Text-to-Video
MLX
Safetensors
lance
multimodal
apple-silicon
text-to-image
image-generation
video-generation
diffusion
flow-matching
Mixture of Experts
qwen2_5_vl
wan
port
Instructions to use RockTalk/Lance-3B-Video-MLX with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use RockTalk/Lance-3B-Video-MLX with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] hf download RockTalk/Lance-3B-Video-MLX --local-dir Lance-3B-Video-MLX
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Atomic Chat
Commit History
-
Delete files lance_mlx/__pycache__/__init__.cpython-314.pyc lance_mlx/__pycache__/lance.cpython-314.pyc lance_mlx/__pycache__/modeling_utils.cpython-314.pyc lance_mlx/__pycache__/qwen2_navit_mlx.cpython-314.pyc lance_mlx/__pycache__/vae_wan22.cpython-314.pyc with huggingface_hub
Verified signature6214757 -
Bundle lance_mlx Python package
Verified signature48bed1d -
Update README: remove broken github link, point to bundled code
Verified signature5ba7b08 -
Add requirements.txt
Verified signature3397cf7 -
Add self-contained inference.py + bundled lance_mlx package
Verified signature8c5231a
-
docs: full Phase 2 sync — TI2I + X→T statuses, sequence layouts, perf tables, file descriptions
Verified signaturea7691b5 -
Add T2V samples: ocean wave 9-frame strip + key frames
Verified signatureb37687f -
T2V verified end-to-end: 9-frame ocean wave sample, updated perf table
Verified signature76c1003 -
Initial release: Lance-3B-Video-MLX weights; T2I working, T2V pending VAE streaming cache
Verified signaturefd60d09 -
README: honest status — weight conversion done, inference wrapper partial; document MoE-gen + QK-norm gaps
Verified signaturea7c0c8e -
Add ViT config
Verified signature9f80e19 -
Add ViT (Qwen2.5-VL visual encoder) — 390 tensors / 668M params
Verified signatureb2b3b86 -
Initial MLX port of Lance 3B (video variant) + Wan 2.2 VAE
Verified signatureb4404b4 -
initial commit
Verified signature0865f5e