A video-to-video lip-sync model based on Latent Diffusion, featuring custom code modifications and post-training on 3,000 hours of video data. 


Here are some examples of 512*512 version.

<video controls autoplay loop muted width="100%">
  <source src="https://huggingface.co/ReopenAI/LatentSync_FT/blob/main/anbei.mp4" type="video/mp4">
  Your browser does not support the video tag.
</video>

<video controls autoplay loop muted width="100%">
  <source src="https://huggingface.co/ReopenAI/LatentSync_FT/blob/main/musk.mp4" type="video/mp4">
  Your browser does not support the video tag.
</video>