Skip to content

deps(backend)(deps): bump llama-cpp-4 from 0.3.1 to 0.3.2 in /backend#177

Open
dependabot[bot] wants to merge 1 commit into
mainfrom
dependabot/cargo/backend/llama-cpp-4-0.3.2
Open

deps(backend)(deps): bump llama-cpp-4 from 0.3.1 to 0.3.2 in /backend#177
dependabot[bot] wants to merge 1 commit into
mainfrom
dependabot/cargo/backend/llama-cpp-4-0.3.2

Conversation

@dependabot

@dependabot dependabot Bot commented on behalf of github Jun 23, 2026

Copy link
Copy Markdown
Contributor

Bumps llama-cpp-4 from 0.3.1 to 0.3.2.

Release notes

Sourced from llama-cpp-4's releases.

v0.3.2

Summary

Patch release: latest llama.cpp submodule, EAGLE-3 speculative decoding, multimodal video input for mtmd, and fixes for the reworked upstream mtmd helper API.

Changed

  • llama.cpp: vendored submodule 94a220cd6c57607016 (#256). The public llama.h C API is unchanged; the mtmd helper API was reworked (see Fixed).

Added

  • EAGLE-3 speculative decoding (llama-cpp-4): new eagle::Eagle3Session driving upstream COMMON_SPECULATIVE_TYPE_DRAFT_EAGLE3. Pairs a target context with a separate EAGLE-3 draft-model context. The mtp_shim is generalised with a spec_type selector shared by MTP and EAGLE-3 (MtpSession unchanged). Verified end-to-end on Apple M4 Pro (Metal) with Qwen3-8B + RedHatAI/Qwen3-8B-speculator.eagle3 at ~53% draft acceptance.
  • mtmd video input (mtmd feature): mtmd::MtmdVideo (+ MtmdVideoParams, MtmdVideoInfo, MtmdVideoItem) wrapping the new mtmd_helper_video_* API for frame-by-frame decoding via ffmpeg, plus MtmdContext::supports_video() and MtmdContext::marker().
  • examples/eagle (+ README) with scripts/setup-eagle3.sh / scripts/fetch-eagle3.sh to download + convert a target/draft pairing to GGUF.

Fixed

  • mtmd bindings: adapt to the reworked helper API — mtmd_helper_bitmap_init_from_file/_from_buf now take a placeholder flag and return a mtmd_helper_bitmap_wrapper; mtmd_helper_decode_image_chunk takes a post-decode callback. MtmdBitmap::from_file/from_buf no longer leak the video_ctx returned for video inputs.

Crates: llama-cpp-4 0.3.2 · llama-cpp-sys-4 0.3.2

Changelog

Sourced from llama-cpp-4's changelog.

[0.3.2] - 2026-06-20

Changed

  • llama.cpp: vendored submodule 94a220cd6c57607016 (master, 2026-06-21; [PR #256](eugenehp/llama-cpp-rs#256)). The public llama.h C API is unchanged; the mtmd helper API was reworked (see below).

Added

  • EAGLE-3 speculative decoding (llama-cpp-4): new [eagle::Eagle3Session] driving upstream COMMON_SPECULATIVE_TYPE_DRAFT_EAGLE3. Pairs a target context with a separate EAGLE-3 draft-model context. The mtp_shim is generalised with a spec_type selector shared by MTP and EAGLE-3 (MtpSession is unchanged).
  • examples/eagle (+ README) and weight-fetch scripts: runnable EAGLE-3 demo plus scripts/setup-eagle3.sh (one command: bootstraps a convert env, then downloads + converts a target/draft pairing to GGUF) over scripts/fetch-eagle3.sh (download + convert). Defaults to the open Qwen3-8B + RedHatAI/Qwen3-8B-speculator.eagle3. Verified end-to-end on Apple M4 Pro (Metal): coherent generation at ~53% draft acceptance.
  • mtmd video input (llama-cpp-4, mtmd feature): [mtmd::MtmdVideo] (+ MtmdVideoParams, MtmdVideoInfo, MtmdVideoItem) wrapping the new mtmd_helper_video_* API for frame-by-frame decoding via ffmpeg. Plus MtmdContext::supports_video() and MtmdContext::marker() (per-context marker).

Fixed

  • mtmd bindings (llama-cpp-4): adapt to the reworked mtmd helper API — mtmd_helper_bitmap_init_from_file/_from_buf now take a placeholder flag and return a mtmd_helper_bitmap_wrapper; mtmd_helper_decode_image_chunk takes a post-decode callback. MtmdBitmap::from_file/from_buf also no longer leak the video_ctx returned for video inputs.
Commits

Dependabot compatibility score

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.


Dependabot commands and options

You can trigger Dependabot actions by commenting on this PR:

  • @dependabot rebase will rebase this PR
  • @dependabot recreate will recreate this PR, overwriting any edits that have been made to it
  • @dependabot show <dependency name> ignore conditions will show all of the ignore conditions of the specified dependency
  • @dependabot ignore this major version will close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself)
  • @dependabot ignore this minor version will close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself)
  • @dependabot ignore this dependency will close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself)

Bumps [llama-cpp-4](https://github.com/eugenehp/llama-cpp-rs) from 0.3.1 to 0.3.2.
- [Release notes](https://github.com/eugenehp/llama-cpp-rs/releases)
- [Changelog](https://github.com/eugenehp/llama-cpp-rs/blob/main/CHANGELOG.md)
- [Commits](eugenehp/llama-cpp-rs@v0.3.1...v0.3.2)

---
updated-dependencies:
- dependency-name: llama-cpp-4
  dependency-version: 0.3.2
  dependency-type: direct:production
  update-type: version-update:semver-patch
...

Signed-off-by: dependabot[bot] <support@github.com>
@dependabot dependabot Bot added dependencies Pull requests that update a dependency file rust Pull requests that update rust code labels Jun 23, 2026
@pacphi

pacphi commented Jun 23, 2026

Copy link
Copy Markdown
Owner

Snyk checks have passed. No issues have been found so far.

Status Scan Engine Critical High Medium Low Total (0)
Open Source Security 0 0 0 0 0 issues

💻 Catch issues earlier using the plugins for VS Code, JetBrains IDEs, Visual Studio, and Eclipse.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

dependencies Pull requests that update a dependency file rust Pull requests that update rust code

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant