From 411ceb758840cf1ee2c1d2510f5f445c618baf1e Mon Sep 17 00:00:00 2001 From: localai-org-maint-bot <306269227+localai-org-maint-bot@users.noreply.github.com> Date: Sun, 11 Oct 2026 08:08:32 +0000 Subject: [PATCH] chore(gallery): add Qwen3.5 Kimi K3 distill Add Q4_K_M and Q8_0 builds with the F16 vision projector. Pin downloads and document installation and variant selection. Assisted-by: Codex:gpt-6 --- docs/content/features/model-gallery.md | 18 ++++++ gallery/index.yaml | 90 ++++++++++++++++++++++++++ 2 files changed, 108 insertions(+) diff --git a/docs/content/features/model-gallery.md b/docs/content/features/model-gallery.md index a5af4a904..5e145d82d 100644 --- a/docs/content/features/model-gallery.md +++ b/docs/content/features/model-gallery.md @@ -179,6 +179,24 @@ local-ai models install cyber-ornith-1.5-9b-obliterated --variant cyber-ornith-1 See the [model card](https://huggingface.co/DuoNeural/Cyber-Ornith-1.5-9B-OBLITERATED) and [GGUF downloads](https://huggingface.co/mradermacher/Cyber-Ornith-1.5-9B-OBLITERATED-i1-GGUF). +## Qwen3.5 9B Kimi K3 Distilled + +[Qwen3.5-9B-Kimi-k3-Distilled](https://huggingface.co/khazarai/Qwen3.5-9B-Kimi-k3-Distilled) +is a Qwen3.5 fine-tune trained on coding, debugging, and tool-use traces. +The gallery groups Q4_K_M and Q8_0 GGUF builds for llama.cpp. Both include +an F16 vision projector and use the embedded chat template with a +32,768-token context by default. + +Install with automatic variant selection: + +```bash +local-ai models install qwen3.5-9b-kimi-k3-distilled-q4 +``` + +Pass `--variant qwen3.5-9b-kimi-k3-distilled-q4` or +`--variant qwen3.5-9b-kimi-k3-distilled-q8` to select a specific build. +The weights and projector use pinned download URLs and SHA256 checksums. + ## Cyber-Tiel-Coder Install `cyber-tiel-coder-35b-a3b-q4-mtp` for coding and image chat with llama.cpp. diff --git a/gallery/index.yaml b/gallery/index.yaml index 944c96774..d39d59d17 100644 --- a/gallery/index.yaml +++ b/gallery/index.yaml @@ -4143,6 +4143,96 @@ - filename: llama-cpp/mmproj/ornith-1.0-9b/mmproj-F16.gguf uri: huggingface://unsloth/Ornith-1.0-9B-GGUF/mmproj-F16.gguf sha256: 013638437d619467b92412e6efa281f706132f05ba6d0f4733e50f812c7e32fa +- name: qwen3.5-9b-kimi-k3-distilled-q4 + variants: + - model: qwen3.5-9b-kimi-k3-distilled-q8 + url: github:mudler/LocalAI/gallery/virtual.yaml@master + urls: + - https://huggingface.co/khazarai/Qwen3.5-9B-Kimi-k3-Distilled + - https://huggingface.co/mradermacher/Qwen3.5-9B-Kimi-k3-Distilled-GGUF + description: | + A 9B Qwen3.5 fine-tune trained on Kimi K3 coding, debugging, and + tool-use traces. This Q4_K_M GGUF build uses llama.cpp and the embedded + chat template, with an F16 vision projector and a 32,768-token context. + license: apache-2.0 + tags: + - llm + - gguf + - cpu + - gpu + - coding + - reasoning + - vision + - multimodal + last_checked: "2026-10-11" + overrides: + backend: llama-cpp + context_size: 32768 + function: + automatic_tool_parsing_fallback: true + grammar: + disable: true + known_usecases: + - chat + - vision + mmproj: llama-cpp/mmproj/qwen3.5-9b-kimi-k3-distilled/Qwen3.5-9B-Kimi-k3-Distilled.mmproj-f16.gguf + options: + - use_jinja:true + template: + use_tokenizer_template: true + parameters: + model: llama-cpp/models/qwen3.5-9b-kimi-k3-distilled/Qwen3.5-9B-Kimi-k3-Distilled.Q4_K_M.gguf + files: + - filename: llama-cpp/models/qwen3.5-9b-kimi-k3-distilled/Qwen3.5-9B-Kimi-k3-Distilled.Q4_K_M.gguf + uri: https://huggingface.co/mradermacher/Qwen3.5-9B-Kimi-k3-Distilled-GGUF/resolve/5172f92801f0bec7be7a7c309c4acc2331190004/Qwen3.5-9B-Kimi-k3-Distilled.Q4_K_M.gguf + sha256: ff0649d92045b564daf4e0b9b0e6b21c6b83b2b1b59448554cf778ad206695f7 + - filename: llama-cpp/mmproj/qwen3.5-9b-kimi-k3-distilled/Qwen3.5-9B-Kimi-k3-Distilled.mmproj-f16.gguf + uri: https://huggingface.co/mradermacher/Qwen3.5-9B-Kimi-k3-Distilled-GGUF/resolve/5172f92801f0bec7be7a7c309c4acc2331190004/Qwen3.5-9B-Kimi-k3-Distilled.mmproj-f16.gguf + sha256: 313847bd5f5b54f1ee2224850e8046bb83b0cbcad86e588581b6d6dd8d239619 +- name: qwen3.5-9b-kimi-k3-distilled-q8 + url: github:mudler/LocalAI/gallery/virtual.yaml@master + urls: + - https://huggingface.co/khazarai/Qwen3.5-9B-Kimi-k3-Distilled + - https://huggingface.co/mradermacher/Qwen3.5-9B-Kimi-k3-Distilled-GGUF + description: | + A 9B Qwen3.5 fine-tune trained on Kimi K3 coding, debugging, and + tool-use traces. This Q8_0 GGUF build uses llama.cpp and the embedded + chat template, with an F16 vision projector and a 32,768-token context. + license: apache-2.0 + tags: + - llm + - gguf + - cpu + - gpu + - coding + - reasoning + - vision + - multimodal + last_checked: "2026-10-11" + overrides: + backend: llama-cpp + context_size: 32768 + function: + automatic_tool_parsing_fallback: true + grammar: + disable: true + known_usecases: + - chat + - vision + mmproj: llama-cpp/mmproj/qwen3.5-9b-kimi-k3-distilled/Qwen3.5-9B-Kimi-k3-Distilled.mmproj-f16.gguf + options: + - use_jinja:true + template: + use_tokenizer_template: true + parameters: + model: llama-cpp/models/qwen3.5-9b-kimi-k3-distilled/Qwen3.5-9B-Kimi-k3-Distilled.Q8_0.gguf + files: + - filename: llama-cpp/models/qwen3.5-9b-kimi-k3-distilled/Qwen3.5-9B-Kimi-k3-Distilled.Q8_0.gguf + uri: https://huggingface.co/mradermacher/Qwen3.5-9B-Kimi-k3-Distilled-GGUF/resolve/5172f92801f0bec7be7a7c309c4acc2331190004/Qwen3.5-9B-Kimi-k3-Distilled.Q8_0.gguf + sha256: f8079d4f29c183ffe72a245eedefcf715e7df44144ca906e8436de44a7ac2a78 + - filename: llama-cpp/mmproj/qwen3.5-9b-kimi-k3-distilled/Qwen3.5-9B-Kimi-k3-Distilled.mmproj-f16.gguf + uri: https://huggingface.co/mradermacher/Qwen3.5-9B-Kimi-k3-Distilled-GGUF/resolve/5172f92801f0bec7be7a7c309c4acc2331190004/Qwen3.5-9B-Kimi-k3-Distilled.mmproj-f16.gguf + sha256: 313847bd5f5b54f1ee2224850e8046bb83b0cbcad86e588581b6d6dd8d239619 - &ornith-1-5-9b name: ornith-1.5-9b-q4 variants: