hugpadhugpad

Efficient-Large-Model / Sol-H3-VAE

Other4.5K6Updated Oct 28, 2026
Hugging Face

Downloads

4.5K

Likes

6

Last updated

Oct 28, 2026

License

Not available

A fast video decoder for MiniMax-H3. It replaces the 36-layer H3 VAE decoder with a 24-layer noncausal decoder (1.62B parameters) distilled from it: whole frames with windowed attention instead of spatial tiles, and each native 5-latent block decoded bidirectionally. It takes …

Tasks & library

  • diffusers

Tags

  • diffusers
  • minimax-h3
  • vae
  • video

Technical specifications

ParametersNot available
Librarydiffusers
LanguagesNot available
LicenseNot available