fix: correct facial alignment issues and add API server

fix: preprocess import bug (#345)

fix: use torch.no_grad() in inference to prevent excessive memory usage (~30GB) with inference (#349)

feat: windows infer & gradio (#312)

feat: data preprocessing and training (#294)

feat: real-time infer (#286)

fix: infer bug

v1.5

Create __init__.py

<enhance>: support using float16 in inference to speed up

<enhance>(inference): support using an image as video input(#17 #34)

Fix fps calculation bug in realtime_inference.py (#35)

Add codes for real time inference

<enhance>: modified inference codes

Update README.md

initial_commit