;awiG5RzTjfqKNCiesCOkt82JPuSbf>2%zVKXgDuN|Pee3ht zYc|F49FNF-5wDu+Kt7ltD);x@Qsy9qAp28e=Fwuxx3?CINx?&;<6Ig@DjUcETt;kY;X+<1T>}N``}Ci?h$i}z z6(nbOIrldG^oC5W!me#(mES`u(+|+i#$uqGl@)m8A#*X7C^L8Jk1%VW|yU6~)0I zO_nlux|MjAIR}N;h?j;hs!$(16N#oqmWl5`c*O-&`4^blCgbJ*st3A*oyKzP>31n* z=+)Hn&&FR62d5(|3&aHfPiI#i4)xmiMI}O7PL!>VQ`(d|Nm);bPK%|Y>=R`tGLsmS zigG$?l#_^}5;2iv31um#D8>>dyUIEfW2dao=k8qB`@Vm?&-Gl_<6oCC^PBs=zu(XI z{w$BUe$9);qU2;777AT)w!4bhzbsF%4KEAXkKM22dYUB+mG#UEN{GzqGiK=oQB18C2@8n#9eO zvAADK-zhPQy(hH@U2M~oHoEop}T5Nr3n%;WE^GM$g(m7z{b@RQJS z-?~6hd#QA;%G71Q^OFDnYH zWJ>%YP`*riY2#O~=0F^(JlXSeluvh+Cnrj6G%t6~+miuPxYtlvLp7%oEPn@dKru_pNyVqV}aFgE`6SzO62U~!CnQLjdo z>mRE=Qro`N*IsVRQl-ofkoP)SO7%M%hDn-^(9F+*)>(RP1>AGL>CF|iJS zzxsop^zQGVZs&adwdmz+SV&3>IA=HQG);GP8u(*6PJ{-Jy6o0Eug{QO%yh={=-f3Vg4 z{{xP_Pu&}~iHM34C2&K2%MQ=xBwHete&pwgRNM~*3_8hviB^L5pLUiVM7e34Nl-$* zYzbh*<>=yz(46tr@F~x=I&>)fCXgO@4INyw)oKfA<1&(qtM$nAjyQ4JGDgkf4=$0B z*@)Ki7NlUO5BrJmmp{t}JedR^jqKv>YB|@gcIQ`P?Kd#MY!@IrXu4|>1?>qzI z@>;yfuFJ!R3A}FZ1t|xFmz6GXm`4uUD)umGsK&9DI|~MuG%KA5!|3+$BnUR%#oyN> zLhz{90HGL8Vr$Xb6wC~@0%<3pc7<#BdJ@}RQ5JGiJ!pPVMcyW$T#Iws9!~4-+h>WH z)=B-BX}fS@9I7lFd@I8n1X8SR$8ekRSVX?vab?HxCWJ2>Kp}X@A?~guDis%QFL`qv zHqt*mtU}hl4Ult)2wq3$)+RHj+UGa z#)Lr)bla&j1>Ws%lS{q@xHQR5)>8|Oqf~Y?YzNh9pV(Lx%J$MX#pCy`R#@9fWu?F( z&(KZ1Pj)EcR9?z53S`P_)H{--HX=yTdizx}>W)reYlRkNQoa6}aY}k;2Die2-8vkc z+DvQp9*ZJ=YMGcFu-#H<%Q9&(jTv#cJdz-tLZf*{2vvWMGdlS^*cB-@l=phz8!#B2 z9V}7EgXs45UA1-V_Ntwm`nJ9Q2A(RRTJcc`)&2qP&%t}oC}RWF^I?tS82J$qbQxs; zS5bVd#)<6de7V3~GX2|gcddY*XAY*dkl9a8A11`8_wdcD}TT>M58=)$(In(NhH!$b{k0JEl_6E zMXFxBPKxKjlHWo6Z*eJb9ghe_fk7vmypviPHiyDMWD#3#VKauyw_OKi;;9UZbKx}rR={k}rdeXYw6 z*=aQ5jBF-(b+qKnn^}D7RNK(eGhNbq`+bhDr<@t9e}^fGCyW#Fum?x;z}DZ7IGoj^ z&ES8#*AFq%g0dL)SG3{|`pUVD#VURHYjL-nA#<1e(K}XKr*R#{=>x{jpWh@>rZT?L z@VSeB97B9&EnPdN!-@o*-rnVtsGe*FowO9r7yc9>B37Q_2RA_2FdBe!oPDkufg;8B zp$-_?P+`_(P?>61Xsb02I2U8n3fWc7rbTu$-*k(mX3+D@O?&8IIbvpmUMULrT`pa3REf@bBq2ywujba~ zeq#3vUF?g7kC^SPXJ_K@WYIe|3kwU!APqa$$1##n0vXZgh6( z^UIh$5d836q)jOo>7r=7uUaQ zA?$l@M~h!d?u6!N{xDOqk!qJ?nIt%^%Nt{SVscB=;BiR{RAUZzcrx!3)@oBc>%)KT7Hrgpp^D z);4*PTl46vTdHMzME?Ye1)Djt9m_WUdvhNk$absuqRLhaQZfZZA{)JnBd_Ih6atd1 zT-kGd>{hM|n+F@npyJeo%d$3x;stk1yF?S-niz=M9jQ-ijLcH}8hTzB3bq@Gdn8j$ zqde0HQI9^?GZBs{G1JY6=>$mLVhBF??-~yOwc| zzaLY!hI|A!$BH|*yHw9k0X62i$23kf=)jXF4{J1!UA9)?3-=e-2Ui`Btrobeaci>; z3)X+(n|JImAHAmlUpG@H#gTgkE}K)wC$YhfRPsO{=qa4=@aW;gmz@BPsgF#hu@(iW%^M$OAz-IkrD( zV1Tfc>BDYvm9iFg#qyJu{q)p}S2(-NTa5yfn3U9CP5kf9Y+^74)>LOoJpAVV;q*RNnc}~`xPIBVWkQFDN z56ts)Au0e(Ha$m?@^;97tk)%7p5@Z8t#(@V=%Zy?uLk$nZ>AVIa;%UsZzwG)Yhrxe4d>Du9?`Am z1gEP09znfqc#LCXyX?q>bn@ zZ?P-$Pr^ZGjV@5O?<$Zr5gG-yHcyB#*(lYtXgGeEJnuk!!Uy@W6vg*Pk1O2&X`x*} z{aSRDX8qT`v%OC+KBihdY84u8=OC>xE7tj1mKu0QoE7U}!;&+DRLcd66Bg;U+SgkcYCOE|L>i(9 z>AzLDrD|(W3&$$ScWF69o$Wa4u+GybLCZ#@*ah&!8cdQSae_75d^;5CxnX~-6Vil9 zD9dOlSY+w9_>j_h#AMMAGyjvoq^u1h>MJ~z!M7y<+LQc;o@lLnD--p$gK=eL>4bug)@3 z7xW#n1CR;OU&UCwEyE7c|G86?_)nLN m|1~cC-+Rh`_Y(FEvvb1TDQ#)K^Hl`!^ZOnn{U^GVi~j;fRoWi_ diff --git a/analysis/figures/F5_feature_diagnostics.png b/analysis/figures/F5_feature_diagnostics.png index 6e664456334a0c7fa3c7a9cec02d1153a98cc139..27aaec2c48be587cf3d6e589d5d1e99b33a28145 100644 GIT binary patch literal 89252 zcmcG$XHZjZ)HNJHR8VZ7AV^UZ5fqRr300cXJ0b!K(tDTQlqQG@h?LNKi?q-LL{Joj z&;x{`v;ZMQS^|W?cjdmH`%E=I!q6=I!iY#p7$^>E+<& zDkgOIuF&1vJoet+?p{z~VVD1Xhmf16ov`O5-7feR2KNW1UJwZD1?oQH7hn{K8@e6C z^6r|}E&W1`>VX21&|Wr*a%dM?n=$>|K=RdROYGW1Q5Z}N_K?^=g_T2TT>=kV@az9R z&RzJaRBQU*E6B+Xw*Q`oox8?H_uq?i_m6=^_@Br9n5U=zd#-Tud^|Yle;+5WM%?}H zIpnIc2zAB&dGNeE`Tf7=keBEFAOFaGtNrbjqNg1Zwuc9M8v#+QLQm`ymAaWZ+IHto zwY16x61#ec5J-%U@aMk~w~We?n|$%??I-{DAn(sjHF+%|+7Cs+Cu7iJK^t=_fmfhe z<097IPcwtni@S8k@xDucrnKZ=F{iPLw}exWx8hoG=qS>mRKHNv`g?pezMqK4%#%Nb zWt{o?bCds%cBuqu!Ti+|2zZ{&|9RH z@oP5p2=9Q_hC^*I@NyHxcTz5m6g+t&_}_JWsSnN(&p$h!qYzIk+Ls}fn1ZK)oXbuQ zPrGd@I$Y;q5OQ82=H2CguHeZg%TUT_o#SZFls^_;q?*9K1oxY{i$++UZOH1$lc8)6 z7y05Ri(YKZEv>f@BvG=YOzEIq_eLMLsm5n*2b&tO^akgNnj))jaZ7)qxsfzUKbGP> zP+U!8m@^{|KD~Du#^=4rM2uTsLx@q&KF=ntgC}FofGqrGYn&HMO`zGt1TE|?5ixQ@1CTi^{S;gQS z(^}S>d!ICUej{u_+-rxiYu((T&XHI4ovR+#m#vJJr%p7uW!rHETj0E=X6*{MGUbvu zJENFtdy|_sDWus9GNFAInc~(?LHe&0rmlH`E4>jY6M5n+5_NP)Cj6PK?@1BN$dvY) z^xX&VZ|LQaa$R%q%X%>v`+!7v)E3mVGF155MQKEPq|~7M^Zsf%jmC%LURla%Q0|c3 zTo+6AWd3IuHhBJYi(kgFXFGTOJZai5+lA!KHKGdZE}l1mRbsh94?MgUew=T|yk5Pl zy7;6i(WfJV-j}=I$;wGvrYl_JpK!SUJ3mp5aoEKUVbS_kzfj$#-B12;gDbxN5Jb|~ z%6S=2@D)VTY-nKRY=hfu81CqA=h>{CHVV1DOrTM7Yd*y3un`L{*pBYpF zhtyYay)F`eN1hG1gB!2TpS6CFCF4I17IBS~d8~GJeY%A_?F#<%897?4ll=*tWMG!t zSVdZJ$V83JvxX_p*Rw9*k2lxAGFfyRS6hAiRAE{hCA+_z|7Q$Ma9f{h%JPvvI`C>a z0MW?cmo-*lYMqUnZ4Z4J$_=g<)N4ueQN82nLCB%)kGCDQ7OyYeS*y`V6?%n<+M23& zLd*1yWI0X@>&WBl`>rZQy`+GEBa(&8n_pta7JE}WYo7nS;ERR(1>BJb!Tbok{Av-h zXG8V^!AjxsO5)SHHr*byzfw$tS!BE@Fg_pC?+^7`e=1+oMRkUmu@gV>VKK=201LZ4P3Oy`3>gm)h zxTI3ItmUt+Smr*v>|k59;t+6NNMf7n@T%;rFUCbuP^Fs-PI<#F(PMk=Cl1l>iftXV z-dP>BzyyG(!|$0ldzbdT96V_a?^WXt9_g3JpEv*fqFPohJ7_cZXNq8|xd~?5XLI49 zSRm0E)?-Rg$@hi%uGu#fzZ>jdcCp#EXQ`hlMv!I#QP*{{rrP`<590Po1ULC|kk;J9 zZj6NRN9tuK(lfHeY!F8!`K>ooEkmmD5u)m&As@hEYGz2@nUb;Y*&o2RB>kD|k7GD> zYOBuaF+weoi%4$r!Ujxo);GMvTaFL~tmn%SZK52K&LZFlKH(E5c_|f(gVH)EIdVm- zP>E@6k*sQc#$&gxnDa>U)sfO9ljC}Z3Ec8Cr@7^(YC~;8v!uPgkCdCBEXV|u(d@qfo z=j&eyw;37oli?yQ*mW@sC2PTbdxmL7jH`pp`zI@u5YuUP{>aLR|s+;*9)^9=@9B>8Ht^VeT#V$r?J$v8BkTirHZp3p|F3Jw`6cxcfJ z{^XXpvOJh?(VOxxO+GXPVMBM~OdV#t(ZjkCuZ9fe#BDD2x_mFGQDbBka>o4q?7uaM zUfI!;)nOI3IIRJ;)T5v8Z**2!w6eRlc0``QxFF11ByX2xlEfTG_z{bsp#FW15Wo!2 z?)3h6$1B^L=M}FfZL)67dDlw05s$7g9hUP0m5o1}uB$THbKX`45DDD=&bYj3MUEIMc)nU6%bxr_QJ2RT;|dPu6ZMNN1DUnP>xin6n#cg$r`M(uPgK572cl47r(<#Xd;~_ zWN;SE-amy*VC-gObO`)c`j5$tr>Nn6ILyWNCrFRhSmzBFC-`li`ucP-A5@%_UP7SunqTIa&tsx4dlOlM6AEj($I@OSr5JQAQ%&>~P$PF4=$KY1rGx zO`bN5(4C%7_d*UDZ*nLjU8fz_n{h&0V`gstmF0`!P^potBlS)u6g7E;ol6t7iR&};s1Xo< zY*J_#6v@Hpyfy(UilA48dCQosa0&pfc8vv@D7r7 zi{n(0j>*mdh~Rns)~6)L3pCkNq+^9{zRmv`lz-o*bL6L_&#K`TC`*B=JmpUk`ioVS z$bxEo0xtOsuT=;SIsFAxir$F&h&MTWU0xk6$MoOSTF$(oacS!ZZ=&+nG_As1v4S5Q z>f`MG_;RKrS~=|u<4Zr>*qtXo%R!|dN7>wbI#X?}h7f(Er&n%#$DrcTT@0~5Lke{O zkQkCBjrqC98yf&HkjfvE!dEft6_)LCJp?ndu81OGnqirLIX@n~oORRYnE=KUBrM;z zuS6x_NMpH{9(1ATud3YHx`=6FzR9nitMbu z5EW&5eJa^1v>(hPe)PF5Ae-0x7hu9?@#@9=llsMD z20F4fsCP?@1GpUv?|5F-_`hT*y?JtO0cFLnNLz-!G3pHGs>Ey<47LQ4{}RAyb(hrx z2#~kQD`IAaq8ERdJXNMSQ?JLkEcjq<09!6+XSA-D-M(6YR=UxZ2Rn-8=iLOEzbKHj zj$}y}3yoI_zeFE+Q18@##hx*W|3{8*r&Q+Od5IddEFzX)H>cukLfL4^46oXokF|EF$+Z-;MPQ%y zZEmwmpsDNi+r!oA|5f2N^I7xXTB`53&iQ(^tz1dtM$1SCi}~vKpv>+4GG`teBtQ8k zTWi2h4QJ5mh`xwQ8kDhaIn(hCC;juF(uK*h@&`{NO$BoENR`3udrKKhd)RDJ3%3?y zr6y0SG5J_nfp+8>!%e-*+>XPBjs|!o0POpyK-g3#8v**anj0<29VHcPT5JfSUUvCl z;iieB)J1na)qJVoDulk(Jnj37S5s;BWeBD~Mc|BFZ%rJ|od3%8}CrEX!K<0H9wkTBJ#vLuxo z_eQGo-yy>JKCn&~Im#voP>Y*+5#%M5O!3i{#yN>swx@3iQ1Y-TF7wGJH)N49Z_Id8 z+X42+l+L+%21?Ou8saPwJ5qPO42d`mxCEo_ip;(6?3gh>nAcK#-(S6M%N$b9U?a$G z#nVMzGc3P7drFVxow?bkHc+C+E9SNI>trGw0eCsMa=LAdi&4RUEt`Jv%Q0N)c!LdHW+h>F-ZbM#|mn zDeb4ag#U+tIFFZw|I#OoSOSKGFZk55a69~3(C>0~)Cp|aiRL?K}>HNub44;c06 zUlxJPoyirYE-U2PC)+;uj6ymBPrP~!zo4qARH1rG4YML0Fdtv?9w zK9J*i&2*LvK71PYkoy&ZO2pH&aQ7!q+^{EGeg!qCl@= zm$;avBAPimo1QQpYP+HcF9w(|UDT%MMTo_xr{VC=Vl}Ft0ILvtNc!Y3%xC6hXs7|x zWg1ZM&dOv^_FEf&23pPYgzlG(#i#NBuQ3^dh_!*IuQR0Sx2VOlJ;OD?#6`V9AHCtL z8M`5KZR2sx^VmJmjNlhCjLMA@2%R*Lt}lN&qw!xZq#zig)YD6=j{5F3vqRqQ@*5a< zHK=@-jlOa&^+!BSL=FQi@@~MkJ>qFR)FJhb*`+^D;_uLX=i+IsBKb~uAI>_WtKh$d z-%wF(R_LKWIT&mv{82mKlWp+lsvF3ym@Yh(odDZi5^Py22^E_uc2D4u6z+#3L}9K# zK(JcRf&S#hh`S0bh9$Zkg5z_7173d9KB>1%sxoJ>qPD+&5WcX%@~QwKL3}dcey#eW zv{w-&0A$s4&=+=-{`_;XAw1#W%as)~Bwc>xC`RT5^*=^BJ4Bu3?r(Y%Mk~_sA(HMw z9EkbukV?QJ1aVUtLGZXB0)F^Ho!hJ&EKVK(7|AaSIJ9M`M2j!JW>$XpcjTWMKO$k) ztrtNb&>A>bC*d^qk@=?fxF?k(IrVYJb7km;Sn$3#Ik?=o^4hOAgTA3;yp-3$u1AYS z?)_I0S=M|4-7FW=g%_v|(qgW-e;^1#XC{3TdDBm5dOHNvb~o)T83*+673Gc(apOpr zjJ^3p`8AJ)emKUkRR3GQ+Cm}{gyJlQz#XuB6@YT~7J?6MY+@@v*cf3k{Q=_%^Im>C z@!7xDTn+n#Jn$0vj)5552Zy2C&QcqX&r5qh)5MA$MoRR)kaFgS0tXHh1bS0IeFhhW zs|69o#G(2kel}<;K{}9F(P!ngbU7~O4gO$nD=U@;0D_f^n$AN7u|GfDF{?wpLVy-x z{HO&GCzGVd0O@4JVhIOoi!!UbsY2%Lf2NEAvL~b3_v|9}WX0WPrkn!7p=VDtdgMF8} zHs)ZD%MZ=BzSzj4IFHRnh6wL-3=_TA!gtrrlJ4MSafF9nrzGqLzTn(M1jz1(A=`2X z6z&Hgk+0dgXriu$WkenlMpb>52RaDK1WbdQloPtD-lfrv{HKck2y_vXs5SC>3yJ{k z_w}ySuX+9uwOMLVoDSaQAEKQmT8Q-$`q)Ij(^QJ*k_{L^dh9zhAvMvX zTl(Ygi&S1;swYSY7@n|1XYR)vXB0Z6mJE>xrhNBLkL`S(X|Ljr%ypXH+rnZ9fLC&^ zu|uuxaT=;8Wh99AHK>W)y}F-Hb}40g2BMHi~2sG``*A zQoK$w28={vAz%a*r?*;E#JTH`#fQwH-c$RO zCi&24n@u*rJGmcxx_%7@;^ae4|0dv|);w+iZf|XOWByN8J6Jne!X6+q7)S>I3cF>u zfO;s~E9pA5?BACidU$yZgxMz$$^IW!-ho8=iCXgscRA&P#%pKZaKa9%@)z$st~;rb zCDVlV-QhQ`cr?E7GpThfapf+()7O^_&j7&PoOF$2x%Cz5yT*oeLJOg4s6d{81JQYT ztj=NBouJX4ZQGariUN*+h9qm(pZN*4-pD((_aIv?cpT(|uT@W)F04zeZ`Jq_ECMs3 z79<=s`iX!Xn%GI?l4Yk-V^Od+K(|ehSRWY9nt=V?*?N|wbfj}Q)0rh3_!ab}S3Zc^ zsH19XY{sc-c*kW(p~b~ee;-LlLoUdhC$R_^@~>5QbDe6`p8uYJ$A(W{Fy}bl%ryetzjAf~bV@b2ZhfwntU#6CTow8XxT*iM$ju6OM8eNrvt}^HN`=~!#o}2hf47EGmK*SV0WX5D2 zpYM~JrP2!tXvy_f(o7h(C$&mtbfy3t@)Oty(gpL$pjca_IRV6-zeNL}sok?TnuSJv zcp*GXBSIJNJRY%lePZadzZW^JBCrpf__2cnG_+ajy z+$0Q^$I&2Qd-?na&H;=9^b8s7&rU$cVm^T_{c^g|gLkICm5BF#S?bk~2Wo@`ytFltbcVF=I9fJ=ol-pQ zSM0pExrj%9eRT@MAt#Z%`rA7zTiiiEIr1(v^s<=k_LRs6K@%a+&-oIM6Z>I=y51n+ z+g1UZ9WMZz)C%A_sc^l`&q9qx%AGy z`U>d4g%!;)rZ%XEw>H>Qa}##c%|4w~<}H{0elDqfV(BwqfgS@Cw^wf7>vAA#Oe&;? zRDtdL2=up|jw2<1no0)tKEK$os`!jZwFtPi7Jh>3@0ASKg^T>2TS4XSb1x90=3Ms|{{o1=npra)8hwVd1_g6zaI@2OrnlafUJg=S@oEI|nuX$KdcILeRLFMIR!gml zq!#YLJN~%JhRH=o7$|6j2C=28mi0|%%k0wqDFaw3{vKhp;SNcy!?74YBpW3MtF`I1 zO)b51YH8-H1IcG3fVAo9$vx$~&&J5FRV4wn3R< z5hIaf3G=Sy`Lp&`?j%RZ~M?&~(A zTvP?&k451Awu8=5f>(AGC?L-3$l-8BFTHk5$BOs+hUGK^k{cxUboAwggZuqlaU|2j)l@fMf zB&P571u?^Embv%oGjjhFWy332ed3xO%=&*kNQfCx`KUzj0BlSarOZoiz<=;n7~xxF z#H;?3EV78#g`mvsQecCfq@BCLmf_Gty4Dl2BOUXE`|7Mw0c}8032DebpSf>&nrJ1I zS3wtQ zXE1ibAp5p5BMKe^5_{xI+72YFoZ4{6qq~ z1Y_eHFMM8wPOvK3BJ#Vj$al+(V$}7U53VeWY15=tnl(hV2#v)&veSJ-4$#c(I1lJ% zQig2)R-RMY4_em!&)yMZpqe+CeClRMC!hJmn9PYl9t{<00CM^A+G*93+k53;|2_lt z)){M%^ECc==kgQ&$m5W6!Oe&vyY}N7^qb!~+{RJXI}cZCO-LVdPkvwiqe1qr!O-nU zl~N3jA~M!%sd8b{&Ecn~K^wcB8Nqs3V>jaK2vHU>x+DI=5VOBMtkb$q$_#(WICe}a zIB+$TA*zZ#lkB~7xhXWe=lYNrD8oAe3HyoSU+&YP-M4tz=#FhAUU+!Ufq*8uHut3`*tBGV@b_K}zI`#sJV0@=(ZBdPM#0OWsLh;}e+Y#&@}_V7&& z+L9&e<*WLu-aNi4aGtS-dI{%negKWxi9 z!)7!6%}<1HfS|~!qV4bjaKcxAhip%j_$2qNXVe$O*N$I+yL?}e-94_nO(u5VIx1XTzGP{XhlpHZKxgAX$j$F z>0k?9cMLYqCp%%2;S^A4>8G+)TODO@@{X8adPKODhd!=^Rm2Sau(D+Qx-4z8*Y#2r z-`s1m-y&)FNgZx~p59tH`wwGak^`22VoX>Qfs(7kV>s;k*zS>z>bv2*T*h|awUFUCc z7k$(CO8xlGgw7{O^(K|JZ zE;H)Y&RIh7Cdyp+d(d=K^wgN&Y_MHBz%FCj-t3&!!J3;O|M(-}0P(t`oUA4Rw4x7c z4Iym;oOwL+UScln^YQMe zWoI!yLWfq;Z5#=YO45;YXb0eT>;Uk8A?QJ6Qo6f8ac$xs(mV7-zyaXa-W_s|RGDK~ z_;qZef?&^tSJZXgl6F z_Xi+T6F58Q+74f_TKs)%xh=g~mw3Kak_{+=pDZa_&1<*4#$u}gu*M+#raYo&XHmdB z@Qa!vZ(6>I4?zLr0}09v2^)8Tjrvl6>WXx`1v>eSA>UQ5v4#XMIkJxoa2YV}?o;81 zQ}C>>)BwfHxeg93Ud6>YVnCs~v^8pCnFccOMGL9643|f;q{hq zg#`I<*L$F^aHhNG_%xHF?MI0Ry02+DNX5jNE6JV$JTJ$bA8-Om`}A(x$lG2Z@Bael z_Q!OcETNSyP9%dY=s$Kur4dK~{4+JWaROqu^cUc()?O)qTJW_pN!U`Z9pOXE44NxZ zfH33TIZ6Jnf0E;>F5FO8-u)brw>myr{h2uFlto!}3k97BksK3YJ9v=;dp~>%Xy_lQ z&1`#<0!u}#TX@J>K)cjf0+ir5tCr*S`#>HK*tM<4K0Pbf%xUE5E2vzIeCoCz_XI)q z?Y|LL30f_hM2(}@`)jhkYac10l_rQRtTbgEv`s!@u2c8OejGni1WkSc40>=pnipnS zU7Z-yHAW3C2ZeB`5w(dX`4`XsSa~p%!)@MD+Bpx}lFhoLYQ=2p=C%}21Ji*SBak?Y zu#`vmjQ(7_qNB`WuLD|orboD1P4srZ6Cjah*toZVw)qeEjtEOusD&9FrRkB?`|OA<0XO( z5oq2#;?T5jnf$dP(0;e;7i+r^f=9+mI1|c|29fNe!mKGs1N2$;Cs3`k5T1NR;IZ)y+AE@#8vSzPl?;ws^MafEjin?qaQ1GuqBAZ z#1HRXx}0#q51LD76LySOGVvjhD6yuN`CPON+nE*!o>=})#l)xmuLy0tJxc@Meu1TR z6te~{sd+!R!i3hKtlM3G%fZFx(Z`1JR89rv7T!LYAmmB=y7}XXiY4e$tyz{&T`$5~ z0ozYez|QItXz8=wm4S3y+ipNVp?p-^^C*U2m)V4t!j(BM4}`nca|!VwH-R)*hcPIg zrjq!=#-kmDuT5mHn^Zla8kHoe>NQ-qBd%Z4dXsDF+J1%mVfMxN4N9odO)6!8CpEtn z2BaY+LpwvVl2S2t)HsuLY~L(q?dZBB@%+1SC}?fDDZ=HIo+r|##Y>bp)%-UXHlA3n zda2=eKHjx5 zdg{i2U0#jRNXWhLoDyGS2g7>kLGs?5)7p5oFC6ee^sQPxSIU?NRt)JN$_cbw>;)|R zIl0);)O;Ew=rbF7xAzc%X#gKgkY^N(MDmm(tAjAw)0mQ>E@mB>^MJ#lT%3$rANJv{ zK;EmuXdA_P&?>$g{2dK6dyLm1sJ#=nQvrc6q4<0i?d7IKBWfdf3wJ~nJ8rv8)|I9~ zeadF7hl0z9iWq+?5bEOfdubu1tH<>H^kvcOs${ua7~kVWfnBHb7HdT{Y|xqU#->!^lV5H1ZMyfUy;3)fNInlgVO}Tc zJT6j#Zj5}%?ov@>D)253#h>!`alKL*- z4=Pu`tD7cki}Y{w3b((Yy)k&6l^jz^?aBrvT$Sk_P(4tXJ%CDC5sEyFQvz)HJh3n~ z=9H(A?LH|s^Uo1Z?d;)YL}R-AAZfHtn~my0eM%c#uA@zoP22V@be1+@Y!s@;)V0ih zAN~RwVspMljR>VvvItnYA?7jH#m^MveioRtx8k)+dCL(t4X_gu z-Ylh?qbBaI=~8?gSGLxQD5(qWhxR+oHRDH`PVt?~7wvTvo+ddoXFHFSyt{)Y?w~CL z&C98NG9-H>tz3`5sEN7Va)>&e5VkJ^OFB$JiaGF5FJ8hq-(v#w7uEZhm8N5Er4y2Q zyilR}gp2Wd(S_=O502UqMn25yJ}(bk)U6rf0Bl^dJIS6L%R8%? zCXXwiBZ0)V%iC=}R~OdI^7%3~&ssKzkDa{4U|SV4elCf0QJsAUafXw^iE}d#-P&&% zE9rZxy3*aH=lprsg5jt))Ehnl*N;_<>@BiqZ4PTP?T)G~EZ;xyoS968~%$cZ~ zsmk)yM~->>GXWbJL6@GdlXjW?`gQGzlo+UEGW^=1YF3$2S1WhD7rTGO8)kEk7N=eq zUGd0Z-eAS~;5NZdHME&_@tdutVjMcgg5&hfb92d0JPa;Xsa_Op5fkwK6@Y+R)YBb` z;>p`9)*Ab(e#>%p+8H@Rw)+<7j<)bC=GMxRCxP#>7HN=O-{w+0d3u`USY=h*DG##v zhbB_K!rMu08=W%LGC&z5jBnX>ZyrYNWjRB93-djX1~@=1;Eke`(d-Lr3Xw~F8PR4% zLx)s-IGBmOap{<~;Tuhop#TRW{7g@#PlLT5-+n*H7J)&H$VKb zA%|?uklu!m(6wf^ODU&MCx5-N=I<>eHyX}TVo{}v2k`w=Y{XLlzND%Zd5cFk<~4~` zs487oG2~1qa02!LOAGTS*@4*@*T$(Wm~`lg^EM6AHe-2$MSzN*xwdb1>f0<$SXaj_ ziMeD!Zv)JHa{EvYLyCfjydhDe!;@t9R-M1B_I}4p1++qjl-nD3SDD%Q;rj67?0fx8 zE1KyQ=fC$uSVJ>6EwoG7NIT^g94Y<`6us!glG^Un8ky13g$m}=t*Se_JA0h`=jNYl zJBCFN2r3f=dEuw_CuU_V_or1PGC&vVk10hh0=PUFf{^A8md}URq^yfh?8_K`QHoTd zGp|)E#2X%u7RAAl!@&q#9h`ASS$tVI4{3l%D49-hz%;o|Lo@C&ryz&9Zw}1$x9PQ% zZT6b(mOd6F1ysw-Tt}4zN#m8%O>xy1;aZ7>T6Hi(}$F9+*jkQ)& zD5tJQu`*ayWmYYNL|;RTI-cwB!P`$6M{BvRpPqM20 zoO5P;svUecb$@GM_46w;-!wf{vCFvYf7W$X+)&GgL@*UEd*rZ|*GySwsDiiC1-Gr# zeouu8OUw!ISv8-X~6EU|@7-^GE`lG6fa~LU~v`RVdQ&KSp)ItZbx}Awd@*ci!72mfoL^5=}6aGvahu6^Y{yYzDka zNT_iC4;v=vAgZmN;&ZkDIW}AH6B72r;cfCd!$CdV(U0Y&$;yk+?7{uDOZ3*6WZnmn3c#vdLbEs%1_6jMqncREL3mD!Y-KkO^8yXbbt!kl?oHfAZ;=bot? zdkj|iccJgBX}1<#Vm6&1|~Lr=FiT`a?SZO!BKa@nai` z32@h!7sE-c_Ji~az?d!*eg=zeEZ8dEHjx-=ZW@y|s3d?^4)fiAwk`EyrdwuziAA2yRwS?Kw zQsV13hq{!@`RHKaXk7vl84|t(y+lcHNRAFEUVUWT%FI(%dXO zC|ipX8Zu1pyKK(QDR4e->rcTg9PQolQ+>PlSktkp;o*R9PX~TV|C$%xxfQbrXzq4x z+dd;v2Aaeevim2RxI2Z5#p|+KXu=J$Xxnq+N=cip%NwST%_{&fX)Av@s zD`yuos+Y9t7N=y9nATPa#yC!(=5f`kwycOflG5J~9D0CI2tCmew<0(4up|ST@OQ|m z+^}@+)~nw_=FA+_s4j4x11;h>3hL)zaq~x>^uv)-X8Ye^a7g=t=%P@%zeK>kH2g+5hRZN9BMJvw#tpMTY=i z;4{FooO&Tg-ode!ceBL5W)+T-1#W?m#8M!#J_9Cc(m~Mjz}0pP2yE`Jz{SV%*X@RB z>m$1q@HZ{MIwkIDjGAKT+N2g_>zY1YDU>`6XoN(V;u})Y4}vl^~Z5%m!fu3Lp-V zIK#@v62`OW_ct`Y4uN3{R|?a5G>K-IeIg9S&H)GX zGawl4w1fZ(zmqy&K|U(U26Q=>Yj#Jbnlupm&6nKdp+%s<4hSJq0^@oJe?S}c?5yne zTZ$!J7YZ5(x$?+3V;C6%YfabMYq>|G!p+~{wBU#{b zMxZ$FlmklOC4vI92z&r=ux~s*1VI}G{1QC`Y|LagKq1i>306{dE0<^T& z4+zs*zO^^7^}j8`^|#KO8#qrlOH_bwI*)`kYsdToLVahtlpELjPV+3)qtv+L6I_G^ zBPV{SnvMDH6EbJ8=YL_{fZZG4ywABW_WZ{=@&Q0o*QPH)zO*OG?f(e*unL{p051E0 z_2Fxx4nx;lEU3HfLn4Qi&~2a9kr4SKVCClY1yYtf2K4Jm!N-6zn3=o}WQ@Q&=wOA| z3pWndHSpDw8ww*g3viJNE=EAz@5H0Hr;Dj`JZ=~u=We>=aBXTRoE$n=Qc`!;V&=0y zp3sE`9PO14!1K|LXaY)mr*`JO^Xn48FZc;{1Pqz)+JvvWKK~9;bYD6LJhr|$=dIcI zD)^gKuiJ~asfzO(+}8_EkV`CoILe#HFfCrD%gf^J+<(l?RlsX8`H6vQoIE?z@66&D z^(qI>!Vv+3A6oe(a5Q;r)Vt(1m}IbL$x~(G&rsv7lHdlkf_+=YxJ=R0+zZp9)C+O-! zn^hH>ZkCQL-%0QtTDD7w8+*W?P&T&V5DYh&wT{#LH}J{w<>#GyjP*+|Clu)O!4E@{ zCob9dUrG=dUY1Oe9l47ml~5cCQbKgt(-sf7>ROi%623p@X5TA(HvuT6*#=cB*p0m) z4RqHw*aDg2R)(fs`$v9D1_qLuawb#x59!R0ZqbeVosmfJb(wCi=aeo-U9br(xnmT_ z!QZQj*`%EI8n#Q4d!5O!%GtAPuvDPBDG;cW_R>+%+bw6XvMmYSa%))kvc!JSh)>u$ z!*)}twUH@vuyzYJ;FNNA<=|&WdGS0`eDkC?T;9)LBYILymXO7vL*nVT^XJhv)>O`P z{-|Jc-zY@2oz7g47LunHEZk$Ihbbk?>v80vJnU+A6-FdWrZ)iLO3FWdsUF^}=3VAc zf;Ti{wKqGjXB=O(XjI*$HOdLn{beYX8k?_!w~-iJ9or38%|_Y15oLjr)7q)a((xUdo1UlUKw!qU+; zz*JGh9m`Yj1%b}R1&u(v56u0sj(NL#!$6s@?JOcijZ5)|9b}(m!Y$(dj-N{&puH#S$1f@oxj8s$De=FLoMG^{6yUL+C-AU zS%|_FcBpp=Cj+}Z*M+GohDVmXfxH|KnII=!wJ+I(*}LAIngU&76j$I{^{yOC7fnwx zjl$`YCtaVObvplKDr_lROIB|yJ@%eY+o+UV3(7iJP;n9tG~H?7d!D%>?b-Di%;g}j zz=vMZr3izOiridJ9>>>4j>g|E8@^$dVok}>bI|k=*c0rag~YIHiON6wdMraaw~b!T z%uP<99dV%)_!6(R&t8uZwSkM7Eg`tvAXecFZ}(rl$QaZ#bfq~4E<)F;%(|4waEU;N*XmcEY% z%{GQQ%yU7$3J-zg;E#oMUAGD1hMY42Q(nD4!R%_|Lw6VO4wsvd>wMrG{(Jo2|Hyxz z_&*HECW|I>P>26B~bK^w7DkG2p4AgO#09J$u!Li|~347hL2 z_fvo3!5nan*ai^S93uH;1GYs@oVg+*Xk78h613SKUX*>GEui*HB09jVUMy1=aQBSe ze$qH_04zDafsAYXCHbN@`=&IPsq6ZH@!yf+hf_N8l;?;0IAK_JFIWI;{h|X)+F+`9 z0Yo2X)S|EQ|q_tZBf%HHqkX*#sO*XfH6=gA9iQW0n6H_2j}bIKBRfj)zl#5yzty zI?mfQ)ml6U^FMCA>8-%qMjftQioKa7(&V{Vb6E_yLB*+F>dPWe!*1#2Dc^281<4h4 zpFCO$jnlB6G?9GR0XFs;HeX#CkiZqfOC#1 zja};!&`0ZO<}hyU)81?2gH-P?4xNjV4dz=1(<}Xbh1U8IRo=h`7mO`-%@Nc=^9Db{ zjAZ^fNLb@$=UMrWpk?|9fOZA|&XeJW{X6ZG{UEP>t$0-ZPplL7!1c=jf(!-=k0-t3 zRhvY5f*~m+<<6r|In;3mm*XYJ9Voz9`_GU>!myLv-U2g5NT5^2#SZ8ct#2OUO7I6j zM03WJ2451kn2HDEmInZ}6YI4xrSS(q)wmHS&$jf>#Qgh4s);OQl<6ZFnX+sBv23fw znb{mWT;k@VU$$ic{{Re=cR~9h1!Vmg(*~ica=XAb~%KAHet_R*b{dCJJb|1?q#0DMX6d?j!V*DGd&s+R#SdcYeeAY4gyx`Ba5 z;zwXVf6f*3SCJ!ZkO+h|S9#QAgPR*S&`sx9PeEds9XY5TCYXXcK7;lKqxUk@T--hP zVG+q$0x1WqJ8ppsU|5q9yD-Z>jEw28#G?2MVYU`$JN zuhxS3JB$#>9b?GI^lW_>EU+5Ob{BqmJ7(G*(i+_QM$2@kxp)92?YZy?<2l*i4`nZd~TEPhVS`D>c=pm`P+sbfpqMccKyUp(3KUvwp#Ow&HqE=#0 zGb=9o+`m<;D(|FR%?l0Bvg>G=Rf z!^#=<^WW&ZSA07RdWABm^>O#Z48sEG0f|mPtE0#fZu^bVg?-d7ev`B@=hk0NRW3uK zq{G?tM3nPAQaZ4{z@HtBX;V*g5U*025Q#p^jV#og694)F*!_^7dw|EvmZ~o{gPSV2 znswN+(?ZVmAZvliFiRo69l1p{u^qEgRy>!oa=Hkw35HeruJ|}l*8RL9^I`xA*XH@| z084pQ_KUS10n9WEKrj8&**D<({-4@(-2)hZMS%H5N}jTvbmwx;g=6b_F3smP-d%ge zsj^;_A(C(~wl5(6`v-9SQJ~;{4QAJV1lJZ0y0Y6~EzUx$RzZI9?f;} z&%n$#9&Z$FAY|xLJUlj2;c~Ns?fR_~s4J43MZ|bQC#WY6(QZ2A2_9Grir}#C(-wexfw4{_Yk}GG&yxROaCnWk z(38gB4lo~i&)RzUzvH7HJ?6is0sp}G?T=uxd<+-A9JP`lHFI+Rwu8k-6U*Sy&%keN z3(5!aj}P;E(Zb8X>_}}q=XZXc%|r@KA~=V$nq0$PD$^{2LNP$4k!`_fwQri^sGvgr z!$^tC8`QE*_q9w#c=f41XV9HQffVP#%twISNBXQ8#i=AS&O6c(jEoxsq-` zRLPEAec1!tfR%v=eOqAU=-4qI0YXdZXE3U5`}cQ82&TVbCZPAsxMxpPFc|!`1yVe+ zHKB}3ch=Ylm=sA{flh-wGUeF+MVM)c{jSwZ7bS3?g=eXt_V=9O8NTu%wh4dzpBtcj z(avlF=2eVz>HeSqCZ7Y&r9oxbZXjnb-Q1wAoTI zxLby=sFT(K1EgAI;ttgYw=z*%%oJr`M*t4?U*N_UGAO~43M}wj+iM6rYNrBY?9Y%7 zD$#Pn67NlRwg7YCMt7I}&J?(D1_~;G22&np8JF|-oIyMPOceM$V71De<{-&bE4tkw z0DxAPcmfhuG7uypz&ZvDOJX)XEEx>>NQ3YTAc#8iNRa0C6vjLOZ+}+?V68LyDMK_c=R45#3smRj#0du*<1r1W z5B9gro9j%%$RJNkDyb>rKwZnArDyCKnt)ifN7F-_reRz+G1?ydDV5TZF>_e%+hK9#Y6Iw?yk3*eezh_>nC~>rtVyV4ns! zNMw@Kn>4r-(@JVH)V+N3^z{4F)Oj-MR9TD`bvMM2k<1U4Pk!gQdP-glY^ncwSTX(o z?`Sh0z1r=lucy_<>X`n^kgOlvx~vPw0C0iXulesYJb$jT&3A)R^a}uvKH#&te|YMi z=R(Xh8r0sroqgbG8EtC=PW$SBC)Aywcugzf*nc0+OMI#+BWO?*zX$C5)Y6^5v+vfn z_gx$ewLAy))i@Cc%xIE7m{}(j|NG3;Fa7T%n(NO&#ECHTX^Mb(B~d^(Nl|e)xGTfo zo*XJ3rN{vWlMl=tECJXe1pxErWIvGW|3B8=J09ybjvqFnQc9>$MkOK2R#u9Tm6@cF zy>1Osw$v%gXi&D09SRv)DN(j0vs8*?rQ0Z?=Y8p%-*}#Xp5>48dQ}JacU;%?`MlT1 zVS~s~%K~k7A5RqHI?#%ZxdHx?1n&{=djen~CAbjyXAIdYh!cie-Xn^Xt2nV^P;g$B zSF7;A$4W^2Ge5BpkvSq%xGrMxj>n?3=#{V&GF+T6;clEUmK%et>Ty4;zD@Ua<9krD z)Rj-xZc6B^LMHg(_V%` z(GCKB8*~#^mssE(E2dq3l};F?bKzPBL-rev@5@U)vArmcltYFiBX|?2v0;iDGp_DW z6l!r<%V>#o)tEp*rZ;_{i0p7edi2X}D8>)yda8~xYOKX1&s-Y%Dhc9?eHAZXoO$$7 zvM04kRtIS(3VAE-=9%t0lz#OcUJwfS0!_~TsE!ndMspC^nm&E*M$=$ zr>n30xZ$TU1+gRN1=ZvtM zm3gD<87#L-p8c#rZB<#@J;Cewd0lLDGT%!$mcN*)(_W8}cJ681z`4(_9B_6p1Sp}I zMwz!x5dTb1$X!B^tiEGel(PC`L1gPsYTAWGyaz@5xor7{N6@7xUff(twP~hfw ztK2+OG^Ta2u&_8id}?~>Df3OvqsOH%tTFD@^H+gVmy)RN?ii3$ls=YqSmV;k1=BK* zoUz2`%!61CF111V+Uy_nXw==F5&2&8*e&=jZbPl7;V)jl=S;?N6Bz$oj$DhS<7M*N z!+u)7_8MEzO_&a$$oUopV|evW4z|XbZDp3Hzxw6@((##mvC(|Q9R{^KQihD-`GKlK z)!ZWbN19HYe`ni}CA4B2+Vvt+X{?8W5!h+^?PxN!H~)RHZKR&!8tD^n za7l_c#G1ebxc$Z3Yc|WtjCIPIsZ*b2TsAc-;*Ca?=3Qx0DtSi2eRFdQ*G1N^adHiD zlpxLfC_2_E{d{USFQ}!KVnVV{-aHZt8VIM8NBOnJX%o4~fEpG0pGKUYunc9-zxTBF z;a&47dhpM=z%5)MS&KB04krGg!oGc8BRMx`8cH0$XJ%0x0&)U(z)o=ev@df6H;(iE zp*QM3;G61FVw+VqxjGjsWs$p&YibqFb9!n=@q@ijr2}}9)WrnooE+Y5e==177XD-e zsc@RtrR+OehbyQ!3@x5aipp9BdG-kX?`r-Z>W>ERiWD?t4B548GZE<1a-I)Rn<^N* z7|b-J$Vlv7GrpvkbuOj4eSG#$MWBYRSVKi)eA6~!Uf=SA}Z2Mm6u&SeP)AOur%rR9`LSNrklvQjrv7~)(RPp?kPaDl)9r zy%WmhiC(u&t*g{SI(&@q7BD>QIx*{0?ON}5lZ5}EArk1La*1SLte{~nu4(l!uk`6I zieDG0ULH*ZgaVp^DiW3a;l+wO=u*_fWuT`m?$2Hwji`$Id)Nb7gBb09cs5;4?ksVn z^B%GiBc#}_4Gde2>O};MlC`O&Kg~mgR;g`VULSUbhl}=8-o@}Bw(>Lf50)iU@BOkj zdZts&2dLzl;zn3isMj83Rc6UwNY&Tks{Y()G0@u5llwkJGpAJR0iaA8nh$B!!SF5@ z%sxL+E>7frXySZ;GRCAI`HbzkVFE9H-C` z9N6#sINr8C;*bP|lm-LS9C)=~YY11EI;O;Qs_thvpY;`XK$q_-4eLepat`-P9M_zE zeYRuTXhUdG%E~b;^F@b0ufsO)wSu{pdlv@@Un6NZr0td;x+Y;2q1214irY%QZv=8X zN7yym=AC+ray$p4?YAEPgHrn8#EBuT)LyBLjex;U0{yhQq_;+t7dyA2nj#4RlNVMU zdLasaYi2J+h5AWQY-!UD*cW5361iTg>#%!XF4Jnf7UF=y3Nul#Rj5OO?NQR{B$mXu z@tgUP*>T}hsm)$j2d@}2UQ1}F(TYqwdw0NNW}0-J)+GcF^v%^YlXiOZlIPNRw(%{q zNM0Kpq#hC;O?f>~YsxQwu$ywYds9yWqX?Sa*OK`r+oGq<3oR6rS&x=<50?k~B`)r{ zr0BmeZNk<`iiJOF#^2s6Cg$eB5z|%misV9x94mYoX1+Y2D{cQzr{Xx8e+Jb|Vv}3c zrx~1kRy@|vYW=KbXeHRIx6tOq+aV*>u5HQ>+!4^k(^QD5Jfag1?Mvjh)01MKMbk zQEjToT@co||EsUVLyi&`rnS%q|7y!3?3w?_@0yNAE!o}fu?ryAE?9W;$VUF|PP@lk znZOEZuabof;2#SwoDXS!WqX! zj;4ZSnqarSYM1`SllQe9|9qWR&;4+iFp)kL5{`xAD1S%)`N!|=uh*Y1xC&F#_D!KC zh$0?oK5ZNZzZcJZ>-)NgLv$(bAxJY?!ndR4o2u8xe>S@C*?~5k89|BRa!FX1Zz+O^ zFI6to$j+>q%{T#gU(A?`L8;KYSSQx=%rZW(Cs2aVHVtHX+uB=v<9J=^emGZ&Hm88} zG_O)`2|n3p;vc{$nw6r#ZMnN*>Q;P%stI1llLJhw*Ff=@YtDCt7xpQpNh^0 z)7_KUam2|ou%{IzRPWwyb`B4v*f%o5mY)hw#q2)e+XkP7q@-P=7{MIJ9pE1W3xx(b3#-pX1*VsVylo%IM12kXT?B?_>z21YN(`M zNp5}C)Nf)#u<%A)S+9pPk9cH;(fs%>!?XuIC2C@|YfnfGAJB;CNLJ^InIfr-<_&S3 za#PFGRwV5Mg7p1Zf5j=;g^Ti1r4?mw?&cAg&Rkn<<~<#yKk2$oKUexi^W988jUF6r zQ`(w*%aQ5Z*L6XQrpCR-DOaQJtsz?y3pZ2T9`1gw=t!9i+Gc@}-zb>{?v z)h6YlS1|cyw42k5cO2r+QcYU3nfayB_ZLZ)!GlWY42Rtq&PhcXzWH_EvL3LdO978+ zycrWy=A0Vpi_yw?q^cE9_c-Ty7NxFx_U(fvAe1rt*3HG0)VyaG>YvG9GWXrvV0}zU zDNBm)9MR#QPw%(oW6C#1Ho(k3BrJ+?U1V@-+XtpMf;TosE1Sx4ogU@kTTQbl7?2cj zrQPFOffgm5EkwVnknzi@cTYOnPvusgN+0_TaJ~`s0cAe3MM_^r3+w_VIgoMRo;Neh z9AZzM|I9szg@ibL;%Kk2*g?wuz}>nQMLntvXQ!eepdHJ8>y5frvDctKziiA%;($*P z5C+yQ()F^E3vA!4RN9OW8ZK?Ut`|!l2?eWSX4(ty#iyf9djIA%vxO^SVTYyRGs&jI`)&v;Yvswvw{`g}& zdVcUv=uW25%uD*+hjpjdSqM|on|03{Z&(l&+>`Di6v_O0#5X&VM)`Pge9Wh3vKOnN z#pk+(GK@$V*4V8`8++r?M|&H{YmdlAe`s}gAw+1%*h(s?7?;6fFQuQLHm|Cda>Hn? z-a5n3siiC7lk?HC&o-=mw=(aCUDkzlLHRq_^(Iqo$31Q1Y%6`_-^Ck6|#>`0i2KJpR zzmJl4l&1CggDjQ2x!Fav=By(XnyDGj)u}InYU&JeJ%yyp)vJd`G1{h^zhM;)i~GF*DVigIh&e;nMrLe78Dr>ezy zoIVrEm^kj+l64m){q>C;&EJuz0KdZw>u4r(&|-S5k&4uF@6l!EKRtNB`%~}^mkr5- z6gLOg?}Qh_L96EJ@3c*NcSw(Jq>@`tGvJ9RqOvc2TBwVw9rKZjS>8N}zP9^UKC5Q$ zp?_m>`482r>c4$P*sGzyHiv$+Sy4cZJk@%MT4)gGm^5l23FI0{!v|EjJsr@;#vETi zC}?CpBsZ=Oo$0t-8)G7X%8n?*#5eL(;pACoDpKV-&Wk*D>6ny2RT|xJ=#4x&PF~Z&19WfpfwxfGy=*#{{u@uV4)F{V7|(PsDn!0jVvz$ALL-k~=hAR^_P z{i|O+`y?;BJZPKlwPW{sGVwNO9P9L;PtUnN<6Ku5?iI~RlSH!r%ogv1ffMkZoARj1 zRjUQZ*W;Y;W?NXIWm+mUqPoqeqisO@$G9HcN&*?H4$^Ku-yNwuZ?IP=GSa~Q=Dw`b zya9b*?^*sj*O*z=_+MNLBWBAjCVn9uC!ei6+pvCk*(GuO;YqxivfCPm`-MHL7s8|h zkC_UPf)^C%4>{xdJ^Dx9paW@Cob4N!?gO?6KJ<^Xq*U-@I$5sZ9KPTI5|+CST5e8> z8piMJLHd@vZ$Q7*b#p-2e`whU0xkIco>ZUrzC+N;FxBu~a|ds*pZ49T5XotH`hYFV z`uj)9?mIhUzvv! z#AMdMYij+TY2Oy7325iURFB(lVh!1UtNI{V*r!XOW_(+jB0{+3-+Iu=vB&8jl4Ef0 zyRD+WAw-|X25AQC#!sQhAiSk{gOo)eog-70!!l1`qjZ%-8#dGRX63L}d@$43squI6 zYE|y!W3zuIH{S}|wnN?}@70?KtT}qs#$7*~46-b$ zDm!0N{C|hG9$^_IHu}6!L6&WR{bYe_itjYZl_K$;c^w+(Y`r^j%sTzn(rEO(B$u&* zh0w0c3H@E;+q;tA-3E{xbywiO5k2_FqX+-j^yw%`#zXbhl2>)EW-1QM!;Yb0kmeD^ z(tq{uW>daZ5_JFHum6{Ng8#mqQ2mFV4mQU-b4fX{?3gD>N|`6hihiL@0IEmWM{wpb z5PdQnw+C@E2r~Y9wf<6>ho5od=Y1Fjl`RPt33x30wxRpfSN#H)ZVBX0Z^-Fv0X0wZ zB}vxC2RYmxvAS^H@c;4sC8o@P+$-18uHclgwiL>?#X}T8_thr1t>14x$FVy<_0Sy= zPJ8zMKJLaq@!$XnO#%;dA!P<1%?_s8v1uY_a|sq#BpegLc&MmDF66k4^+;2zm-9y9 zy@&?k6npadm1yuvx+j##O<)!VOou>SOCwKL z9AHD$a9he%w6~#-ZOI3pO4thDECRP!<~RB2MH8tRL4jG1SI)aq@w~jaQ{Eb?g$C{) z@03Aj)OW7)1i?d!>qN-A3B%~4=siqaAcvna+V*{ETLRT8yv~(1K^q!{Oo!;@ygsev z2=&;a(Ei^gTxN8u}?+tOelmy{F!!wLietQxV;5CShm)ymVUWo z)`*N$idv}Po@DLY@DX<>!jYt~v8@v1m*FOko8Xg|Q~@9%youc@d`%bK`0NZG=+qLsIZMp7TJ3 ze(y55bGWCsUH#{=|313ars^nFWm*n(AYBxJ;8UCuwzrAubn~9mPmh;hXp4g!fpB=o zfkXm?uD0tGG*qi*VKByLvZj|=r++#X2W>*)bPw^i!Zz!9C-Xr!p2@_@xII$mL*$h| z3j)mC!Lkh;a=yi59R}+N-;{h8{GZca@tXG>yCKGK`A$}9(j&hGFZrK?2w@WdNho9Z z1ZqhAo-2&~B@J&FlP}8Dmy`bu$_(|<}#d-!4!`*Dxw2}HQ;-IP%ZNO>%7 zC)4rc$X~7W!>&h?(Pd8mKnQ5}mtz*;$`PT-3-gSoaj4evt)Wa>b(i+G+&L#Rrlz}h z=3NiRQRF7HCaRl${Q`t9Ac0zJRg}2f_r{MJ7T^YMun&7_+n(8>6Z84;Zvv*GT95_s zzT!NE<0X`~W71AGh&jn}d<(yG zRj2S9YItn7>}0tuY%EL#3*PIw0`#L(!&~8px4<$&6#VdBQEDl@v8Jrv@X^SRBsC>o z9osAo%`G~zNvLfTzL30W0NU7C9E&~Vz^D44>~Q)12>?hzw+I2|$${A^WA6E7Jt^OA z0}Y>dgDSN`!VrS>iOK?MXUzugDmtzD&utp7NS%a4r5<16fTkU~go;^tDw98&j$Z-{ zL7qe@5<iHXS`G#?c7--( z3fKV(WsKlh1`l8g?SVv2PzzseFAHtc`BHShC$RFg`L0Kz?>jv`$OQ+jfOzK zS2=7o`lqRh{{A}(hL$Q{Dy3#aRJLzLP2(hK80QNvO`%Vm#7BK@jBYnfet<=s;Ie7< z%)EC_;6AxsB$~f_;SNG-o&BOvXZ0$#pvt9mLhE%@s%0xa{ni3Vt>-(=^C(|s{6RE< zzo7*7u+(!G8d2+fOtwYQ!%Nw)hZwM6w9l1J-vE8ngZ|k5U^l;O%-UK}AY`-+KDI3h z@05Oz_2xixk}zSMo9s5!ypE?hHc)jjtao+~HEzRnrDwIBkGU#)Hl#X9qkw?n)7{cI{sjzF=Sja2-gZe8@v?$SYiq<7XG(ed+c>NB3Y^S zAZuovaW14hULZ0m1WZL^|6+){yHqQ<+AY!J0sOhxG-gI{4^KVASEOM3S-c}&A(|04 z*OTARmub}uZGfl57YRdo?I5BJpwZAH8PY_6Cuwc&Jtj>Kz7{uB`BH?@f=@sajL;9S zchEagcA1wubnZ6FxY17`t@v*QMGY49x)T0t%`fkz_y2NGBbnO47u;}T%O&P#FM(ES zgSpctT1o(@O+-xggV1w$>pd%6v3<*oe8C&STrx(WbiXm2w1)sKX!I}LUVfv$Vn;Dz z(=9#yF_5_y_r6vsPPqyEiYpPJ3MOv1Z6YMZ%sdd+vv+` zgd`*+Blu03rHKHNQ`lXjaehWYusH9#h>wv4B|Jx>H`Y zvs=i&;%nuGzgo$qKm30mO%kF-)Q>pdEm>0tQyU?u|7Pk%Eutn~3DqKeob z<>Zr+FBYd|*2l^CfRXBulCvM5F%%$_4vNR=qwmH^fK#iqa4}u|H9EcK#?xP0#Pu%g3H91=H(} z>?}!fGXs-LMT-#q@S#|Xb0W^~cBFGt!OujYQte@NBx!4V-M{MHuQL4arm!aSM~|OI z`Ml|#o|q+K0c9N8iceWk3C{AC9R4~G#8RHtaEG+^i7*wC*#L_VK3oapAZ`{;X{QM0 zC{o@qZYJ6~Qp0qTNdI3T0Y_(%-h zR?Wj+ki*@_^%4x>m_qr*j68mvz-K)AXaauTLAAIbjz(l#i6Oor)8=Nsb2p6Oyb_Xn`NgSuPhDU$wRl?YDw*%c6&|} zcelir@RX*4!t`m4Aw+9&%ck3-d8mzPDsqQG*ej}c*oBH zeRp2uc@p$@la<{)?VRzujE+mVF868h;IEEOD6M*~Z6V5v&drrK^^MmVtG4kO`nwG= zCfD=g?oQ307Jhj7M+q&>O08R~YoWUDLn-ybxf^ZeZ@U=6uQ!|hfSa!8iV3q7$eZNy z!weQu(G`Oc`dnEC+Qw=kMn;S}(fsQeT--M^tysX}3>gacL`qK%>NK>;9{g(V&9QlC zPr9Fyu}==C}=gENSk^2nt?+{P}TQmt7M|KKn0Y)y;F6<9!O=vbo>Lh^v_X& z>M<<>PllTA(jBX>ZBvlYrCCzgY%e>!->4ImcIb=oYpv9}AA?OJ z=NEvY#$cuNzR5D)d7R7fp_F4q;HDTY0g#$SpG{DSk?~L3f<{?Juk1<%+0?h~fda*9 zDE!yglyC*N3c(R|YNt(OR9hTIK$N3K@_4_2;f$32aqJ3K?(cnsGLlQJiHOIwYQe*8 zv`gG<#E5gRE3$%$E^f<7nKC=!E+{F~weatPI20+Z>P9m^AYx)1re`=TK25u~W1mrx zw%br8)Qf3|wp=+=a`C>?hr)Aheq&KHk=D~kPrGT}s^-vFQ^zuJ@leBS;G%2=g@fOY z6caH%q{bAb&-0BsHh=J z-58>V3c3Zth(>C9%6xOolul@&X0YRlrd0#2K}?DfnJEuAQp>(2KNZqX3kVL|D>o$C zI^TL+Of$#$v_{?iRjmF)W*1XPM_?KLw!QR+to;%@MHA2qRmx~565Hk##fkV_&+ymz z>AVcygT9+%+y-$#%HYOHVC|pR2dw&!+%IYPbby|GO#&C>IzbJ&qHD;sw?dJz) z%Ne|Q$E}&&bA4wGA6iyx{M~D5SkjHW@9mEV_>gb$&aLCh0o*;xLDutsZZHndKy}bO ziI2TQ{)X`-f40|W=xGk@IJDe)YsxQU_8@48LgH1nXWH01_$?_w&`Z>2zW&%4I?@nb ze31T-$|1L(a|Y{gkqD}b>2S2lAXHfjPV|uJfwu}T^j}3G;XL&Ix)VWZuuDV%Sm~B9 z9lvG0VksK-tr2O_A^#J$uk42Scpe=Dj|KI$li%?sY8k*eZxLNFfh4jc4ldqZwX#Yo z{Sd}ot*EIvb7alN4NKq4Cns!_VOb^I^)YxxFbj>B$w3fa~gg!UiZE;7br5z*nN(v~}YHqnV&5`-O^Bzm464X<~j$wFDnKrLe z+VOP_p<0L`f;8nHH7_HNwbic*xySR<;2&fz3^H1vGyxXMmB0pXb%Hh@lnyw3-v^JM8c-D1X|p>3@MPn;J+-Qe|w{)zxdzBC#J6cQhV{T-R#W*fm4aY zlfE1p_t%7++W0RtjjnFp|G(gL@OFX(17fmqq5Rzq@S*znSJc+rLSI2T-0lhNx6Uy# z6fQHwpwQ5Ks=*G@1OcuNeAU_rs$N1|Q42J|ca~hyL@eeGW{@;V!Os&;YFiYB z8V`8n$rbb42o22BxH5v>mqhs|P0HMT76dWe=MM!j0UlQLz%5U?XM-i5U<;>pv1zeu z{4X4*P2g9w@hT*{s}Wno(l|Wqm3bc;q0Jn9OhxB2Cc4KIzvTXMTV?4#K50_rK!~#w8W>Y{}D_CJN z^o#8H6xgiZl#}yTk6nb@0?W7&Ex9&Y^D&RYDje4_MB+!A2LlZv>e=5f8sZ_W+Be!*xpUNW#o5=9}oHU>H2PM%m;8eZ3(7ekW5u09s8;c zA`e@3${#cpb_Gx+3x;u$$%??bK2V<8q4O>3>mmah@j%By2e9BO%=baxut{5wuUgR} zr2ZB$;4_+|f3OcfBpd6cKy8$yolS{-hg7L4xpmFvi_O!T$Xh5#83CxPEo3QY0m1bX^;BnuIa> zBt!wa+*q3nXf_A((E^PR4d8kDuYBXSHUtzzAp(v=O|3Drd09gF^*=p^_zJZkfxOWN@7{^$Cwk)tE# z$lEJ)yB_VgeeOfgk119x_X0u$Tkk#HkENwBLZB>}c7d*e@5PCDF=Gfb_%8|G-MES2 zMx24u%?cBS^Vpwxy>E{W+Dv>^T$ZgG%nBb-AGnNhJR?5apC&uF%EN6MU=R-z00(CwyNj+1cO?{`{%=nlY zXy<)ZdVhXBnSz4joUg0(?v(6F^|nB2oLT)5(=ST|7V!dY&=uQS?}c$-8hX9BFtp%z z94?8Ijh5t%R%B9}UY-`DKEtr2`C74~G3}~fIM{7k@23#QNhg)5)lXLuXu;WSYGeR|Mzwfj4G zM7&yl`&IApp7E!*Bu+QtD})W@<9b#7+F!(iNW}E=N0+u0SC8N#?PyKM3YdaCayRHq z`GKn^o|xEkNn9PecGm}`&kXwL*HuY&YCXho1_ydM0=F!2F*Va+P*8WXo&bWsYGYih zPFmOwi6@Ube(wYXm_P9m2(le^Nssptmi^g*`@Agj%70w*%_F#?@7FE@eYipJ#im(3 zx-V*WCAn80ZBAK$bey0vWtDVZM+xLOYI34VI9rX{^mXoN=QZ&7`O&RhpFhjD3hBKq zs9J8iWToes|&fhCAi_LfOKAEI{4?2~^1DCpG=B`SuY^+&O3YDLT>(F6`U z=^wdT^ri2poT+$^2Y{Ojb9k0k-yJ-n?-3TbW*ufTJ|PmsS5e|AO=6}+E~X``Z@+w> zJ!sYDH)SmQsweOLsj;5NIymXHGZ+f6@9)6F{*)lOoJ~xP{hf{$1Ztw)%xz{f-{##q6S*>FLVDwM zx{dmtdCfs;6~!6N=>ro#-B$~%BiJf7{~sTOj?agG^_`PvT=2mdiLx+rwiC^9;$(jr zqwqP(i&dV03yy|nOHzIrNi*U$0x!$^kGHv7x#0DKZdDG=RWHk77n|WK1(06!CUx=q?@tkF`Ja0eI?zh@k5`at#<1LsIqIfIs4GBh09zFK!D9tTq1UvADhVqdg(sg z-#@>qMs>z~J5pKn*VHVNQ>w7LCZMd4qJK7F-xjz(Snm+`n25 zv9jB$W=nOW{2FV4WHh8;cInAa25TH(edzzk&)bb#15JJ7RbD@i^g68LfXgb;FF||A zK!TfRT5>w+Rjh0t_l5fgkAyk1S*wWizNAyAEBTyJ+S^X-YmJ2ZFAC+UzjEZzvB)2h z1gzc5LrN?@AwG`V%aGdql<50u6HS&@3qJ*g3=yS^6VL(Dsg(F_dY!?<(?U3R%t8oy zJwuWU-sYZgHox$Z0Z6q+=xu4jLdHT2bDEo|HuzqVpY7nc*`Hd8Khn683LL|3#9`rh zSk2L2y1{o0jY4j$wDXawmZAE{1|O*NXOdVQVy^4EEhA{AvVqD{FA%Vi!4i8k>#*r7 zzzVJQW#p(tn<Au$8W+D>dSAz~i1{Utn=`la3fYWex3OowFu6!t?`CpX|4b5bz zgstCw)N4fa|Mu{VY`>2+DTrBP!L&r-@Z+y{`>#(>_R->u-BbpR$03Z~>2JISw7BJz zeObLn<~HIRBc2kEXg`m7FFd`*iJ7U>Z#$fQ847Zg>*fimKJ@P4N8?zXg$&aQ!@Cs1^y1w9dmygY8adA0TgxT&k_Z%vo>#z3 zOhnM;BaI1)y?tjJL-e9y|JKS{qO2Z7 zk1d4(>c!o>WP%6h-m~{R<(PCy^FRhln>V20@g*FTA;PO#5M6PnTX&Jj%drc@;E!nS zdM-4XN?iDE8Jn6^2TX*ZYosS>B(;1|9v?(Prj00`my$;$Y+hQ@(=Yk#aP+v;Ep7`} zu;THZ(g%?CC}83YCg_l?>7wJEAiGCraL26and(;%d10n;_MN=|Ird^$-|*X<*k~Eb zu@Y9>V=eKDtnUVh9GIf4T9@jkgnI zh@-wF%+!Q65~6Qb{)JWe{fPgZ-cmZih+bvdctIM?{JkC=(h+}*%C9q-tCk%#NeY8s7a)ljX~Mx`hE^)w45Yp=2O z7F_jA$CTHJW_`$+`stXf8%z772*pC%;6^YTVICb{H z)8Hb^Kz5AnNUAjshif+qGovmPCC&cEbH5zBTi{7uI!%zoA(D{3TQDLrWsESfgV0`F zdtJKP!)&o;@Qmu0xTQjuW?4Q4)ouD2wdYJSw^VlZy}2wuA!epEZ+O?eL>=h7Jd11$ zgD4_y=KeuF-n&1;!go57^1wmB-j2d*MTw zXwHzx?1{A5kKsj`jUm`H=6%iU(nmv=^IUjZd6#g~GIN#DTbzDk9yRa4K>La{zW(Su zy(!sfZe-?dfKOngWQI!v!tK?KT0Ctc4h~Ety(TIb?1F=D5r492RIk+<^lDN4RmRdI zfQ=aq0Z0W4Qg*`_Qsfj|U$Xn?L%!G3(kvY-Ycuv8h!*3D$Wlr{B1H!QqzInjXOVqP z_?yY(?Dx)oMjDN+)g0-c?@ZK;EDBhea1p$m;wT}_Crkm08s4qVT5KucWFl}yjh*y0 z9)TC6vmo`Dwt1&ol#`*&IPoCcB~*DT;Z4a-v7{eXn)`jx)>LRlNW1oR*@f3k{k*l~ z@c8Uz`MMzm-foZNCpmgN>hAyU(gAs-EG`X^K9!dhA*iCZAYLDY-7D-wB}$2a^#;G& z=3+XJq<-$GTF)8TX)=8Wn1y$JMCP*StXfi9d2mO1EvMlZ6=$l=*rM-oZ9;)3X7nAe z&Uhs(Ao6wB1xifR4jBopNGVl+697!Kk&T6^uq8?uY$wHGqv$dAk6t*%#R#ttn))%P ztZSTT!?`c@>o;A9AvPmcTk*X^Kup%o`b`(m@x^S~UR0bqx~>1I1|E-l%ip6WpcmSX ziox9;?JZm%b`q!EUAAGkK6q435R6^&A>pA97(~x^p2noAn6l zivsGsjLBf#s}*W&4r^`^0~eVbw`8$27008HOuF~<8Q$WtMRv6b9p|36vo>;f|Jb*qaIxg&PWNX^qQkAT2RbjKaQHy+Skl$ zH=#wobVmI1*@o}M|FW8XQi)bM&p69ya=zTq$sfOY@+_UP=|M8(YSz=34{3In)ZJER z7skwK7=2q!bE70RWSEhCg?>>zEdQgbGlaLVjz(QFC+e2Evbzj|%c?m8#w$FP+q2o< zoIa&xTgBc=JiUj|jKyC+IBy0;;pJWfc+I}-iv8l!|LWmQG8r$DeWb6nAim6&gvSSZ zh2Gb5$`ZIuCKW9BAwkLxH@N7L(WP(8nB5mLz5hCi!=dNm)0ip&iHnT;obT%emm zaTIvZx>AXa{MFd)1AJ;mwf|O&N$E_&>zZCA{}@TG7dUfpy)C`6 zx*a5Hi$HF~$sK>y6MnU#I3HCy1~V&(s@*hvw2ho>4@gdvlGL2?c3dv`3h0W!N}8WW z8AG&8*>Aj#@4*%;TAHg|j}cHVLNgfiBInBWCkc~OEftGxgLLsi2!~LH%@;r z@^+|M*3BGt1fvuE2o!k_jpn`AHozLTHA`6k`>XVxka3IKtI%tHnvU~ZTxw!ssf2igaGE+jFsYWmgA{MWCUr^xIlY+D~XM+E$i(Q#W%}+!>EHGCt zsoHd-L93}dddJKky~w3VLb&c?Be6)Z^%4yF#DQ>CqPzSK{WIMznWugoxm7rfas-OimsYf#h-iAZC2Ib$ktQ_LaQ>3w(wFwXhb zKlplKL$7S5k9pej#?!~paNQ_*QTzDaQ^%3yg&y&-o$B)-6l#`3YR^&gc}_&>3?;}? zFYUh`=KppOo{&oE8NQ@$-_H+T_1b&+@T%Q9Z|PIU@Bd$s&(Hn8W!?SlTFObs_n+LB zl+$f;@}%Vh0=(*{4}2<(qQ!mBqL+GH{=ZC?U7grwq zhcDs8Rh=_nBlW&6gMW;7;rJjDyzJ1_VXuno${&6@!iPTQ@KHIQ;8njm1d0H*=+B z49id&_JLV5U2W+{+{e^M>Tzi#pWhny#cs{!Js-|ncvYzV+B>y?9*K_8y07;SWbvc3 z=*hn;cRQd@_W~7l-Oo|ABh8} z_X)6%COGnSz+REvXKZ>7bxv1jwri0!7VRL)j$+w6iw^S`rzcAira&kc)*%DfoC?|@ z8dPGhEgqE&Tsqk~4P!Gn>S9o~_FgS$Mj0}(GJ^kbLLc->wNfbO*aY?UvLMkVHGgr- z@@)cj8U4?RaTHljdJ~vJAmW98{E^u>`OTR*PJj z-28{EGF+}=eeW@->PWz*0_SGDeg*p(7}oWXap-vG?N9N)+g{JEs)45O=~x~@*04P4 z?7qigFf01p+B=Rv$t1eVnQ~+{9T6)4p<_#bK?ZA|rLb~-S ztw_fkMmFeJ2WR!~Yuvo|@B3o*qedz46iTEt$hL|UqUg48Jx(9J1N2tS)BA}jtgsn7 zE+pN&h|J~;=Gr|eGw$26LO9Rg*l#LW9}?_RNn{OO0hhxhF(Yd*=_*ft%Bp^(V#jXZ zN)kmH&_{osUEl#lZc`uest*XDdH2@(55O0%De}Uu|A>}E8+AXpl@Zh|p4E3b|F(N> zVWoNYx%E+`eg~rL-8%7n?-i1(WjEzuhA5tg7q^@jAp49}cz%-YpTz}I4#-1go}UQr zhxa34H)uZC-OjiqF6@2JB?_u~`6ERpVC+vvi!wd=DliEiuxMXPl=yQ-UL&Taq*5#P z5`EK$L4uZ_217*L!#UjVlnxzl3))XH+qNR|W{QMGeN9`rjBEW01DO_*gm*>Toz9?b zui63+@xm0C#x-tKEKCTruI~!2pZ~6SC4^J*iu{gsC#Ue*DZkmn*$MJVc1cokPS;Wb3l?`gBOVFzeey}yQm9S>u} z!&(n*MTmw-@KBM7`KXk_J2Cy*W)jZ26DN!Fr(jQ6KDPf%( zF>LAM<3#L)O0QXM+ftvrgKGORHO;0uuDhTG+k{t6x-DF9AL>9oHj!Fqhfe6KIqt7S zFq+rMp%`I2g_eWhA?G+~kDW=3s$+RK1a+wf$T#aUz%=DarOJ#08^A!WQB#3_k zheesa-)%@K8^gv8yLag*}Ssn9NKN=eRRq9G6%HBOCAxb(QvWO%a`3+TlbcP-QVcP$RDKGGogBdQ5Mv zzq>C_ z?}^x-Vz6>CWAtwR$q3qZOeEjraQd#GVe9ka_m5qb*9QbS%rg}VrqrC!)aiGQPH@YZ ziW&@Pb9ey(YBU)7SO6?hM8>U@7sAb8LG*N_mf_BN1{H-Kx)kX9yL{zcjr)@bS^4ui&Bg7bnMV{U&|3cvTTwKdYm@`01V}4pkoi zwoWv`kH?lB5EbaOpA2?Z;SX+%`kD6NZyXT5CqQ{cEk7+G`oTr4_*t&0=FrTIzvmZ; z3E!)%+W$tlJqT@Djqf!mNSsriYD9fF0H zDB#mE0I*20_hR-W6h20=B#o1l1l=-32rRQg(()-yZ(3$(Z`t|X_N+9g+cgo8r_Z-i z&^Rk%;RzNW;<|B%9m-&%h(bu9Ss;P}(UP0~FM6WIA@~9Y zagl|T9vqmISZryMy=f3zUvFn`_}{cf`~PO029egn9noo6(bXiCaS-2EG^zzURR?EM zBfwRwyWq}RQx7M;$E*K>IDv>BtjGe@wG7(1qKlWn0<@*k(ztT{y*;X*L5SOt6BI4j z68p-227HtbAXtj$U3d%ounuZ;BUaNWA4q#08?#)8(xbJz!iIEx#8FC006g}?n=&we zVGE+0pXyxv8(+M_;mt!jQ3xZ&iN>2WWn&LeQ{5#eBPX)LM9uEN_EUfz9Xki#76XrT zI#H)~skJ@C^O$@w(i17kNV&V3eBzl_+ z!{|+`0GcMFQi~E}9eJtd6)9XejBiL(k`+%A;EyQ_V#GhzpY7*|vwN(jR#<1lOSDE- zRqzBwwlmRah#4KDlKC9{WFRbYVIpZeP==Z2h7J7Fx6Y1#?YEm*r2_}ct+)+&Hw^H3 z+;J+`-VrdoYK-f=GT;xc#f&lMxnCnL9Mwmcu+n?hn|Mt_OF|UUk|LKB=L(*PtkMAxjWn+ z!6DU;)TXfxAmenoL^XKu33{#I^-HbV2Z0-aT9$Y&gs_m?p}NGvY1E+owXY&9=Tb-{ zM-qNZU!$M^Ax8Q6h|CB{gb(`@QDqjmppG>kK?@uW4iV#d2D=ikvkS4VRbK_t6)j5f z+A}>e3!`~D^4NZUA(?Qjh3pKm>#C#p$AK%_!F!noOByyq{B*I}=7mpYJO7aLDqNYI zudp$<{K7>FF7YL>c68?BbnRB>1NR{d{x;-N3Ox;#_{JuiowjTF0tY>$+4z`25U+4@dZ?7}1 z`&r~jBnqX#Z-~rD`-%JTr(y8PE4Ty25A~5w4t(4Cehdet3HURCnDB>dIi(zKG~H@z znd8BxcpuMpW(%1IWGXRd+D}h&;|=LfHMDUb_7zyyts$UfETD34#dNv%U}7Hk6LB4e znsgX$losfHEklOqE-uVkv{1PY;lkC-thDEs8UmkR`M$d6oKCLY1orXOJxC*TtNz+@iJKKcZ!=2~QmZ zuDp54)!Ji9aq0}Qk@Np{LJo|}YVW=xK0 z*YL;I8K<`o95IH7TmC?8(gR+G^tcZV!H(VyM8y??@7b3y7MS&VZp^5zRX{LT>UYBwR z8%LYYBTesv=goBa)V6IS-EyZfpPCet3a4acgmH@H6~p&hw=0D;KJ7Y{^uvy;?c=5p zg4{FL>|y&V6sxwMlR-7knw;{A3BYwO(5VyhkGT*T7sqDFm;@j#1)cMqMeGj@grt3e z0%dU{wAN^34O&+?mt3xJ}^~yxPCC_;0 zL3PA&x+rf0U}6Ucdr|-X`OX$t#;%>sEJT;yowc3V&hfaMUWRQc3g1L?dA8*P*Q;t+zfAu8)Wp3a?1LX#Q`PmckIw4flc_d9lWlTgm6b z{ch1P5;*3!sV%^8y0m6S{I79l6&4E7NZ6f)<$7PAn|{uDuF;}5KcBoqBBUQZ>mnUH zlR1e)xFgVR()8}ZRVVER#F4F0%at!~yOi)Gw8+kWHQb_NjVh&DXsZyteG5vjT?v z2F$>yN&j9C?wEEH-*qv0GF!M+4v=5(ur6ffv!Hd`8WCw@P15}diEegHn|jv=xkG~_ zNP~!uZc{RM4HKT5UNS~Ocltnl&j+RpB}BL+^F4(tCFZ9R#L5Xnqybkq6YP*K%3zO? zpb$J^C{$mj5qj9n>^%B-+DYYR^%6Us{^UZG+NTw*oNrrt7s; zjEnj9Ds#VIw(_hWu}@DeJIwCVO05JdFd;pwrCrp4vx5qDpqKBeISN4WGUmzi^zUJa zn4bo!HFYeP)fSIU2B+f8vD|w>6LC^8{JM3YcG;&{Uc2)S?K9NBYleM=UP^e39^JGI zTB!(A2*rK({Xl+G_sz{-d1PRKB4ds>#3u(Ytger1TZ}WH;2IAFGgb&Tn06i;AKBYf zU;>BZ|3TS%2h_ZWf8a+Hx?C$XiAb4^(~=U>QqdqZY15!Zr6jJ|Bs4|4s7Oj$A{rVR zqz+}&Nh%=~g?_K+bno}~{r>*@oj=@Lola-GKkxVRyw;<|@i`fUinY3()cJdAA+RSpVY| zIux8uQ43BVi4}J-+EE_&^F8f{B9c8ZlPK7Pa+yz?TC%?Fgr1!`q-ee)d5~)33NM>u zc0X{~IUQ$EwVe`(JJ8gedcD9V&j0Uwj-OT-#H8f)(F_!rZ~iu!v}`xW=Yud?Ok}_8 zvan-P#vRG^6|-Y1iqSPm*w$cY=Rw|x4a#T&@8myhox~r!z#CMK^G}(Pt<`uLE_dx| z0LR#u;(!J%9lWQ@jz!z=5VMuDw?f;26k9G9^QbWot~aymjpY;up?AkM0XyK?fPUr{qjfQ|G5;kZ3sr8iu5k@ zY%xRc&C>3plY6q-{6X^usD-||1>vup6@l()X1=Zf_@0HxInIghtzneUG@KkUxh7o)w}r< zd^tWOB_bUIrhLL=0$CDvB|Ac&bD;O5B_datPOqNd4PjphN3ZzwLpb_t>B}#gmknVx zDUcf)S)y0twxj@kxzm=CtQG#~7+m3Ag&D%2G=xnBow41Ljxx{u#o*|tJgnlF@I>#{ zV?SR=oom6a4`q#&FakXxy zXxJ1IIQhRp4a=#W)go3~6uqKKp2v)t^Xd+&TvO9OAf@F%QfT3LmCx$u(KlV#-xA_z zsnnLu9gw8!{`!FE3Dt@>gKsltbd#D7>vL;H?^pr_McOJInvB}`7RLT{8OyMYz{TxEIBkt zY@6Ea{5Ul!ET}&Q|C^1B93o?n)F1Z`kA1ocCgH z&_hS{zKefSdG^ZEwITU_*~Mey)srUQP@8B%A9Bi4^xGN_+HtAx&t2?!T2k#>o$|m@ z&ma;h>fhdgJ&WJjGJ7$K=HGw7q|pE*?H!;gYtSPvH6MT>?XcSh9GlB0Vn?{{A(CXf zC@p~$Kto z^T%SOMYzPgq>SRDU*7&u1q}G!?A!g0<#eH15lG;V6W&Q{`zx~EQ2Gce#rN~ihLf2( zXFry^0woGY3(rFL8g^ zj<=`31u&C_V0iqO{fUa6Z`Q4rMIff}SjtgYZwF2l7^wJ-wMR)7zjnM;7iY=s!8=bI|upuIn z!j>rC08myEg)09_v}3vXm~vzaHu)P+snAH(57veB!QJ?Ge%1~$C(uKtcACIp#g(ND z>8Gs2^XvTm>p+VRxY_1wL8Tnn77{3rUO6d!&)pPABu#@trZWz?AML-_QUhmM-w4SY zTjNhSd?|xAHPeOmb&!}O?^o;o*tVmey02kT;XgaD~(egjkM4{F%)=ZE{>E~B>9uEvs1cc=U zM6>tgfkn1@_T1qOO-1c3z|^w}e-)39VVqGD{(u#wRsCV&Y{5xFZE!kPy~9szntC{M z_8%K2%Pg*(r%2joh%VX*teB%`ySDJM#^4(BlrA@@0yQ6G;>2L(*pX&y*6|2AP6Efk zM7jY@i7P)rx>qeOno*l7y15xCi~L4rGGvjXtD3z*T}}s{*BG7mXydym}(_?{90uiJxAj*7C%&BlkztsjiDPx%@ zSg`V&)&D$Ump{#5WgFVU15h}#8i|Xn!Pr3Y2l_RL5pjf7qNSG*?-e$mk#oH!U7Wm& z6vWl>`(yN%|CTO9NyfL38ZY{l@S8O!1z5o97}MvMMb2*O^oTy11xd9yWDItwnDUq{ z)X~SAh&W%q<^VxYan;V5F97F9QHVs}yd%N=R{5T|(pc9CV*xWv(gpHHh|PWmKKWUo zsPphK%C13AD*yb$57HbRj1{eh)2fCBMiG={dh-|5?eb{} zRX@L5#Ev8@*^z}9nvov?w$-_rfaxNMbAcBt632cF5%|*`;|xg$tX+%Blz`EG&r2WM zaFK1YBb?-Y`8g^dQh;MLw3nHjmi`1%eWZ5$$fIfQLd$85C?4UkIf3f|ktP{j`0zxj?lyO)F; z@$5=?u*UY0M9h;LR(jB*)X-F%vlW21$&c$Tau9i{&H)+=RXvl23jG6>E_o~u@vFgF z*Q?Cr=F0j$H4ZPq6)w8O;Nh026?mFVVhTyG-CaN=)Ir&qK@Hazj;LKr1NwPY3A#T4 zd8f$YtfnnYMU4}pGrxh{5hpB+6{F@H%`@2{Eba3+F19*M&wOp{(7IVG@ei)I&gs`G ztwUApp2<)XY0s>HE?%6$H>s!5*JO%;(^(I0b>oPH-LvP+-=x-OULGjIx613z36C9{ zCK*UBP*V>};oIahb<*+y&-#Il-U>=F?rv@ujY{Kt&mX;5Yj@;sYgC*)TL4|!8wZ;~ z97Oxb*`g`D>h1gFzh!mC9=N<&A5$%#X&h=Cz1wq2<5|wkt3Nb?YHV$`Xt3eMg73MK z2B}Nd445qBX4tq=!fDqzE@?D|fC9o=F6Wp_KD#`649iy_FAlo^ubA~*t+L77yO>j2 zdSRK#Pc#-z?&Pfx&0yM8m&UburgtYasw!ew-|e2q7}m7tTL~ae@z!=QN3VJQY4;y< z7E*6W<9{P{@a%0_oY~5}nqrPcMeZ&4{dkYJt8~9vxABw6Jpa`R-tov8%{eP(H|%V_ z_ywRfhRjKlFYw()m{7V%@ht)xYodR3?FnEb#7#aF`Km&*CN%!mh9zS}Uu zWF^Nc%1%zL-MY}+rfz-jYSX7bFFSrZc3npA1m3iqd4bZ?lfosi4=XH&f#72GR_G|^ z=45?o=39IFhW6JYbdgb=vV5xzvJMA7G7#iV-eh!uIk5h>VFaB3(g0RNZJKT)4&U}z zuP=jPBOU=7AU4EGm&+4)&G5$^pVrm3zW)V4yv*3@oP`OH(*9qv(fM64I{&RYz)qch z*MY{r^QztSU+&wUEJ>M=T!UG_F`{l2Wej?{W_tovA&-RXhbA&W?l(n zhQ8>9m;y}DI<(&4dtb6B*9>e zb0&pe&SL}Ubw8UPFsXC-N_6aO$4cHB^pxf>|}A^qj$u% zMHmwIMx+Q$TFYn^e*2DZ17<^5n7w?Fq98+AFa#WLj_S;p=PgqxFlLO>^`W|`6=ZBTJkHv!Sm9q&ED)oeTtFmO>v9(Ok^-BZ!s5Vw#hd)G%1f{=8HjodC=K6kd04P zw6JRav?*Ky#dEZ7Azp*^(+XGB!ecJjY}thIp}f=0DbgsXc8ZNsgx9+Sy22D(=6k$s zg4{9(SP$BR`kFQ*DlFIR`Sq$fs2#+44l_7#5eB(Pf}K(NT0XB9Jh3jy3a8NbfJ0YC z2ENQT()VJs8_11W(|dm~aq)itflEmj83D$=J;_%|(YT*71gHq%nCWrW$5dHU_aTsM zKhSn>m8Qa_D{+-djTcj#Yy=x4Y#lOg&uH%Yr%=XcUfAFK0+%LJnox*W!yCm_!$o`}4}kgINVnU$!t^uM6g2jo9`&MQ_8f)LRM1$N$LpK1ECH z*mOBv_a>F5&u>h&6=aB-it%4Mc{<^5A^FvH8=lOu(uwf}L@x;2ot2(*Ap z5>1>Fv?;_eN!PaN=bdzcVyrKQw!_*;`taCdC{u(&Rx(755EtheU7d?Mw%ijtLgkDJ zWFzV!rCDVa54SC=Kj>y1>{A8TBaJ1jbxe)plpCDT1o9xF1Lf!kR56F!P6|X{`c_)`I0Py@=w=Om39>%9~gLnf1U>~QI;W^USAqEAVN z-XHJg?A27sZZ6g{;6Z2cWfUn zz;8T|aR#6uZ@cF{dVQXzOY{IaiA4g%lhl0n{v5F|x-(uO{N1=sF-NFk1|wj`p{D}& zR5V4SH7c*0*;L?e--I4}5NfI@LfufbvAQU=8oJ%bN2V|;bB+DY4?mf2R}(4Ce)QKm zn)oG?8s`-~eF_(+6vX|!3h+L(V?>^4qQXKb#>6;EVVnI(SjeKtU2p2SV1(;?Tm@Q? z`lMRujIZ5bO=GAr+kJZ17@<}l1UD23jYR92eBDAyXa*Cn6p7U^p^Xo#a_qC!cE{RX>BJo8N zqK=K|G<*%g7F|Y8q#~sd4-&DBDUq_;k0j>`_OXXg4|J0b<`%}LuEqfIwWhbNHW^_q zpF38>mHP$%%20zan9<^hJd`I$4g0f2a+kdk;Xr(pPAv6QA6Vh`T$I zutO>J-ny%c>!yFON<@tF-UM_KxN!_-E|R?hY>{Du{#1Z%QTN9voyRHqm?ts{&L0Cr zV<|5kSL{}^5pXEG>rocvne2oI#+Cq8b@)CxSI>os4me1Qh zWBOByMyHryvqNRvW!T8wC4Y#f#=4ycx>-cY#69Q$mEgp47A^aUkx4ZaV#0W;yi#M* z1=QiM82IyG<;)so@uWCcP)jdt?9&@OF1)tsiu0Jq46rJ@s&e}Q5Q|}TXyk_S0n^JZFyl@ zYfFCY#}*Uz58j@jz~xloCa$IfIUY%-iO!wdGigo-r>n&BE$bFUybGdoYQwC8{_G04 zf$@1IlGV({tzuTpwkN@(T0QQ$S=gngjH^Oxqy&Q>UEYgU|1fH6iE%Ay4KZxcihvxN z)y)ZsLyN{@gmg#S_lE~4ZF4}WuW-UA52hn-H zvz&i~xw^Bk;GPd{n~O->s#*Eturu^@wDMvcje>JJxuhE@4eaKwHKub6q{O4^4lh?l`BKfHDW$CLa87wMPD`CA5CAe%*OA)!1KrT zMC*^^Kt8Ak-{QrhtbpwYxVGZ;iRD+Z21UU!Qc-gXIqpeUXjd-WDS$4 z2DZ{)dGtF`l;!ojyqvzJpV?Eadhb%2v!>6;=eUgDTnrn(R)3oy@!LmEoYx{D$Vq;S zm0F3(Kler?Y)h{~9lsJO2PbX~jV4Q6_{Z(y0sSkm0Y#%Tnu^VE9k*RVtBFE(ka)Da z`)x&q*N3)bu&o=`|GMU0mg2DSCeB)~^01Zi(QiMK=6pVXt?Wb{6(*P!Tpq)IvrWVv zwoDtNq?`TXDlMrGL)K?sgR+{&b0P*svOyrZncbnpch`^e=g5cI1s`o|QjVe=Qg}{Uj@BBUPXv6dA3W3*5 z0ybZ2?;V$O+vDnJ?GH6?^=L3*o6m%@v@1L`{HB>Kt^ZM(8s8Z*c) z#8qr{Z&%+^#wW5)L45@A0{0S6uEiPoRz}kAN}#ci1r&~M%=H?qbfu%O{gc`M^Lm|U z&UIXNE~e$lnyVJQMitFKj9+!FE9}cT`z+3*Q9JL=7H3_-tsD_C0btl+U$rDouT<9w z&kmR)`FfJ*ht9KBx+pXIwf?R(NhsLDyr?c_Nob;_iKyBnuG*hDM+PiNKP7wY$06m0 zq9Za2+za@mz9gzWMhjGsS|0+veT9Dpq0wu?oPr=>)9hAtKUemN95y1 zE?1pACq)XYNWF+_h)l*Cr) znzscLnUn4a`cnbnJ7pHUVl+c%r>1tnSCk?LsgNYuUZKm-4`+BW{9vfMnmp0maMK*88IV(jU!q9vJ$WZO~ip_ z-##M`x~qatUWpB}IvJk|mJ060XB)5sTaX+X&ge?JxP{c}5U0ncO_qqZ+Fg7;R79CQ z<^%67VcY%cPX+em7yh@LtvkFJYNM|pd!8Aa7Sv?6DtfAk$jo1q?wc~j=wTv0L`RGG zRkP48`6)-+zI}MYza%{IKhlIHg>)ynqPT5#)6hr?R+>}P)+ zd2;N2q^x+al?`Y4fkIo106~~C;b5bSZzD`6kE^#iZ|11p$W z03G~X0FHCQE}%tbwwSpGo38_mm7u1bo2bp?WnSx@$L;y~!79(0x2SIQ15C4cqK*7J zbi;SNbD0QPC#%668mZZ7Sakmvg#4o8=fF~x`I{LL2E~=O+bH+4Pa4;9L`!u5uzC_ znSAyjhUDOnlg8NbOiIVs1?9luy+JrP-S47@2$P%e1iRg*Jj~8_$HCLFw-hfu!U8>ex1#KkSiruZk_NU5?v$I(f@|K`1RD3TpY`$|A; zzyuQWI1UFm8{1^w-~&qWq2V{GrJlXJ#zs(2>Vkc)h4et+TL~HC`D~EG$?I(~_=x5c z*x`rbv!_joZ_$dxeWRfZXL%pMs)^V4RC z)-y;*00Mh{dc4_A??I)+9y^;FKnV5#8?AE^&`pv$dYyLlnWOx~z_jY278Z|!a6qNi zmf3o9xM3)v<}As|)cR1qcnm^)7De6|^ct6);K}^unpYlZ3D(Q zvz(I8&AeJ|srh9?_}twK0+JDCZ4+fxLz-M_WOEXz-Qv-EFA{`L{+##68)uxdpQyuI z98FN;MRN^>-nI&LZtH|r9{h3=8J=rM{y=;HY37LvF8T5Pub4+Q^`x`veW0_Cy%Sye zAVR#OfpSC=7?!n7@;P%o8y*`uSO>apa-K+HSjuXV=39H3$gKh?L~&T$O~UY?}%uTW~bZ_52R^0aWr`#1EbNGtohMK9z$xksPy(-dce)7>*e3Go^rqt))B&sA}lbMNFJSd z$eU2uP1;_Q2lFl~^{ozoroi3?FUR;?6>eTWDLP%S&+2^fEZ(?b*8PwuAGwSpiQ+Cp zg>xuS$JzfYb+bQS(*X+1xCRn~lbe!&0JNzlfuPP?k(_5R&e_=L?b~&=2ZrE74bV)| z#2VJ~Y{r?!^zy)7?>qPErq2_9H8hUMfjLbPl0O>Y z0C#`6$36-jNvnQPM!(ls;?O}6L@?!+@z-8Xm?QOWkOQ9b_!qH`;GAk3=~aZrv}Mt} zMmzimHPFB>;Gt&9@X4m2~5ZSEVHE&f8ehD* zOZSav-B|B6E4EIO34Yz6wY2}P#%G64z17UQnzJa(oiifam?n6gz)5er5h>EQ5W)tyxNhmj@ylyA$Dw!YqV@F%MiGN z^4{eehN3jwHh_{rSVonWoaApT{moa<^E?Y29}t(>`9%{+t}CU3H%zxIy{eJr@I6#v z_6KuIOcHTw%W^IG`84Fv9M?9fg+T?K7q?x!H3mp_#-CD+PlUqBhjqo7Xq<>dOW0}^ z)_4}bR9yrpmvHl8&;APB=M0Qyg8^fTf>*O?Mk*)hG)&OG3lI$T+)Y6T(a*x zg5=zm|58407ucf3$1qwc$9Llv>R!PcA4Z;=F*%t;zRY&ckU4s$dv>M>2@upT3&=1` zEr4r{znC5$V!|b&r;_MTn|=A*6+>6aLvJOzT0w{DCIAI`%o7HnvGA=s??mZ5tf}I3olqc3`=58_x#(KWolY&q|ul9x9+I(uQ#lHD7J7<7|%cvJr z@yRXtm~f#+dqteh-du-$-ACwrrfN@bdy3k5b`$}PaVw~b;vB2MWVK~RLJ|#DBa0_1 zZjk0ld3_dG7&(311?b(Iw>hJWW-Z#lpA%&G$gw1fAcy9Iyqj^l+G{snpXlTmd#b>M zH+}y@J*KlM!=^}gVNgTT7oCME%zfx*pi`2#{sqncPyLnIr~QL7cAPBXnrQ*%QUQC# zy_tV3uxTdhonhxXvtDvkvhH#^8t!=#|6FZ6S)jCgY38xq%`cB{L%2CI_2*v-}p@$*KRjXB2}A|NDQvuj#w1d^5%P(V;O3pfx@G) zG)kdG?ps%^d)=MvRj(4eJW z(#)3H4a9K?7%E~5Y63DV!U1Z~fe$PB4Upz^#uUEHk6ad%n z9~6|tP(wkA9i{#lp8N2(z8HXNv)J%o0*tDDwK8M#MTp?Ek%_M#HX`{ckfVp0?^J` zUNONcKw)bO$=yy$@@|O5P+&SQX=XjyR%dyS)6l@l@@VO12+VK7-#5y!(eCJP9wk8( z;k1RpdnvJ47IPu;{`rnpwQn@ylkSNNXyv#&OyLtMgn^q4qL2*x@5GG9%06A7ckZ$-3UjX^^^DN<*}-`v6?dAAi*nI$0WJ{M99eSS zD18W$?Q<85OE^OwhG_IntZ)^Q!20Q*hy%S}7qY{Eq4Uqr`6g!jCd;JXq2V`ZFu65< zIpm(_TA!h424-)#fW|f>BB^fCs)ZjL9OBTovd*93QlfDs9p?g>2_80r;mfG&n(5&< zj=JALg3+ldc3$`0c<`Z#tG6|@&DaHJKFiw^lK^P8w=4vic_J25XA0U-_O~f*AbGT){KX(ueX79$C zu2-D=sL{YjCRj^bvi1s>aC^U@+U91lrN0WW6HPrSo$&CqMtNFGIpA4mSKfijMGKc1 zX50o7*2<#^dqJ1+;*)64rcRdr9Qje&ovjOw>byL61adGJ$-HMrAI!RQ`7gi4kpJgS zl1bJy;@-lF08GIeo7lNU?@>b=voi@wX>x(x>7VHmQH}S}S@UYznl3cbL6MtP!+W27 ztlw03=Wypol3^)H>$RdL3RIH+Q06M(2>GV&jm`h`ZCH`;-44yF3`!7}Ljv<-0U&8h zzu0y3fs_fuH~vVyqoIOkqJ^2Ly8AncOcT&e{aTCG)dq9gK%BL0;mYaS+u8S(Jc^hK z!GUTC*CWy{UN>0tSsIP0&}<_s9We{O4TfK~x(`X{?E%2szs%F|#!tszi+DFo0N{NG z;V2lkwC$W0-?Hm?Az$w&wdh5?O&bai2pQboI>?oZ`92JpcwuQ7jY-qxeq}?BkkbDx z(LhFU?x*DI8o7MWp9?buJsYzvH^vw`Rfa8D@=Rjuo*zXkza|(+MaOLIsU$Vu+d;wQ ziPz>?@bddX?8Qa^N$y}%EN~7&uhU11AVF>#QMxK#E8@s9)YfURc&M z5%=Gv*9y@>ZPBQpAhjFs;XtM5y>v1*^M%7{W$(Z2V zIO2Iv_UP4Zoud%~>Vs{0q@N__t0{y1LInU%pF@!DZO8+S{Z9KRq#g2ud#G=8zS$unU^LHt z9Gtg_`xnmNgFLeqIgM=Zsit$U+v_l)Qzs4KVl_~P!Yo4)G{?+HiO6}GP}4?}uUKE0 zzylSI9E>Q!w=#S4^S1rezx-6MOn|$XhF*^qEfS1MK zPk|gw_^W$_xGs0*w#SzZ&&wq9v4Cmi^tO}_@6G;e$5|hvk4mmScQ}NG3UjFj<|$)< zX;lYyS9?p72Qu6TbA$fcEIFGa9p1e4@+k-Qiw5}vg`U6sf4rL-Gj2Rdf_rBXnLk=G z8#LE{*5Mw3R%J@JON0F?`;3)hU`X${e|Cs3VqO1@EZckiN(t?YC2RAsrv&6SZ3u{- zykqDs&h!*V^G~yre9}I}qaVHNxf#q$7ZzFmxW*@|UYYX}iJD7WQg1DL*Q{+aIQ?c* zMzzMhbcxHpMfs-v7XFV|z&?9cRZOvtg(`@3X7T)Ef+yxII{M`R1i(4HCQCl?8fLG` zE^1`Dwr{8!4-w?yPY)Xd}hub&%fDm=#ybZ0!A5Vw6X_t319Pglgh+NG01->B6(7+*A zFR21)kV<>;M_a?bHFrR9#fFc<>yLV}q>+p7@0;G^5gI1hi@bsLPmFu?x~#*QCpGY-{!&PvB>Yyt6J3lRc%&nX~IvhVkCXJ7`+&HT_9WA zOEZYeWsG`SXhNxj!f$EzbA$BFcDR@PLO8>KI*Y^R=LL7moY}SZTWl!3Z^Eb3(~# z(_Oc7{gK*vzfjZgbh|g4t^E=tP7I^1IfRpqxvFc%5Z1_2@`Y~M5ra&fH6KICwOsL= zEsk*-o8UMq%usuYS#wLgxBX_{z-Od3%oaSow)lIamFQElqf>e5{7E`l>zp|MrC9vU zDau^XQcjDG(RIH?GH+Dg%ucyl^I>SzR(SYDw`@(^CfB`Jv33$POA4$xN$1+BP@0%$ zSNX2FcNYXronsGt{jqsF{<#H4tnIimuq(FbB*xnnP@$5)#SVdE!*{@OK)o(tI;_6( z`1fC+nkeMHsYo^02e~wbaVzx1w#lv3NYsPy-c?zVhHSc^b#RPIObekl?9d&o$lm z5Zy=Vw4*5sF=WO>@bpgr5*!Nm7^mqCGYU_cp;T?@D<89J(Hab5Ots0@yltjZmF?EsD3AKl$v7(XPO9l=iS43n;2K$*iiZPyeT%otj zJD{$u`DdtJ&Zqw7hyN>b36qyWq;5!+3H zUqo7f0{gK{%o&Nh7vCGyBDPX<5&p7T$b>0ZO-iv5-Qn0-im*NwzZe|{cwpTAtQZ;p zO{QVaEnv!jS9Omq-^w4MXi^OAg;|!$$&IiX{Qrkp|r^pUYSd>habdc~$SqA`x&ENore~oEQrYy&df`Acx9YD)7nwPqXa)c&0)m@PRWy zx|Dm+?-0Yq%^L}&JBOkK6y4f_+^mCiQd8+?bv~D9EIf@EQce306Cg+@6;xAmbNr}L zqM)=M;Iv-~Kyb#OO^BiCgXl3ybh^AX`%As9qtF|9NAoI?S*=Lx=18J(IQJbSKmkJb zX(gXI_7P4ndOumFC3oH0oe^@$At`8hXFbaMHez0LF%z}QeyjUxI?U)+$(vIfE6rK0 zZ#-L~AX@_xe`y^3m<2aK#s86W zIqbVBsn))@a*II}>8(6B337P(=ii8^L&LqiRRa-6qLRpX`#nZ;4GHU4@)EAYJlw(F750d%(fpFy@*^LKr$k14wSKS9%vYZ!( zaLwYa+<@(aBf7?MBt?|grYQ;uA2M!c$pD$r8Ad2!oipk0PVyO|_aFNyIEkDCVOFH= z4LsgEe$QDjYU$`W#WLWklznf%wJ^9rI22oPoAb^c5;Qcj0LVoQ1sPHW6np5=^x*>1 z7k{?Iaxl6#(_(L`#%qT<;rwjDIFAaEBd_Mj2s&PhwiY-Sc+j76RCyy-M!X;U@1s1M z>Le$lA-2Kn?WTypz6c}zJ3TqExMu2U7$Q3*#zlry;vzF?1QW5Wtd?!?a8F)9-a-;3 z_euPH_{0e<9BzC$9N}82&fPAWIC~3$FqHEWc|cdpjZW4uPsdZbZV7LmMub#gk@r{f z2VXmUf`8;^TOJ_SDe*?5%!Oyqbty^?MQT(_a7SO&sJ!1|`2Q}tJEpgNOrsZ7mX+}< zoEQ706#V{BR_Pxf<#8<~>@Qy@0J77DeCfq|{4?2J-S$-Z4+H;N z_$^=K$XSKuk%0^B<2-jsrQMjLA(#uq(M&Hde)e7TVp_>GuH{VFRtjGZbD@9 zdkpoJ{MIY%Wn@)xJ}<+)xM+a#F*raB^b{35(=p!f<7J1FlfosoMrnLd1i>YP_V{k1 zZP9CqxK8!E!IC1glEB6G3{D`O?LI!g!O?Ri3;y@)S`92N1v|X=;ufZ8gYjJlGRQmEoNd|hC$rg< z*JB1E4MToC*P6pUpE;>`n?MSq9H?QR4EbGqp79lVLr&(bwq)065%3Qs#g$J;jHNe} zTwwnF0sFEeBwlZZwK!|*j6=i@ECc6gX}=P~K*VrW6WL`3H;H3APc}H1>J+t`AO`ei zrW!~``)X8ZIM)C7EGt8D^rY<2{ZrA*(V&)131%1^DpT$$dX2hH{_+*Bd4jbKD0^v^ zg7Vz*TE_$>W61toaN?LK`=}?%f6u1D_jl2J$W4arN49xf~RV1VJxgD4VWR$M#!2%FM6uaX7MCy%%+Ed70< zOVQCk5)ExN&py5>O(pSe)skgO`w~>HOuPa$pI3_Cj>pAL63ukuezZb&(`F8dy2BV^ zDG#Mbyljab3H49LB_an~uGOY|xb&vKW3k!KaUj@7V*3&WN_``;e6FOgT94D zhptvRHn#EaymNfuzVlx&HdNQeN~H2wFSylyr6DF@wdjhM=bDzXf`Y30n4Gb{oF`G~ zJ?*KDk$+Fge{ZVVC`@L0$UK%vRQPY#K?>ua@tU6h?!q{${C_xvXqirl`R6aoE;dtI znW(_QtL7f~P}u+q1t<04H)Il*lKCG9U-SmQz`t~aC`;#=!zRYrh3x-pp821@^#3I& zL^WIUJ&U7`Zr2?9?>?d`{|)cPcmDsd8~;6k%|;>3kdU9n^T$Tym1c@r|Mz_`7U9xo zW1#qosaWJ&LlViuDj6iLGch~-4-MiWZKfF}f^cp&3V-c#efCk0F&oA7dyW?hy?D@p zR;YKIIiavXXe8H?T9cD&4=D1MG;~@!!*84mdab?TZZ+yDIfXFL}PW~ z$LRDGu?Rg(mYl`NS=kt|WUcWY`0+0io}94v5B5DKct2&$iKO20HxRMt{)_hMk_)$# zOvJS%;U9;b5oc5Mt$Gzz=46=hY9$9AlnF7EsnpNPAM>XHF?uN8#V{fLYei(92qR2s zc2E_A&d6CCz64-|!aWlI)DUt4R;A)B$~Hi2X)HG zzFW6`Vrn2Kry6DQ^Ai6iJ%uQe6vTS9Ig(jxcnE=VtgexPj>X@iXq1Dw3xaaWx@m08 zU7{_B%`H`MZnL4xh5VZ+eu?Lip_q5!n6p~sqKGmdW8W;eM8sobn^~!HL!p>NA5$7WH(uwAY_5f1Rpn{ zS7B+(UBv4dRV9z?pfntm8=lIi$Z{{JQ~WT4b>Te}Zt6Do z?z+rLL4kQ`=I4`!7kci_S3hjL1cPsiRg$37BA+oEsXDn^4%2u{p;(HbG*9JH$~cvl zdNyL6C0^rsU4$ytLYd}FdcVo)>p5W9>RcxRyrETcn)O%1wE6QJ ziP7#o7E3FUO>u*Gy4G z0Q=Ljwp~6=;e++0Rql6(m{ATo!_eQyckU%`+$x#&fDj_syqM3$LO0BPy|znNM0&q^ zbH)8lCwWThvmv&Qklefuv6r~ztn&QRyhX&M0+u%<;>>Znu%qyNn{I@+&7WUC^D>vu z=_Ot3#UWv!j(K;}+Z@)G{&G`30JHF2DCgq*kZUDxpZ3?av*N`jvuy!|x@}`AenSR+V+zZXZY<$Pi*BHe*dIYB>+_+u~RB5pp#b0pY{xmmuy}|4!+a8&4vg}eE;vHw{ElO>W z6ckBZv}hDzEpXM&U+iHFGQPFOXr(IWUGI;(6G8o(6PLx zB(vA(-ddmou{2sG$Njgr&{^Zqra@R+9Wq|ZLFqn7BR~)h$P9nisJ|KNtMK&)l z$}AU}VUTf0%X#WKiw>Ys_Txi#>%JNaV3dp}OKCGOzWpr2h&uB)OBiz&Vp;A&V?Z{k zSIm9yIHYbb;eaBGZd#mgf3(r?$)-eWX8nc6MDOmz!}uS=Yu=7->w~W~Y?(iOr(hO> zJ%uV8(iU?^7oAw1Z!V~k*zW+??BaR6vy8d+(4DdoWSA=w(EKs6fM@5d7z_d=jxSD8 zQSJnv=PLuz6l6o?oecfM<0iM;$qjYS=GG56T39CAo22f!dQBy^ zO^5GnLu?UDgt=nI@ksDh9?4!`l3~4NZ)3jZAfkNs(|RL`sh7+cF6B0h`93rgg=+hh zQJqlFbjS~{8twb+vy!}8!-3xu8Rrdec=orwXhxH_Y)3-we%@;i0{FWdIS9xhU6>(yn0Yl`L-rXN#f+Lhxq3b@sFrE? z72vYWooT$-*j(RzdWJQ2nx3-c;cs@+z8a35!q^W9^OjVZT-v-r6Uv}@Vgf|a$q8!A zMK{_luTs<)k$%h{5=lLPZ~bE!PbKg|9(>RW*k*MHZxrDiCCo!!^hP_c;2RmgQ4e_L zGCfQ=#4iQg`l_epzefYkC%imyk)VZO6+s@XlUu~BnoS!GW@*l>+WF*R!)tloZ$y<0 zr!}8lxN{M3pLsK$Iz({BZdw}6*6yiS?^)OEZpg|G-w6AUuFM(IvA#bDBe*?At zUh^MyWIcV~O?wss(o39K;gs!nGwVS$=>qd!)WmP>uv5}Lx4w2%g@68JhQI|t)YoBH z7{0MD^^?+#N6&|n4qbW_2k=6>U18s&DKU$dF6uQrkyX{zY-*sg=6eb0lL=q|xh`Gx zZPaY18kzCKx?h+!>h0=|`>-|9l$o$q%$}LMOW}RH(xHfP4@mU9rUa>W%pAetod*5# z-bydYABh1*SkQ?Z9^sX11tS52+qZMnV7o=~YPfn&z=|?;>Ow}CV}X3ImcW+RgZ|Eo z&+U$v372@7V0D(=C<6JX_8QNWotm#eKebbz>9rp;lFAfonH?Y?kHxo~a`L~FldieH z=w-56CGv@#eWZi+4Z1cJJe1>IUZilw=eR`eDZ`k5>bFw%d8^Ke#ldL}X)#*geUdB$ z+Qe02wdQqEIoxloz!$7L{Nb9(dLTnGvbL|jcrmA?2Aqxs95mgsgDN8Q^fkbWobBc% z+l5}&k-rP3U7o+7(C!q^f+FpEMHp z;7thgl;n;QQt7YotUUZIc}4R3A>Y~jT#O$D;37^+Dqh}NzU}y`RcDVVxvso}YKj%{ z>Fe$e3cmo8Ud#bL7rZK{XX6x7AAa5O6s$J9MO1Gh_JT@2u z{Z=t$I^LNimC=Q}dgS1*IB@qkX(xET%C(#~JSC#!R!-(*?VM0t z&nz++P&h4*L1T_;=9V&n+KLn zQ{DDX-budCO3wmcu#ly}C`Zn_uzup+i_heA78zn&?yC>>cp2;G|9VhBdg)%J;ooUX zaA9mD%AX9%HTbAqCNHtzwL4|ceQsnl;$ls(INpkEzaMBh-4A661Xun9YTN%fv@yRF zQl0d+{|&vr-ryh2_bSv^wbASZfKzdrKj~NuP)}PHcHg`Ig6))y>khaia-P0o*^!9z zoybkA?KP60n_to?vAl>x?QH>}fztSPe&YU!GW9r{bp@Ssp@69uj=N)n=0u=)IwQ41D*D6(1&8Kl4@t^JElR&3 z`kZj+#dtG-wYxR(Ht4xo-W}v3L`7kJQt4e8znAQ{%3`YOvo4cZA3=N8>58U_zv2 z2S1V>YC*5u3yl{Gnr*7Lk3&QnY3fq29Db}#!oe9|-FlC#z5kYLqQ`SXSOK0R;}(F3 zS0TH8QbIZ^=?2rvPM)k3wF_N7m3npxpyY8HZ~d`-Qmhz>4yJ^(NQR>Ie{jN zra|>bTepEoIHta+AM1w{HJ|Rm2#_vN`Y+4?Wy!=lK|f%Kp3pMl^p08$Sdon$5i8xO zSqwWV{l8ku|D#>|KXSDFb^M%%k-33V*Z#XWJ)-{qG@t}-E3M^6_L7YI`*SOzVReeR z^%1o+6WDCXm=9`LCj!exhEPv&ZedMb_pNHMxz<8B%{eP_5>$MLBVp+&_bVnP6bWJy z_70Pb2yK_+>wH=ibPMP^t7M%M39CbeN>~!(1y#)i?|7;?07^dm=6^b5fI}N)II1NS zIr8KYrbG<{2Tb$+js8Ry8-&hWvqL>K)EJG8P*%=^BLglP!Gg9Jo904HM$*~vc^C2J#e zC+N-E{CFVZHkA)x$=zJIeMOwdMeojV&CP+DpA}V6%ySg51xOo>grw@UlMxf-80TiQ z3k~(Qrpao<7c);vlNOM3K~NTgl6wkKew>;(M;EP94Oyx&KY0GPUy#B$5hG3sSYtp4 z3!PVHQW!)WGsIM6HfM6FM>!6l}e6@>5b-B$EuV+WoCLt|o0U};Y$#@&P$qIww zBB$&sjVZ#G+)`pvcKO2E2bFzq3KjN8=a8JA=NLr=KMh^%+LG#tcc^v>*Cc|?-O*#T zwh(Vd;Tg0kiBH#}tgO!Z_>E6S99V0yC%LydrgD$mg(tIYNFvPmuAQE8qeev!qCoD9 zb-p7E&99FJYo_eO6cyzyhh-yn{$qK&--~(?*QmPYd5h%4N zOOahPSofE=g_zVM$!Q79J%YhYDG(gP4$qQea?2GXwTImM%elq!oX5r=^T-|8nqI`0 zWH2hbCNbCIo#Xx7wF}RNqb6Y;#{%wLR`1<1a1RZ!Q10`O*ADS6y$tzG3C+2LS8U`} zgAy;(d$KN6KKW&{xOJCBkNjx8z3=26N}z^dgj3q$FMUc zjwom_RU)_rhIf(I*CJkajT}pwv&3cXg5<&T>OIX!Qk2~_R{%F-N=U#D`Tp4=_V>>!(!U%oZE>R z7|Khxq~8@rqnh}Acl1$AP zBD+F&a_w~Yd27UkVJ2@nJzi_ja9}$+=xTvY_g8Ru#4b6C+^oF5W0t~x{yE-#;PQv_N7KPJo z3yJ2KX9l+}ZAWpZ;N~K?oYjSLdQerp4kPP(mib?eXZ& z2Qc7vR@oN|n2{79d0r?qHtE^%VdwvT`TT{g2ZLn<)Xrqc7(1>?TQt=0V9J5g$;G`r zE$vFC{9|w!2koQHW-bdD;+EytR_s=}>k6Qw&-=%5L^CN?bH& zN&&bWL36QY(AnR}2-u$1jrFFDL-dLIuY+16={S3tuuOh6WKIUi8QFhyb$l|r=lC!O zj|46GG{Q~&ua0^vz!hB4vJT{iYR|9JYr_Op-Qmam;6)m=X0*_Pma$Wa%9zV&vG)7b zaVhChTaUyF?$Xfnw*dOp>SIbQ(Nb=tr9h$Fz2Q`8(WumOgVNmpz`3W2rOv%lOmQZK z{?fhnUelu#yY}~R9XHS`tf4|&&AEgVo>Vs(5L-~a8 zG4&^Opc@VQ-Fz$Zy1Ppz_wQ)l#?FBnYj)Db6snPX$)Q%EC#T?at`fYpiY?-AvcS(A zOUn4q?Fm1>4?l0NHH-A!((vyqL>io+?Y!r!Cf}#<-{-mE+rQM0zyJOJqUHauzx6+H z_<3QEzsh3xarqM{tCEWBo5{|oqW+g;0F{ogBwNHmkT$?y+1rc!Bme#9e@X@ZTbsep zB3wf4_)(LuJ-aPzcR3pm$2)05P2Yh_DiHbGAir(+$ezdmj}qvm~^D-P-LV zi%`2fJbwYm!KPb*|1uH-##k>Y8l{Zy2OOv@dp0ConhPEO`}vNP!o=!~n{KU#Ey`O; z1e=e)+8cytKqA=trPj_?I^^t0hP~0hh#UzJfc(n#6L9#i!UX#x>a>rVQ{QQ z^vM%&&F_i3G0hT^F*I|jedsHIRdDFP^8Vh(YS2#Kv4j_`O;8NLyVWjQ z@NH0jaL#hC^O3Bs4Zn`}y(Pw5(Ape5tCMwJdzn?UY$)fvf3o!?3gZgr#W+IRXJk(j z?4teYaPBQ03hO~4>GgqCsXzo0u8)C5dtIrt?wB(`P-oaoOvv8;>*Kr&{(LW%Q(?~L z(Qs=1h3$9mT>9(R1sSF1lP1Hjzg#^+YB!H9fGCe@H+k{5>F0Slf= zfQjyWKqD|zmL{O+vTI#g{oWqiaYpo?XP$EOmYnIQ*;fq@bW^pce$fUI*Rg+c$|)x& z6Qf_ZS1jVS@i8f1NiNtmY)InAM-wtFj~0$mtFp7_t`()XJp+f37Psa0cu4!}NqU}H z!z;(_6Yj6oe6HKQ+uV^MQd?Qz>k*?xhLjuhqH&SH8Nf2_Dj!BP;)FWdL`~=evN& zb9}r`gG!wEuQ>tGHc#6CpGmnR_7ZrsYo8)K`}1e^b+@j;)_=;I7V&%F9uWBJJow!W z8Gm6sfNOON07eIN<)lqO0%9!;$EXD8(`5k5&}wBSo^f_W)jh=N?2m^9A89CVss;TF zSSW>CB?yX#=nK#I#p;L`HvF~R&8gid>Tuoi-#7cc^I!AG$TBG2tp6J60sC)6&wwR$S_t&GkyCjfqeF%*RjWPLK#!XO zu0jr|#Q*Z%V54#U+|Ji?q+_L*f`Kpgm=XWp?KX>8bXK(b0LX*bnUW+k8qU`^j#K5QdpDMNg(dqwx z3eQAcj~eZM{|;KgPUzO`5!oDhG{N-h#T+GsamI!P$bH96Dq2w_=K+xO$T$!8b3^wu zQ{IH8zi71jt!{A~*kDPsj;OCdrk}vNDa{9F0tiZPB@*R8LG1axbzdKl4FCa2@tSyC z^aaShfBi3^9YM3!rSkwsS$?D_fI4&;355VbKx!&Y`1?nt6rif}1Y{?;6xQTm3^W%u zM#yOZ+Wr#&yvicW7O>Mb`$$=VGN%*i!QkAq)jI!6*}se9xbX?|YJtMf;4OzCLD%a9 z7C5&f0RO4IX+i{3jlbE+fH3|IVTYl_ZLqTZQa=k42y5^VkDOgWHZmThobJ6>-+S7_ z)b$oX5;>rNo~r260#3wtH6r~Vf39yopf#@QD)49kN>QT8<9Xzp#7`g1hn}~b!TRg-V`Xq9 zdT<@|5XG-4{)m@aK>P)j=+Me7kBSLEVs`&&^ro7a9=eiI-T?bRSa7D{51O#T`RJTt zOkh-8gkfQxcaU3@RU_@V z*%%znFjDX`)h#lHuvK1@IzXc{+j!dp_Z$sJp*<1)R;MpDlN@0KZ`=FsPDpx=xRT{> zmm*@KH4z)vUL~DV)x*WS7 zRvsby#EhQFBKd!l+F+v#E8g`BINEGS@I|`b00wT+y`Ze&gR1A(_~W9R(3S?lckpo5 z6Xn~0IiczoSS~v(Enx#F#j)LOfcbFHtUVZg21%6E9E}8eL*yEa=WVTE_@8e=g!Q=@ zFrIQ!fsZ-LOsI2KGMtBz`W9&X50-a>_4%#_u(-^(Ac956p_(uEcWy#f2PK8@^R}j3 zU^%s=F2Lcq(?$g*Np|V5R0Nm>o;2_%s36^}SOQ+b;ZruM>2R&6g`0yB&?cA>`g{TQ z(2lXkMao2T^Se7CSG0Tr2?P-_9NrV+E|(6$*Qe<*iyMQ7=l7g}X0}5M<`Ia(;T{o{ z0u0{XKP8PM4R|-3z}-lz`ro|*>uV1RzqF}Gq}(@f;wBZ)!ShGv6X4JnVSPPcJyU2< zc<3UBH_x2bf!R^+Gvscc;X!UpHG;ecrv|Bvh1HDFWlihGzzdfF zujMxv0rR=MR0g$O5EPItRM<6W2tg_M%CKAM1fG=QD>!0=Xdr1t9s~|uM9~aSTCc|Ij+){FX6~G?i9V5UOc)YDrqXR7E&q2i8ViLW^6Q2h&01b+*wtJFxYFF^iUVb|dYYpDXahRY?sk8j2;r>@R9yY}u8K{36uhM4 zW^v^RAhVqS-25W!!|3Cg^T^~q2=sD7*9lAOzy@GN!gB$dP5}QQ@jeh3C4aY2YMBAn zCnG?cHr^OLrov-)VfX_w;%RFd2a}JWcTe_gi}wc$w&>%zPfCh`8HcliV^3{x4KS6V zswm{L&_IGaBY38U`>flWK%=kJkQH;3n$SdYQnxLztObQiwM{%Da2x!tJ!saX7T4p- zgt6dufa;`e#^40sjlegxVJ$Or6yhE_^Z5>dT0wBg9b4wX{QmM_-w1yn1@ja-h6F>P z!aax$1z6*>8_;A~Kn=ljJP%BX6MmuqR z9d}W*MbIpk;|~C^ED~E*?otVbpYx++WZB*o*w3)yor5o+D!#%K+GDqzLc@a_uz2Lyw3$bXde|d*qE}ZfHo$ya zM)BVK%d=e!xJrk)YJkvl;yYI5;Eb}1lA#Bzj%^VkoLh0`D4uuCl;~t$or#0t?zRqP z^AQ8Bf(9?Pt;f`3R6X8(yMGeaUW9L#3Tvu>9rAv^PbzWY#x)MQj*t0La?bUjEBbv1 zr>>c`Gop@qryv@gv!>=nY_o8qN*D`57Q+JSc=24r1KSUE4bZ$ae=*JBh`JED56QZ$ za<`fLARP+V&rw`^>78KV!^nzwWsU^h(EcX2Um1^;iT1g84cJdq+?Z^7e)9#{prf*% zl9kWDV(ZKI<7Ure&6t&&*q~MD(nzuP z4_jZ1af{@6w=Z!y$lN5?);7Cbx%wcz3!TC)WUPUF&`>jxXe%<7B=q~ThW%E z$r2*c0V$XLj>9EuYMhvYu_{}M_DCiT8|3WMwNkmU)+)5gh!C(?FJknjlSSPtWw*5H zK^yl`zicfD`P_=4>dAsC>GON1s?P8m~Il7+G_iq(L75SZzN0?|)>q>(nv&_fIU^96d1NezX{97?jQT4DWW# zMXE9(K z(##rDs4Bz}`V>`x@*)R|dYce`LO^Z{=Rp!Hxqq?uE{ugXK}A=IvFy*|aXtY}>iM8# z<*<`d8>(QU@?}B(*4zit;fpWeQW@P%JCZkUk1SVT-9I?~Kqhr0(qQ^<$ASoOG;V$;f^z%&13P?g!nJQ9WrR;@_ z-1#9dO6GJA?TX;WL&uG+DC}gl1LUI`b#7KO#U7C#fpP`+Qp zroE14Yz(+4w{bn5fCSh`B613(lz>>O`fmM)-^TY(uH0@kYC#4I>hp<>7I#);PSLA&u$QOGxxwM9OQL&W77iG&hs%KEBx3FyT(hmhbzGbL$TX-KjVe zU_mx}xe{4;p;k+*Y>Ty--F2MzcA8nyyyZwb(JNe8iBV~iHtg`0k|K4Oz0iOhjK>CXMn&(VsT}|? zAja%QCRe-LihJI*zCBmi+@6|PHdG5pxHWbhyuqlyVbpz~k=J=or7@dcBbd` zNb{Pxo?r<&?Z|a%r|j`sb^Eij>2=s<-^B_ZSx5b!Tg6sA0UD?@dr-{z`dxlzD22{N zoenV`BFDmyphx z{ZOdddpNwE(hV|;EjPcS9OH?Q(TLIww;i;)T2ryrIwlcIBdT%`4tUR{wsCC<`+;e_ zK1x>-^NWs#Rkw)~U)sL^0_*3;mhW!n+qJxQnZp(#`F0`SL)n$+$Gt)bakwHg`e0MnQ+?neut^ zFUT0yeynNf$5@H@#%=V%3yzDCgN#o~d$O7S`=2K0M2Q7J<+wC~tTo!45?@`5R&zM2 zqj@zp@9!19sk%GS%7FdF6AEB(tsYM;(u9+cS}g&pzZ5g==-szeSVNt3KdF*NfCXqJ z65RK`Adz3P-bD%PMoJZa&X#st>n>0v{}JdomMB~{uw1gBSlbr#T3*vIUu~0YnJutf zj#_YeX^}^-O32mb`*?g<9Dn@rXl=PBdtE9>mY3F{#7T;roRd?V20Ik$`-#MB?7=@q z4VJS!+Uw1Z{`-^}en6U}r`(gNq*s5p*Ciss(~ej{O^d>|Y%~<-OXNQ%So$)N-v>$- z@0&wE+rs+rw`De#mY4$p^A&p`AJmv*)dG@m>GLVv@W}EU9tiu(x_|0sf*4#&6QF=o zHp8-Spd3cYR8_m=L3r4v^}Ie1=?t`TimYD&z9G}N?c5XLPlNO;t0j`Z8Vpslo{pMh zT%pr&D6nfObSZHv6E^J*Z8=tqBLP1)0glyS4;2CvX4J}!b5NiUQT%RQkV2%^=)wAO z180}O#bMMoj+gIEep+%~D@|CFmiloh+?P9qP&lBNNUVs_J zFpWi(T1!EwYsXdF$3}P;A_^UJy#@4dd&;gr=u-C2SGQLIi-CDY)Gp3C6MIqiOMVl3 z@y``|WSDF2GiHnA{a~Y?A~h0uh`jhRT+4A)0%#1+i7@bVq>O_Z-aDAh?du~=uOWkv zmKL1HuLU5De14L-K!K92y88<2s_n6!-_ z+*EL&@>}yjTmOsv57%jH&*m)uI^eI9U*&#DCwG_izqT$;qWoEe>$jSyJ=TW^?Pm@n zQyUfqHKF+$(X$!MFlZiGEc>cz@UN;QIRUReg@j-!#sdih-E4BBd*`d?T+y(sBlTza zBi6y`mBM?Sh()$sY&}Z2dlv~n6P5Wu@}Khpw#<+gcldpgmB9Bb7vIvSdj!9`#%e>8 zS$zZmXX>u!!{r8~qAoq3J1AC@88BmJ4)fu)RIfW;yAVktdU`TL%N`;=tW;m!gq{U~ z@ol^u(gE@Kh<1I#oBP#zkK*GRp%L$dRwCJE=s)n($kwp_ON(Hv@Oonc(keiZjch@H zz6iH-ygUP#sYi79x^Lrr;fC?@dXUTaDzA4A><{w%{%-bG@zJLDPHBDRiUw~G(h25x zE=VKeuU#0apRNArp*YXbqmgyMFCOZj=v}zev1=rfr)wEF0+703TYaF0r*{dpJF%6lPm+yD(^C#QJ<=l|XLl0!kgCjum~VSH)C$ zvG=B5+RRL-3o;UX;9`$^kf)O?zWJVVN_|<8^y%0m7BiLRj-qwcs*%~sMPU2Sn`48v z&o2Gfgzj+lif2f+bDwyGQujh{HQzuJyLj_!M#hE?*`t5B!c`7z&a*u z3>-&w&4pwMea7PtW;Me;lV40Xj*EXPW`0v;xc>SCv^G;qJ^nnX2fZIAamL{~^^~eg z_SN|qMXCyJJ$N-ZJC(<2Z1r!t3y@$``F*~(Moq<;x}1FKg(MACtg>c>?Dsbv6K9VZ zQ(_w$*7CWK!LJNnEzfx%Z#LFwI6Y|TSZ}57Mlno;@3qUojwfds8*l?cgm@+|+MY~dqM=8=~y=$*ucw&$D)1|jH`%Db!&#{s}t8m0f? zJ*^RN$49vEU3AW_gmAol?Ty@V3tNklT;(i*5C?Nq=@h!XJ2Y9@T~?y~=$X`%_zC?D zd5rsb$wlf2v8m+p)lxeSF`nGTvlFM@ow0f1sb~nTd5v>LubWfK0{KS{R+>K5dxkvP zt=59!ZLz@gmehP>h5Qvc>#uD0k{6yTkZ`*xkM-p&>V|gE#`4}wLjQN$^_=-cY@&9M zN?1%!XOnm{Xh%br7osq>kLA^GiJg9%-Qu0koAX#i_Rut=@bPmua%BNm0E?9u&$Vj? zVY?lz_JP=8!*^i)?;x5JmzxgFfWB&=zK>-$FAx;01lwSg!d}OAJ^#4j zWRf!L`+Sjdyj!8WYQ96viY6(E^U#3rGCa+Zds>E5nr!S;noJyn-(PQY38&Wpmqpn6 z$~a&wyO|@;VuVlfdIjjZ2FgA!%w2^7&;dyD@?3Wer9EVYo}^WmnOWX!6c+mt7aoG$ znD@)_($%m(*VZmXIdmQ5H{>xqBB-BhkX{5FJnx5(O6)+7&MK8;h8J}%l|MfHN1I09 z$(rYxiv2w8E0I#UsbAdZKxiM#y|3~lHuMXqw_FLEsp~3O9c1L_2tUnAj|kTQ`J?j& ztZim^2l934<|AI?`!hDp)O&Nqwh(JvQ*Mx{`aL_|f;DF_bm~sef|TkE#!`l?+H%FV zR{pa1y*zyTrzG_NxqMwnF=Q!30{ z4;atkj>{HkUg9&CR_1YBOpHI}J{>^bP%u$<--;8*p=ou6df2i0+6FBi*S9LFc8^zn z-%Tjm-Q9ZFMqYe&An6i^5M5@WLv|Wl{QJj0;VmnF;Sg4>9z~`aRk5W`@!+PwNAY-{ zybXpMVGvMZT$WL?V9s`(pE{%_kK|k1o+?gwQBlT+vtR;#9oS|^ zq5OH>8%VizT!++vkIQGg#^YzOJ+8`8?fpLwY?-@d(x@nV)l4pC_`|H%Y!sa$nvnf* zd;Bz2cQQhEdqFh~Mc8tkSglD9iRa|$KW&t6)ZccI;B20Dx}!|T)y8L%hx+T+$X%9% zf%J^jcJcZFx0&8ILN~d-31HeU{a7XXc|u(^f|b4yIMdy4oUx!rPlI--kGETy;h`gWPHdSiTJC=n@<{hcd!XvtG83 zdBot%h>|F?K+7zDJpV~1_Lh0>fN!amBDpwC)P$g@of2{2A#uq;=>8nF|2GwT;>t=` zfmX?vM66KWOEaFu)!(vpl*DwJS?0|Hi-R~_+se$frMddyYX@HR1aD_qS|xw`Kd@cbZuda;fonR;rR6%LkhvtYdLx<4n_kD?>Gk=kh&<{$d5zz| zVJ${pKS;o}w#X_WzDte88a%Dy!?CT)I&ZF$Qc6@s-&g9xZ_OUi3(EFtYo4D$LgwnglwHoUec?*)KZ!2M2 zX9=DLoTJH{Lo>X?Wxe1~x)OAl{&vvKo6yhpbo6mDzN>Isdhc5? z+Xq;o+HQ5KX7RjbA>+9QZG+Ia(M~PgJ7t%6PEADK^w-o^QuXi*RL)p=$DNecz z3V%DV1N?rlVyNqd?s7dc`<0vdaUQ+rPjdcBzZaV1W+qlQYIYM!s!0Pmqs^~5Q~D@1 z_Gt&t8#>D16t87A$seno%&05o8}lsOXBsx1*z(TVGovJ$yqNU{i|66!q-R^D+`lO= z1+&KpP0OipjjURQZ{}5T5=K%(GV$cuJZFv4$&3$wz(8dr{DV5Bh(Eb~-99{I(+H7Pc2FF)#rS(jn zKZiR0Zt$QHyMtQ-l~N8~iuh5};uJ7`Stq|AlNvls z>bZjYQx{Mo6VIb8E-RqYZHT+?P|~97El-O2O{w%(MH8IQK&ohULeD>d3Jy)X{Q?1V2?9yXiJySdx7Y@h zHHZ{Kn_-*|LqhKfm^UXx5&_CSC<=aH0nX6#36vr=-O^{6B*=GC`pU*3BY?IP0Lt7T zFq;3&fIMgZ$dvOd)0y4`81EiF>ot9?_2lem?dyPIVE+lw5FHaU2P}HOXA~OH;mliL z9B`26+)4%)&i&w^O&F^NG2GGbvTMbp1<-hYhT5*x3B*>;PCwYHg{asJD!{6hq!*r$qWT6|fmniS-?-Hud;mH8M?n!v z^5;f&M-vQiUO#r5tGo^>`H?kw_>G(4&v*DagLrJI5(ohz=XmAk@Z8W%f5(`V++F{( z6acDtM=O{din+XlzYvgy=Ifh5c3Vb)CeiRCGkA6eCayaXogh-qb^HxnR0rmVfotj1 zodn~W;e`SJz&cO42A6#uD>wbL@n?aQnGQZnOAGKF@yri{0jl$n4C3OtMZtyX&v}ab zTvmVo3PR?LbKUTZlfIn+IkUjGzz85D^U68D0}AP;nPO;VmtMQp&E!Oqd}crI%QG&^ zpt&PoQOp_fmPQx^9VweNI${7WHuX)=;qZ5kMDn7(>H?tW-Med>DELu`E1z&S46tm$ z(%0<=#~_nAxMLMQoZ1{8HxCJ#t`QuNC`BWCcwsnyfT2Qq65a^|#^{En9u>YJMjFh2 zs{YI%g&=bn_zS61aH|c0fBkdjUK&lGaZ7#=i~(Y1@kp#&KMKh%f>P0D7=HlR)yVh$ z4lQT~Z%b)##$^UX%!M2^52;M)8LvNM|LKjC98g%{-5^>@4GgCCG})##ZIgBe%VHB( zAbvfMDU;VKa4Iyr^YszOCNX%L+FyYDx)eY~y(iURr3j`=1HA0!PgbY@G{IwPBGTQ! zD*jam&`D*!okG0f7F+0_)hzlH{RtO*8hxDi7nu4IF2p{HniJ6X5!1~I4zTn!wFKtV z>_Q#KuiiQ|2D0>rY2fUyKKDgW5O%z|ZkJ*k-?+9kvboLOwqZ^|P$xyL#TF*4Y8DHH zt}}OV1nYqKs46{V78K8fIkun_NNNUysaqYDB%q%&Z-I38%>Y3dSplxo)KL&%K)oDD z=!4^GI8a2MkgJ(==&yCqtoshap7Ut0!XRtO6PfNr61_?b{0$qYiaoQ4{`3OUBw=U(S|&uXpbL!B~R|-FNfK=XS@A^yq}&(!83Zzx`{p6tHuQ$D?`dTS3WW zX6=8(emWC>TNGdNwUykN+cqgvmJ^T9C={u9Y=YET4kl@j9Ub3de$?vI4r{LF2(*!B zN-PpqEMhhrvPap9kl9N1iIA_SSLmyg(tE^_M>ut+VBIKCn}=PcRE$^{XA}|X#(9e( z{GsoCE;%JnpkInutGm(9-{O5a+fdNstxo$De)tnl76J+;<3wbLAN2 z{@0f=mm9o7WtubWzn#~7pZB9uF&7?abE+_ydzZb#1J8CqD-ujyMw*2Kd0Y-1!($GDLHDQ=rw z%M(fQnf5sj{Wh{jkFy0zaV62Tlak?i9rN2OYydZsL^C4Zj(w1!BOyg9nNBPn{fE3< zG!Zqy?X=?TpsS_QV#01jHpzPV1z#E{A#OSlJ=}?zlnWB`IP>Enk*-S`y0>xjZ3}Pu zN4+&A+yzY=N@qeTDW=fe%VmZ$35uJQ)WvlC0LsguGy}d?NqmojNE8W9L@9++&Gb4i zuQgMc4nLnr@+h~25R)^S12C^WG^=@7OH9DLiva@1@th)^>p2ZS>jE7sGRS)pW6Wbd zTx*ePoLR}KDXg2xe%<-` z#7Gmu!Ppz>Ck0m;{YV#k{nD3rd5!k_4`~oYc(x4;87w)-jX)uq86B}td1;Uy)OUehbExZ;L!wf&+E;>wlNe;O7+3M)j z*{r$n1Xwy+>oT!$YbZ_C{p$6Zt5PA{HBXb3H|Tb<#0B(Va>-r8t^rl;o|SER+=h*& zjEXuJ^6&g)V4E~P?=6^V9KX#WzH>DBScdaB$jgdH>YYmCh(42f4WDo(3wQ@!PHCRg zVbgt8)4q;(acljgg3zAE$K);H?fWY=pbbpRui6pIC!Z?Lu{g9cqFFIFp38I`BiIMB zr|cPai7l2Nxx;@Ylllz=`583 z(>?{>>dx`u*>{Yhrq>BxYQn zFfqEkE5)1c*gi)z|SCo^n=$c^4(-#5R% z*eGpA+eT9o6zunxOU_uDqx%*~4>D}F0ei*u)wWE=%T%L4%puU3ooBAVJT7^P?ywj_ zasR@o0jgf{|I6Tsx4jD!H2!yo}0>oUjw_RU*>6CR}Rd)OugtZ_IUPUqYluv z$o`>_aQjWcG25o?`96Q0F&Lw@@%mu;PrpAR>li}~n5{MggnBcg08-Urc=AbkTmqzk zX>I^R`2tFh;nz@xho8lSl6s3-1d_YL)y$Ju{TUc2aW-kVl);lr6gnwx2M9cQAGKj1 z#Eok{6%4rv{;{=i+4viS@PunN4>%|kcbLwl4@srNS*^ysT=GM?8t!Zz=(!%rsZl$_MCgCEb(Dzou(_~&6&6f6*0gWiEW|> zU&1J9w~`JE@0FG84TPl9(IuI%mNQu#Vb^@{_a)9F6rSd`JdD)*nL4I%L2S zVW``Xrx2>KeG$3Xh7~$iTN?0jwI#Xqm8AHpfmqwXKX8vK+tjL?3woIF7~$@n|##Qnq}W^ zHhAVNGbM@BmR7!D+M*|nf5{)E9{mLDh#9_Li@KIQyk~J0AeHDM?e68V&H1~uTq;8T z%D=9M$!msV3BO}!=G>cEIJS&@@lte*(ao}@3>3!OJbcyFY2Wqrr;aZ$TB^DufYRhs z&*%GPAu z4{}ia%O4=m8|>o2a}&VpqZVhhR^jKDc`WJJYyAohEe^*ieX2*oK3-F{Y&=}i;PJLV z{M4WTUKx*J}gfz8J{k?p40V@8@V&f^JK4(HC~ev{GRTm-Eqcd8`(TF z)s{@n;*BI9$<&dMpYU>F>U49_W4tgo8)B+WTNKZ^=!fR{^x#+VJ!E4f%=w%_E$4_kENss9*qc44957;Bdv!Kd;G{>8(d~;Yo zkT6oI$8vbR#E~D zwDz!X*QDRLk%#h5DW0d*lx*%X(*w!AsZ@baQy`Iiy&&}iIiPU7ZsiW~om?i=uPt#i zo&D?*PkFRzr?ExCCRx~b!2(*=3nixa~j`5uQDk04g)%qdW$Y8ubRUP z@0q#p=#I4N%Y+@&*GccsCwH+Al;T|TJ&Z9$z;kk;1d(I8OBwvkYin0C&+A+Eo!6?H zvKM+7n8exm(zBno@{&o73A{_@;nT#JdNql0-NTAa2VNm)gf(2lm=K7nFymEO;7wtQ z+x3T-WB%aGWNxZD@2@K<2qZHUya+wtr&nk`t|T7GNzsKTC2R6hb#XI^=}WBC2NQI? z7V*US0R_h^f};Jx836;b7nub;Z1Qsb(Gv0oi$0^kCTDo@w>~4m^R(I$rl6(>kZzfG*)&-Q}mF@C2*>J0*JD z-=R_<|MtLk{(!h81Hd>*AxR_3f;$G9NQ0$IY-_t00}18WOm$s(sjrMnE5{kk?YU8> ztHFUCI0g95C}Dgsw5ZY9u*O>a%PMOYWZ4>bJk z|9jF$`sh=$bARO`@UM$nDkuJz!1(_Ui~9c*-|^oU`9B1T{6DDL54b0xLpuW@=BH3M zj&)@%4^mFm2ButmTasKUL?qab62n4qpr{*Kn1a)dRCuxK1oRvJT6>P15>irJM?Okv z7XYB$wR~s8A2{r7As^ZqC5)l~@4LCsQIXtiPl0}}dZH6r@ED}`I5!R~aLqKufR9f_ zQtc5dzqA|JLeAkAq!Kz28JduUbij)_{2@Rme~5vr2SmIuzD{eOL|fXX}tf$m~~KnY*YJ#yan6$*t%Y~riHiq(V( z?MDM6b6ZID;~Ma(2dmeGR>6oDFqm&ahI44#ggo4lZ_6Tipks~%_`=m&nbf8qai&*6 z_>E*pc0k^?s`i1oOTL8>(5$2H@Qt^C+mt50XNdqt6jN}a%fM};Ro%^UuSoTWIryu5 zM4@;Q50nj! z=fJt}Jr3@;4_jT4=J>CVEn-^#K%8R4%3c6EcM|-Tp`@9Q-;pg03bmAOTzt-EA!-aZ zQNs0&@F7kigR$!a@Ah1{O=v;l(M-8xdVQ?|^uz6ryb$5RG&Q=y^()gUbyF^Ti0rLrJha}izgn*-XAqv>4d zlecjLzoFxU1mI<${3hoRcUwE;RZUGVSPQ;=>f$%}TF#@`bY!p1wFsK>JFk)BCJ-ov z7Z2y(8&G;g1tx7L(x*PjB2UTowsN!ZxYW%Eb3L>NTjGS1){l9df$c9)--*()DA!J? zy)#?;>;JX*2>V6!C$cyizV~=AKlO0I2E?+Odb?A)LQwluG?!~b53-ZX(KiMQ7B-7% zUk9&<-kZpr_$2*p|1ZypCyX8#=122ZP+22oorhxtN;mAy2BV=N4{!$1;XcjJ{u?$F zl>lyw<(bv8E1;}Zp3;q}22?W{xRo&}7tdfj4{zV;u?oxO%sqE-!p)*gA0EpSABgJ> z7yHZ~*tjf%zcN#++>uXud7njH&k#%y`(E8pQZSitvs!i#O(kN(phHxqbsKTH79fq}>lQwZUKJ4pk()~Dd*4|5}!eRIn zI}71}v2Nt94SQ1%OCP+g#CGB+9Y<5 z2-X9^OeT^mSfzxA-8aOiZE{n$076~mU(_vu%^M2Y2TC`Gwt)z7FcgIW}U|ZO2iC5-lr<7Q}+hONWEQ|SpkxOAgzTt znjYl7%9hRFPm-WV$8RdcTLfXVcikqIq_7=71j6I+-z#A85kSnHOyBWwfN9*KL6qjX zz{7;LU$AXn-R?KIg?!ermH*%<+UM>MO>lr7C%&2jOscApA5E@4yB&BOcNy{f1Agf} zHXQy_fG)2h+pE=VAtx6bSPJ0jGv5Xo-7@e5TQVtHFd`NqC4kNQ419!`+wqmu2GT>H zk`h&~T{tSyeNeeiUP^sqK28d3I2oAyA&%^r{wmQcsB)(MnXd08iM}Z6ar?T+g+jt- zWk=U|y%g;aWR^?>5(H=@ZlOiGQ)CG(k)SPDFWYkU?7%w%7qGtT8MLlwC?Tk4D}ankh{!`cq9${H)Suc* z*U?{^adq}QD_3p(J8P+cZ^myRzF%M(Tnb+QNK0MFIW%a8VtR5m4OTfGG1rzA$C!xU zt*|h^AczcG&9U@-G&_mntlniXtr)c+Aw&7Jh59)xSQg_{)CxLupUZpuQ)Pm~za=iuN-4MQAeO#|)d-hL*gxQi7aRqwy9OdoIf|-g9 z&SO3}Q z#@aWEZx-xP#Y=*fDj`xwt?oh()2oy+cU?v#yF6Z5gPDhON?Xg*yi&>>>w9m#9`$I; zJ??Kn<s5Ox|8I(n{qshfD|my z0m_;+?*wR=hY0?3^1)URG_{@rQCYqDEucyx1qF8X_L?xLfDX+MK??cRB@Boct2@9? z#dKdhI9{tgYw;uZ=X(!4qyc;6B$NDlHh`NN+Y^X z&=dweqi4|Ox304>*HV=(S1++ zZ~$`seL!jqkDXw`CEpy$Ksfon3-1E>zK)WRGr5IELkA2Eq;lw3&)ZN%=rI`xESw() zw%(R5D8Jx0;e6mZG%?)(RoDs=={aT-cr^SDjF?*)0FCs4Cie|b%R!A3Oz<<`V*^T} zwzxuQ)%QRW&Ka9V;Dh+S|6#!42hT}~tk-5vii0qhc)jIfLS^SyGfX4cI|_9B%=G@F zI02MEA<97D(~@e?wFduQ25oJ`iZ6^dJ~g3|;OPn-gsa%qS$Kpj)qO)~($rMY^|hp> zZNN@ndf^!fZL?QE`3fdp9i!Ab(cuQ*aYuP(Ctu;|juT6)geS1EMrxi*D1aBMW4z8t z{GV~SB%iOj{dhVjw>JE4wY>{80iSJwENnG$<7mJi6kbdJb$hg?ET9|K?ksZ!c1hxQ zI?&$WoeWnxjo^dOeRGEanaj0He$>}|V+%8&Ex(KTzU(`9p?oV+6ze*kCwzM*F-t*} z?ay;#y9(0f;nHB|4b{n{L0LleuKQqY9`gm#pH)5{tURTdYcOq-I|2ETRxHXJIb=;9%H1OC087pPkMl$m=H}*Q zxQjVaEC8lE8TQ9kFxtGqm3{Dp8485t%*&9qZIAg2*lZEh-)Jydm+CrLl)&~;)GZx1 z!|(-*(qaB*@k<)9IsLt~s|0~YOG33<$x9Ssw$q_u&q=@)pZa<++I zyFN_20~2tFl?HY=v%l&=PIiYYVa(^HH3L5DHc;I-L{yIdknd)rM zYo&mdiG5RqeB!s>qMyQ2Tzo zRk%6|e!>(w)jz8myTUB~fh*2JsMe$ zVsu~gI7SZoD+2v^f(S;zRDZn$ZrM|?Vn>z~44ZSY>ax6b0l%2xlRTmBEY;ZT-#VFh zv(&G6T#^V2-;vvX1uUDXlyUwiAmQ-Hc^1savEfrE!SX@g{Y`lDgF7xb)@%CS* zzw>*$$ORW3{BQbx54JgZ>4u3y$awVl#cXP9KD>5o6=jAkCC?XDbbx1x%S0(QpoN9{ z)}QUlb9l;K**{tRgvw?^cAl`$dLlmXb(Adzd*zB!vnX6(Ozspy6n62S@^Rw_C*eSP zdbV8u5{5YQhU);1c@26lA-N~4S>mci#CQIkAzj+VdV3^y@-MJh_aDykP!*Ivx~2tl zp^c72(=*4!)9BY7Z>yJDq3n_A`Yv$n`b;#)U0eV)-<4gu%Q^v?3=vN1{_%Zp!!qu| znUP@kjxneN5ggzf_PUBz~Bkwx%)|i4%Ud zVZW+ho~(PFhs)&EdU*|$#R_sk^Iqalron;BdI9ICJB&TFrGOetz95WDws;_8I{1>kF#P% z@v?l%zndIV*(H+<6D{8-O#+bbUIXX%Y?XcLUK;likddZ#{!{3AX+#Ri(NhX9*P6hT zJMc6l;V9Ter-29l(Kga!Qk3@7`JN|zt4HB#t|l1Cm4-CQtyRTV8!d!qDSIZ^D{}v6 z9vGGT8pCekNw6c*|V(xeCUrW!get`8V}2HX^nVoggiqIHE_ZlgdcErw z+IZ{;G&Pyr`wTpv)F3_Q45M=DUhm4<6xS!3y62VO)M|U)<37bc4a7ag7m}AQkzP%0 zQ15izeQ~8H*h{Tjc!;a(XdtzZI4+xK;7jd2we?Mgen2+F5&&z^_zDQGg1BcXtko-*>b z53K5s;i=oxyOuJQ*)tk9?tQfLWu|}VPv@Rw1%z}sZ`1J)RVWdEf}8EjoDC)Z5p)6* zM&C`6Fa|Px{#c)Qw%}=h|8`Pdl-g0*kMbotqPlKh*0Fg*Z@+@%xt!hIseBIE+*$m7 z=Wt)Ir(8^DbEateHK+~0EF|iiKTkTVU?|=?s(6L@DJ!+7H_LM#s*vTTxm03Sw?p=q z(4>c>nkMn0iSMOqHzYXmF9k;QZ1Isf(#XeXn6N2DRIjjw z)C<37-chBTd#PS1Uih^kY-)Jvr{~*_6n2*hufDwzU(D3KoQdBT6Cud78vQjd-Fh2PQu$Dvwz@1SnxYR5Cdv;TI{&Gog`FOt@Dd~Q2-BVTS@*yhi zjmAV!yJ^a4$9r2Mmuo0_fu-XVdtV-bRr!Kb#+PgcBEaKhzb;<{sZzp3Y#?a0Jy;c& z9#hkGeO&8CT^vQeFkE(A>wH~_oMV40_-vf68YlZiJa3(h30~WFFIj4^+!K(fI76fhbhg_!ru~r5s9ozl7u++4m!m?OWA1?ZE7uWN-p8n;+&bC+{L37{u@ z4>av2a1DRFidnM(JBoe&0siqF;%1R*nVVK?9?yXXCAL!0s2O=QK)RP7$CmkL{l1)? zr(6PnIiqoALp7AhFBp<>_r3r&wXZysOAKsyBLwMZ9KoN5?X6$FyG872I#wq_Zw)(A z>13z!HqU1rGIC};5Zp^^63)1BcAWIU$Gj|=T`XnKiV-pKd)HIIyrgX=7BMTbbuCFH zjVRI)MXML5OZ?8+caE%<^K2OiPS)cbrK%IL6Yj`Kmf8ex@z^b(yDV|{O_(qMH+(*p z_PwdFqDW45bQ$!T!v3#c3_D#r-hU!LphvabzU0P+-ZW;TrAON_G4KAKR&AXjCcqyf zlxNsx(J$#y0wn^DK;b&+9I|tX$+A^OT}IM26}u{2?OMWdTc-bp8c!MIvKCrJ*clP_ z@U5MP3Iwh0}m$|8IyJB{EA5O^N-VKleoUtlp zn#b~?J}oY#j?s48mXv(X^ZA$@E;%GF{K0U4g%)vcZ~kn5K+~MU05)e{!!~z#Xfa`h zh}){+HFTtVO=YTOOx)wVkB*^qxhjWZJm;Gc2~7?k`uG6bOHS&lXnKZKm`h-W+A$&4 z#D>LQjiq%y{PasJrP9o|N}0Fy(|4oZ6^{}+$MFms@bj&+5z@zLgngP-0==+dFQ)r= z7sy~C=k%G6kYt8c-o8u={E2Qir^P-O&=p-#&c&F+WB!aFz+a6tMO5jv>AvaSS4E)< zygs^#eY4iYiSOk$EvbB{^v8$z9ek>LggPz2PVR;kHUYU`AY?*y?M|%enKx0myadu< zXP)^f!*dh6&rJ#xeFINGn%h#Dlo8Ln|6Ks{IgTI0YnS`6#Rbi`y(Ev%BKQuqFYnoc zU=PWYa@?ri(7S~LrZFXjsFTE$SW*Bn-GvlH*pxZa$EO1J3X0DA476e@0@oRl|T&VquF5ofPx965{`C|K52+|NRtl`*@)*$A!{J!;*47X#P z_v=jUlVjVlC2ybgzef3s>bef}7VJR+2P@w$==i0kjrFm(b7+AEx_kU&O=(L->XFd! zVaSlG*ZPg-ZvSfC-$xhay2YFU$EB+BW12I&W4gdUyXfK*T+sN{)y5O^(mf5u_?_$_ z{NhWIg}LVbWwi?(zs#bqT)3&HJHRS}27YfFR;!yoP3F7z8e>W_*N@{4?>>)`zlr=uTwR>Zk{dc-=#jx5zB z;_I{(#lqdVImOxTQ3+A}Mt_C>*zu%%GfJ4F7u~BB9d*K6cbmFJsPAm==pztDhofnC z6Vg)+hj>ceb%KOiK^U98nQC5GsE?5+fh-ZO!ymb+z?^rv+iZB!6~2S2#=d36Dza^S=~xr>{3jBZa=%lWP%+q?>ZLc%h*Lw<>&g$+ z?vsgrrh?NeA4OUl9UQH|+&+y;r$w0R47pcT6Frfw+v}RSIpwpfiVa9m(8;V zuIgxwG)wWCOQv^gn(nkYTa zU%jXdzDVIE-n;mnXrtNs!^;_tIkKH7aRvuY&mtm&E)k>yB~B0ml3T3+)@wuKkB1a7 z3AIY8-8<#61qq_%6r#v>7G61zq=&T7e(c}9o|ko)CGF=Jvd?KpmFy89)p~(d4a*%X z?nYQ(V3H`|?h-W|+h{1`%>XSfXtpC^xv{Jpni5BZZRM(!e0bi+tF#N8Fc|D{X4A#tHHBjCyay%J}ulqM2*pQ}oaiP!LOo70?M>0t2BpCr%xQW)DIy!G6tzV&B?}<#jt)e7hZ)?0 z0rTK&rjIOL`K?Y&FX)p?fQPi9{_$K78Sdh=rYvOQt21-ceusBf2!LSlU~Ie3Xq^}u z8jD=g*QGTjejfcY833HWg={DoSEw9jY`H*4ykYSPM9OW=n;=|p830u_OZW)SXHxYr zunx)08->Z@2~`UqgkFmwwPanQ45X11SPER;G{W+}E7*X04arY^Hw6W>HduJ}v->Os z=xV?FCx4U8z_H$s5km5=Vf$GJqn-M5R2xf>ja>h>(?l>%!H?KrkV=8RXl3Yeidq;4 zaWq+Hd7}82QlPE>ji5@1Pvv(&Bi$m8DceICdv#T%|Bp@rfCIK#1teHa>L6T%g&@GD zoy`QZ_4?O<`>IwtiA{Kh$&=UNa`lT5B`9Wo!5!_qjkESj*?{eDWB}D@0oc`|mpq1y zy-peCdFFJaLMOU^9#OszzI|}w>B4gnn&CA@W618T7$-GOqWxB4&M2`*eM$rDlE!jWwYSP42H(C&SF7gvhY|XY%Z? zN}p0lCK-&+19s)vH_xROX?^E|12bXK`9K>z4!LfomUjiqD%LHCe0iJ3W>=-qsv9B) z`jighILSyIZdH1vd428jR&TepKf~Ufrj1a5TmxH?%{Bcnb7S22Pq}>jxcbPmow&bOHto>DmlgBF(admhFU68;G zx*Z{YxjLftCce7t}j{f znXtknEBO@*t%w!0gnXkY7c4VSOydHHq*K$7Wupj}3qP)om?e>?FVH+fH%?|WlHz;<8BL5yBWGCh7fhJCG$Stl`#ST%aL>`k?-ER`}k}o|G5sT~=%5G{TXrF%zR~ zwl<*jjuK7@PO*s7()%@5?u_ph+>WyYo^~LHEB)@|OJ15~#Fq_q*oZPH>Z7Ec_K1FH zMZ(6<-?r*#u!YiV(nC))gqPZKuZWV$Fl5$*kXO!m&b=%xoWO3xGL?LpMaVUyL*po?tv& z;`ozYS#IAbNdAvPt&M!HYXloqXPVQi+w2z~q&EfCv|*Tw7bA9@&3jG;uha-O>S(Bb zTM#;a47@LwK9u63i99|k49V^R$YMG`7K4or8&?vrh;?#p;{pvM@7W8`kw-!bieRKB znuQ#xQGkR;ybMN}VR$?=cDcr8t3PxrGn)$o{)ue3@ViChv25V<(kl>i0+Lz`ext5@ z^K#kG|9nS24lVmv4Di2#8~?wSqyOUa|0mD*-%j#<+NXI9+_$UVFRBecg!}_n7Y}F7 IffK3!0R|Ib(*OVf literal 85326 zcmd43byU-T{69KD5K$0BML*C^dTIKv4#a zF{vSpxUc#7et+lQfA2Z>oV#;Wn6SO$^?E)Z_1rscO=T)dMoI_-LZzyrpbLSJTSFiu zb`<2`8K+>*b?}e4r{XhDJy%;#?-%Yi5RDg}ubf>yogJ*$ylmV(99&%l`S|(y_#d#@ zdwRa|5P$d(`hR|c&(+=Tq5B-^U+^jyU#S>-Kp^y2iGR-7WsmDZ&Osom3Uc~BA2$~- zBwqV^fZx7D$t8CWUGVa}YTg|RwMW4eSI;YcAo-wwjX(FUV&cc^3N#9GJ?E8VwdHhF zKiqq2=&nh~Dh_lV4zl&K^%fcG!=XQlwa-sXEKambOqn%rRSMC92k~6~d%TDjk<1|b z-|vEHqQD{i_n=@8<^SKWAukjj{h!}qi=ycJKmYNXw&&*meiQuY&P~$){f0*AJh*QE z^Y|QLdHH|8mc2>)|LvQwRlKEUc6@(>*HOfICa%QpDAOBN+h1$-^}Ww5Dn~ACh)Q(9 za^#~J-NSiO2wN?DcUZKk(zGj}RaN5#@AH>S2Yc%ibOi0V(EmBk*HdHp%EMJ=?L%)a z(O=zHXO#(DVCir5K>xJyueTi*0MCjT*S+NbA>y_9Y+iMbfa?y2TR?&*F$r4I{K_%6 z_zlYRYs2U#@#?9MdI~%lpXEBersLGSyWv()=xR`AVnDjZ5yC5b1YQ0vY(ef7CW!g z5&pZ*DipJIHubT8{~V&T=c~cX#H3))LE>f-;jKYufpa2k>fipm3wLiuw?Id8Lb%P^ z8cI!?s}7gLu8>p}Ki}7q+q`sc+q^35ZaNk71QT#r)zms`@w_ zus)g_cDNENlb0#j=P#Y{HIhDqu+@%W$-p0+5^xvbtQG;=Z@5jGD!9!%TI^;EpQa2= zmg)^l9V}Q6BywBx7*^QFTAv*4$D6YzoHRk_r#nvY^LRoyi{G~{%~-?A=ZbU5jwkm` ziuFr%ZNJ6c&VsI)O5$g%SBo-S>Ih?}`nJ3@%-$148qn{_Ddv9Plgq|@r!dVfiyOwc zfu#F*Q)GqlL`)IBxXyDfJ8%gSH|L=FL7Bz3H;~V;M7yWKe(dw{pD$Ok&~_sq9b^W( z!Mm$e!w?QkA*+wa&}aFv(vAZu4_#%tgO0bA2@4snT?~JobrH5C2pwvcN6k=&iK0gA z%(r-U_C`U!#0SO#b0-q>&90*0lmz!Trs|=-;BO5?mK1hpCW{kyFxm!lYLA&S#d`nv z{Dwoi!lZem%e!$`*m1H3>rgW^nR=_=E*oyk67WY(uhrc~uT=Nmd=VmBW@+NBlJ`hF ziNe3zOQQrQ;0+$iE8tL_SfjJ>tOF76t%YFBx=xexY(xU5)=@$X2-QAq`P~B{Meo^pTmEhDhW9TN0xVsY_Kl#a9so!74)s| ztZDy>VDNc zU6~@#bi2i{3s=&~F5O5aA)`!2dD*_J z41~_6N?+H#R4M-^M3Yi9qfMJP&L!|@os-ZJv;ODDmmGK>pnx-w%DT}(Ej!D!pKY`%Y_!V_@-Q&*?K!s>|Eoeb#DlS zd#s-|$ocIRRu=)g5odFsA7u5O>$-&2vfqvvK8vnf#aOK@$=S&=gF)NLG=_!)dpBh)cN`uLAL;RM_fw1fYoh(dGF7(q9= zsLpi~?>t-g)wZZ>yN^j^S9i9mR!nQIYi{Jp8_KkIbewaTi;bLY05|k+1B_pa+GeDzQ=d@2TLA-$%TCZ{!T-|!M|Os%oAxxs zjl^c-k;arM;T^NLur`)2Wh@w@lvVYt_D-IBI5n;4Qh@Wqh@i!e>UwEx0;smmR(LJjJcv8)u&VyV9gN=3AOzv0X!Lj8T3-_t&vbBG9>C#;7K!iNw*Y|Lgs9VJFF1aY$VHqAxlVzxxU_G@T~B z+To}%be*>e+Ig{UDll1QZ&Y4-aLEtDOK1Z(lkk95)a9dD*NZRY2!Y0w=lCN`wenn` zC&7rR)GJpjTbi~R<>#dE{lh(|NtN>3%M5qa61iy0clEmi;q$9&Cx?Xf;_OWk^2;}A zX!W*E5+7Pf>yi(F<>P`VtW(NNWeFE}>sdvzjk3@OTgT$4mzw+^oX`yi`&9Aybce5p zi44kruS7sv-(PxE@~->Qo&(|lG|$?6XJL~4(Omf=%}jBx5jvZ%_MXr-&-EwbBDe?i z%8fshU>=(@{1QBdi7$5ckrwmcqYaEKgd7b#X6>o34_Dq1_wqj8@)54vACbUCO$EyT zFe_eiLr+cR9N<5ZUYQvxkY$gEyvA(>%F{-@I(y|lqs^Z+VaDwToqqdu#_Eg=6-#gI z_i9FB1z?~u>3O}WJ3oaYyXq)$J!&pqc79ZnJw05V;nvwHq53gBbJNh22t8vnuD7L{+sKHF6*yaT~1CtwT z^-DU&s-1@CJd6RFf@uVqy{gduhSTWf?BA@u2i*e)nV4X^I%}nSq`QxdN09kLDQ?1 zHoyDu_2R!~c=1Wd<>>MKqVzv}bn*an$apM&Iw#?$Kg)*_q7+X3`89I>GL`afa|uge zmZb0I0{6Be3~#X88F+G(Wdp!l)44=OVaFLZg-47hj3#`SIE&6PdYc-0g??L+CxIrB+JQy+fkz|vv#LXtJ%bMP=?B-Y!#;BQ<1qeOsK0_o9+&nCe!V2xsOGcHM(>^^ zcU*p}gOIE|){QOqJS?xxaJ$%g`V$RK%*)E-))BK4AFF7}To;_+{5Fcj@QAq=RI+t; z!GE|sj!pmb258Rvo{#KbSKYV%EgoFeiGMEf+=ziML)5itP{_})u%`{wEr$%AwEbTR zT65I5Qhwydm$~R~lo^!&+?Oheid2jCwQG1em^>WgeDXWRVpE|{JPkWrNF;GE+Eh~YT;CjLf>{< z13{&ntIk)BZNk3gD!2LdE#4NujM|YMdjMOgqU_^7E3?91KDACFmY?~et1rA4+>}2x zDMKA<#&dhPXz~wQUj2B25MT};2yPWH2ff2jWslZ~sGW_;v5`^9K7Dqgl)*x(m8NM47Qx=GuNtph(+;l}m zdjb~QU|8f(n&A9J+mUvDr76Qvc9r(J>vAfRww;ICB{r7+kr>qNT@eT7IP@Vq%=H(t~ z?8?*#M7sK4KxU*dn3<GO2F+w{ zt^O?8#AQ1`+IA>)EU>CAyK}zad86ZDL(o=zvrfJezMLcNAp2F|P%6L9R?u*|u$R<9 z+SVf9!0*8nY!~>+>+6l^k<5B%sCluM2~_Ld)q79o&U)X`N#S&`G4nCz)!}pBWIM_k+NZt;Q>})b94#cU@1~Tmx4;O&ac304?k2 z|GT}S(dam}>BmzE!*3ev)mkY^sJemUUBUhpc7K*gwo5RoYq5P+#QUwJ+gUn%T<@G&S}y2I=baW37kV{j8)T zoC)CS?}A!2JMP4oR3U!Iqk;bs&g!T1kl=s3-489jk@R9L;sxZ>S)E>D($4WNV&0k7 z=GmHuL0Up}$Iknc0B>;pUu;a4T6Y@}iuaSQPqetLJj+&V44(^Evnq8?u9NKGrZXT= z3Ry(WRf7J3u$3-bG@M_l^UBK?k=xDVnvloF)F}$hcGj6!wXUW=%jRqp95DAi*+)Vh z1Z~nwut?bJI+3J!_d(v~-lG6v=-i}jYZ>pvDGZwHWojed`LrDDh1HQvr>#~r3qfM^ za9T{~hQ`_H$#8*c!q4BJ^gPXvP`6?||F2j59=~;mQu7xgd&$eZ#Mc99PVBj^l*ZJbOKupif+BA5RwXze#Z6thCU-1 zp^yDE^o7LgSF?4L)5VJ+koeLti|lTmL53)4Z{Zr*?DmZXBQ|;LdlI@dLAxf*ns4I; z$K8&ncI1Rjq{!6j5rV8pAlMU zQoud6`RwVk5OBZt$627~aPG>quyW3_j%kei{FDeO)!T@TkL-u+|7jtzLZ+Ol;+23r zr{*D2f*x^PqnzflR>@Bn1$ZtEu=ThGhtr!>x1M$QxLL7M#{DeR+Wb#MVJn@WnC?D1 zm9Q^*>evXK7tyOWmy!rIW)C$`3zxkNxyxz+-#2JTvBo$DE*=uM3DARQd z)!azVK~&!A&lTE_<-LR7R2>1B0J{EotV0S=$)}t0B#>TNHQu1Y4T!2r?s`hb`@FD-rei%!tzkEO&MUsou< zn-s_04!|;p$%XZ`41+GIW^zox*K)&$v}8_x)dzq+@s2S{)dfxaJ!OLt(C_}Ni;qCr z%bE#44~Yl?r$qkIgOf(%d53jBwM5x-7#e0{`)mm2oT^{7>9%Y z{&}HSsCHd1_X%0HpayHDS&58 zHWy=nN|NTn7-mk|e+*Z}6&v_P>eQ z^E;8=T^)8p=+$hhy80h%HTIWx8{1fIbd*ciopQ=A9sp{%iGIUTJ-O7j>iZ5Wy?2n0 zuvR7jK`I-K1}kY+*Vj!d9P*(W$ZB(E2Nzg;*p_$GQ*@vo=k7IX=`^;tp~^wi{sap= z**7BidLqrdXYo@uHGz9$%F>w9;VE=g)r)!{j$pUj5l)>K8h4K>C?`+JiL}h0H_yoy zj@l5^L>lok&j6&cN57oa0(cEGXutplg(~B zzf>iZtAi{erMo-P#;O@2l_w7#0^P6`XaMqU*}AMDM^E;iFN&^Z=`MwTUKB0soR)gM z^>qR6STZRB-W`EQDPkv~&5Jj5S=IE3%7Vw~vG2T5k#v1aMKy+|lQYRfiRaomHp9Ls z2D;p!+ftfYk_zNO@fF3IKFr?$}&q%LKzx@pW<+$~V6t3Mj zd@$|>HahDV=;0AS)IF?VMOssS?FqZU&f>EiqB1`W5PC0o*EL@YfPf=*;Me&~ACHT; z|N2JDAMp22PPGL5$g@{ZOX@m6l|?GmycS&ny?_YO|AC>H(iQIa9c8lK= zIF%=XP#U-dRuWfGI6`ta%B&h}OWIs;IuG+CM|cd%de==KGl@7y*aPsZ0Hi+BEp(Ru z*1MTj54(>e)<7r1)}5XZl>Gy@IuB4Ph&Ma_uTD=1u7vKqZss#<$H@|%-WX;Hk}a@} zmtwzV%LIKoa0lAE@=&IP=$#^~_Tib1a3ECsO&bHbK>?76JD|6tcb~(Fnt<@yHZ_an z<2)b%or^WXt$>O~^;aAR$KXl<7N7;x_|Johvp`e)kJG?VDez{r579KTI|WV!6~^^S zmv25|TN5u-OR76NYXv}^NDvaSTR@*@@Rab=1dGrPGy^%AY@lu3vYBsoLRLjwU=4UQ zZfB4#h@=1RcBiXTrf1 z4KB7h{=1!p_yHQx!?n`;I(e_n(15?5%QnimUQ(@i@TzA{b@8d|QkRrHLCA5EYu@WZ zo9k$&tNxz(F@b(T9xftiJ4CUL_*I-8_?}C@Bo2Yu1(dZG^zw+A@!EFioAD;a=B^`k z&nzI~vtX>B>?l4WSLwBkBDn=1d2Fh%lVLB=y%N@X(f~&CW8?`U;vp^TCWi@*Y;@4s ziFdoHD2!9{W5gu43D9RE9vIinK)y43FT8MlJaxM`!wC79RmxAOuP2q7S==W^RTNL@ zVm}W$;d`PgW+n|vU^cRMKQ%A^z@o|77~uxIk?$@9BFz_Fp6d#|qeEs5XzGS99B}nJ z;Nn{~=URIW`r|hl)?say9zEp9(y~ZynC(Vq+fw}L>T&CC9yJY_yJv+sj`w-PpLO+W zpyR?jCD0oI_JB4j#NSbr#9U?+;xb32h+9snV(}KBi4p1# zvvLElAF9ef1&c2<#)L8e26)lu`C#`RzNlU%|00$X}8AB{pJY0T(t8 z>~B+nq0P^+bZRma+b90wQE5GQ6Ph338=VuDU(HK_ku6(o7|EcN=0&dzDk1JQbm+Sy z$sOEB_n;s?%UtSvomD_{&-zqxx@CQ*X+8?;|A#uF1wjg56>>!_oNAiAGP%+y8e$rB zvY)kEu`;<7+vV5o1xRDOIl9voSzE+(H@Qh0^7o zmA|CR_tHu+A+qP}ko*OZw{xy$7VEpU0G!hjZ6yRi5|8F~i!m4JAM2P*fleIZRMpzY zH(}?=wI_y+F+vWB-(XJgKii<#EAqc)R^EcJ#_Az!w%{^^&H8x!_MB7oEc$ShJb2V0 zF1ac^zwJV$d1u?n0@AKxF7FkYwa_;$=r8tI$)D?kufnch_t`#}v5xhfqs5v)%Y$U( zvZdi9ERkoO+kl(d%|F|hYVofvA+Hm6^jOGI+;=K5^E`MiZs!$1J#YQQ z*n(6@h$i2nJCI+R^VvqEG<91+!~xT^>6W006Yki z#odZ#E$fGXWQtn>$4CwlDQ@4XoQJV`O0T))YsSX&X@$8$Jm|E8in3*76O_HL^MnoH zv5dx?S>ojAp>3&D*J|9Ezf%sK&wD^CRSij2^6q=>t=xPj78C1{8{>3hc9?Aebirsl zp>3kKvVN$Iunr)z&O4Igion)|?j1DQ3?{44k8gc8&W=QxE~|!II$1oNuq*K{(&DnJ zA8zTmfa|71rmdZL=O6@a@atk^4J6S`(`P<8@eYg(CqyQT+~B|heNSi0WcJ}{e)>w= zl@n|`)+wl`_xCLlmB2)^azw+{V*3R%Mcm?Jx3bhxCcyBb#BPYflpqO(!WMB=c~>TGG~9Coq&)i z^5+_tHYtR{yiy@^uZR3WN#-cJmgH`m>zG|r(7m|7lzk5@x)$sX0eQd%%Sxe_`gefB z?6Z|^aIrVz=&I1S9x{*P3&tf^gOf3EjI@|`5*1bps_$v4U96EIk|m-^K(dkY8hzp_ z(ic-A6_Ey5J}#6$x2k%#;#0hL`}@fd%gZ!RQ6a%4ET{K!O&=-g9D-doJ4Ukc$&0EG zM{~`f^tI(ei7YPurep8qTG{qwbRMQ=Z`DP1N{qfl&oJ}gbXF)v3F*Kb z9Y`U3Ge5&r^odqJrS+Yv{%hA^4|qmt4V0U_zjK{gM)4&seETl@`{yk)Q+wv<2v5(6 zO}-csSZ}0D;^MaAe*5^FWEtpEbjPIqxGa5yc=lFjC9p)yQAW;T{J{9~7~N02Q2`TT z6rwpX-|G;lw&H2CR5wK^dUls6>1_IGDZZYTv5SwYTl=0%-U4xa}27LPm=zUHt&j3-#>o;Ru*`Spynw z?UKyykf7YbF0uAculO2@0+cohecL) zSUT2m4@L{p(}xa;n%`0oz>{*ZD>Jp$c-%`MbglxZd2q~Y?`|{pYiR5wQ<0O8pc3Ih z!R9-9761U322gIpq=NuA^M~>RXx<222Cy%Vt-o00`7tLYkyc=l=AC|I9jCN#mNu`& z%7@u-eA>y;q@ETT99lwTJ%lO$Jey*3?=+F`x^1M7wZ8fhdctV`!M@QT?|b^m zB4||jto}t#2qct%pLc~?gi1@0wgw#5rSV(8&5Sz*_($Mi!6UX6V1>=bR!qi4PUiU{ zjf^{u3FSuBO$VbJ_qE%Z)sajd z6Yd_OPAht6>^tC>zCef}kw%v2unPvA1`v8GHTP8$IMR&FeZStlD@?J{AA4}#WKyoh zXWL3E=N{0Rnyl4B3VOj(<$1Z&tQQfX4re%ok z>g6RH01+3yzGDnJll>@~`9)T=vh5}KinuJ?{BsES02#cJFDvi4a?nGhxql^9qHlJ4OGot#msy)<0|H#J`QyJU zv9>^jvRxa^%|Zh<;aFwbYIW8W2D|Z9G2mk)pi)Sp4i3}?!2vk4l;5Oz1!r`F7AW=XF|TFUtEHCJB3&P4oMi^rrnJuRQG#7?nf zGuLn?8e$E@|@ z=>|?r!eC0A7t`n3F<}I~C<9x`CMl1%;Xc=Czqa-)alTGiyGrLl4cY;!okqK1I= zvU4=zK?_W(JSy-!Gi;AsPY4Ii*aarn z4%+ck>$a|=x@(WegUQI|_6Q0avVeA2c~0a}FB6T+m?tr%pbpGM-Qosoqau{w@qFwp zF)NAho?m{~y)>PU?BcnZ|9hjx3xUa#qf%o-F-P?Dz7z^qa)=Z^m7n!`ax3BEW4F&9 zraBk#9V6`}{M4}Cw)x_NuKcz*w{zEREZ5xt%s-T59Z-~2N0^y{UUr@!v4>C^{hBIn zer-nEu9{D1>c>jfmGG9V$Pm&CexVJ1*-1PGbs^nEmuD$5uoy^>_)T_xGJ=%->&ij4 z3B^tpez6N6%+P?|#x^fCrOEO!so*T=|8BJCdA(5Y^~`XCUqR&^b-X&OJ_^!}GiM)A z%tuU`x8!8HM0o1u2Df5l^}Gr~xphvM752NYS`a8Wq!D0YxB^WAe`cO<5e8qihs2Tm z07SlVa4)Uc+0-W9iRNj|(9ap|S4L^SPz_6ND}A%)bVI5n7A-R^7j|9XI2E`P6yHa7N=^4ybGeq~%M!7U%9rYxvk(lC;*@ZpK>A6#7%{LRch?|pmOa`GC zvPr+-#JLC|!EMsPt1Y9lB5{^OATf7`SC8#mkZ%dJ;6Qv!_~MjOUqhz81u^vmomS&A zuH-jc8pAzjsN}1sC?)N=$#zk<*SzN@^4tx&ws)St9Xo;jfp25Phq?+e?z zynag}@*a=zjlS#MbCXw#tLnCalBRARwc8k#qkgCqPVIE<1RVH-0(jfVto8wxPx@NG zpDVqupMqQ0vi8`MzvKZ+^pO$AYuzuf+Xb61E#G zyGf_%ruxR{MrLT>#wS1xH4EGyh4`dM=mo$06nsv^TAm#yoUT%M33j_2abHS^(nype z_x-xa$`MP4;=T;#^~*_BO+W4xI`; zeiC-)`02+rJHTC@eU`XMmq6PPGHuY008)rkw~;X&S*Rz~veLAl$M4p>_ASrfR-J>s zZ;s})TQDLlURjuPrZOLPe5d(GS zH;%{QZ+cg=$RWB+aKnu(=}~lxns(6kS9+21#dh9F2s!NMeTtl(^Nd2*C%gn^{CL8O z2sQ>CqrntO(?B=!E!ejjA31DJAl}7-%*nQRj(Y7(f`xG#{x(}{Vn{a5myi0QZ|DqbB{{mYbJ1l; zP#1P>gx+~}*k*a}^8!IvEQhEqwK2rJ$9$Vy3xs{Ua6e&H=~&sIh_~zOZ}^+Y7&&&? zsdqVb7LgAl7R>AyFKSAOWK!UrmN@9 z2Twl9PU@LcvAI(wby?`IA#x&J;+YM*f#2-53nyukSxi~C&8B8(dM->Us{Ac=j#ZLMg?Kh3GaKuzPfRbLFV6UG_SH#7A{lcG)4 zi=N)`_ycA1+%$#1Sp`wCTNVd?#DS;#(Zt%T=*iKwgj8G664D;C4u>ZPvJGa73gNm3 zlG5N7zq=&PMMIk{6spMl8Pxj5PROJ^d$S~W)+O(rii#pjxsU4;#r=npZI|?u5-%!< zk^lv>^W7yQ5~W@G^|O1@_KCZx{MM2(6->V85l zLV68>@^^B;Ftudz_>>xQSEK6uyZgv>4blM9#ASu3l$6Ogj zYrG`0qH52nM<2>BOQJ`5eN##O-qpmplDG2Ay}2Wtv`drg#hP$fZ>wy$_oF@n8i25g zKm#@o5gS^or#2bl-doPy$mseE$;oV`#`}NcwLM?tdYX?7m+^^otbbO^HwcAkZos7u ze1`lf><_>q5ZzE=cu}PgAqtC7C(69|CsMIhcXK(QdspmBzL*9n zS1AQJm)2{@nNA@Mz6+;_uxu*zBw@yJTxFMFCyVWqo(F;Xbo02nd%LBxy#!Cz%ls7r&COK09Vtyy`(~@u13WbAdb?xdq8l1yu=GS4uU zz4}4-ck4%}z4mFTgi9Z9!*bLa;y&$#RJ_%Vel7iK%D4KhBD2oGvsVdp-M}4qm&}kq zVg*yV|EhresSJTNlF@n5vrVNdsD|VAKRvR~5UokotYp7tYnFd>?UgoXlX$|^_tLBS z7ZpXn@Eqo3IA$tGhiYHEGj-0O?>(P_=a=&D8zfV9Q<79X%7u~K(E^ieYGmIgFIH&! zK#d9`g-DcegA(me68g9?N(6U$vZKXyoZ7kjqVLT-PF@Qx)2NsNsn>4dWjp2Lb=+{$ z_QaAEmxspr{7cHPCoPdm-nY}PA>Ig#3%yriR^Z7m6hA6ssE`e*o0;4$K%@Yfo_xWz z1g*3Cdk?p?#1eTLb4_-V;tuQ+#zQUBL%+I231ssbO}Ue9jH_&RNSoYzgG0h_%($Hg*a>d4;VwKv;~EUf zMAR8nFWxQwkn!-h_@xyLv5|Sg4S}qol5e}uP|>8w06SvjimM7yJG!h>^Cih@N_Y|k zZS5{lEC9NL_s~a59z|)l(xluw1|&B;AZo6Ne8S#@apvcoMN-YAmiz{0A@J{i!zsFa;{ z^kCYp6ut%0H`s6Ec6D?2&Ai(_CFR3lr<2J(cG?(0a~rLFb75q#3>Du%rbIK^4rP|+6cwV}}y7ky8&94|v? zH;G=Ufcd({*~})E;!U51ymn{(R}7&*@tb@kXf-BEI8(fU%2rT!0yldgLer z7w=9CAcsg9;EF1MD1F!|0z;fQ;6ikEx3~xv;d;(+3TW{e4A?t#ZXm%BOqg2I%0ZgK zrVQ{nVhRIq2ag_JvNy%wmO0r=ILY1#f<`k6|NN&$;~2C+ zp{MdYewzH?UCWQROre2&4(;GD-{Wtw0Ea9OgyrJREecfPBS2G>HkPmI8{eOU*J zeqwfWNEMbz$ezX_Q$0;UF^XA2WUY3C(F`dD+*6AN+$D%^J2;Wqrec$3mn<84KGPaF zBfwFyGwt54qsT2(=RB(TI+ymRChHLpT4g5&SO0i)fn>|Md8)-P>%N^LWnDF3rs?Pb z(hd$}Hqx!9(7G*o*wXcgqe)^KJqK>%ch*Kt@s%iVw*Cq1Y?AO?8&$!8?E0uQ@nDtP z%@!XzEdo*Ct>>+=%IZKKDH;3GoVcd&#IT{8Y{`l1%Xv$v1RiB&_DtcnV-U2c&!1TQ zWyMHVQqrkp)BaACP7h=uHdSQCxMYLG`#>Cb_6Wt;Sz#Z@#wb-h#A<}+qKOWkZk=3v z62$nU1Z}FJ6U10ra`htP?L2{t;w;fTM$_?11|=WE&TyPS=H58wWpv0Jqfz6Y=6AUE zUP*DIC2@=m#ZKMH4Kpll9EclVy8?bEIaF!1ow1yeo7Emgqw%S1EDg++G_msw?I_Oc zNxh2p8~Zjc3)VYhWvNbDh^{kp^vqn!RW2pt!qDzNaL=@0C-bv-e7$ z?Hi(%!rQ=1hAu}M!Fe=d!qUp0Z_i(UI!fuv{Rvg}s6Fq3!}t0VfYC3WCNtNe~|BmIDh0Efe||&Cs`7ZBJk7` zTNGbYb+=eLLBHuP4cou-*KX}5x|I<08=VI9^~odao`j9T>IB56F663TO)<__hb5Iv z2i)6WA4NAyCnTAbdOamjOq#Eo3(rV$l%?n3b&&fT=vODab5(V^dmra6gP!NW99btn z8Ga*bi$4*X??l=t+!j6xuiPT@%q~N=x1#g%ql_D;#ADlvwf&2lRG@`2yn6m?byj(K zE2C|WyYjS)M{PI84~_0!FW_^y7O%P)1`^^5^w)X93aH-zK1`UruKZcgE+UYXQjM#H zcH`@H8hT2}7N)C7s65R_fCu&j!Y-EvM0f{7xFZKP?Ei$5C~fSKLKL(@D99{?Qoj#h zh)kU4ERYLA{ho~%SCH8whtOCv#$Sz4dgz6nnP! zx;U$9Agt5u_z6g_DZ5z_!~{3zqJiy)@lQ$A)OQPj(CW{j7328X?%#{V3B)n{-^NrO zzO=gidmj81JQ?%A%F%6_4x77kv3&3C1Z8UmS9fvC9r6V)%;fCl-|8NBqhV2n5f zV$Xh#t@ZYQp?%`J{{L^7HducPTrgV@-Y-&1;^6>8ZAGAM6oEct3ry+SIn>cUMgvUT zX@J(N(G_s0ez@~n1CT#((*x*m{ow2&kneKo5#Z4)O~H0M@xIZ>gcx`-_)N@qr=S$X zGY)qVGVGZjZi%@Qy{<9+if8=4CWL__KGEo8M7Jrj0?ZtK5SfadSVy-}s|3L9s@4C% zsi6kL1>}^kW%eiaR)G;L0CZ7}%`b8S@=O=+QZ*#}5VZUR@c@ae49@I;@w=j6seni6 zGcW`QWh;^mA+ZHmz!v0uY{8;k1ul#zI1540d__PapW4BsKbBRLD1wKif>DMU?5)SI zK2g72M^*#VTa3L4m@-FA}TUr)0OvjcYCom>W!7hTDGW{!qc>VM^KZ{LPnu@c`g zxHmVZ3k0=ha3ERkroFqG;hGNOX|s@qfk);Y zK4V+TGC}xRJFm35D`153r64)TxH%$38b|0Y0$w1+;71Lm#`Q(KIG?rA+q-)cTJta< zq8TSh`FkaZdDt?21AG3K=i(VK?=pbAng&K`Qh`sBc9}t72F+*MVhJp7(t77v0SVcs z2#|R~z5M>+t8{f!J@7?rM|0#>!CGjBR?Z38cx|+G2c4bT0zhjEh6Y#Pac0#p9QqCW z<2vo;EgAhqUf{q`LASFw_p(IqqY1jrUq@U2zDbgSQ01A%J5=C>pp~S0R)At2a(3ZbOj8*+RlFiS-v#loM*D$kf$l#|j_D54vC^`6?@jBaWH;w*o# zz8X8SSB7R~txv^M?*=7Iw~S6lR1c)99}yW=r?EKj7&{p^@IeH7)|U_b-$gx=3P3+rf$>$ttQsR8otTt|;DJHj)tyd;=!0xlHWodYf(QVK ztSl(TV+)L6*U34Q@PU;juVmIb7yMMr_Q)37l*%w9jz$;)HBaENVL%Zj;%@hPe zL+x0-_ZLhrNgj~KO5szXv8EYK_G5SaALfKm?GR@HodRv59RmfGXN3l0+=!#HOf|j- zi{4UU%)lLe>C%oBA&dsU1;cwj| zhuD9mT%}MbJCbW(hNju{p6`>wotKQTsu3ldPbD2OYP%(*JHYa;7xmv*yZlC}m(0&z z{}x0+x$vvdj^Xca(;9`nP66F(T!zz;N@V^Mi1cO~;7VuNfQg#eYY}d<_fhkPG(}+K zIZHMQ^KxABIY|8{oftD~Q`WV}M&+wBLLgV)&*f8I=TTrx{XLft*5b^j4t)e|;~SvZ z3xF7PKElb=u(&yjK`ze>ViwP^6RtZG|0rEp6J=zp8_m4I-h z5#wBUz5l0zCW_x#9M)i?EQnYYq?kI~MPqMWOqb}1=Ae&upEwm_>{N-rv;ArhLL*G3 zudw^+{H5uV^N}Qis0D<}%bNR(HzUq=*p*{u5aMTJZRtTz3GV&ZH&Y=H6|S*H<#dnX z?&xd01dz%PNG5aSOv0{9S^fZ=17E%xH};9Qe^C_7i+`ajP%XvY?YUXN;~$F?h1z;O z-TVSpq)zIrtt6T=q=hZ2{BRRNrfaIrA^zN=a2NzCUZUl--$rrZ^uvn>$%|HqjPOHRgvfqaenkeeEP zrHm6>N~#__CB0g}o5!BdA-zB?#2WZFSM>pCf>9Hg@TRgGuKgrz9p~&|TTuW~Y2FI> zLM1x$n@QBALC!wOoH*MLJ{2PBq>OQIV2er8rv}a^vfEfvF#skZT+!)n!aHXDNtbp2 znd?zc7y90DxSFvFSZnj3rzGX*+GE%)cK{T9#OY)|HzqX1%kJdjRg8HU9~HK;oXqEg zOn*Aj|Ms;oovi^PH4Nxq64od)Aau!V1CZ-U~Q z0WQMKa!jaL`u%cH{EEPJ$$Ibw#F%Gb#E%C7(8h*Q#Mz!yu`C})^4l+yLuXpyJ0;gS z2+h)@NSWyjn4O0b&~y2V?cV-bMvsa>$?`I1Xe>@&`kiECzRmE8hK?1!7tb48)MSLM zB6s7qRwBK3i)urcTZmZ?$lpT%5B(dH*P5?n8v`L;tOw7fqc2xvQh$XG{kH_hH4@dp znoJXTu9m`w_0`P(GG)xw#((JzM`mU z#VPSVDZx5+h6x-R&a+C<6w~~z&WeXX0~`iW>;C+PBrO4q*c4S?Z)m}?8jQBje#mib7dEoVIk{`#aL0ZsUd>*icr~ik+y2*F zC)gP4mLsXJ3w^1Co@gM|7#WIbUtPGu(vTUOj#ey$0DbdknuU^3s76Q<;8F8TkI}^W z9-wG8pLBs}XA#iX@U~!XP7lwaA?i^^S}8!mV{pTSPmgovqrh0MIs=5Q@VyfgcOo}w z@E^8+FkI#Ly#T|CXD=k6`nt+;^wIC>rrf3_AE82Sj!_6D(U01@oT)QKf6NeeN+yoa zVdWF7?+^BIgV6+_w`h@cN`UPQQDIE|@5R3A+sfLwi-UPY;`}XBgbqZp@O$9lZ+eOE zw(0?YdM=1q7XkFuG!#q1&%>RQjsPEakw5Ig$-An;29gC6J<(vM9fg_uO=43#~=d?8zqNyx(@#pY@6V3wl*g zPq-g^o_HepHO)_@*cIun;=nB=surzK6j~4QMRjqUT%7PcutgK_;=avW$c^CJzsPL0 z;oq!lYJdh?*qfVuMyQq|RyE*DJC|&3g+#zMWXF~qfAX)%OZ_b=l=!Gb-NUutt3>2s zTq7H}J7Nc~bB?}^bGUQUD$gWd>2D2WMjFTgJqtkc>JA6S*p@2ILJ|MG!X)x4{00Je2xd8^s3lL4(feLz1yO3Rr7L<0_%u z4)}!*ar3WD^CJeEswVYihv9mY)s5i!+2vCUvF`wrn?Cl7oP9cL5iJUfU?>@wJ{1pUND!VmHvi3sIGXFlsTnGfegdrTh`Ph3E$hWMqsmXVxQo)? z9iVJ&kKK(7ea8v>#UqlkrH{&c{5;>3bcmzQTB*I1N$|ujHxoB+*0@V$#s}`gvnU)@ z5IuC)sg9%NC!n6?e$y}maZ29%t4uq)wl{+$t;zkQ;Y${)2HPgtcb&-GUA}QQNGr_! zG>d}v_C*A2Rcb>;jHLO!yZ`;7e6r$Z3H9kldycgPh^ujSUILeah6;2z2meEN)h_`J zr)s>v;nny+b9Kv`m_7C}U(wq-HhG;%w4vck3JSP_?m|%a%~J%H`~We}wMF~+Cz4%8 zh!CMhE|NbvLy&6qhjT!FY7e|e{AJ9uF6YedbN$gaw3%8d3H9q2ES1L5XXA}Wq!V8^ z%n|Kvccz>P>v8$n@uptn@#ANiHp#iv2pbjDkdk)J0|q;7Bl{s~9j#=;kr#U`o~#l8eZbVon%0}$`nMR(RnzZ^ z7}fYJSg7&fXtd3D#pJS3u}?V~oX@pG%y2Ezk5q_5!|-lZt{uTu&JeuTjJc`#u4F#4 zk@;k3_9rH>+Yt0Q_C;Te#NrmOpx?aiCs2!8yM~5j=i34Q4R4|{Wh^y30z9tf1did z{O)0dhZHy##ZU?LY5~^>C8+-#Z6OE8H~A|GP>4@@5~ZR-_4_(VsHT=~Ei%13B|c)a z*%kgRwr?e)_fRVcMMAZ5CAq*QvECILhrTx%WvNxJFH_ZDE2$_Zr4{?PkOw4|)t#3< z2L*WUrG%tyzv}V>r+Puht+dU_EXNe4lQcv7yUV;2HZ7!9C45IE8im5LeFXQ{)_)gk z(q2|Yfx%ar3YpVd;~VTFpBvw4KLSb4&`DudmDEtflRxhX+}ROx>80^mtK06>$|qOF zmEgvaw%)dz*}g>LxL<=Sw2Z`N;RUY@Ln1etr|zi9!Q&7 z%JPu=hGV6xuvU;LNq{m_!(u85Lfd)hk35$AOcjFRj?frIz2e!nviN%~e zFUqy0`VZ>mQ^$x(P8x}(M@_&8O8H+u+i{{fs85Q{grQJ z&SjZ0XP*^$(qPu5kSSvnMWl2Bx_kvn zclK)rh!M9rp7hC7$ibwwUo=b+=YaI($&>xWD+Wo^_`Z?J37{)48l)|C*-CnekFVe^zn^j4P0+4npZ88*0Cr!XV@ zVt#M`#}#pCT1~F&hL2BF@1y;}5OucKF8h4S_z^-=LHS*;w;O!L@$9%meqyz1f;BZH zOY)mpsraES{IUJY)|(yoqWoVE)4h+bV;1h*DEv!Vw7dv3gPra$p! z$NpV3ogKldXYeBQT!@njE0h&%74#mLiZrP6}=^>1mNyTx*P z!0Y(f!zz-Zi?dyL7pKA2^3GKG{j!~ixE&ixQfEBh=Vll`>sZ03$KzMf_bijtJTLAe z+bB#7ao+#9$b8Kk?Y^+8#7DZ1(8Ma3Kj1G6O07-bG5xqcq$%R9Cmn;>$_36R9*Bi~ zgCD_^3Jzb%@qT9??-TfI%e!+Qo%DjcYU(C z0G=90)4JWUQCv)Y3(_YwhNBOkPpdPs&G3`1y3imS>G&IX>!@C8;~0vq{7m@^U#_{r zb&UqQjN#wIRua#uLoKUrURD+5r!ebuURzDg?_Z^XRD6i}&Ymgou1&AAv)qO+N$kY4 zsbNj;55fF}H-%^?)>+z})zljiX`9@{tc3Wxrg8A#f>=_4sp&aoKZ*<9{GvbB8Es_g zot4Z5e|Fx^_Y&4;AD13RyLi}CiRM1gdqP|FM$U1`^$prrZ^>q;_-=fk3IC1GxySD3 z!$b{uRT2Z6+qVt)G?xB;WYJ&hidY#F3x^jQxf01Y+Lc>3Xg;1>LVZ~Rr3a%=QU3+p zP@BqoMrxHwytfXA-Z*^kQ6u~>v*RgOm^LIlzDpFo=2weN1Aiqq8(U`i4DK=LF0|g( z?LHzObt=t8-=AHjU6_Q009LztKZ`6Sh|xSIGhJK5YzKY?=bvKMB>|n=!@;wi^_M_M z;_?leQjH)}KaH`DI%u2UQXMXqIGku8?l<4rvot`Q`c?~QkFA^I;bd>6Z**jdv8pY{$42dY5(_8 z!oFCr7Pt?!K+>kIKA8P3uCkqV(=PVUo*EQN`MQyUH~Dumf}DnLr9An5%=G;5EpaV9 zwzte@zTlwCZ;z>p3rLf9qK~c!7kH&-xT`8=)c&Hr>!AZY)WS5I2j2AbFAJE@+-4Ow z3o+FjIleBXq2bMDW;0GP{bFgx6_36yyI-(#Fui82HrEA%rQ1S~FSccNE!Z>?`+m%~ zWog05?Pv6kbbIGz$d;XKKflQ!%Z!P_dw)2|dYiK4&37wUtma`=vLV*htW(_EYO3S` zs{{3&p7gu0^^RPUIgWit>%?|UxZE~fzFH)0g9fa1UU~^LK6sH1P(!**l6FDby^&*5 z!_C^jd4x{&_MdC$OczUTcpceY5oR*FdqN~~&vB_S{YU%j1x)sI2VJ}hqpeZR;ai7q zS=K+ZPAqx&<-uWxG^g;C7v~|Q+#R)vLwDCgwe3n_w>}@^O0?s>++k->AfY7}(Pqm> ztzvvOWo~~^hCcNwPWDKmvk6_T8{`qxGTf3L6{xbjy zQE2hgUC)GVNxXVEp(y8E2E*<@=*spWA(LsI6fm($VE;$B%8Gv6X_fK<|1TT9ssJyq>l#vGH6+kg!hDwFtS3 z8E+v4IrRW&jf=ttT6(X=H)P8Nt}MEIF5HLyv2}fZLpJ1Hdr}qKzz6NcwIoJAOG7B(S{gd6?mB$KGJSYl4}TVCj0}Z2bj;TuZ~ipyMp>%1 z%~ZUv>bGa?O zT3tcUT*Foe8MDo_X1?pMp?Y2lo9Ouh`)<2xcfo5Pk8PK(i1m1j*TBn)ixk=XPP3^rD#Us3K(ZOoN`7B4E^3>UMF9y(od(zZnX}yRCR_8d} zN-y|{sY@{|dXtcQA^$waEr|c;-A7uj|GfS$-FXwoKX&d7c{e0q+MO5CE|R?xHOwN> zT-v4DW#o-`4{rV6cQQkSgpEtu0JE2K3hI<|3jbCAF4Lr&A%*%jns`Lj0LoMD zYPqqy)Pm>&ET>WK(BKAY+vPp)IWv*=D1((QMxoqZCvJ8-iyZ&K!RkxEOc8e;co%Tt zi>e;cG7Yq(TjNn}c8Z2?LliRLT`SaJ#!Eig5zwn7X&fG4XO<@q4J~*o0h~c)I!AJz zwYyHP|1pVL)Z#>ejiot^X5ni~1(C)SysdKtjR9cGFxe0c7Q=BA`VHP~oo#MrhrcE5e>JZv3nfYWR$_~)~q>m0vfb=Qhd z)utwDPu;%n&lld>^~ZUZa$9L))mI!vNquhP{in;ULehl1FK~5)c!! zrZR7DNBCaUjY02zkCBC-r0_VLgcFOFQrm?(AO*B4(nP%#Jc2VsT7n?E3Yq#Tfx%PX zo{ki}*w9BJVKvBO#_(7$iT!*Yt{0Y54~DUHDm(x#jGpIbB!h}pggX#HcJP8!zymkN zJee|OUn805Ns3fQswH;~&y>t2NO9rX<+T{PnT$Reue zpGPpglvLe33vF)o2IUt)kR5~*o8iFC@*TOgxpfNG1a;Ou_0;L>nH?e4jY)@q%?RC^ z#M7fs*aJN$t#3yN5p{FTH678bjWv7cRvfM+`er5{9!-+q$K25@Y!N-w}5AZ-ZF}MWVi%>fNqG=gc2Hq z*51dD-x#=AiasZSSzRJFc&=NrikfiJktXHRxR?6wn%}l6N!Ow5%AvZd9EZAt`6GBd z?a!}~6qa}xxQNRXZqjbKA%0GB;MGiiRk-3L~M-%!-_Q1JcUFFDmls2`%KkA{_zD+GTD9SJ4;SeRN_x*dd}8-fu26HTbbND3QS8l4a^?w%H*T78Nua}ab8>=l)x14#!gk)9 zO|w};dz&z~!rLw+Rp?T=3Jy1$*Q=*auqmZ$#rvHrH8%LlcmLme{E0KE4VF@MIncjH zK!fb&6{Fx;9PNG3vBF~Wdg_axi{&}>ESuB9B7mq9ppt;T(I&mbDx+!ip-Kdkm|;%& zvFY8-YWnPI?P6V>U!1!}xc^*_Y}$|boIxUHY~KlUnUVdu#Qe)N4{~k?g?Q4dH37%K z4yRhI0J$KyB^7)NNeOEEn2Hm#7k&)cV`Y2sv4t;|1wCd@ezk&Iy5<}=;ak9rZuIv_ z(txfBZHt6gV$y1=L;S5RtipnK)W+VtA#O|?WpJ79_0y(*9%{m}5Ma358yzF>82F!t zQg`$4zWn%~*wknMUY*L=a69DRt8h4cQ9tY8i#upMMaX@7^6KcBdBXT^;gYlnVXNR2 zIq?qUkk@*Tj?)NKdrqz@RGT*usZ|OprEeF5(njIQtO+RUS0=(`%I}scMC6N zbd7b8ISjYrYjwythh--mNJoZK&{LiSw^)Em2sgG%#r1hdQ(77FWmkQkhB~Vw`hGT8 zMlPl69ZWu7KoAbyrg{0qGQLA3Sb-yyzhHgxBcad9pxeYeUTbglT)%1%aZvi&$3|nR zV}VSv34MA$GZ4hPPsh&`-Q0EK6Teff0juPP;9-1 zb1>I&*{>$P=b+#&y)?`92Iv}1d8It(u-kx>&W9ZZCkvdiikK3cE>0fiSf_0C+Dnmx z!aKSTH<(C;6~?|4ALe7dFYx5ksRLiYo6hn-Ev5C2`t(Ub(?q^H>CX;!7N~ggK$)(b zG;MOmZTZsosulUsNkNNzJ~XP0Zd{FTF+HTLXP|e}G;}h}=Sm?OqCbB>zTt05EhUP6 z9^c8~bTq|cMvTB&Lwkp($H84UM4;YbhUY*UYgyMH|JsXvj|=r=DbqX@V?OdhaYb2*I(P9oEME~9sv`f@)@LHIxGHd35P`y zpJKLR!f^QU1i?SQ8z%VskIrbTHK7GzT?VTz-TMX16h#j$pY6rG*HFUMuBYBqJ(ygf z#8zHDjyNs2_@7q$3FI8JCqTXSse5X(zR=i4ZS5377*bE> zg_bGL98q&!Dbd zs4#LDC6esX1+lOdjsyhLu{4OJ8oW}-lOG=T9h$De>g1Ra#oygt)co=uXa?k?ZXS_$ zr>${>dE6MJhl|;MEy^VUpjyzGMDTz4g6hd|36Py3F4+rW_h0%RSLY^{%v%Eouc4`E zS?B8$4or;eQe)I~>dA!AR0iNm_QRN+D;)RC@@KKL8i(FLZS8bw(j*6q!(Plc!Po9$ zlna!HdJJD%Fi64yI?U`_w`$dChDW<68ro;_Xa93A#la`vOM52Zs#ghp4Qwb&M;i zVz4I3HSHq_ePQQ?8?<2J^A2HnQhT9&EOM4>o?hq&@#DN!$ZXF#&6q#2??seimuuf9 zVtmN`X4G4ryjzm=BD0XRjdaKIlKBrm%N3SW6qzihEcet(u=^p+i9o>)^V&bX57KAd zWuQeiKI!%Zj=GjOeZ3ta>zp=Q3SQf9YyiBW8)K-6k6s2&W@!o4h+S3JwfT>rJL#9P zABedljPrC$E{(>$2**j?2}wo;PWA19YJXUf&_Cw{nq}xWQ7x-f+@5#Q6 z|GcpeCs&}%#c%hC^CLn24+WE21XatL5qdPpTf}lt^THT%^C&zrNy8XSsrmC-ZS<>L z!6&Idzt#B0lwu9iblHj>3|O0=Ti`spG$`BgN})1GOtEqc3(@+7*B0A49-X z8F&Kcx6+5m8%S|d_XF*_jtk~Hk3Y`ZzaOOnBfCfSodRV$;hz$@#;t-qz`AE-<%J9~ zY_OEpoU)G#hXR3`bJ&TjhkSM7fU6M zex9?W2|Li5X%BBnZ(0_r~sR61hE$j5F8Z15=(!rE<>NoA%(O?Z%I18{{ z`Z@NE?v9uMIWCg8y%M4qX=o2y&aGqKsxqVC@iWWw8)_eFtrMg%b~D?>lgy|yTdzFpzwwv_)`XL`CRLw~O`7R9NO z%f+sfenN_#J=bDn@7hv?MM@)KA(w{?thgt*B=P7k1-xb8AWi~CxV6M8P za72?!9v;}D*!Dr=Cy|-h#>E3;D7fDO<&@=-Kta=xEy^r)Zg3NR8hx<`ozkGOQNf>y zj+lqEFxK{KANm#H!07(=*kzU@;!VDXeBZyDo^OZ4K)V8{owEb)8+M@+d65{je+bAi z3!RA1tnzq+=%^k&+%*=*bNiZnPXiF2ZEsMO2XRo^V!N~c!pN|@F3Yj3sV~0~ z;?N(a=8-me@Md`{)!>bmeRU4os|dy-q4CMhBjaSm{OFU|XVno0FaCQ$*)o0F#2E5< z53AQEx+C9lI>UJy7s;zFZpOM+*g~!b-Jii+58&wwTD{SVrNJJ_s_&5Y2bHl~<@}U! ztcJ)13t8G^4(laoM2JQHRM?2o9%l&XY{V~&(!^pjH!bQ1o)KNgCNi%pOsL6yMQG0| zZa}v=eVy|g$v6Nn5i~4I?015pa*JsP?i4QU-#e9b-88V*O?z65W}V|YX8zQMG#(L9 z7PPh|4WnPQ2ism6L1&`o#%kh0%_zXJ;3gL>;asHcvtuK{h}@>W*-D(^T(nwqUnZrR z>}0fB&cHv~B6#DpNyXp6hZbyO=HM)ELnBgt55Sy&d!G@QgQAVhVX+zW{A%@T*k9hF zDPAA3efYQy8ZVDt>wvn?*S^Xu{^vq@Wdg<)ZOnW&89AJ4uv9>EGGnza;Ki&vd5c-Q zkS@MH`aPrOZkrbS2Pp7z9ceZ9WtyOKYd7yFgp;H7n^(BqWPL2ap|-t{^SpL@HYp)i zTe9j1aH#Zj=*8?rooo)Ac5*gN#SmHeaL!NoXJxj@7^N9x+Fj_au_O-aN1kg!?=4KX z>MngI-}-BoI0TtX3bJ#=R<649U#I7V+x%PG$~TE6)r7!wi2g&LHN=9+h1QoKyx88X zB2Wyy(_BMNpcrGWNAGk#{ct@|@#gZfxQ`4^l8A`P)`VZc6q5l z>kyXG3J~8@#=3ePG6?lk$kqhOpC~VzR9VU2_&HXTUuF`8n#l1}&pq8^=6*RRmz4jz zEJNI0yVizl1+5o*dACc4OK4>E$$$S7MbOZX*$lJNBDJ29l#cOI_x!iHfgg4L`xrPu zFz7-!zX&)*TTWRPl7#BdOHoLPkqk3_`rizA)BbhdZlk(Y@UYC8F7NXmXLwjBd~(AI z!*|=R{`rIun_K@~w8Kcl{`{B!-E6r_Mc}>C2rPO#k!VmbW#ZLFStkMGLU+*V?tn}) zhBkfjo3Vc(bHyjC0!J@=DqrgOtyAXC##3m!ZrtzHN#wD{g)_)1+H7&+N*#!();OHhtfIU!5OkR5b4gO3T;9hEUCs=!5jSjt4B)`8ge0b+ z1O)$EIr2aH>g-`6^(}Yb)~zq^_6*PqGiQkhT29Xk0LvC6mm#Pg`l_5?oe2wg@T&vj zC*2uck8{Y7dJHRVtEL5v&v)w&gSxnwNF{YwBgZ%-p-BvK${ z*u0J?XFdZKHB=06^NQ!uK?5Kg#Cwr&r%2Q4TO|QhhvtmEXPu*h5?3?i49G|w)yoVR z_r)ib3`awI_>~4+{q9l?Ope*~J9o4|(oDlGqLE@A$-93>ne*Ur374U^xRuW~Q$j8z zSJkYS3M7;e;NRI2 z$i1hw^ZYBeUILL42?1Lq4%Sy04@TB#tzzcky#KW!@hVCYlO$WgdhydQ7jY|UvlGcu zw(nh#9Gi(ND3KVYp#SO;h4XLOUnw?#TBQ5q<8J!Nuv(G6$8z+r8Pw~dz8DGM6WIs0 zF`I#yZeKW$Fy+?usHF_h6%MqY;;s16xm(lQBOr_t02rGiw$T;B`bR|LWIF>~3D?86Z=($HxKR4?ibbfJkJw zyN_};A_dN;+B?%8z1WM2O_CJ%hmanXYzDbCelJQ*OL{>%fnz-HJ{nP3Xx3b}T4=%* z`DrDH!}}Ggn{UG0+DrlpF1fwUQNIqnTgMsrD+H-}?QRZ=PO2VPpth(ImoRHyf3Hr& zwhgP|HCEcARM>-1njMr}dfjgsV6~^DD5xvmuz*3v9043_t9o;@E%(KBfel_xx`Sl4 z*ipAk6!A!*3DMcr4)m}XTS3JH-kTfDn5ilY*1vyaH3QG5etD}XRvV%;+NI0>n$(EI zDBkwDD(B-OOs4QCh>sd||5j9a7RyVukEZSOn)$+)#WTf%6Ul&vO1lR(d7b;^H4tyOUTp1d4T*DulaPu$y+L|EY z_4^VjrRn?iFcrdx`C!Cu&;=zJ)`P|~WZNaR3jItxyww*%L&d5vA%8Kelg<6EH-G4U z{@VQ^ZDTv(jV=^rT?CPOw{+^RJFktP(!;pEH&9s3w(fA&XUq3nUL~sSDo%~=j~8Am z)#+Ohh9#mJHxJ|9H(msB@>XBtFL34j?%L9(IvyHP4FV>&dw5gy@k;p#nhQcu)3 z9WJ)}4d#?mS4`C`X5Z0)-M(ItLk39>JL&Z1if2NUuj}iUciHMSrF^VIR{@!`b zitnHk{kq-E9Os86<6aw2E$>%S=8|5vRyb?Vo8rIrqI~u=POjIxqT%?qytgD^s(gwfb&q zH?N#m3``w(15ER5nSmk;6k3}C2{S!CU|4gk2&eeo#$qc)sqVKx%WOqe_uS5~G0?Es zwYGV=EdY4i@ip(-kwq@O2fYN{=lU|gMF}5>&pg#W%h@*z9_eFnnYeT~^edCn!>8FY z_H$0Yx1sx!6N}k1m0K{}{};GMpQoo$0ZE5{dPb`LPe-ZcP9hI(XGK=mMTy^$Iu@VG z?H9^lssVJUsSzSUutz& ziNWN<`4S1h%YynWW}H4+@>B@cIeTH+jC)TiCuP+`rw}6s#^5KQr^)$zXMl<5S&z56hyhd~RJPzZ< zn)3RT1Xzq_Z+<&PSc$@`rM_WMM|QgyE@PRIBE&`FE=h@g;*mEFZhu9xZ2v6@KRCdV z5ik4w;|80Pz>>rp3&KPI&__q>HGAi&Yt&%?77Z?h~K-@%8(B5?t3wrCOq0>cSg*_@foifJLLT zRqF3PO(!6vs!LB$6a$&~R@>AXjr4UEXUF?Kh_u%9KhxaJ@p$D!^veq5^|~w3>>j2V zvR&6tW~Y9CQWKe4ugawO;|>n9gxky91SI$zduyZ>6HE`ptv<`jn|eS$+S)tv+`*pp zdQ~-1IY%Xzrd?#yVde~tPxx?^Y1J>|cK(c8Yy~1By2XzRabS0txey4XRaa2-0q_!O z-o#nK!Pm69twa^q?J*YQY2tRNt|?&|g;|9ux#`MgcVU9gJULN{ZN(Ss7_KRn;N8A|7P zM~L96n4CKu2et*UJv$n)Ye3hMIQQ<^tHrMR;vV`TGr|VlLBCEAH3pnBZRz$*6m?e} zThR(_m#ZB8{CW#GG{2>+S`1mDNmAdM575J@CH2htE*bH;ysWYR(Ui46YHIQM4}pGy zzAcV*2c5G&dW&JI7Z0~-z_9DGNd>=W3POv{V6nceJavy{@Y02`m8#D2UumZd0G?G! z`y}2jsH$!Gz|Xem1b7y-+T#_PkT4{?kLhI#d~0EfiPsLElG@Z~Px}1Ww60z6IFfHo zaz>Cyv&cBD=Pw?-=zZBtF$xQ7-S?Vwq}5-nC#rWZou2z*@lci)+W5~RH;6*}m)S;L zcPsPZ(0t?|KA|+WIBm-|#m*o7EFFtFVwl8*{2#f_OD_I^+=;q)_6^K?KYs-Wj$Ix- z#|2B$KCMcn>GlThd;ziY==x|Kw!QVgce9qK^8JRO(L*TtavJhis0HVlDM=cHYkcU7 zexEMKvm(`1AEjJ>oRw=Jgu0g_&MlBtFlfH%lLY0l4LL5Xkq*TjCxf=RuQ4dm9(&;#s7Su>!~w$ zVy-`-_WZnOiql^4QYUDulY_Wd>{(7XxcL&NS(M`^Bsct(czL&R$@;9O!vY7Zyjme^kdA}0-t&UX4l zt8Y(g)8TEG;Lc!YS2pWX_k(C4T-U;vUxJ9cqc8}CKE0?13yvWQ?0z)HO7XdzBz?p- zDn|k^8c6P6R?$_#^Qe4GII$I;KzlrlNwL(PCTO)2eXNg$83z_y&A(eKv@;K7!-6B{ zg>v;GB)a_9=4qHrYzR|p8U@Rz^*-1Ygb|Pe{_<%B4#?(}{&W*w6TEZBisR|-LEoN{ zDh;u)1_J4!F_3f82<%m7t@r8Rf{o0^CH`VR2-BN4STnob^>T$7o*0(1Ma0su%G^5)A9HeGhKEj@*`ooM&`S0ion5L_dzVPNRyz zj@9bh*PY5i2dP|v6FZKt58=HROQFzt5Zs#6K!;BEjV=3vN0v{meCPV5`#_Io>!}Sg zk88JwIS`e_?Q-W+-R(O+=gr5br@j6PE&yLo>CGp3(1TLgH~yK3i~`n@*k>!ta&&Ds zNEl?l5Uqp-h_Rr_n*R(iBR{IY{nJ_^f2RD;Vew~bA?Mw(a#%X{*?D_X@n1f5XN!r$ zT?RRaa(f9wbI$lZ%Kxs$=RLl4>sB&?>XXMLk#@t4;JN*ucgJ;}A%oTwe@Mb*mKdQ& z;t^I@%8{ysP>;x_`!wA4=i@`Z&v8j?KbpC_EzzuG(=rO5)Y%H&<7B}#fmF0N_8H6$ z!XIT6U+=~V$9VC+-YUVr;8f9t*GZa#-1EhFup-SpOsj?F!@$3~)PC3NE^xMkZqbuu zO3~Ad$p@{Lvx>f}!nWzwm)rIsqXL{?Dn`=H{U(4KyqaZSbScMu4kLJd-+e7~KL5NF zFDRL7gu|f!<-2sFRu&oXj?R)!TrP^@x-pQTVz7{&&Ynoo;Cgx#eX1Ii&SuZjNe6>- z*U2_gB6(#hg5Tt+0d0^?NFofi&5>C1Blwy_!!empst%&}BNDc4!@7gSD>jBoT=3sd zV%(Csw;ceyhtRn(>Ul3<1NMw&50+O1j>||m2!s>qIohg`;GiB9(oLEAVqruIbcW)s zVB9!A9qV7l*o$GGLB1ECq@UR7Jga6QhN20*wg^0zC=6w8rzj(-@!(H4xKah0mxY8T zA?FdK1Im{(s3y1{*`g4BMM`Y1%;6GrvZS>nE`B>1)hvzy-emY(teI_NQUHTz+I1;P z^t9i@lSMhiDIHC3q$D$EJBh^WB!EA2V2iyUUikN7%5A}cFhnwjevRxpGAd@@k2l+T z!7bE|DI^Gzd3paV4@wjQ-H=9!!iacxaN^{6PY#V<;P< zkmOFz&#^Z?D2xzcEiHj~ZSowm5|P_+mpFt{$4IA#w3kX+p>BwRD`wm6Gk`m#0QNGD ztjg5BB*=*-J%6LIA$^TAu*Dn9UiqH4PD4B%-+iaiSknZw`uRKepKYV4m$rL;OUsV_ z*1a(uehwb;6h7+Nipvc+U^EG5OHpMkqd%VF&m>&sMmn)%FqQZ}&-&lL5P+5(g0qJ0lClMOv~4@A4{{0Z%K98CY}k0tZe5s) zWA*Gkn6pb(jrEp^L@ygv;L+B0IU09$nbv>ybjXs5%k9`aw(KuwMo}QNiJKiw<#=~B zM&3%nl{km)jdBG2+dPbQ5#U^5)Q~t%1>m)sbM=djd~E@7&ie)9~!7 zR=3ef6Gm?qi&9bNU__2+HPq01WSOe8y(_?JD2a@4_dwuu87?oix1EiXe zVWhBuaN<*FWGqZI$s-vrpZX9k4pkD&2*;I_D?}bAQa3?@sLW{$$h%`oZ~QGvri&Vi z2q~xdRH3QIltEOjkw|vhr%#dgmfYyG#MwdY(td1H!CJ!Wrg)0EB(3?=L~wU2QRoh#Lt^s6Dfb zvxaQ#w|7ON>MqcBCSKhWcBIMu`!1Q3k$jCh88UWD&~>ea(KqL=p4NY_LVZ`G%Fko` zetM@t5hvlj7df3Y@26o1=f%L(mPYn5xipY$6c!&MmA<`=I(>g%9^A}4)QFVq-$0P` z*DiN=Y{G0e&;M6AYL?bf_&RarkT6YEKGtG}#pk?PZMdE(sv#ouO2J9tGRKh65cdJ6 zfp>3ruoFOq%(^3uLo|fT)VYK(=1cQ(;(vCg>$dfG4yzs{Q%rI8RAC)e5^A#fYYXG` z_Cnq%rW&B$Z{xLtedro%9?}NTKsvbi-fQGBjmu({|A#NMi4Y~ zTqmu0#Urfk>c2}w{ooshIRfL&mi%`J{xl|5{_`3Cr}sU~QyaAK5bdqVoKYmUummPa}~fQo!<23-UX>u3OgMsXB#Ok(jkq2aH)V8^$&iO2Apz{lG6b zMhva^F~hEmQ*uD+M#2$MFPAVT+x`YTb|7}QN$I{n?1_gfUWDt}AxAXF5G=wS3Gfua z8>ufMV+DxvMlnQUcaY#l;g&5VEO#aKJ%KDK?vDtuxqUG$Zkjqgd+%1!gwVa|9JlbxzuY5 z{lvDQ6`wh@C4~<0 zII8Zx)e;rrWMm-USCSB!8Qt zAJx~ir-oh2v5)bdbB%2k$Z7TSo{>+>#A0kR%H= z6+2sVuWuYJHe)cy#hL#;Ih{ki$w`Xk;`!|O;J{B_eQ>Tg57brR=Sjv2~re?ahUI;Ww39W~5|s=~qM>UI3P()RokLDA26;sXUh zS?v9#gT>E=$}r^fB^kJjETqE3V01ewSchO&u&bJ4dzEf8TT2=I|PL~`L5rjmh~C&S@EZ3#sNUv(EYx@Ky|2GOg*x5ec}t^LWlxY^qj^LWA|Yx zxrZJQb=w4_<0R@Zy&^8J$0{6OFoEN+qvj6aT6&`iI<+}tIoB?(4ccjfZCl6~fFx;A z|D1ETapzbkG<#?Lw-I-f(SwdX7D~ryBDttqjDl@R7J>|yln(`b5mn*OpwKYKW|Wpt z-~P@o$OYsT#h>qilTAJn*Jq-hw4U3q9r6BWob_L50vMFo07wqkG0{NVM-yiML#F;H z6VxKBr8y(%E~va+j0R@3rZp#CTZlf6GTYOM+pli?CUEHi;0?Au2b?pc`rcxb*!n*V z7>bg#{1nkz?w!Fwc$7Oh_%i5?UmjG66Vo$W;E_RyI32AhKx7(!K~emKxHM^^_lsU1 z$oP`5T_#)MOjyYG6yEE>ujSej8M&X3E3s9RI8(kv!g)Z?Xtm}TowO#uz`HGD3=y{d z&`B61X}gz@XdF#9M|{!V>|YgU>7AiR*!-W=i12ksJ))xR;?d>s3mNP5V_sH6V8Be; zyVGAZd(PunW%y&n%CKR=dd5O$&%t2;Sp`YyEIk(P`2EiYODO8+g+6z|H9A(NLGQJ) z|4PZscCek4y6*bz-q1gwIkwRHeA+K1JuVf;q`$`ilD$0=xn$zoHbX7tXz6r&@jG@_ zNM~%+&i-61Mq)kEZ^E|wHg*!ES{w7D=n!jIC3JiuNQIef_pP?8O#exxPA3WvgKYhL z#V@$6-2FG@*z`P+_j>P7-K)3v1X1RI#MK}GX`uZqUoEAjzA$*agh_lus3ZIobHV<~ zg|hW&aBr}-;QJ@%!Q+swF8&`HMa7qNq$+}d^@u8NQ(l_aEmBUCmLxg??F#n@%X7s> zw!c}r(>|#q+Sp3n#`x_n;*iLqt9QnDR+P2oQk~=4?n?IH|4-VyO@g0GTRH8+k*g~< zX3orZs)qWJAcdq;o1_mXc#ScOVj>mG5N(atmHT9ZOEy_SCp)SPKrll50BA zdM94TbuHB%mN+Am9|2(6NFVuy#`cX?Y(`$rv@h|z>OhZerZ0|M_)!JMkn=>*nNJ~$%-deW)Zm=LLJ|D=X&8aJ z5PWsU_~OY({bY0dCm&g&oIjuHT?>TMW~_6c1sEEE1X}M* zKI**y7w9aKDlamvCg)ec(M?ihVpo2@zR?^$N(zbdk)c|vVeoFYo~Y7Ts!dmk zOxIMl_q^AP6trK`j@#LXO~6@Vf8hCWb`JN#b|zh>{iXAAyAS^JW>}jJvR*im*cnyF z?eNOY?lm4DM;17R%O;btAZpd6;BYy7;oPC@hp^hIKQeH&it%dcCHC-nEqScgS_w}mGM;Z?fhl4!U45ov^(V6<+)C;h& zj2nF7`NY|507z?|?Gz5tq0xIFdpxd;n{H>4hwsFyM_7%Qc^oTv_1{K5&9W5#@`%A; z@(CeE>A~xdt2GnCV8u-y`@YbI&t(0KRaq>`J{$!-<4MbUUvZ^V$5X9GNX1GJrz7EC zZ-G;`-d>K-QU?k73vF{x@#FWA3QRxd=0F841=zKWu=9Xp+iQRe!og4w@?0c2hoU%P zWj3MnHiknYk~ExRA#JkRu*c3u?FHgB-C?r~IO{rEX?lVZ@YOJye1u#>+J`Ye_jWXK z2nECNM#Cg_z&YYK<~aK4pwU`dH<@D(+o<8GIWCcgX`q`LBn!pqMqEJHMcFyNnUON? z51IX}Vt#KosPno-7pRO`jkWiY_PqxUy7E3M6tDE-=^LDT7f{}I5ts)cwyl;qj^19> zWqXKdoyg;0I)i*I7l+c(zyd*8+k^hIo>AlA5DXV^n&my@}M^CFy+J3Xt({7Zc?OLVpx- zoCSS?{y_y67Yu3;pmQ6R)Es4W`7|BBHg1<&vW1_}jLR^8yv0tJ4r^jjl9nVZ_PpH@ zJMJ3Po7He%y2@j8526{jXyceHrX>ko?m@8lS40w^Ei#Rj)k$JM@xLOjTV$XBVT?02 zZSLVjzi6<*+9Q7@JYS9|^BNyU2-%u2`z5`#8k`Xf3O7R@e}Dyevo!A4N2-)}BPv#O*t(h?Mh@Cvm5;Q={&eYI{LPOs7GBOaqa3F!wjQG}5|izM*I^ zFpki;!1+WhItL=%sk)_*&7dCp?A~)aJ5|^aP@7GV2lH7!hIv~nj*#=Rc&JwnyIYoW zGH-W-;|o#rN}TaK6eXR`R-5`v_ib(XdIgUsBAzK2(qNZg9qO7E7r>;{;Ud7P;Xr^_ z9$x`1jtsJv_FHb{tXQn$3nQ2ak;O;D_pK}goZ-j!C00y?1Dr?7HI-yT$t_uaq8B?I z)ZVm94um!?dg+!lJ}66;%VXp2i<4D*hrKNxSvk1LQ$Eb=rJrafC@b-)tbMlnI-VE( zaVW_!EO$xgN3${=)R)y<5Jqf&lb(EWWjVVoQfm zPZOk@Xdw8DP|8P4zs|GYh$!^IM69)TXQa4CNk@P2qbj7S77`{_tr@nB z52l7w8taH>Hjzo=NIaaB$V}n&!c`CtDhZrH`Q0nI#LIt!`(}<4VyWl)w=8|A`IFY7 zgv?n<(%|7Grn8SnaJvun*2RuYD?nN#eMGf-94aC~{WR8!`||EBExc%YKB!xw)cmnF zWQ+f?XSSLvP(H|#WAm7%&d86GfHkFwS>-y`9Y^nrg4;IQe$Og~1ltr$=;TKjc4^1S zJbdjY*lu)UYuoax#o7Sgld-#znR_nkIEEydhWm34KKMLOGgOtW=yMZ#iSC(QBgcH> zK>F+e8ZO?_r^h8uptn7p=l$`Gjy!ixpx+ZWaZ-?6Wr6}=*y32gz(!K>pWQI)-g@dH z^RpH_tmMC*x}*;)1IsmHJlWG>LVjYbHsizeRx$@YrBG;gf|zOFiRncSKmP4Kolvg` z;xe`&?C(^VfU{XVq4-rf+QLVW#gy};R!AZc)ZXmM%@Qbe3+Rji&`nSjJS|;T;eoG^ z0mS#gOKu)rqW_3 zfDC@d^yLK!Fm4M+q#E7fP4Jd^pm}Y`9~{*xcdx8u=jeA3FO~w8>$Srb}fsec=O)#&spHkmU?>O zZq^EwDPv+EdGpYMO@$gg?V-tuLlJPci9*Wyx06RzF)?|Q-@o&G!t)z`TC9zib9PfV z3Wg_NO|01xeGHlVyH5-`_A3k18`XX;)k+^{aeUquXFJjSTXbNL!`Q1Q)cEgSK+)f7 zUSC-u(4;i>pwnZFHku<(eehHFt1$B!q96n?IEA(rGJ!ovBE}yl4zV}?o?LOe+-W}{ zO_GYXz0A9e?S#dOwX`jX4I%L*t@zIEKg@-qdV5VBI0YIkLNRQdHML(}mQ{(kO`2zy zpEZG4OPr)PT!?_QAKI3MfVRsmld6&W`J30+$eiEWLI;Aq#K83~Nq_I=Oql%p45Z+{ zbYIrFYI51n-OK)-8pO$Er{f`7p(&Jk=!1X5$?a8pM^kEo6`7cq>fL>W!%cdp%8T`$ z`ZBvlWV?m5*WTe(j+>CY^{DDT8JfYd1b!P&A6zzY8j(U*c5)D+ytDqwU7XJY4pCHm z*LoL&-P3!yIFpP$2>wOJXb;cS*U5TAEp+frCBjU|{bpFJC?psmFSxJ=3p@l(_K?-d zKTS2@(WU`yC=IS}Hanh9r3kNPW|O>Eh1|s>W(MuUGkH8T*M!xpB1gWE#F~pIxGDfv z1P0P(6;XaWUic~a`tb-IONqL$?plL&Sdb&e?EXD&jNSk!4aB6`@!8&vyM8+HD3_vX z<=R(*Q&Xm)J!4F=u8SW!LT#2z+YTK^3dr-5@y}aG6OMS2bQ@^YuEMruybfvO`CseZ zr5&-DnT0b+tB-)<_PK{)7K$;|Lu3AnR~F z3(4Jw5u~<;g=(#RL(=>mr}QbxHx>16YgD!Xw>`#RM=VqoRU9)f|1#czE{24Y-c5bS zA_3)t2RNI#h((HxVYXJPaKz*oL5N}(L*v{%we9AviE8dXji#&P!z?R{Q&~zz`y-_LLO%odfMnreD zC**xtcW{n<@L$>C$n#zXi%tEj6I(kV*4W#bkvERDMxhdexk%_LHmtoA--v<2dik=Q zTf8B5mOFoZ>?@AXmPyN0KV)Zb4I`J#ztv$MIXQXNZ-@FmoOlvOEAETBZuYuivT)xV zSJUlW#OMp5`soWqtNh3!u%)PmYCiTQhHO%V#*bnlH`ylaV01^u;MxD5&Y{nIA}beh z2_g*$OmE1lhP1*C@WVL}jLeB^JmGrAHpwor=N60|ciYx*)G_FG5hHPvt)OQ^Z+>J= zgjfKLT0vH3yq2XwnE6A!hqrgI6L4D2J~roCnCIi1C&&P@+MAGK8Dwm7|KPiLUp1zu z*pkYQ*eY^|uL?5-G$M0k4a!l;!_&=6s!frpk%{gvz5*LuBEfv7e5PX%4vS(4;L&U^ zEmj<0n{3$U0dVEx6z+nyZ5OIO+po^r?hjmUPOd=WwJOlJ68x=1jcG*a_~;!Z;r-~0 z(pbIvc&4#!-~Yqdo5$7IcW=X66f%WSLYdM)8h524q=8ClG$#a7={nAF$2D~%_kOPOdoDkz2o56^}C=^rRpSB5tHM0gm1 z>C^w>3;47nY2mlWZd%t&96%O4on3Ia9(h%BqV)Z$>$s3EV6xkDe5l)$F5U=)07M`M zBLY`6J3!z{!B{}{4uI_G@yBRlC?OI*mg$CD5h6|*d~THKY(uUuT02V=_KKyGa1nMx ze9lp$&>e|YY_Dh7+?*0ZS&GM5=$;lWSju4R)3TjxbXlm&R<8X?wk~Jafd5utqQqWt z+_frc--||$SGxg_zUXOTCR6tes(!ojaWC)N6Au})Ml(VdeMXOJ#{Oth^CIxXBw@ia zJoAQ8mD|IQ<-dQLf>-N;A$DoG3Bhd@$wVBAfa+qJWl+>ph%7q`U^>+%t&+$RNS-Sy)Q5sGmss!-9h^83<+##>K z1qC7@3%Oi*EMv*0zmrTV2HDj!1;%ILf5$0fz<#(3s8AbToC2W1vxl0IK-#ql7sQUg zKhOrbh748SmcwwPP&hL|zqEz$MIQJJ|1eL-7kkcYzrmu3G6c}&0?z2D|MQayhdy(@ z6DH#KHc`-nzJn^@Hj8F=)*)S-zPb=IL+enbP`~8QFMKMN69CvW?0(00GlOa zMgkln$QPz=K&hySUVzT5RFb*4cfAkxU5c@~NAPmAjI}1YzNu7~I5 zpJa7iSP!ZYK~@PDvK8u{Etn$PIAMHNE23r#N!-bSXH*U|xch2HhWf@!fh>FrXU=d! z00JM`ANYwOla*yGe|z6Ol3u3o%zs69Ga(VS#Y&Ovt1iuXZ6Ij?wC%QNGx3Qbij&;X zTQV%vVlU2P1R7B|ZcjVvW*ni|>f8QKl)C}^T2pv;?KKlgp%B4U2U0oqI2aUWIqa+| z+23)ezBEnW3kN8d{@Fnk`$)=`>^8SE6T6APYowa&Aqa}1%%Jt3d%YDm=P`-0K4KRc z$8m`00;ke@bKhg^%^mUTKDWYtpkA2&`^+&0d^49MpxyW8_0HYEvt81ANUAY#2Rg|y zao80ZJ%l>(N`Q5kmIGl(qP;cDZ6t-@3uufdW9ScD-Y&w$scN{`#xI-62wXlr;LI{0 z(ivuX2h!%N5y|^#1L!$D9H;BH-^HAK`_0An!92jUT&sV`x-=cR@_bNs?z<+;6XXz^ zIY&z1!{)2(_nn1f7&=;X<74H(`t6FINAhmtu$psYn`+It6dcTVio+2UUqT&U9rO!@ z7}nv$NJ4@^&vX}wWukA~9=jXo3P9bt|0x@8pr|rrfC|d&)=-_llp(Sd9;_c^3uS2Y zjkA7VMg~F5G#$N^Nrj%~E-pl!V{|Pr{12nCxnsFGT4_4=hBV!n6GK&tGR;W!LSc;i zRu_VedLm{`wVx4mjemMTL&4FzeVAN0=>DT*DVE91^Jw_X2w+Jm{0u&R^%GsDiA0zwZlDdXvYy(pu3Zm%ABO%JV{}F zzK0~bA0BaJUR=5mW>m+p@iSfy-%-u${c_6nRL>dq>}G89~RsSqML zf3|RwLMYD+?)PnH#j(R7n|&6@9v$z%18z3F;=PuUvM&ktrdJ@$gFn7#7HSVo|% zm~|HkoZ%p(wO|y8cg9m2mq*l|`2mLFx4^`QQ&I{AGQDX^E2f-hGPGC4ul3W#Ma-Hg zqg8z3^nxQ#-)uu_-wP=r1aHw!GnOFF`Th#z=zE~`)g#7Z?^OV{vGq=tdHb569k6S_ z{mC+ThcR0oLa3a9=eH@0cr_&|3k6Kf7Un+7Z(bv|j<#;~p$Bx1-@raNCh_x|>lP`W zAPR94da8tN0XK55bt5rD;6-2EFOp;COg<@fB-@i)i|*krQr+WdV~bpFE3MJEjpq8- zeP+ne>CV;7%0_%z6bDq-ebvZHL30I3_IhHx0ob;BAP_l<(7iEEOP zD-f>^>Q$9vO{H`Z%+noa$#Ivf-p8c9+nMkU!0_F*Jm;kjM99>1XpNc5?LLoNhX1(% z-v_iBlfD$J#*R4JvUaFtmc*7@itV8CGRMr|-m-4KUV_MFm(qi7%k#`$ovfCtR<$M_ zZ~ixcE3!a>;>g@NMg3%+SHY9dZ&JiQcj0Sj>{<@=S&pVbwNR3UIV7069c-$w!9BKQ z9HOigk~JrShGEbAyth6Lbu6>qrd;z3>r@$G0pnMS3we+d4$DbmKgmx^;2HMsRfrZd zHjj7<9d1h1p+jBI#J2H21scrRnvs>qMnf8Qa{tcp4DCL7zHAisoMyu_o2E?u8GBuK ztisC$4p!H=kh;6Xw2Uc`C~`tmSIqh$_pi|w2yn8M)5RmCH~Kb z-Wpe*th=H+3#r1M88u z3n~zBxbd|LTf^)Tr9jd?viW>CW3F%O(3)>Sy~A4moj=4kZTGXG%uG^MT}(usza)DD z0w;TwDhg?t8SqO=uTAfx>#Ts9edBkGX9TWF`+BO8e+u982fCN@w^*J_^qE-BW1@Zv z7alt5oMM6VZ)&Tq_zL~H0#VkbP&KT5%UOOxb2u$K!dF^iWm~3IlmCYEqLUT0`Rvyq zXKi*F1@dgV;81s3c1zPS>6(Ii)R5O!CpU)m*|qL@8d!`Eb?FDIfD6S@QC*LZ=I%%q z(beq{ZXP%wFML6R1hjoAFeN=y2<^b3&XvP#(t|mJKC$&`;gSGEoZGLdR^r zlx-Ct>GTsFJI_w7#I7x`ilb!ZxF*(xV}D7tykp(dP*G!PD#TfCW%u5Cd!5NN`19k* z%i*npvOkP=l5Ap`^^&~aujNMeOA@&jzo8L?Y=krQuY6k*(u)jJ?-9ejOY+iPFH6ld z-LXb0LM>co0bXuaJF$Jt6EG(X;rbn{3#D>?!@;Vl*jCo?J<}CdQa=f zox^kX8kvlYJek(qJmtmC7uq@#xfWg3PW9(Kdp0I@+cxi$^OHmtEuApKcagt;i1y;< zX0?$g1B1hd3+1xp^0V^gRD?1?0*x>4{RKbeg? zOqgf6mZJUGKVG>m#P5aKEQZ1Cv#S)lJ<(c#)#8u z-4Ks?sqX_LJrCYgTff%qb+K*C^`8CC2%9fboIth62Ho~+ z>gOXb8Mw{9_Bb_NM8WKedJVYchZd7$lB-Q?_Y|jN+F%59gl)!H%P%d^g&v@>7cl|1 z@D6nPqmh%i`sg!BxT-?j-ca~`zQp9`8d#7dzUf7=0JkQsniQH~U~F3c{0}nun}+UE ztoT$@<~f(G3Pa?=2Yaath`+faZO7f4bY!(u8^kT6Re+>xX5|$JM~WHCnd`5ZibO#* zYOSjPR#&Gy-M0*hmwKu(iDd3{ z&G1L1QQwxD+k}ZL0>mL(J%B@TKpn(9IzWTYB_9DC>OJP!t)Uks z$(J}=9`O;{G?jl?`o*@KZ(esrC(@?4=UFCbOhov|d?v0R^+mN*$|pxGFiJahy?;vG zou~!daHNX8cnn5i2{rwMYeaq8A}m7@ zEQ03ru3iWiBWRhgjJP_E_bSlq!e4gdi^oHw5-!O~!{y@N$Y&qARG;H7gzV&YgdfJ zEiP8k1wedXEQancqDXAm7inrHO{qD8%ga}how_58h31f^lVoV2nL8+;ZtaR6G$vDS z(@;Lx;T&;8ZhK!t43q3R7cyRt^TVw;4Qma>KJwZswy*g-o%3eyQIZ)PYa-8fB*CHM ztcUQZTM%2hA$e1&I>YwCdq)J|G5HdQ9q8UcAb|C0VZ?(oTW%(olAn~uUYdq_oiZzTF(%O@$6bqF*TUZ zN}1;pPMZ;~L3U}4azJ^`WI3`;{nL-$xDxoc9ch3mpv)Lo&knO)SUkt!4nCEGv5BOf z$s_Tipj>HT5jHV=R}ZM&e9l+eae;2_IH3ST<2^o+i^nn^2#*q= zg`$=(rP_p{Z6>2{>`)leoRD0BHKZ(3ocOJMcL14|WDugvWlE8!D_@A8i?PHtZShXd zb@c{8v#wd-foUqtP!CwImrmLUidFz!b*2hUS|-2$bQ9^HpFo9W{Xq6|K$o4^gZ&U= zP@b_d3?gE!FAoRI6l;W%YoekJm-r;k)|R(hppdT3vfGS&n3`K&Ooqt*QQ7~D2$CfB z0xljGt*LO*2I`H8TcrTM^+%@!TmmV@LBq(Q@j%B}d8iJ-Dk^W~L40*aGNUoZ*-30o zx2K@v>QKzOFQ6@M?ck4!Ts=WZf7nkoUC z2~JK0;=yv-TtFLD=_xkze|e9WWk*iiDe7X@8bdh;u)UVk2Xu(KbAEie7hadIH zWvg=U2Wet!SqNjWImdoCa2QwXqD)1+VW*ZmZ&2+1VmT8iqVfCb&7|U5sda5jnl7{% zmBhe|Kp>Vzh{QMLoA71jOnip%{JGOH(vOoa5;B)LpQ;UPPT?4Tz1oc23c!orDGcKq zlc)XFWX7;550SYPA0M1b5&8dqrT9}Mrto#a0u)X>{@ncXQ{vmRQz`1PqeK!t>sW>6 ztLkNFGrvy_NX@-P`G9!(C%G8{o6GOj*rjHcP;w&V5W7#eH}%7O#x6qS&--pOpoDT1 zGC;=MFVGU)uZaI9d?Ppm9-Q0?^{zySeEwblw<#_U+8oV12y0Fh%vgt^szY)fVYQ@g zit^+v@1|JwBnZxl3g+?VkKV_%DAqH(`$tO)z!kQuH2H1!jv&r{_P{VcvwNr8=eqo2P%sT}^q_Q(oYwaSCw? z1Vg)re_kJsqAG|MD)Ak$X4;@Zkq6tGWwyMq&P%uG_>e9D&LOGItqI5n-IpdSN5k zZNfVBB&SEydHMG2r?eG>xa=`06qOo*T*(gYtqylCiH+sFSoH&S^UPrQ`X(t#aoHK_ zs!`&Twvr2Z5+ks4NkrfPrL2<(HfYh+qN#J`p90sCNUCT?Pza??uUqi;*Hw`U_#Uk% zD>=-c21(&<>by9QK!K$ZmElP!c)KZo^VXe}>l{GKbMmXNDnwJ$!x ze0gHx%x^7}hfzH9PW_h#i`1;;2cPxe3t?wE+%HeINvqm?XDH2k6RD|zS2lGyhDUCz z%~b1nQ@|!U8V=^ilx%lBO-5nh4uB9&jw=pnka!(pZ0+_VkbZky{(;;?g(5^KHsl?= z&YENrQ&I-G`YuXssz{K%?zU<()wYhSm|AOY9tD-q+?qWN+M-ViLm)v|$w4_W=~g9UOhW;`+eZtjKRxf zA=yP`1if2_pBHI;c7}>cCy3sLd)CR1M_=Tr*(+%ItZ8q!pE`ie#EwTdXaWIje0S|+ zZbwQ0$b!Y*7_3l+ZF%0Ij;}68-zdiloqY~@*Ttb_UACQ>8p;yvZgbMONZ8$Q!*qTO)| z1be(YIp zI1WNVwzHtkcd^Bq#p$zGY!AQpg1I-xFKb2JU}R4|*Vfw~kGUShH(T{VXQU_qtk>vc z#yZE~r(#*_1?R?2tLy%`hPD02JwvJ(`MwhRnw(J0{+fqig39~% zgm}ax*X)4=l=(Cpr@2_f8i~aI%Z3c?wR|=M_CltBmzj3S$**uH&avs+{tAHWeP5OL z?8BR#kHK%q3{UJfNxmo>!!MABJ2pBrMOVa&|E4W^~?#SV&Qw?CCJ!y z3+hF$;(&idr3CN>c=gg%6+TOJjyt-z}AS(ja4-Kh<>Q0*DBn6p zvq95wk$!|Cp{>nrj=q(_nOpe<=BgdZrm*{?e}HDs|JXH4+Nfw%EdPgJN!144nsGlG z_C)kQ`c(-P=IBEtjf!x$Yr<+1f;^)Z8e6?ZO;iM!-lbB|FJ)-8ZrXQOOtwm%$^8+8 zBY#2C7v*`dI-SAjh(@>H!w{E$;j5qK+UeZE{2vUqp7k~4dHH{>s$B8k4K{uajgqc#n+DMddeeMU;(O?)`v80y*a z0kX!i9&Lg`{L^g;#U9+o5ge79AabcZwQp9qQ5=)kWn9+Hs~az8MlkV8?o>rP3`qQx zhxl*D)EjZKF25sqU+lpe&j84~ZRaWTKmBE{)!mUGT3HtonG@zK9@&QKddPL9;a;t^ z!k7Edr1rSA9BE%cPGk4(ti+F`g&|Ir$L_b6UWUim%l|$T@1#YSCFjK?ED_iY&su14JY{ZpG;p41(2Mz?^V$Z3iUh%mIP#K z$RaM8_YIrD&;0#{27{cXH9gBH!XJjIkmdg=qyp^;fn0U5{=dH~A~RsQvg_YZGA1bh zYgX}}c7^PT5+D11k=^a4Kx%;o;0=Xb=asM7|^P$bGVB^9C6AOzX8 zZ~lG-Cste2%i(tfK+pqt{Oj>2DkX|8O#0bQ{N61} zE+i9(pM@sQ{h_HXxF_<<=K3G{cye~LYA(9OP z>Scl5cV-=Hf9GB`;$_NY1S-|&(l-*g4ufPxA!zT{KmOoEY3#*3eCk8j#?a*sEj023 zeWYKgZnSn(%yVTs*0<1wN>SbTzP~BBfakUa2-foh(9F^E#-}QPM^&ODD>x;L0qvf- zR8ibPKu^7$z77fWp)rDhjxL1bfhl+3v%$dKRiZ-}ve)O*y(yAJ!QeEX4jiV4f4aKb zp2av3B~KqJ0%#x5!UT&h427P_I6iEFL|T;4W<)HR5Tst2y?S?cIw9grovjclvQ2d{ z2~~ohERKcVj1idd9In_er$8Q$WP4E;iJw`M!*Zryf19#UtDusrdUR%H73RYa3S@fz z?i^}G5^>wQ`w+z(e-cK^(Cm;`Ec|0Xl?Zhsc=%2beY*guhXzt)Uz>0pE{bpr3a@Oh z!;L-77I!sG=gh^#lgN{BNOm)TS3Pukv_Sa&DW`r^gD3zKZ+P61EMZNs*4a5LUF29K zJIUIICxy0MXLfNoPOET$lx57L@r~{nBBd#}y_M3$T=|ilv$gj(u3I_M^#avREsIgP zJ3m>d6w^tBkn8;#-MH1wIbJ&f|8bUaQ37Pr#>?euBRFlVC`R8Uzk_Lfb>l1*-45pw z9Q-{zwZe9NfQ>ltM}Z7{Qjr|l42l#>M5`(~&A1lwvT(q49tWep9*B*ErGaek1Iqn7 z$UQ^4%%dggA62sTLedf&FFebxIkQ=i?2NW}_vFbSoH-`|2OsYdHPwpeP^PfcPcGoG z7jlY2OC@sbKvjyVH2CXU?^clN&OlPG@P&?EO@&KZ>?$Ffpo}gOE>yyLQ3dMNS$)Hg z1lw74Ugwr^Th|@%T_tdkB>t)c-?EaCe5d}I!!vH7A>m^ z%J|S4dqL0&OiWA^^ghXvb577VuI(Ij%r;=b3rVL9#p7)rJjF=RA(=y=>|j?qo*kOb zy{&F~TBiQI{vz2`fX!PMB&T135=0|^kkmry)0geg=t6vZJ16BCm7-{VvW?i!D*@3I znzIX%klQOySA5KDkc{l5)lVg~zK;{O3_b+!cV2!ppwvj5A*x&1A~$U)th?iWx&Hpl z^YSf}kOAZCu;OJo#=jr*N8Xy;v5saw~fjmZ2&>KcJ>$ z1nm%K^~HSuO9;GuZR#vuQYy!{829yEQkAgwYE?7=ilJtUwXQ<0R?}=A(4ljO`gqOV znQRk)76UiG&1~64n=MqHFXXisTdy@{vBLP|FfOfFmZ#(-?JhBaD;+XO!^&uT={r44 z){!m7_>IVn{IN0#>H(SN*#*&UwY#@ZOPQ=MQrcIPD)-vze|m={qB?mInUuu!jrZV$ zxx9Uy8>%7=illk%sfQXK0m(?KDYML$#hS_RBEHpFwDIFno`7iE{9~t8O8ArZPR7DD z=9uK+t)v-<>N83{<7*hppK74e;@H?u86ER8O1GMqr#>jhJmhS!K`&f+rNYPLwl~h2 z8<)RYj3f4V#oJuHgv81PUMs5^9FH4a5fhi{*p-U?iB~@(E6Qak{h)VY;i~)7Lb=A4 z^BUIfA^m|M!YimH*NItm$5dK#=*|hheqY0xdjQyH>-i{C?T+di4TQt#@H+E89IrXG zN4wcGT!?uZDE>07v=;q>v#6dk3^k(+BQ-lVsbRkS|x32j+jzt5DN7n~&`eOoB}&o~8^Kt03QvQx&Eq zGXm{2pX1*P+)BwU9DyUn|CUg!a$q%ub?M=||u(w3Y?&hB?sB$QM zMR42sw3Sf?e<%$e#~>?Ch1XpC=N;1cfD(}dWgq|l`uN1k##>YIB9?`;OYGxoX`6@K zHr%z=PbB>_p{r(}2=kSOIb~D}y+El}8*Ej=lozyDhgp#XEYPB4tb*Yy$j(91l-P%5 zJUHpQEVf}0fr2HddTree{gb=P5Pc=u1u{Dl+LK5iObx$CbEX$0?)T)2jul zIu4qoU0Ig@Q zn=uT_w8Y8(t|5l{(?CH#{;TH2-)x7*t1a5eIau(7>b(EwZ4cVhNn{M1$cfU3!Ex*+ z=o0Qa;S`(sa1qg{Mi#RW$p7JKdvOe&l8M{lPkFI6?mtwWj&96YE^gglO?qzuu8u;P zk3uir0Ndx1&o;m}sv(YARsIuG;nHpB+;k?5rc>g8cttrxl*Z{NLJ1URw|Ad}dh$Fa zMCsx6RMI|RT-4$)wfZAzj@w)G9EqEs2O1GBXjq@^YuX0<@5?@7iZUXhRlMR`grfi97Kq|wE_RM zVpPG6quVSBVkl2xRrkFi`=}dHiQf-?Gd?s7OOX7-;Lgb&bw=aU&OC^?6X zXjq;CLpc6h7%c`|C?>-Ysr#2Trk?pxe&8~Oa_?HIGlg+#N0|+<3h|l!_qrR(;eGTp zooVG`j#$dV@#WifoNcy|gICs;!MFt3{OA=IbOSop#$Fs05nv7bVrV$)QQI7=T9yM^ z>gh&iF-Kwv@v`jN)lN#u7FYHMxkpJ2&{huRqb2>HXD6uxMu=s2b}%otXg5IJ!&yi$;Fu@ERG-82 zG2CQU2b@UShS!cq;2;)>z9@uPrSXSbBnHz+Gy^9UB$vvB;6qR@6BP&!r1d=L#J?z2 zuK@>G3o>ooP^Gr^S)0*2s;qkKG*zn>ItqmT?2Y6#e*6Y$g3pH4fF5Ufkl>qU2t)`7 z5Ni^2FcKrOeB!mF|ySSmwwMu%*nDO3L)Xs-DdHPfmEH+i)CZ=spQ;mVMOQHn=>TiSSiuX=q zWA7UkYUM-U!qaf&uDH9hr8*!jKtEzj`=Y##g)38!MsCJRVEAgGm+0CW1> z)r4!Z1RStWf;KVSZ&QZb#Pj&OzRIV;TnZO(6u<6;VjwT7#?;LkWBLVC+|^}TndJc# zM%i(u9f)_&+7YA`&RRP5s(npDGIDNvOfcjStAT@2`EY)8!rNlFtZZqC7P|{`^F4AR zj%5~I6@5XDr5(oW)&>#$K$V1N?x4+M%p{Lu_V<5^^NNHPIvkBhV`nP!5zJ9@ZC3yN zY<4#;;2mdj)4YM!jU8gA4uYc3DHd$85%*3ycLKA>YxS+pcY}SId3+>)B3%OEMF8dx zyw6-QT2y_zVXWfdVg%fuP2=_b>~SX3==XileXrF-jpIu_uz>H?R23kXM_*$P+(Qd0 ziI>u4t+!0tg|;ynNJ`HLZAJ8aouMnR?CO?=g|QLKa2Vgp{gND2i9u!;Z6`iDBU;fl zmLnlHSO-bc#G^CLKSL8!g(*}Q!kLZ{lq|_fhma#5?Fe?7tUSHkbb$MUHay}AR zOi`uyk1=Z`U(51jsZkjPe{AKqj`x;-=E()f_2Hywi7(Dme=e_on*k{A9vWG;m>8(Zz^S;jrMFo!rZHfOB#6LHl@2JA$`s# z!ILXZ>`krSbp5}iAT#RI-dZeRvs1}uykW5Akt%HDzV5l1n=fR#+uH|Y zXlo`_S!iQ)RHSeyGuMAWDal@@tx$h8*bdBj!jQA)&6%ou%NOzLD~%Z!lXQ6un5fiO zkBGdZVf{`zimnFq*Es<)tZ6a7aj~0{D<>BX@fI^~Z8Z9A-_7;AJd#^V52+}(U(^4p zx>CAs@HP23?ruULt?!%A(XzMiYWWoz!>fiV?^D;rR_i=8!PnP$DSf~2IH zV4rHwYwPdo8t4t2-E59L$f%q~Q1+BpVyW?mR=pey&17PLncrb&_v~BDzAGR79hHDKbvV=7ZToYQPc8EN62KRO{`~*SeAwl&*j^&7_Uvq<=Igho^#sQ zp#NBR%ln1{QDWL3_gs6FGI=U@{BN_u_O`(_g?DsDKOP**=|Lv80bEm{?%qFQsprXR zDQv#pb=byE$xJ#ycPh8eysu#zF;7*{TJ9kPFiIvLRAY2SawVU_z(?$n$+B?F)U7%| z*jBS;Ib{5aJGb7l32-<{*&@R?b~(!Ym^hvxnmHx5USGF!PIvdzefLIZL(2Z653O(9m3 z6(|068AaXX$o0{dzzgTD10rBg4uSge$P{6en<=}+L=2G`)P+Bez^Sw$Ki#TAptdKF zOzad6Pk8{GX?tW&cZ(5M+}d=5nXJryiGM80 z+fe6NZ^@|jdKbmxZ&A$S;U5-VpkvtMM`E$ZX}6p_xIz7PPGTO~2DbA3)Xl*l5yL50 zh4Sr&H`S#_*q!d|pT|92baFr|T++jlVs2vC4Xn#AM*BIZ#M*}^o93J; zs-B!2kg(b)WXjCuo~NLir<|gohV<=s%1IwhGu?HsSzhgP!a9<=YwW}r$sL|ZHEe8a zP#y4}XuIRzOAib|hpaYk7cKOxg&>gq&U4*oEP3Zh6#xtR9ToBDOmi%j%K-{nMQj6=28aqq@;>HYdOwvdxQjX)XdiSs7QuZXA8<;37O{eE;_w``~>5@q3i=4$li|vY|H6?`M*+Ip@swJ z_!B!^&xMxv_Aa!Dt3j$<)JZuAr$FCDDXrv?D?t+fU}}7lJtTxa{v;07sL&2|LcgMZ zkWS~0sJ1i9Vx^&lHj0ikoWvaUW6hBK`=i$&@9yzLC(1#%+sb#h0&`Es7|Z1ALegmj zmXS)vUZo2Py#*(zSLEb1bZVZKfPGqi0TuieDYXqV>fNeK*eUsp@K;*5JY_BU4;Ug} z776(A+jRLDCx4FDxg@2Tkon1P!6hl&h5mdtH&ZRc0melEri{Yd%6o!|hZ5>ahnugg zAD_`jL8R1MV?lX`%h$@=NRrQQA`*!BJY0H9d|q+B;~6VQ2M(iDTw77(6e^;%5|p1s zqZjd5ayw}C4^4n0J!FleGpq~bqx%$7Q3_LLiL}vvSlrkONF_MaX(WQ{AHL)ne?KSP zD*qv17&&Q~2%7Xb54&d^37P=l+*@UiBSe9#YiNtl&<}n0|MZnd$c?lc1gV+l;}Aciw(4zdQ= zo=gDqb8sA~Y)17P3av3aj&HaM*9J+}bjd%7XQ+bcla*ylqCo@-%46`kN2!!Y-(ZjP z>o5v5pmeP!D2m(i+HYo+WSd`3!F@qO?eBerCljntZ`{T01m4XSR2vZp6duub<@5qs z^6X|-(J~V(P(Pw_6CpXYK>GVHxWKdHt~*EJGLIxOl5$I7?`t?nmlwyPPU^-K4&z3Q z3+EAZwoW`PE)>%;o3Sp6$kpjgoFyLc^#(cT-IM}X8 zXr}5e^dC;oIIPI1V}BKc?t}orSWyRV$DLB|h*5B7Cuu=<1o=ikLZ^(EwQaN%C6R z7r^ZB;m#oGe91r0!D9{Jr&yy|75LlJHnstD9oz#-YUMbuXDCzDTaL>^NuzgP(IA7?z$5&6pfQJubm*O#8&- zl6kyJh_n;gDDuPgXSAL9a5fzzu-aT|L{^G=|&-<-0Q7|%Mz#6IYFW~B-XDpVvV zKYNjUeyCmqq|fowX_i7MXB0^lTeYY&>1`!s(BD>xNpgr4idL$6VVJe2{DZ@r`?+T% zyEo4iv$pN>d2P3DbRn_d7IF z^eCX-Dr%iT#C%ma4`3HBl$qnw z>}W0(rXrS9q>v6Dr3Kr(Dog=dPbZ7+q4)AjlUMJhuW5sK8^^|kA9LmR=FtQ_5;W+lCgX+`sG+EXuvV;+hGRMa@^#DBn0 zUAHQ9C+%w#^7Du!Y0#a{&0UwV(RsiSUn3b=I_7A?3Uk)*fmW5@Nx#Tpqb>{7 zn1r2cY*L#LZo>Th&fx7;R9>cvmXAP<>Q^t#5`M9Xp|U&py-Bg*z;8X1H+zwM<7E- z=j_vfeXx_kdug9{6nU9J^fn+zxtsRGPnt(3PcGtjA5;@9 zRHbkD_HnSU3_ivSvIVl6AUv7h(fP8n9%CLW%QCij&J^RohcW>1*pI>@t&MN3ACrQE zB?3G`BONaCh&sMBg{hoP+N>z>iWOK6teVOQG1m0#!D;NQ8Stw9?#g-HdX$_K1>H+S z$FwUKV_9jZl+7Mjz%DbzLJ}Tiv8PkR5QuAEUo$y1F!v3%W~1>?W9YZ$n*1 zh4n?X(%Jt8XOYPll;OxCSVhm<+tIKZq zhdJYo3yfL~I8QcVaisYAv-|s>OUVs2^z4hgTic$s&Vq6j=%L>EA6892_iwwjH@`L7 za^f~{Hkj~dnPp@jk83wkXm5UF*4 zZc_` zUm#`vpiwyZ^zLVH+nfy$DtV)qg#UTM%BlPmAqnD}qVZCwPI_`nAV0uw0>#;wTnT1Z zCCjRq%MB~E4~a3781lxa0wb_~3Ad`1K`oLgulM+VKzDV1Un5B}(3D0(eReuNgubo7 zTRq2q+T3qjxhmE2X*pMt28ayVIa21*==fxoZVd64weTz;E{Rf6oEx*tATnH@UNHcL zNk;{E|5deE*DD8F4%(w$mA( z^#qrGVxi>+n1ps&TA<-%)!L$7D+n>2r6FWcq+B6d3(?`Fxi&U@%fIO|0@|)8N^2M* z?`m|-#t{_^y6RDrd+YgR{4^a;vVaf{UWBGP&!H~$l3CI50^+4fx=t`e6`pt2Cz~4> zrzyULB4?9}%h8HgQIq;)8t`;liegdaLskNPFAVe<)q_L>9KeKhc$)E%@6h8x)xm0- z|4I#_g&cFhu!=-8>y=!lRG83B4N}`LSj4WK=jnoUpQewiuxMh!-PN4;uI(^Od+EWi z4=xr{S|KeHe$~1rmX=eriI^ zD)tseV2_HR(%yJt+^T?Y>7pZQswEj)Y&5JxyGc1uk^@vj`F_hMl%i?Dravl-mTl+KppQ3sMOdU-?IF&)X851OYi#UA zuRcDyR4@#;>#Y?T5GxYFV3=HhkmE{F%img#l+K*2mhr21kt7dp|48voR=nvb_D!?~ ziaNNd#_CrwTG&FXgu4&m2T`5)@}3M^3Q7RPC?HTK7ojiK1?kkA^X!@jk7jxKl_j z@s;m3Y+rO__PY$r*=lluApS-FbVh$#{oZWi9Gwc|nWm4Nh80~0itM}@K`UjgXS_yA zKWVs|i)R`ls7(x`^!{na`B&z)HT3O(WQ&wW)&zoC zM6#lin0QitD0$|6 z-l=ys%}lOZw7kype64Z}XJtF2p@ou?$uKvOQLx5ED8;;R{>wf|(*xQs(Omh@-+5XM zh8#+%?rFSE+!WIIauRf%(-L>@6qcAYKojldfZvEv52Dz!mX}NNYWJV69XZ$k{$S*! zE>*j<>L$qeA|PTDUs7Q#xyVN%r4WAEo*!ffdehF!(2l+*xzQ@G*rfkj>g_L*7WocgA0 z@~IhH0AGJvL~4sq<|eiVZoj@Box=!JosN`Vvae9tW5r48KX`}3(H3=w$zH;of8+5% zv+}$z^%-##!a3zaYFjHfu}XZO9fyzTB_=;7O+l@b+SsMnR-_3X)@yB$`icqPd5s*3 zRF2OaZm*D-);$BM$MSUQPo9=NuF+bT-H?y%@f6{{Qa2=d06hv%0j+N5Q^`xN>yB>B zHjVQ-5}%D!yr_?=u5VO#h-na zRuLFpv>$1I>ya5liuHD^aqRyQbY2xu7Wsa z&$z)O6M@3bo*asnvZ3xJy2GWfGl{7b(l&gcxonW+R{U&KB#qpft)l0+B(K$6nLKBi z`yxl*b$3?rKi`+?8im9Xg0Nn%cCL$#JBtj8x15(~lCE zl=IE;x3lkX-uHIL>f?|u_qZyE?|?jc;4oN5*46-t_n-R>rbY2vO+3{va3Iy#8okaW zg^rtB*KIjtGSqTyofA}EtPQ4W6+f>&@RwLRjyLepArFeHn$nay%}T)sYH4Mao#4it z*A&XPPh^w{6^Bp241e15%cG;%c`3wa*yg{!nE#XJYg(cS$k}6E=JJx@yQ?W32^p+iz>2W2oGK>UPlV`U$X?C~P-*vSlMBEGs|AO_tam zS+F0oDvtt-s;T6b3b#m!S~5iP1)LGFIsPYD(bZz0(*#P)z!jT(0;xAYu-zs_><=f3 z-&v^H{iu=F0Sc6CkSz49A&PiRt-DDgt$> z%>NSU5jb*B)T!*9=yw7uQzLf#a&CJS9GN!jjss_|@>fB`LUg}9vMr4s{Hi&Rm#O}u zZQKrMWj-IiA9=kyTI$LT>?7JL$dP3PZX3$bLZT%2kd2r)I|gu{4qj+M_Fw%0`mf&i z7LH2$uaFd|1>ol9qE_re`>rrQFr?-Ob)bu2RW=H7?mHKypHY4G0KOH*jFULlB1kX{ z3^D>AdIYo&N4l3$zUn|PS+t9dXzoydLH#s7nk&?s(_$f^1GyOH(JmdcMnSQobRn`C zlZwWz`CxqnFp48prjV@=Mc2@W@O#1Aor`U3lKCNo+5;h!$ZZdd$x77&ZLEXqw$Em~ z872Uki)uaLoJDJZzW6%#|NU`s zJu0%QKVnOqTDu!IjPE{IKgjS6e(x)z(~NwUS)J-ZTwBQ1+6U6jZ|Ve~j?0s5{{3!| z+>Jd|A)e@1NokRckU-icX$v6RqV;mKAtgB!bEhb0(X2DFhE`$auw6KkavHJlUvsKX z;%hE*Dx?j&&r;y=yGl!(s*o5_kX(~V2NqdMDeUa^m+W+KQ4!?!A^|$d;INIja;WbD zZLMrNSijSvgrMe6MMwnD#pqQ{hYjZaAq7oFZ{MMUXQHxKg9j9~^_u1SOG*Ij(AZ$we%7ssOkx_nH$y&AiK5ognVpVOF9c^ag__Xp{DhX}hDe$| zyN`*|kjscBAL2&$Ex3FGCpc{_Ih^qcklI^|3tT#AByUGB8}$>lF6Oj<$o<}UfwtTe zYDsuFWf~D|**tBAbM8O5Qkypjna8l6%S==I9hghH?-6{%P6Mk951m;IJ4+iXM2?_5 zsru+%Q$}2&AkQ#kPr{cHCKcTGk$V^p#yG_TeDV zj;z=;CO+F?#C6VOnD8ghQ+m#0i_L^TEDDor2K8T^UB%ZeDHp`u|F`+HMj#oa{1_iJgzwO{NB?%~|+bdgLHZ{V>est~3=TABG( z!CX)vwphTB?*owcI}KvAod)o_BJ>xsWW20Dz&inO_CDX~Q1l;yWcp6R@ts5xl4yUx znsi`C?|;gSRu7;ls?hYtc7<~36)kbp6UCp*JbFyrKLq(KD&KM6-fd54ds(-9JJJ`M zYKJh4GXGrhJ~Iws?_eOAv_~^SE;f7U<+w(W2+c^d+C)!_fWDi(I6>Y+l{5Odo%P9Y zW$T@?vs1r55gl)|U(SX%*dx3+XCr@o#kG~MDzl=Z<)^yfLVe5Sj?X0d6jFK+z!X8qhq?~VLspUERD~7E zj1W8KX5Fqoi4HNVZK4bJ{FZ>3a7wAzUjF9Q&!z1I_pR`=^I9`(vFGvM*T-12u)(%e zY!a#7nTyx!*(uk32wUU0VAK3+Nu?{3r*s$0*a2UAmS;J6ICzxip~04aXMhNv;Kf=~ zX}cKvZonFV0XbEP%6duSPMe+2NAQ ziGb))OelJ}MmPTL`O-A6Fp%2FcA-$|fmlOPe+z&Eia;b3-z8WDL~0oCUEjV9{brPS zqIXc_Mx&AH4~79JmGQC_Q9%wyUt%dEO;qx5?OpH^Sh>*v3}R%}>1k=uO> zkKP=6v$+`o)43H@Nr+7wteca5nChal% zb~wK;>Pc{>S+-Q;)>zMHPf=AnnET2*NWarO48-@g=R5o3H!MYDE-z$CnBH`5dFe%M zM$pYa7WH}_?%VFKE`rKEDoG-!eZiM!KCJ_Nmo%O{Wz6Vo&^lHn#m<nV%qKcE9pF6rFAzoniTx=7Y~x2(g9B~dc1E3LV&bb~N$5N-VXEDUY& z38ez9-z&J^X+5AORy!ipRc&5I-?>xiOYpDi4-+F~|NUL>|JyaoHftZ+xG)&HH+3}D zZfXAutgDx2b9{=H8M=jmm$|i{{S$5D=_f4a=p`VOwK(QDWYkLQPk)W(j5EA7ng%!3 z1YRaR?xl5d|6NUtT^j%4<|4%%eQzt=t1J%~6ydoiDE@sqdgTxO|042%pB+_=*-+zB zBs~2I!@soi#_0xthD#p0In%G0u;#z!&|B%s3ogES|MTzsFXtscUm`lg=o4qe$$So;88T~ z){xVGTaW|2p`1&T5>`n_cOg*al&8lPsj-1C9^WmBOjsy6kx!2ErZ5RmTJ8fKvQd)k zmNRD{Q;g_DqH@B?`-k}NC`?cr_ILnXd5EcllTYU-LR%>a8^z^TK+Gwp0ul|5R(Ox) zyX_XK6%3Ki*vNZ}o^=&^?1%{r*`pb?T7OJLX^pvG;J}HGw;2? z5jW0qf%nsMu7@@*xnwqWGd9915?GElK@P(1z0*kaTB+BN*0Hf#OWp}RI$-}s0 zdbq(?sXG7D&qjAoreHjU)8VTUIr;yT_U7SG_wCsF7{Zh8FvltmQ&z zL=0I5sjDco*s^cg#+H4bX;=1rFlk5{j3r@Y8U4<;`+mOv{hs4Fp68#Ax^6e)^La1l z`PxpC*`D z5N-k`+0_{S3Hc2nw;Y~y>!hh1B2lMW_~Q5chqTa6Ly8p|2Y%mI5IXEvbAM;7vKuZH%4gjz0`Mv(eUy%MVz30xTavava3#R59lLO^V zSuBu=D0N0a-bFjjFL^BfPPKq{^BxSb` z9~z!5+36VW`GS8s@AY(su4Q<%((mWI#A5N80lkL_QIge3Vo7tlu%qzhBe&)<^!bf9140)G&#JHb~B<%>u$zq<_a#4CtvC!;94+bh;E*gWJojK|d%hiY|wvH%D-U{L9~ zs9ESJ@0_OcHY-WWG;PDfQT0E|w+7{$A)#vv{Fu5=xxLdMxSOIujr#cWZ#Y2r&<|iB zwn=;hqy)CGqi-P18ax;NivOM}|Nbv$Iaq1RZqvS1*aV_+z-R)-Fug71G>XVwf;MJ~ zBMlbQ-4+T=zBWob%L(e;=4a^Ry6rz<8$=M1yYXNN`h&0U`Ju5xj_x9-1IlgV{@HWZ z;xYWOu!X*C`p=f*^aGJ5F6bYSUS@$CCku`dQLxVagqcZlsG^%exg+tYHgW)U;!o(1 z52437+zaz0(dYw<>!}dCDhdGkdYCM9nN0tz{OI)#PMDjrq)V^YKQ1dXhg>MeKvV1) zKZ_3KkVFJ*Ao2vL?*=5b+>r+w2NWg^LI!WCO>o`2D} zI=8&Zuhm@&*sOs%Xzrda!VA*Yo@hX+^9m;nq2K~J^fhu9ir}4uiE-!;#-;FTa!ciCYAKl>k|M-oW^E)padIg~ZkK!ih|$*d<(baNo98(U%bco%2EaxD)muY*VzY|JlKzDVAo{fT}3!72Fc2s_JUP&nufYcADU}_$w#C)c?H`SQ6*9Egw8Yt z{g?v5Dzo+u2+NHid=e$@H85&-PmCV&LRw-p0p}@yO^lR&sz(0lVPyFy|6SYf<`3=g zWWa^Hq@_{jaLGeemP}%KyyIU z*C9*(0@@~C^@(%$NhzqYH&gPQdX4<3V4v&)#w@i^B++HZpbBxjcZV9p1GPb+qT1ThG6$5lREn+7=$FfC7LPQVyCJk`0Ea~$aT zFc0$=Lj2FPtsqzqjhm?fHh-qmtj7l_kjvrwY&3@(KIkt32pJ#<@wNxiu?sau*5(@; zdGT5n$lSHL_*U9BcW?NJl=if9Rv=TEAIO3>-kk+?Qt`E3s4c7@FQf#vr)D8Q=rysB zzMs8OiX+{3B>c1`e01hO1%Lr}fWtXqns9k#90JMQ2SI032$*GcY!57`pgE*d`9v$q zI`b8*3tyulue1nv4P?3$Sim>R6M-7<=m9iBoNiDLV#YBr#F_NLGP(_&hXF|nk_z!4 z?^H23&hr*Doe6m;g@URE5~@N#fNMae!keeF1mk-y_$5~Xli7#4J`a>c4RF8T@vdYz zl}E~(LG9cW4fBGFSXZHQrNH|*YwCKi7(kKmJa~@=g|!QS8o!{;`~@|XC0tbA@LM&y zECVoHg?K4&H{0YlxkF?q1tLA`rW@vK07$Vy*Tv_bTTTz=AwwLakFTqI9NgQMAjb=p zuTRE~S>6F@(bd=y02T}3&_3U4LJ^N^&i-*ioQ_ ztIMKP_)Zo#OcF9?lo9Kw*44oZymd2W5W?RI;g0c>md)~bk1~t<2~E&0-ljmSWcB