mirror of
https://github.com/comfyanonymous/ComfyUI.git
synced 2026-02-02 17:50:19 +08:00
The WAN VAE is also QWEN where it is used single-image. These convolutions are however zero padded 3d convolutions, which means the VAE is actually just 2D down the last element of the conv weight in the temporal dimension. Fast path this, to avoid adding zeros that then just evaporate in convoluton math but cost computation. |
||
|---|---|---|
| .. | ||
| model_animate.py | ||
| model_multitalk.py | ||
| model.py | ||
| vae2_2.py | ||
| vae.py | ||