feat(gallery): add Muse Glimmer 30B variants (#11446)

Add the official dynamic and 17 GB K-quant GGUF builds with the shared multimodal projector and optional DFlash acceleration. Assisted-by: Codex:gpt-5 Co-authored-by: localai-org-maint-bot <306269227+localai-org-maint-bot@users.noreply.github.com>
2026-09-12 22:33:54 -04:00 · 2026-08-11 09:51:58 +02:00
1 parent 0df63de1ed
commit 129f834039
@@ -1,4 +1,209 @@
 ---
+- &muse-glimmer-30b
+  name: "muse-glimmer-30b"
+  variants:
+    - model: muse-glimmer-30b-dflash
+    - model: muse-glimmer-30b-17gb
+    - model: muse-glimmer-30b-17gb-dflash
+  url: "github:mudler/LocalAI/gallery/virtual.yaml@master"
+  urls:
+    - https://huggingface.co/meta-models/Muse-Glimmer-30B
+    - https://huggingface.co/meta-models/Muse-Glimmer-30B-GGUF
+  description: |
+    Muse Glimmer is Meta Superintelligence Labs' Apache-2.0 dense 30B model for
+    autonomous agentic work, coding, tool use, long-horizon reasoning, and
+    multimodal understanding. It supports more than 100 languages, interleaved
+    text and image input through its 1.8B-parameter perception encoder, and a
+    131K-token context window.
+
+    This entry uses the publisher's higher-quality dynamic K-quant GGUF and
+    official quantized vision projector. Automatic variant selection can use
+    the smaller 17 GB quantization or a DFlash-accelerated build when it fits.
+  license: "apache-2.0"
+  tags:
+    - llm
+    - gguf
+    - cpu
+    - gpu
+    - vision
+    - multimodal
+    - reasoning
+    - coding
+    - agent
+    - tools
+    - multilingual
+    - long-context
+  last_checked: "2026-08-10"
+  overrides:
+    backend: llama-cpp
+    function:
+      automatic_tool_parsing_fallback: true
+      grammar:
+        disable: true
+    known_usecases:
+      - chat
+      - vision
+    mmproj: llama-cpp/mmproj/muse-glimmer-30b/mmproj-kquant.gguf
+    options:
+      - use_jinja:true
+    parameters:
+      model: llama-cpp/models/muse-glimmer-30b/muse-glimmer-30B-kquant-dynamic.gguf
+      temperature: 1.0
+      top_k: 64
+      top_p: 0.95
+    template:
+      use_tokenizer_template: true
+  files:
+    - filename: llama-cpp/models/muse-glimmer-30b/muse-glimmer-30B-kquant-dynamic.gguf
+      uri: huggingface://meta-models/Muse-Glimmer-30B-GGUF/muse-glimmer-30B-kquant-dynamic.gguf
+      sha256: 513109c8319115f69eb09fb7b118c97c8167d15bc014fd7670d2e30489bf106c
+    - filename: llama-cpp/mmproj/muse-glimmer-30b/mmproj-kquant.gguf
+      uri: huggingface://meta-models/Muse-Glimmer-30B-GGUF/mmproj-kquant.gguf
+      sha256: f48b452316f9b213758e8659444029b961a24a07f99a1abb2a9f88b06f7c00c6
+- !!merge <<: *muse-glimmer-30b
+  name: "muse-glimmer-30b-dflash"
+  variants: []
+  description: |
+    Muse Glimmer's higher-quality dynamic K-quant GGUF with the official
+    quantized perception encoder and DFlash drafter. DFlash proposes blocks of
+    up to 16 tokens for the target to verify in parallel, accelerating output
+    without changing model quality. Flash attention is enabled for this path.
+  tags:
+    - llm
+    - gguf
+    - gpu
+    - vision
+    - multimodal
+    - reasoning
+    - coding
+    - agent
+    - tools
+    - multilingual
+    - long-context
+    - dflash
+    - speculative
+  overrides:
+    backend: llama-cpp
+    flash_attention: "on"
+    draft_model: llama-cpp/models/muse-glimmer-30b/dflash-kquant.gguf
+    function:
+      automatic_tool_parsing_fallback: true
+      grammar:
+        disable: true
+    known_usecases:
+      - chat
+      - vision
+    mmproj: llama-cpp/mmproj/muse-glimmer-30b/mmproj-kquant.gguf
+    options:
+      - use_jinja:true
+      - spec_type:draft-dflash
+      - spec_n_max:15
+    parameters:
+      model: llama-cpp/models/muse-glimmer-30b/muse-glimmer-30B-kquant-dynamic.gguf
+      temperature: 1.0
+      top_k: 64
+      top_p: 0.95
+    template:
+      use_tokenizer_template: true
+  files:
+    - filename: llama-cpp/models/muse-glimmer-30b/muse-glimmer-30B-kquant-dynamic.gguf
+      uri: huggingface://meta-models/Muse-Glimmer-30B-GGUF/muse-glimmer-30B-kquant-dynamic.gguf
+      sha256: 513109c8319115f69eb09fb7b118c97c8167d15bc014fd7670d2e30489bf106c
+    - filename: llama-cpp/mmproj/muse-glimmer-30b/mmproj-kquant.gguf
+      uri: huggingface://meta-models/Muse-Glimmer-30B-GGUF/mmproj-kquant.gguf
+      sha256: f48b452316f9b213758e8659444029b961a24a07f99a1abb2a9f88b06f7c00c6
+    - filename: llama-cpp/models/muse-glimmer-30b/dflash-kquant.gguf
+      uri: huggingface://meta-models/Muse-Glimmer-30B-GGUF/dflash-kquant.gguf
+      sha256: 27d9a805fa29b943cfb6ad4843367cd4eaaaf06bd452d8cc3e00a2cd18a677bc
+- !!merge <<: *muse-glimmer-30b
+  name: "muse-glimmer-30b-17gb"
+  variants: []
+  description: |
+    Muse Glimmer's smaller 17 GB K-quant GGUF with the official quantized
+    perception encoder. It preserves the model's agentic, coding, tool-use,
+    multilingual, and image-understanding capabilities for hosts with less
+    memory than the dynamic quantization requires.
+  overrides:
+    backend: llama-cpp
+    function:
+      automatic_tool_parsing_fallback: true
+      grammar:
+        disable: true
+    known_usecases:
+      - chat
+      - vision
+    mmproj: llama-cpp/mmproj/muse-glimmer-30b/mmproj-kquant.gguf
+    options:
+      - use_jinja:true
+    parameters:
+      model: llama-cpp/models/muse-glimmer-30b/muse-glimmer-30B-kquant-17gb.gguf
+      temperature: 1.0
+      top_k: 64
+      top_p: 0.95
+    template:
+      use_tokenizer_template: true
+  files:
+    - filename: llama-cpp/models/muse-glimmer-30b/muse-glimmer-30B-kquant-17gb.gguf
+      uri: huggingface://meta-models/Muse-Glimmer-30B-GGUF/muse-glimmer-30B-kquant-17gb.gguf
+      sha256: 7e9b74b7c8875e9e265695df9613bf6290f2392e479ce740495a129019c488d8
+    - filename: llama-cpp/mmproj/muse-glimmer-30b/mmproj-kquant.gguf
+      uri: huggingface://meta-models/Muse-Glimmer-30B-GGUF/mmproj-kquant.gguf
+      sha256: f48b452316f9b213758e8659444029b961a24a07f99a1abb2a9f88b06f7c00c6
+- !!merge <<: *muse-glimmer-30b
+  name: "muse-glimmer-30b-17gb-dflash"
+  variants: []
+  description: |
+    Muse Glimmer's smaller 17 GB K-quant GGUF with the official quantized
+    perception encoder and DFlash drafter. This is the lowest-memory published
+    build that retains image understanding and block-speculative decoding.
+    Flash attention is enabled for the DFlash path.
+  tags:
+    - llm
+    - gguf
+    - gpu
+    - vision
+    - multimodal
+    - reasoning
+    - coding
+    - agent
+    - tools
+    - multilingual
+    - long-context
+    - dflash
+    - speculative
+  overrides:
+    backend: llama-cpp
+    flash_attention: "on"
+    draft_model: llama-cpp/models/muse-glimmer-30b/dflash-kquant.gguf
+    function:
+      automatic_tool_parsing_fallback: true
+      grammar:
+        disable: true
+    known_usecases:
+      - chat
+      - vision
+    mmproj: llama-cpp/mmproj/muse-glimmer-30b/mmproj-kquant.gguf
+    options:
+      - use_jinja:true
+      - spec_type:draft-dflash
+      - spec_n_max:15
+    parameters:
+      model: llama-cpp/models/muse-glimmer-30b/muse-glimmer-30B-kquant-17gb.gguf
+      temperature: 1.0
+      top_k: 64
+      top_p: 0.95
+    template:
+      use_tokenizer_template: true
+  files:
+    - filename: llama-cpp/models/muse-glimmer-30b/muse-glimmer-30B-kquant-17gb.gguf
+      uri: huggingface://meta-models/Muse-Glimmer-30B-GGUF/muse-glimmer-30B-kquant-17gb.gguf
+      sha256: 7e9b74b7c8875e9e265695df9613bf6290f2392e479ce740495a129019c488d8
+    - filename: llama-cpp/mmproj/muse-glimmer-30b/mmproj-kquant.gguf
+      uri: huggingface://meta-models/Muse-Glimmer-30B-GGUF/mmproj-kquant.gguf
+      sha256: f48b452316f9b213758e8659444029b961a24a07f99a1abb2a9f88b06f7c00c6
+    - filename: llama-cpp/models/muse-glimmer-30b/dflash-kquant.gguf
+      uri: huggingface://meta-models/Muse-Glimmer-30B-GGUF/dflash-kquant.gguf
+      sha256: 27d9a805fa29b943cfb6ad4843367cd4eaaaf06bd452d8cc3e00a2cd18a677bc
 - &qwen3-5-9b-defiant-fable
  name: "qwen3.5-9b-defiant-fable-mtp"
  variants: