Requirements

you need to do this below

conda env create -f environment.yaml
pip install -U openmim
pip install torchvision-0.15.1%2Bcu117-cp310-cp310-linux_x86_64.whl

mim install mmengine
mim install "mmcv>=2.0.1"
mim install "mmdet>=3.1.0"
mim install "mmpose>=1.1.0"

conda install -c conda-forge cudatoolkit-dev -y
pip3 install deepspeed

git clone https://huggingface.co/HVSiniX/RawVideoDriven && ln -s RawVideoDriven/weights weights

Inference example

CUDA_VISIBLE_DEVICES=0 python eval.py --config configs/infer12_catnoise_warp08_power_vasa.yaml --source new_test_data/vasa_case1.mp4 --driver new_test_data/vasa_case1.mp4