TalkVid: Training & Inference
Setup
# Create environment
conda create -n talkingface python=3.10 -y
conda activate talkingface
pip install -r requirements.txt
Training
bash train_stage.sh
Inference
bash inference.sh
Notes
- Please preprocess data under
data_preprocess/before training. - Update all paths in
configs/*.yamlbefore use. - Outputs will be saved to the directory specified in the config.