Skip to content

feat(gallery): add Gemma 4 agentic variants - #11622

Merged
mudler merged 1 commit into
masterfrom
cron/model-gallery-20260820
Aug 20, 2026
Merged

feat(gallery): add Gemma 4 agentic variants#11622
mudler merged 1 commit into
masterfrom
cron/model-gallery-20260820

Conversation

@localai-org-maint-bot

Copy link
Copy Markdown
Collaborator

Description

Add two variants to the existing Gemma 4 12B Agentic v2 gallery entry:

  • Q8_0 for higher output fidelity
  • Q4_K_M with the upstream Q8_0 MTP drafter for speculative decoding

The MTP entry configures spec_type:draft-mtp and carries the mtp tag used by variant ranking. All SHA256 values were checked against the exact Hugging Face LFS artifacts.

Notes for Reviewers

Verification:

  • go test ./core/gallery -ginkgo.focus="gallery/index.yaml variant invariants" -ginkgo.v
  • go run ./.github/ci/modelslist.go ./gallery/index.yaml
  • git diff --check upstream-https/master...HEAD

The full go test ./core/gallery run reached 378 specs: 375 passed and three network-dependent specs failed because this environment returned HTTP 403 for GitHub/Gist test fixtures.

Signed commits

  • Yes, I signed my commits.
  • Documentation updated (docs/content/) for user-facing changes, or not applicable

Expose the upstream Q8_0 build and the MTP drafter for the existing Gemma 4 12B agentic entry. The variants let LocalAI select higher fidelity or faster speculative decoding from the same model family.

Assisted-by: Codex:gpt-5
@mudler
mudler merged commit f7ad3f7 into master Aug 20, 2026
56 of 57 checks passed
@mudler
mudler deleted the cron/model-gallery-20260820 branch August 20, 2026 10:53
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants