# better_wav2lip

**Repository Path**: shaneluik/better_wav2lip

## Basic Information

- **Project Name**: better_wav2lip
- **Description**: No description available
- **Primary Language**: Python
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 2
- **Created**: 2024-01-10
- **Last Updated**: 2024-01-25

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# Better wav2lip model version.
Original repo: https://github.com/Rudrabha/Wav2Lip
- [x] model size 288x288, 384x384, 512x512
- [x] PRelu
- [x] LeakyRelu
- [x] Gradient penalty
- [x] Wasserstein Loss
- [x] SAM-UNet: https://github.com/1343744768/Multiattention-UNet
      
Each line on filelist should be full path <br />

First, Train syncnet <br />

```
python3 train_syncnet_sam.py
```

Second, train wav2lip-Sam
```
python3 hq_wav2lip_sam_train.py
```
Some demo from chinese users:
https://github.com/primepake/wav2lip_288x288/issues/89#issue-2047907323

# New Features: DINet full pipeline training
Original repo: https://github.com/MRzzm/DINet
- [ ] Syncnet training using deepspeech, melspectrogram
- [ ] DINet frame training using melspectrogram
- [ ] DINet clip training using melspectrogram

## Citing

To cite this repository:

```bibtex
@misc{Wav2Lip,
  author={Rudrabha},
  title={Wav2Lip: Accurately Lip-syncing Videos In The Wild},
  year={2020},
  url={https://github.com/Rudrabha/Wav2Lip}
}
```