Skip to content

README.md

Latest commit

 

History

History
58 lines (45 loc) · 1.76 KB

README.md

File metadata and controls

58 lines (45 loc) · 1.76 KB

For use with totally-real-news-bot https://github.com/jeddyhhh/totally-real-news-bot

Modified to save files locally rather than to AWS storage, installs and works the same other than that. Just follow the installation below and it should work

Thanks to Phạm Gia Linh, source peoject

How to run

Manual Installation

Linux:

  1. Installing miniconda, python and git.

  2. Creating the env and install the requirements.

git clone https://github.com/phamgialinhlx/sad-talker-api.git

cd sad-talker-api 

conda create -n sadtalker python=3.8

conda activate sadtalker

pip install torch==1.12.1+cu113 torchvision==0.13.1+cu113 torchaudio==0.12.1 --extra-index-url https://download.pytorch.org/whl/cu113

conda install ffmpeg

pip install -r requirements.txt
  1. Create .env from .env.example file
  2. Download models
bash scripts/download_models.sh
  1. Host the server.
uvicorn --host "0.0.0.0" --port "8000" api:app

Docker Installation

  1. Build sadtalker image
docker build -t sadtalker .
  1. Create .env from .env.example file
  2. Run sadtalker container
docker run --gpus=all --rm -p 8000:8000 -v ./.env:/sadtalker/.env -d --name sadtalker sadtalker

Note: Remember to volume mount .env file to container

Test the API

The API will be hosted on port 8000. Go to http://127.0.0.1:8000/docs to see API documentation.

curl -X POST "http://localhost:8000/generate/" -H "Content-Type:application/json" -d '{"image_link": "https://raw.githubusercontent.com/OpenTalker/SadTalker/main/examples/source_image/happy.png","audio_link": "https://github.com/OpenTalker/SadTalker/raw/main/examples/driven_audio/chinese_poem2.wav"}'