Instructions to use OsaurusAI/Muse-Glimmer-30B-JANG_2D with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use OsaurusAI/Muse-Glimmer-30B-JANG_2D with MLX:

# Make sure mlx-vlm is installed
# pip install --upgrade mlx-vlm

from mlx_vlm import load, generate
from mlx_vlm.prompt_utils import apply_chat_template
from mlx_vlm.utils import load_config

# Load the model
model, processor = load("OsaurusAI/Muse-Glimmer-30B-JANG_2D")
config = load_config("OsaurusAI/Muse-Glimmer-30B-JANG_2D")

# Prepare input
image = ["http://images.cocodataset.org/val2017/000000039769.jpg"]
prompt = "Describe this image."

# Apply chat template
formatted_prompt = apply_chat_template(
    processor, config, prompt, num_images=1
)

# Generate output
output = generate(model, processor, formatted_prompt, image)
print(output)

Notebooks
Google Colab
Kaggle
Local Apps Settings
LM Studio

How to use OsaurusAI/Muse-Glimmer-30B-JANG_2D with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "OsaurusAI/Muse-Glimmer-30B-JANG_2D"

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "mlx-lm": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "OsaurusAI/Muse-Glimmer-30B-JANG_2D"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent

How to use OsaurusAI/Muse-Glimmer-30B-JANG_2D with Hermes Agent:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "OsaurusAI/Muse-Glimmer-30B-JANG_2D"

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default OsaurusAI/Muse-Glimmer-30B-JANG_2D

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use OsaurusAI/Muse-Glimmer-30B-JANG_2D with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "OsaurusAI/Muse-Glimmer-30B-JANG_2D"

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "OsaurusAI/Muse-Glimmer-30B-JANG_2D" \
  --custom-provider-id mlx-lm \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

Muse-Glimmer-30B-JANG_2D

JANG mixed-affine quantization of meta-models/Muse-Glimmer-30B for Apple Silicon runtimes that implement Muse Glimmer's dense multimodal architecture.

Bundle

Field	Value
Source revision	`f84ecc3a0ea984a4c04542a84269e3d065350a6e`
Format	`jang_affine` / JANG v2
Profile	`JANG_2D`
Bit policy	4-bit attention/lm_head; 3-bit embeddings and MLP gate/down; 2-bit MLP up; FP16 vision
Effective language-weight bits	2.96
Indexed weights	15.89 GB / 14.80 GiB
Safetensor shards	21
Indexed tensor keys	2272
Language layers	52: 39 sliding + 13 full attention
Vision tensors	809 FP16 passthrough tensors
Assistant / DFlash	Not included; maintained as a separate artifact

This is a post-training quantization baseline. The upstream BF16 checkpoint contains no QAT weights or QAT scale metadata. GPTQ, imatrix, and AWQ were not applied, and the metadata says so explicitly.

Native model contract

Text input and text output are part of the source architecture.
The vision tower, adapter, projection, image processor, and video processor sidecars are preserved. Image/video execution in vMLX remains unverified.
Reasoning uses reasoning_strength=low|medium|high|xhigh; omission defaults to high in the shipped chat template.
Reasoning output is an assistant to=self channel. Visible content is addressed to=user.
Tools use the ATEM function-call grammar. A Muse-specific incremental reasoning parser and ATEM tool parser are required.
Deployment sampling follows the upstream model-card recommendation: do_sample=true, temperature=1.0, top_p=0.95, and top_k=64. BOS is 200000, EOS is [200001, 200008], pad is 200018, and maximum length is 131072.
Cache topology is heterogeneous: rotating KV with window 2048 on 39 layers and unbounded KV on full-attention layers 3, 7, ..., 51.

Runtime status

PARTIAL / text generation verified in the target runtime. On 2026-08-10, the exact local JANG_2D bundle was loaded by Osaurus/vmlx-swift with all 418 per-module mixed-bit overrides. A request with no sampling or reasoning override, Reply with exactly: FOUR, returned visible content FOUR, a separate non-empty reasoning_content, and finish_reason=stop at 36.87 tokens/s. This covers coherent text generation, the native default reasoning path, reasoning-channel separation, and EOS handling.

Still unverified: grounded image input, grounded video input, ATEM tool round trips, multi-turn behavior, and prefix/partial-block/suffix cache reuse.

Do not interpret repository availability or structural loading as production readiness. This format is not a uniform mlx_lm quant; loaders must honor every per-module entry in config.json.quantization.

Files

config.json: Muse architecture plus per-module affine overrides.
jang_config.json: source revision, profile, capability, reasoning/tool, modality, generation, and mixed full/sliding cache metadata.
chat_template.jinja: exact upstream Muse channel/ATEM template.
processor_config.json: exact upstream image/video processor contract.
generation_config.json: upstream-recommended deployment sampling plus the source token IDs and maximum length.
LICENSE and USAGE_POLICY.md: copied from the pinned upstream source.

Download

hf download OsaurusAI/Muse-Glimmer-30B-JANG_2D \
  --local-dir ~/models/OsaurusAI/Muse-Glimmer-30B-JANG_2D

Verification

From the JANG repository:

PYTHONPATH=jang-tools uv run --no-project \
  --with mlx --with numpy --with safetensors --with tqdm \
  python jang-tools/scripts/verify_muse_glimmer_artifact.py \
  ~/models/OsaurusAI/Muse-Glimmer-30B-JANG_2D \
  --profile JANG_2D --dequant

Korean summary

이 번들은 공식 meta-models/Muse-Glimmer-30B BF16 체크포인트를 Apple Silicon용 JANG 혼합 affine 형식으로 변환한 JANG_2D PTQ 모델입니다. 52개 언어 레이어는 슬라이딩/전체 어텐션 구조를 유지하며, 비전 타워·어댑터·프로젝션 809개 텐서는 FP16으로 보존됩니다. 기본 추론 강도는 high이고 도구 호출은 ATEM 형식입니다. 현재 파일 구조, 메타데이터, 사이드카 및 일부 역양자화 비교를 확인했습니다. 실제 Osaurus/vmlx-swift에서 기본 추론 경로의 일관된 텍스트 생성, 분리된 reasoning_content 및 EOS 종료를 확인했습니다. 이미지/비디오, ATEM 도구 왕복, 멀티턴 및 캐시 재사용은 아직 검증되지 않았습니다. 별도 5레이어 DFlash assistant는 포함하지 않습니다.

License and use

Apache 2.0 license and the upstream Muse Glimmer Usage Policy apply. Review LICENSE and USAGE_POLICY.md before use.

Contact

eric@osaurus.ai

Downloads last month: 229

Safetensors

Model size

5B params

Tensor type

U32

F16

MLX

Hardware compatibility

Quantized

Model tree for OsaurusAI/Muse-Glimmer-30B-JANG_2D

Base model

meta-models/Muse-Glimmer-30B

Quantized

(150)

this model