Instructions to use mlx-community/sam-3d-objects-bf16 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use mlx-community/sam-3d-objects-bf16 with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] hf download mlx-community/sam-3d-objects-bf16 --local-dir sam-3d-objects-bf16
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Atomic Chat
Commit History
-
Replace the MoGe-v1 depth model with MoGe-3 (float32)
Verified signatureThe bundled depth estimator is now MoGe-3 ViT-L (mlx-community/moge-3-vitl-mlx-fp32, float32, 607 tensors under the depth_model prefix), used by mlx-vlm's sam3d_objects runtime through the shared moge3 model package. config.json stores the MoGe-3 configuration under depth_model, the weight index and conversion report list the new shard, validation.json is regenerated for the new bundle, and the card is updated.26f21f6
-
Keep weights only: runtime lives in mlx-vlm; refresh card and validation
Verified signature23e686a
-
Add frozen BF16 MLX SAM 3D Objects inference with async streaming
Verified signatureb846e48 -
initial commit
Verified signature7f3ac1a