Implementation of CVPR 2021 paper "Spatially-invariant Style-codes Controlled Makeup Transfer"
The pre-trained model is avaiable at https://drive.google.com/file/d/1t1Hbgqqzc_rV5v3gF7HuJ-xiuEVNb8sh/view?usp=sharing.
vgg_conv.pth:https://drive.google.com/file/d/1JNrSVZrK4TfC7pFG-r7AOmGvBXF2VFOt/view?usp=sharing
Put the G.pth and VGG weights in "./checkpoints" and "./" respectively.
Environments:python=3.8, pytorch=1.6.0, Ubuntu=20.04.1 LTS
Put the train-list of makeup images in "./MT-Dataset/makeup.txt" and the train-list of non-makeup images in "./MT-Dataset/non-makeup.txt"
Use the "./scripts/handle_parsing.py" to convert the origin MT-Dataset's seg labels
Use python sc.py --phase train to train
python sc.py --phase test
python sc.py --phase test --partial
python sc.py --phase test --interpolation
python sc.py --phase test --partial --interpolation
