This repo, a fork of the original project, has been updated with a few minor modifications to improve the workflow.
- Lip-syncing conversion and voice merging can now be performed simultaneously with a single CLI command.
- I've also modified the source code to skip frames with no faces, which helps speed up the process.
For setup, refer to original repo.
Original Repository: https://github.com/bytedance/LatentSync
python mergeprocess.py --video assets/<your_video>.mp4 --audio assets/<your_audio>.wav --out <output_path>.mp4