TLDR Index

TLDR AI · ENGINEERING & RESOURCES

MOCHI VIDEO MODEL WITH OPTIMIZED INFERENCE

github.comGitHub repository

Summary from the TLDR AI newsletter

Mochi is the best open source text-to-video model. On launch, it took 8 H100s to run. Now, thanks to the community, it can be run on a single 48GB L40 with no loss in quality.

Read the original article ↗View the archived TLDR AI issue