-
-
Notifications
You must be signed in to change notification settings - Fork 84
/
binding.cpp
1205 lines (987 loc) · 45.1 KB
/
binding.cpp
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
#include "common.h"
#include "llama.h"
#include "binding.h"
#include "grammar-parser.h"
#include <cassert>
#include <cinttypes>
#include <cmath>
#include <cstdio>
#include <cstring>
#include <fstream>
#include <sstream>
#include <iostream>
#include <string>
#include <vector>
#include <sstream>
#include <regex>
#if defined (__unix__) || (defined (__APPLE__) && defined (__MACH__))
#include <signal.h>
#include <unistd.h>
#elif defined (_WIN32)
#define WIN32_LEAN_AND_MEAN
#define NOMINMAX
#include <windows.h>
#include <signal.h>
#endif
#if defined (__unix__) || (defined (__APPLE__) && defined (__MACH__)) || defined (_WIN32)
void sigint_handler(int signo) {
if (signo == SIGINT) {
_exit(130);
}
}
#endif
int get_embeddings(void* params_ptr, void* state_pr, float * res_embeddings) {
gpt_params* params_p = (gpt_params*) params_ptr;
llama_binding_state* state = (llama_binding_state*) state_pr;
llama_context* ctx = state->ctx;
gpt_params params = *params_p;
if (params.seed <= 0) {
params.seed = time(NULL);
}
// no need for a rng
// std::mt19937 rng(params.seed);
int n_past = 0;
const bool add_bos = llama_vocab_type(ctx) == LLAMA_VOCAB_TYPE_SPM;
// tokenize the prompt
auto embd_inp = ::llama_tokenize(ctx, params.prompt, add_bos);
if (embd_inp.size() > 0) {
if (llama_eval(ctx, embd_inp.data(), embd_inp.size(), n_past, params.n_threads)) {
fprintf(stderr, "%s : failed to eval\n", __func__);
return 1;
}
}
const int n_embd = llama_n_embd(ctx);
const auto embeddings = llama_get_embeddings(ctx);
for (int i = 0; i < n_embd; i++) {
res_embeddings[i]=embeddings[i];
}
return 0;
}
int get_token_embeddings(void* params_ptr, void* state_pr, int *tokens, int tokenSize, float * res_embeddings) {
gpt_params* params_p = (gpt_params*) params_ptr;
llama_binding_state* state = (llama_binding_state*) state_pr;
llama_context* ctx = state->ctx;
gpt_params params = *params_p;
for (int i = 0; i < tokenSize; i++) {
auto token_str = llama_token_to_piece(ctx, tokens[i]);
std::vector<std::string> my_vector;
std::string str_token(token_str); // create a new std::string from the char*
params_p->prompt += str_token;
}
return get_embeddings(params_ptr,state_pr,res_embeddings);
}
int eval(void* params_ptr,void* state_pr,char *text) {
gpt_params* params_p = (gpt_params*) params_ptr;
llama_binding_state* state = (llama_binding_state*) state_pr;
llama_context* ctx = state->ctx;
auto n_past = 0;
auto last_n_tokens_data = std::vector<llama_token>(params_p->repeat_last_n, 0);
auto tokens = std::vector<llama_token>(params_p->n_ctx);
std::string str = std::string(text);
auto n_prompt_tokens = llama_tokenize(ctx, str.data(), str.length(), tokens.data(), tokens.size(), true);
if (n_prompt_tokens < 1) {
fprintf(stderr, "%s : failed to tokenize prompt\n", __func__);
return 1;
}
// evaluate prompt
return llama_eval(ctx, tokens.data(), n_prompt_tokens, n_past, params_p->n_threads);
}
static llama_context ** g_ctx;
static gpt_params * g_params;
static std::vector<llama_token> * g_input_tokens;
static std::ostringstream * g_output_ss;
static std::vector<llama_token> * g_output_tokens;
int llama_predict(void* params_ptr, void* state_pr, char* result, bool debug) {
gpt_params* params_p = (gpt_params*) params_ptr;
llama_binding_state* state = (llama_binding_state*) state_pr;
llama_context* ctx = state->ctx;
gpt_params params = *params_p;
g_params = ¶ms;
const int n_ctx = llama_n_ctx(ctx);
if (params.seed == LLAMA_DEFAULT_SEED) {
params.seed = time(NULL);
}
// no need for a rng
// std::mt19937 rng(params.seed);
if (params.rope_freq_base != 10000.0) {
fprintf(stderr, "%s: warning: changing RoPE frequency base to %g (default 10000.0)\n", __func__, params.rope_freq_base);
}
if (params.rope_freq_scale != 1.0) {
fprintf(stderr, "%s: warning: scaling RoPE frequency by %g (default 1.0)\n", __func__, params.rope_freq_scale);
}
if (params.n_ctx > 2048) {
// TODO: determine the actual max context of the model (e.g. 4096 for LLaMA v2) and use that instead of 2048
fprintf(stderr, "%s: warning: base model only supports context sizes no greater than 2048 tokens (%d specified)\n", __func__, params.n_ctx);
} else if (params.n_ctx < 8) {
fprintf(stderr, "%s: warning: minimum context size is 8, using minimum size.\n", __func__);
params.n_ctx = 8;
}
llama_context * ctx_guidance = NULL;
g_ctx = &ctx;
if (params.cfg_scale > 1.f) {
struct llama_context_params lparams = llama_context_params_from_gpt_params(params);
ctx_guidance = llama_new_context_with_model(state->model, lparams);
}
std::string path_session = params.path_prompt_cache;
std::vector<llama_token> session_tokens;
if (!path_session.empty()) {
if (debug) {
fprintf(stderr, "%s: attempting to load saved session from '%s'\n", __func__, path_session.c_str());
}
// fopen to check for existing session
FILE * fp = std::fopen(path_session.c_str(), "rb");
if (fp != NULL) {
std::fclose(fp);
session_tokens.resize(n_ctx);
size_t n_token_count_out = 0;
if (!llama_load_session_file(ctx, path_session.c_str(), session_tokens.data(), session_tokens.capacity(), &n_token_count_out)) {
fprintf(stderr, "%s: error: failed to load session file '%s'\n", __func__, path_session.c_str());
return 1;
}
session_tokens.resize(n_token_count_out);
// no need to set the seed here --- we'll always set it later
// llama_set_rng_seed(ctx, params.seed);
if (debug) {
fprintf(stderr, "%s: loaded a session with prompt size of %d tokens\n", __func__, (int) session_tokens.size());
}
} else {
if (debug) {
fprintf(stderr, "%s: session file does not exist, will create\n", __func__);
}
}
}
const bool add_bos = llama_vocab_type(ctx) == LLAMA_VOCAB_TYPE_SPM;
std::vector<llama_token> embd_inp;
if ( !params.prompt.empty() || session_tokens.empty() ) {
embd_inp = ::llama_tokenize(ctx, params.prompt, add_bos);
} else {
embd_inp = session_tokens;
}
// Should not run without any tokens
if (embd_inp.empty()) {
embd_inp.push_back(llama_token_bos(ctx));
}
// Tokenize negative prompt
std::vector<llama_token> guidance_inp;
int guidance_offset = 0;
int original_prompt_len = 0;
if (ctx_guidance) {
guidance_inp = ::llama_tokenize(ctx_guidance, params.cfg_negative_prompt, add_bos);
std::vector<llama_token> original_inp = ::llama_tokenize(ctx, params.prompt, add_bos);
original_prompt_len = original_inp.size();
guidance_offset = (int)guidance_inp.size() - original_prompt_len;
}
if ((int) embd_inp.size() > n_ctx - 4) {
fprintf(stderr, "%s: error: prompt is too long (%d tokens, max %d)\n", __func__, (int) embd_inp.size(), n_ctx - 4);
return 1;
}
// debug message about similarity of saved session, if applicable
size_t n_matching_session_tokens = 0;
if (session_tokens.size() > 0) {
for (llama_token id : session_tokens) {
if (n_matching_session_tokens >= embd_inp.size() || id != embd_inp[n_matching_session_tokens]) {
break;
}
n_matching_session_tokens++;
}
if (debug) {
if (params.prompt.empty() && n_matching_session_tokens == embd_inp.size()) {
fprintf(stderr, "%s: using full prompt from session file\n", __func__);
} else if (n_matching_session_tokens >= embd_inp.size()) {
fprintf(stderr, "%s: session file has exact match for prompt!\n", __func__);
} else if (n_matching_session_tokens < (embd_inp.size() / 2)) {
fprintf(stderr, "%s: warning: session file has low similarity to prompt (%zu / %zu tokens); will mostly be reevaluated\n",
__func__, n_matching_session_tokens, embd_inp.size());
} else {
fprintf(stderr, "%s: session file matches %zu / %zu tokens of prompt\n",
__func__, n_matching_session_tokens, embd_inp.size());
}
}
}
// if we will use the cache for the full prompt without reaching the end of the cache, force
// reevaluation of the last token token to recalculate the cached logits
if (!embd_inp.empty() && n_matching_session_tokens == embd_inp.size() &&
session_tokens.size() > embd_inp.size()) {
session_tokens.resize(embd_inp.size() - 1);
}
// number of tokens to keep when resetting context
if (params.n_keep < 0 || params.n_keep > (int) embd_inp.size()) {
params.n_keep = (int)embd_inp.size();
}
if (debug && ctx_guidance) {
fprintf(stderr, "\n");
fprintf(stderr, "%s: negative prompt: '%s'\n", __func__, params.cfg_negative_prompt.c_str());
fprintf(stderr, "%s: number of tokens in negative prompt = %zu\n", __func__, guidance_inp.size());
for (int i = 0; i < (int) guidance_inp.size(); i++) {
fprintf(stderr, "%6d -> '%s'\n", guidance_inp[i], llama_token_to_piece(ctx, guidance_inp[i]).c_str());
}
}
struct llama_grammar * grammar = NULL;
grammar_parser::parse_state parsed_grammar;
if (!params.grammar.empty()) {
parsed_grammar = grammar_parser::parse(params.grammar.c_str());
// will be empty (default) if there are parse errors
if (parsed_grammar.rules.empty()) {
return 1;
}
fprintf(stderr, "%s: grammar:\n", __func__);
grammar_parser::print_grammar(stderr, parsed_grammar);
fprintf(stderr, "\n");
{
auto it = params.logit_bias.find(llama_token_eos(ctx));
if (it != params.logit_bias.end() && it->second == -INFINITY) {
fprintf(stderr,
"%s: warning: EOS token is disabled, which will cause most grammars to fail\n", __func__);
}
}
std::vector<const llama_grammar_element *> grammar_rules(parsed_grammar.c_rules());
grammar = llama_grammar_init(
grammar_rules.data(), grammar_rules.size(), parsed_grammar.symbol_ids.at("root"));
}
// TODO: replace with ring-buffer
std::vector<llama_token> last_tokens(n_ctx);
std::fill(last_tokens.begin(), last_tokens.end(), 0);
bool is_antiprompt = false;
bool input_echo = true;
bool need_to_save_session = !path_session.empty() && n_matching_session_tokens < embd_inp.size();
int n_past = 0;
int n_remain = params.n_predict;
int n_consumed = 0;
int n_session_consumed = 0;
int n_past_guidance = 0;
std::vector<int> input_tokens; g_input_tokens = &input_tokens;
std::vector<int> output_tokens; g_output_tokens = &output_tokens;
std::ostringstream output_ss; g_output_ss = &output_ss;
// the first thing we will do is to output the prompt, so set color accordingly
std::vector<llama_token> embd;
std::vector<llama_token> embd_guidance;
const int n_vocab = llama_n_vocab(ctx);
std::vector<llama_token_data> candidates;
candidates.reserve(n_vocab);
std::string res = "";
{
const std::vector<llama_token> tmp = { llama_token_bos(ctx), };
llama_eval(ctx, tmp.data(), tmp.size(), 0, params.n_threads);
llama_reset_timings(ctx);
}
// set the seed before actually predicting
llama_set_rng_seed(ctx, params.seed);
while (n_remain != 0) {
// predict
if (embd.size() > 0) {
// Note: n_ctx - 4 here is to match the logic for commandline prompt handling via
// --prompt or --file which uses the same value.
auto max_embd_size = n_ctx - 4;
// Ensure the input doesn't exceed the context size by truncating embd if necessary.
if ((int)embd.size() > max_embd_size) {
const int skipped_tokens = (int) embd.size() - max_embd_size;
printf("<<input too long: skipped %zu token%s>>", skipped_tokens, skipped_tokens != 1 ? "s" : "");
embd.resize(max_embd_size);
}
// infinite text generation via context swapping
// if we run out of context:
// - take the n_keep first tokens from the original prompt (via n_past)
// - take half of the last (n_ctx - n_keep) tokens and recompute the logits in batches
if (n_past + (int) embd.size() + std::max<int>(0, guidance_offset) > n_ctx) {
const int n_left = n_past - params.n_keep;
// always keep the first token - BOS
n_past = std::max(1, params.n_keep);
n_past_guidance = std::max(1, params.n_keep + guidance_offset);
// insert n_left/2 tokens at the start of embd from last_tokens
embd.insert(embd.begin(), last_tokens.begin() + n_ctx - n_left/2 - embd.size(), last_tokens.end() - embd.size());
// stop saving session if we run out of context
path_session.clear();
}
// try to reuse a matching prefix from the loaded session instead of re-eval (via n_past)
if (n_session_consumed < (int) session_tokens.size()) {
size_t i = 0;
for ( ; i < embd.size(); i++) {
if (embd[i] != session_tokens[n_session_consumed]) {
session_tokens.resize(n_session_consumed);
break;
}
n_past++;
n_session_consumed++;
if (n_session_consumed >= (int) session_tokens.size()) {
++i;
break;
}
}
if (i > 0) {
embd.erase(embd.begin(), embd.begin() + i);
}
}
// evaluate tokens in batches
// embd is typically prepared beforehand to fit within a batch, but not always
if (ctx_guidance) {
int input_size = 0;
llama_token* input_buf = NULL;
if (n_past_guidance < (int) guidance_inp.size()) {
// Guidance context should have the same data with these modifications:
//
// * Replace the initial prompt
// * Shift everything by guidance_offset
embd_guidance = guidance_inp;
if (embd.begin() + original_prompt_len < embd.end()) {
embd_guidance.insert(
embd_guidance.end(),
embd.begin() + original_prompt_len,
embd.end()
);
}
input_buf = embd_guidance.data();
input_size = embd_guidance.size();
//fprintf(stderr, "\n---------------------\n");
//for (int i = 0; i < (int) embd_guidance.size(); i++) {
//fprintf(stderr, "%s", llama_token_to_piece(ctx, embd_guidance[i]));
//}
//fprintf(stderr, "\n---------------------\n");
} else {
input_buf = embd.data();
input_size = embd.size();
}
for (int i = 0; i < input_size; i += params.n_batch) {
int n_eval = std::min(input_size - i, params.n_batch);
if (llama_eval(ctx_guidance, input_buf + i, n_eval, n_past_guidance, params.n_threads)) {
fprintf(stderr, "%s : failed to eval\n", __func__);
return 1;
}
n_past_guidance += n_eval;
}
}
for (int i = 0; i < (int) embd.size(); i += params.n_batch) {
int n_eval = (int) embd.size() - i;
if (n_eval > params.n_batch) {
n_eval = params.n_batch;
}
if (llama_eval(ctx, &embd[i], n_eval, n_past, params.n_threads)) {
fprintf(stderr, "%s : failed to eval\n", __func__);
return 1;
}
n_past += n_eval;
}
if (embd.size() > 0 && !path_session.empty()) {
session_tokens.insert(session_tokens.end(), embd.begin(), embd.end());
n_session_consumed = session_tokens.size();
}
}
embd.clear();
embd_guidance.clear();
if ((int) embd_inp.size() <= n_consumed) {
// optionally save the session on first sample (for faster prompt loading next time)
if (!path_session.empty() && need_to_save_session && !params.prompt_cache_ro) {
need_to_save_session = false;
llama_save_session_file(ctx, path_session.c_str(), session_tokens.data(), session_tokens.size());
}
const llama_token id = llama_sample_token_binding(ctx, ctx_guidance, grammar, params_p, last_tokens, candidates);
//const llama_token id = llama_sample_token(ctx, ctx_guidance, grammar, params, last_tokens, candidates);
last_tokens.erase(last_tokens.begin());
last_tokens.push_back(id);
// add it to the context
embd.push_back(id);
// decrement remaining sampling budget
--n_remain;
// call the token callback, no need to check if one is actually registered, that will
// be handled on the Go side.
auto token_str = llama_token_to_piece(ctx, id);
if (!tokenCallback(state_pr, (char*)token_str.c_str())) {
break;
}
} else {
// some user input remains from prompt or interaction, forward it to processing
while ((int) embd_inp.size() > n_consumed) {
embd.push_back(embd_inp[n_consumed]);
last_tokens.erase(last_tokens.begin());
last_tokens.push_back(embd_inp[n_consumed]);
++n_consumed;
if ((int) embd.size() >= params.n_batch) {
break;
}
}
}
for (auto id : embd) {
const std::string token_str = llama_token_to_piece(ctx, id);
if (debug) {
printf("%s", token_str.c_str());
}
if (embd.size() > 1) {
input_tokens.push_back(id);
} else {
output_tokens.push_back(id);
output_ss << token_str;
}
res += llama_token_to_piece(ctx, id).c_str();
}
// if not currently processing queued inputs;
if ((int) embd_inp.size() <= n_consumed) {
// check for reverse prompt
if (params.antiprompt.size()) {
std::string last_output;
for (auto id : last_tokens) {
last_output += llama_token_to_piece(ctx, id);
}
// Check if each of the reverse prompts appears at the end of the output.
// If we're not running interactively, the reverse prompt might be tokenized with some following characters
// so we'll compensate for that by widening the search window a bit.
for (std::string & antiprompt : params.antiprompt) {
size_t extra_padding = params.interactive ? 0 : 2;
size_t search_start_pos = last_output.length() > static_cast<size_t>(antiprompt.length() + extra_padding)
? last_output.length() - static_cast<size_t>(antiprompt.length() + extra_padding)
: 0;
if (last_output.find(antiprompt, search_start_pos) != std::string::npos) {
is_antiprompt = true;
break;
}
}
}
}
// found antiprompt
if (is_antiprompt) {
break;
}
// end of text token
if (!embd.empty() && embd.back() == llama_token_eos(ctx)) {
break;
}
}
if (!path_session.empty() && params.prompt_cache_all && !params.prompt_cache_ro) {
if (debug) {
fprintf(stderr, "\n%s: saving final output to session file '%s'\n", __func__, path_session.c_str());
}
llama_save_session_file(ctx, path_session.c_str(), session_tokens.data(), session_tokens.size());
}
end:
#if defined (_WIN32)
signal(SIGINT, SIG_DFL);
#endif
if (debug) {
llama_print_timings(ctx);
llama_reset_timings(ctx);
}
if (grammar != NULL) {
llama_grammar_free(grammar);
}
llama_backend_free();
strcpy(result, res.c_str());
return 0;
}
// this is a bit of a hack now - ideally this should be in the predict function
// and be transparent to the caller, however this now maps 1:1 (mostly) the upstream implementation
// Note: both model have to be loaded with perplexity "true" to enable all logits
int speculative_sampling(void* params_ptr, void* target_model, void* draft_model, char* result, bool debug) {
gpt_params* params_p = (gpt_params*) params_ptr;
llama_binding_state* target_model_state = (llama_binding_state*) target_model;
llama_binding_state* draft_model_state = (llama_binding_state*) draft_model;
gpt_params params = *params_p;
llama_context * ctx_tgt = target_model_state->ctx;
llama_context * ctx_dft = draft_model_state->ctx;
llama_model * model_tgt = target_model_state->model;
llama_model * model_dft = draft_model_state->model;
std::string res = "";
// tokenize the prompt
std::vector<llama_token> inp;
inp = ::llama_tokenize(ctx_tgt, params.prompt, true);
const int max_context_size = llama_n_ctx(ctx_tgt);
const int max_tokens_list_size = max_context_size - 4;
if ((int) inp.size() > max_tokens_list_size) {
fprintf(stderr, "%s: error: prompt too long (%d tokens, max %d)\n", __func__, (int) inp.size(), max_tokens_list_size);
return 1;
}
const int n_input = inp.size();
const auto t_enc_start = ggml_time_us();
// eval the prompt with both models
llama_eval(ctx_tgt, inp.data(), int(inp.size() - 1), 0, params.n_threads);
llama_eval(ctx_tgt, &inp.back(), 1, inp.size() - 1, params.n_threads);
llama_eval(ctx_dft, inp.data(), int(inp.size()), 0, params.n_threads);
const auto t_enc_end = ggml_time_us();
// the 2 models should have the same vocab
const int n_ctx = llama_n_ctx(ctx_tgt);
const int n_vocab = llama_n_vocab(ctx_tgt);
//GGML_ASSERT(n_vocab == llama_n_vocab(ctx_dft));
// how many tokens to draft each time
const int n_draft = params.n_draft;
int n_predict = 0;
int n_drafted = 0;
int n_accept = 0;
int n_past_tgt = inp.size();
int n_past_dft = inp.size();
std::vector<llama_token> drafted;
std::vector<llama_token> last_tokens(n_ctx);
std::fill(last_tokens.begin(), last_tokens.end(), 0);
for (auto & id : inp) {
last_tokens.erase(last_tokens.begin());
last_tokens.push_back(id);
}
std::vector<llama_token_data> candidates;
candidates.reserve(n_vocab);
// used to determine end of generation
bool has_eos = false;
// grammar stuff
struct llama_grammar * grammar_dft = NULL;
struct llama_grammar * grammar_tgt = NULL;
grammar_parser::parse_state parsed_grammar;
// if requested - load the grammar, error checking is omitted for brevity
if (!params.grammar.empty()) {
parsed_grammar = grammar_parser::parse(params.grammar.c_str());
// will be empty (default) if there are parse errors
if (parsed_grammar.rules.empty()) {
return 1;
}
std::vector<const llama_grammar_element *> grammar_rules(parsed_grammar.c_rules());
grammar_tgt = llama_grammar_init(grammar_rules.data(), grammar_rules.size(), parsed_grammar.symbol_ids.at("root"));
}
const auto t_dec_start = ggml_time_us();
while (true) {
int i_dft = 0;
while (true) {
// sample from the target model
// const llama_token id = llama_sample_token(ctx_tgt, NULL, grammar_tgt, params, last_tokens, candidates, i_dft);
const llama_token id = llama_sample_token_binding(ctx_tgt, NULL, grammar_tgt, params_p, last_tokens, candidates, i_dft);
// remember which tokens were sampled - used for repetition penalties during sampling
last_tokens.erase(last_tokens.begin());
last_tokens.push_back(id);
//LOG("last: %s\n", LOG_TOKENS_TOSTR_PRETTY(ctx_tgt, last_tokens));
const std::string token_str = llama_token_to_piece(ctx_tgt, id);
if (!tokenCallback(draft_model, (char*)token_str.c_str())) {
break;
}
res += token_str.c_str();
if (id == llama_token_eos(ctx_tgt)) {
has_eos = true;
}
++n_predict;
// check if the draft matches the target
if (i_dft < (int) drafted.size() && id == drafted[i_dft]) {
LOG("drafted token %d accepted\n", id);
++n_accept;
++n_past_tgt;
++n_past_dft;
++i_dft;
continue;
}
if (i_dft < (int) drafted.size()) {
LOG("the %dth drafted token (%d, '%s') does not match the sampled target token (%d, '%s') - rejected\n",
i_dft, drafted[i_dft], llama_token_to_piece(ctx_dft, drafted[i_dft]).c_str(), id, token_str.c_str());
} else {
LOG("out of drafted tokens\n");
}
// the drafted token was rejected or we are out of drafted tokens
llama_eval(ctx_dft, &id, 1, n_past_dft, params.n_threads);
++n_past_dft;
drafted.clear();
drafted.push_back(id);
break;
}
if (n_predict > params.n_predict || has_eos) {
break;
}
if (grammar_tgt) {
if (grammar_dft) {
llama_grammar_free(grammar_dft);
}
grammar_dft = llama_grammar_copy(grammar_tgt);
LOG("copied target grammar to draft grammar\n");
}
// sample n_draft tokens from the draft model using greedy decoding
int n_past_cur = n_past_dft;
for (int i = 0; i < n_draft; ++i) {
float * logits = llama_get_logits(ctx_dft);
candidates.clear();
for (llama_token token_id = 0; token_id < n_vocab; token_id++) {
candidates.emplace_back(llama_token_data{token_id, logits[token_id], 0.0f});
}
llama_token_data_array cur_p = { candidates.data(), candidates.size(), false };
if (grammar_dft != NULL) {
llama_sample_grammar(ctx_dft, &cur_p, grammar_dft);
}
// computes softmax and sorts the candidates
llama_sample_softmax(ctx_dft, &cur_p);
for (int i = 0; i < 3; ++i) {
LOG(" - draft candidate %d: %d (%.3f)\n", i, cur_p.data[i].id, cur_p.data[i].p);
}
// TODO: better logic?
if (cur_p.data[0].p < 2*cur_p.data[1].p) {
LOG("stopping drafting, probability too low: %.3f < 2*%.3f\n", cur_p.data[0].p, cur_p.data[1].p);
break;
}
// drafted token
const llama_token id = cur_p.data[0].id;
drafted.push_back(id);
++n_drafted;
// no need to evaluate the last drafted token, since we won't use the result
if (i == n_draft - 1) {
break;
}
// evaluate the drafted token on the draft model
llama_eval(ctx_dft, &drafted.back(), 1, n_past_cur, params.n_threads);
++n_past_cur;
if (grammar_dft != NULL) {
llama_grammar_accept_token(ctx_dft, grammar_dft, id);
}
}
// evaluate the target model on the drafted tokens
llama_eval(ctx_tgt, drafted.data(), drafted.size(), n_past_tgt, params.n_threads);
++n_past_tgt;
// the first token is always proposed by the traget model before the speculation loop
drafted.erase(drafted.begin());
}
if (debug) {
auto t_dec_end = ggml_time_us();
LOG_TEE("\n\n");
LOG_TEE("encoded %4d tokens in %8.3f seconds, speed: %8.3f t/s\n", n_input, (t_enc_end - t_enc_start) / 1e6f, inp.size() / ((t_enc_end - t_enc_start) / 1e6f));
LOG_TEE("decoded %4d tokens in %8.3f seconds, speed: %8.3f t/s\n", n_predict, (t_dec_end - t_dec_start) / 1e6f, n_predict / ((t_dec_end - t_dec_start) / 1e6f));
// TODO: make sure these numbers are computed correctly
LOG_TEE("\n");
LOG_TEE("n_draft = %d\n", n_draft);
LOG_TEE("n_predict = %d\n", n_predict);
LOG_TEE("n_drafted = %d\n", n_drafted);
LOG_TEE("n_accept = %d\n", n_accept);
LOG_TEE("accept = %.3f%%\n", 100.0f * n_accept / n_drafted);
LOG_TEE("\ndraft:\n");
llama_print_timings(ctx_dft);
LOG_TEE("\ntarget:\n");
llama_print_timings(ctx_tgt);
fprintf(stderr, "\n\n");
}
if (grammar_dft != NULL) {
llama_grammar_free(grammar_dft);
llama_grammar_free(grammar_tgt);
}
strcpy(result, res.c_str());
return 0;
}
void llama_binding_free_model(void *state_ptr) {
llama_binding_state* ctx = (llama_binding_state*) state_ptr;
llama_free(ctx->ctx);
delete ctx->model;
}
void llama_free_params(void* params_ptr) {
gpt_params* params = (gpt_params*) params_ptr;
delete params;
}
int llama_tokenize_string(void* params_ptr, void* state_pr, int* result) {
gpt_params* params_p = (gpt_params*) params_ptr;
llama_binding_state* state = (llama_binding_state*) state_pr;
llama_context* ctx = state->ctx;
const bool add_bos = llama_vocab_type(ctx) == LLAMA_VOCAB_TYPE_SPM;
return llama_tokenize(ctx, params_p->prompt.data(), params_p->prompt.length(), result, params_p->n_ctx, add_bos);
}
std::vector<std::string> create_vector(const char** strings, int count) {
std::vector<std::string>* vec = new std::vector<std::string>;
for (int i = 0; i < count; i++) {
vec->push_back(std::string(strings[i]));
}
return *vec;
}
void delete_vector(std::vector<std::string>* vec) {
delete vec;
}
int load_state(void *ctx, char *statefile, char*modes) {
llama_context* state = (llama_context*) ctx;
const llama_context* constState = static_cast<const llama_context*>(state);
const size_t state_size = llama_get_state_size(state);
uint8_t * state_mem = new uint8_t[state_size];
{
FILE *fp_read = fopen(statefile, modes);
if (state_size != llama_get_state_size(constState)) {
fprintf(stderr, "\n%s : failed to validate state size\n", __func__);
return 1;
}
const size_t ret = fread(state_mem, 1, state_size, fp_read);
if (ret != state_size) {
fprintf(stderr, "\n%s : failed to read state\n", __func__);
return 1;
}
llama_set_state_data(state, state_mem); // could also read directly from memory mapped file
fclose(fp_read);
}
return 0;
}
void save_state(void *ctx, char *dst, char*modes) {
llama_context* state = (llama_context*) ctx;
const size_t state_size = llama_get_state_size(state);
uint8_t * state_mem = new uint8_t[state_size];
// Save state (rng, logits, embedding and kv_cache) to file
{
FILE *fp_write = fopen(dst, modes);
llama_copy_state_data(state, state_mem); // could also copy directly to memory mapped file
fwrite(state_mem, 1, state_size, fp_write);
fclose(fp_write);
}
}
void* llama_allocate_params(const char *prompt, int seed, int threads, int tokens, int top_k,
float top_p, float temp, float repeat_penalty, int repeat_last_n, bool ignore_eos, bool memory_f16, int n_batch, int n_keep, const char** antiprompt, int antiprompt_count,
float tfs_z, float typical_p, float frequency_penalty, float presence_penalty, int mirostat, float mirostat_eta, float mirostat_tau, bool penalize_nl, const char *logit_bias, const char *session_file, bool prompt_cache_all, bool mlock, bool mmap,
const char *maingpu,const char *tensorsplit , bool prompt_cache_ro, const char *grammar,
float rope_freq_base, float rope_freq_scale, float negative_prompt_scale, const char* negative_prompt, int n_draft) {
gpt_params* params = new gpt_params;
params->seed = seed;
params->n_threads = threads;
params->n_predict = tokens;
params->repeat_last_n = repeat_last_n;
params->prompt_cache_ro = prompt_cache_ro;
params->top_k = top_k;
params->top_p = top_p;
params->memory_f16 = memory_f16;
params->temp = temp;
params->use_mmap = mmap;
params->use_mlock = mlock;
params->repeat_penalty = repeat_penalty;
params->n_batch = n_batch;
params->n_keep = n_keep;
params->grammar = std::string(grammar);
params->rope_freq_base = rope_freq_base;
params->rope_freq_scale = rope_freq_scale;
params->cfg_scale = negative_prompt_scale;
params->cfg_negative_prompt = std::string(negative_prompt);
params->n_draft = n_draft;
if (maingpu[0] != '\0') {
params->main_gpu = std::stoi(maingpu);
}
if (tensorsplit[0] != '\0') {
std::string arg_next = tensorsplit;
// split string by , and /
const std::regex regex{R"([,/]+)"};
std::sregex_token_iterator it{arg_next.begin(), arg_next.end(), regex, -1};
std::vector<std::string> split_arg{it, {}};
GGML_ASSERT(split_arg.size() <= LLAMA_MAX_DEVICES);
for (size_t i = 0; i < LLAMA_MAX_DEVICES; ++i) {
if (i < split_arg.size()) {
params->tensor_split[i] = std::stof(split_arg[i]);
} else {
params->tensor_split[i] = 0.0f;
}
}
}
params->prompt_cache_all = prompt_cache_all;
params->path_prompt_cache = session_file;
if (ignore_eos) {
params->ignore_eos = true;
}
if(antiprompt_count > 0) {
params->antiprompt = create_vector(antiprompt, antiprompt_count);
}
params->tfs_z = tfs_z;
params->typical_p = typical_p;
params->presence_penalty = presence_penalty;
params->mirostat = mirostat;
params->mirostat_eta = mirostat_eta;
params->mirostat_tau = mirostat_tau;
params->penalize_nl = penalize_nl;
std::stringstream ss(logit_bias);
llama_token key;
char sign;
std::string value_str;
if (ss >> key && ss >> sign && std::getline(ss, value_str) && (sign == '+' || sign == '-')) {
params->logit_bias[key] = std::stof(value_str) * ((sign == '-') ? -1.0f : 1.0f);
}
params->frequency_penalty = frequency_penalty;
params->prompt = prompt;
return params;
}
void* load_model(const char *fname, int n_ctx, int n_seed, bool memory_f16, bool mlock, bool embeddings, bool mmap, bool low_vram, int n_gpu_layers, int n_batch, const char *maingpu, const char *tensorsplit, bool numa, float rope_freq_base, float rope_freq_scale, bool mul_mat_q, const char *lora, const char *lora_base, bool perplexity) {
return load_binding_model(fname, n_ctx, n_seed, memory_f16, mlock, embeddings, mmap, low_vram, n_gpu_layers, n_batch, maingpu, tensorsplit, numa, rope_freq_base, rope_freq_scale, mul_mat_q, lora, lora_base, perplexity);
}
/*
Currently we hard patch the following functions to common.cpp and common.h into the llama library due to a bug into the nvcc/gcc compiler.
It seems that copying by value lead to a misalignment of structure and copy - resulting in a mixed up values that we pass by.
See also: https://github.com/ggerganov/llama.cpp/pull/1902
Keeping them here in sync to generate again patches if needed.
common.h:
struct llama_binding_state {
llama_context * ctx;
llama_model * model;
};
void* load_binding_model(const char *fname, int n_ctx, int n_seed, bool memory_f16, bool mlock, bool embeddings, bool mmap, bool low_vram, int n_gpu_layers, int n_batch, const char *maingpu, const char *tensorsplit, bool numa, float rope_freq_base, float rope_freq_scale, bool mul_mat_q, const char *lora, const char *lora_base, bool perplexity);
llama_token llama_sample_token_binding(
struct llama_context * ctx,
struct llama_context * ctx_guidance,
struct llama_grammar * grammar,
const struct gpt_params * g_params,
const std::vector<llama_token> & last_tokens,
std::vector<llama_token_data> & candidates,
int idx = 0);
common.cpp:
gpt_params* create_gpt_params(const std::string& fname,const std::string& lora,const std::string& lora_base) {
gpt_params* lparams = new gpt_params;
fprintf(stderr, "%s: loading model %s\n", __func__, fname.c_str());
// Initialize the 'model' member with the 'fname' parameter
lparams->model = fname;
lparams->lora_base = lora_base;
lparams->lora_adapter = lora;
if (lparams->lora_adapter.empty()) {
lparams->use_mmap = false;
}
return lparams;