self-hosted/ai
§01·model · /models

Mochi 1

videoactiveApache-2.0

10B open text-to-video diffusion model by Genmo, run through the diffusers MochiPipeline in bf16 — where the transformer weights alone are roughly 20 GB. That figure sets the floor directly: all four cards covered here are 24 GB or larger, and the recipes generate 480p clips at 49 and 85 frames. Apache-2.0.

§02·GPUs that run this model
4 total

benchmarked·~ runs via recipe (not benchmarked)· untested·doesn't fit