Bump llama.cpp to 74ade52 (b9672), release v0.8.26#58
Merged
Conversation
Update the vendor/llama.cpp submodule from 597b6672e (b9621) to 74ade5274 (b9672), 51 commits. No NIF changes required: every header the binding compiles against (include/llama.h, ggml.h, ggml-backend.h, common/chat.h, speculative.h, json-schema-to-grammar.h, sampling.h, common.h) is unchanged. Verified end-to-end against the freshly built NIF: 158 tests + 7 smoke tests pass (generation, streaming, chat templates, JSON-schema grammar, raw GBNF, embeddings), formatting clean, Dialyzer 0 errors. Also backfill the missing v0.8.25 CHANGELOG entry (Hub HTTP/HTTPS proxy support, #57), which shipped as a tag without a changelog/version bump, and advance @Version 0.8.24 -> 0.8.26 accordingly.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Updates the
vendor/llama.cppsubmodule from597b6672e(b9621) to74ade5274(b9672, latest release tag) — 51 commits — and cuts v0.8.26.No NIF changes were required. Every public header the binding compiles
against is unchanged across the range (zero diff):
include/llama.h,ggml/include/ggml.h,ggml/include/ggml-backend.h,common/chat.h,common/speculative.h,common/json-schema-to-grammar.h,common/sampling.h, andcommon/common.h. Evencommon_params, whichgained a field in v0.8.24, is untouched this time.
What's in the 51 commits
n_tokensmiscount fix (#24656), post-decode callback (#24645)GGML_OP_REPEATF32/F16 (#24533); Metal bf16 repeat (#24638); many Vulkan, SYCL, and WebGPU improvements0.20260616.0(#24693) and assorted UI / build fixesVersioning note
v0.8.25was tagged & released (Hub HTTP/HTTPS proxy support, #57) without aversion/CHANGELOG bump, so
@versionwas still0.8.24and the changelog hada gap. This PR therefore:
@version0.8.24 → 0.8.26(thev0.8.25tag is already published), and## v0.8.25CHANGELOG entry so the history is contiguous.Verification
mix test— 158 passed, 4 skipped (against a force-rebuilt NIF, dev + test envs)mix test --include smoke— 7/7 pass end-to-end on real GGUF models:generation, streaming, chat templates, JSON-schema grammar, raw GBNF, embeddings
mix format --check-formatted— cleanmix dialyzer— 0 errorschecksum.exsis intentionally untouched — CI regenerates it after the release tag is pushed.