Wan recipes
Compare Wan modes and options
Supported modes
FastMetal MLX is T2V in the checked-in examples. Image-to-video and
TI2V stay on the CUDA recipes. Temporal –fast composes
with either spatial path. –refine and
–fast-spatial cannot run together.
–refine wins if both are set.
basic_mps.py is the older PyTorch MPS demo and is not a
FastMetal recipe.
| Mode | CUDA | MLX FastMetal |
|---|---|---|
| T2V | FastWan2.1 1.3B, Wan2.2 A14B | 1.3B, 5B, and 14B |
| I2V | Wan2.1 14B 480P | Not in the released examples |
| TI2V | Wan2.2 TI2V 5B | FastMetal 5B is T2V in mlx_wan22_generate.py |
Temporal –fast | No cookbook recipe | RIFE. Fewer frames, then interpolate to –num-frames |
Spatial –fast-spatial | No cookbook recipe | Denoise and decode at half resolution, then upsample. No second denoise |
Two-pass –refine | No cookbook recipe | Denoise at base resolution, upsample, re-noise, denoise again. Wins over –fast-spatial |
Setup
The generated commands expect a local clone:
git clone https://github.com/hao-ai-lab/FastVideo.git
cd FastVideoUse Configuration for supported Python and CLI settings, Optimizations for attention and memory tradeoffs, and the support matrix for the supported model and optimization surface.
Troubleshooting
- Out of memory on the A14B recipes: the checked-in sources already enable CPU offload; see Configuration for the offload surface before reducing resolution or frames.
- The FastWan2.1 recipe requires
VIDEO_SPARSE_ATTN; confirm the environment variable in the command was set in the same shell. - FastMetal MLX: install with the MLX install guide, then pick a FastMetal recipe in the builder. CUDA FastWan-QAD checkpoints are refused on the MLX runtime.
- FastMetal 5B uses
mlx_wan22_generate.py. 1.3B and 14B usemlx_wan_prompt_to_video.py. - Gated or missing checkpoints: run
huggingface-cli loginand confirm you accepted the model’s license on Hugging Face.
Evidence status
Every recipe on this page maps to a checked-in FastVideo source. The FastMetal MLX releases include the recorded M4 Max system memory, documented unified-memory floor, and measured peak MLX memory. CUDA entries remain Source-backed where the examples record a GPU count but no exact GPU model or VRAM. Unlisted hardware is unknown, not unsupported.