Skip to content

gallery: add Qwen3.5 9B Defiant Fable variants - #11335

Open
localai-org-maint-bot wants to merge 1 commit into
masterfrom
cron/model-gallery-20260804-nemotron-embed
Open

gallery: add Qwen3.5 9B Defiant Fable variants#11335
localai-org-maint-bot wants to merge 1 commit into
masterfrom
cron/model-gallery-20260804-nemotron-embed

Conversation

@localai-org-maint-bot

Copy link
Copy Markdown
Collaborator

What

  • add the popular Qwen3.5 9B Defiant Fable NEO-imatrix Q4_K_M model
  • offer the MTP build as the default and the plain build as a fallback variant
  • include the shared BF16 multimodal projector
  • record SHA256 values verified against Hugging Face x-linked-etag metadata

Why

The model is currently popular on Hugging Face, runs on LocalAI's llama.cpp backend, and was not yet represented in the gallery. Unsupported trending candidates that require downstream llama.cpp patches were deliberately excluded.

Verification

  • git diff --check
  • go test ./core/gallery -ginkgo.focus="gallery/index.yaml variant invariants|keeps every entry that declares variants installable"
  • full go test ./core/gallery: 336 passed; 3 unrelated network-backed specs failed because this runner receives HTTP 403 from raw.githubusercontent.com/Gist endpoints
  • all three artifact hashes independently rechecked via Hugging Face x-linked-etag

Assisted-by: Codex:gpt-5

Add the MTP and plain Q4_K_M GGUF builds with their shared vision projector so LocalAI users can select accelerated or fallback llama.cpp inference.

Assisted-by: Codex:gpt-5
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant