feat(gallery): add Huihui Qwen3.8 quantizations - #11909
Closed
localai-org-maint-bot wants to merge 1 commit into
Closed
feat(gallery): add Huihui Qwen3.8 quantizations#11909localai-org-maint-bot wants to merge 1 commit into
localai-org-maint-bot wants to merge 1 commit into
Conversation
Add Q4_K and Q8_0 builds alongside the BF16 entry, with the shared vision projector and MTP configuration. Correct the model description and document direct installation. Assisted-by: Codex:gpt-6
mudler
added a commit
that referenced
this pull request
Sep 11, 2026
Assisted-by: localai-org-maint-bot:glm5.2 [gh]
Collaborator
Author
|
Cherry-picked to master via 3-way merge (git merge-file). All 21 gallery PRs were applied in sequence. Thank you! |
mudler
added a commit
that referenced
this pull request
Sep 12, 2026
The batch of "gallery: apply PR" commits replayed gallery-agent diffs against a stale base. Each new top-of-file entry overwrote the entry above it instead of being inserted, which lost seven entries: - qwen3.8-27b-uncensored-q4/-q8 (#11705, overwritten by #11909) - qwen3.8-flash-next-uncensored (#11832, overwritten by #11841) - spark-x2.5-4b-q4/-q6/-q8 (#11923, overwritten by #11926) - deepseek-v4-flash-vision-exp (#11873): #11927 renamed its name line to qwopus3.8-27b-flash, which duplicated that entry and failed the "declares every entry name exactly once" gallery lint on master. Each restored entry is identical (YAML-equal) to the one in its PR head. Assisted-by: Claude:claude-opus-5 [Bash] Signed-off-by: Ettore Di Giacinto <mudler@localai.io> Co-authored-by: Ettore Di Giacinto <mudler@localai.io>
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.
Description
Add Q4_K and Q8_0 GGUF variants to the existing Huihui Qwen3.8 27B BF16 entry. Both use llama.cpp, the shared BF16 vision projector, the embedded chat template, and MTP speculative decoding. Replace the copied upstream description with details about the abliterated model and document direct installation.
The existing BF16 entry keeps its name and payload. The new variants are separate installable entries with empty variant lists.
Notes for Reviewers
246becd5a88523e21c58e848e791955719e9d1c9). The publisher documents preserved MTP and vision weights.x-linked-etagheaders. Q4_K is 16,810,714,400 bytes; Q8_0 is 29,047,084,320 bytes.go test ./core/gallery -count=1 -ginkgo.focus='gallery/index.yaml|gallery variant lint'passes.go test ./core/gallery -count=1: 388 pass, 9 skip, 3 fail because external GitHub/Gist fixtures return Forbidden. The same three failures reproduce with the unchanged master gallery.git diff --checkpasses. Inference was not run.Signed commits
.agents/ai-coding-assistants.md.