Collection of AI model pipelines for video generation and processing.
| Model | Description | Docker Image |
|---|---|---|
| lip-sync | Multi-face lip-sync using MuseTalk + LivePortrait + CodeFormer | ghcr.io/allgoodthings/lip-sync |
Each model is self-contained in its own directory with:
- Dockerfile for containerized deployment
- FastAPI server for HTTP API
- Model download scripts
- Documentation
See individual model READMEs for specific usage instructions.
All images are published to GitHub Container Registry (GHCR):
# Pull a specific model
docker pull ghcr.io/allgoodthings/lip-sync:latest
# Run with GPU
docker run --gpus all -p 8000:8000 ghcr.io/allgoodthings/lip-sync:latest# Clone the repo
git clone https://github.com/allgoodthings/models.git
cd models
# Work on a specific model
cd lip-sync
python3 -m venv .venv
source .venv/bin/activate
pip install -e ".[dev]"MIT