Code for APDrawingGAN: Generating Artistic Portrait Drawings from Face Photos with Hierarchical GANs (CVPR 2019 Oral)
We provide PyTorch implementations for our CVPR 2019 paper "APDrawingGAN: Generating Artistic Portrait Drawings from Face Photos with Hierarchical GANs".
This project generates artistic portrait drawings from face photos using a GAN-based model. You may find useful information in preprocessing steps and training/testing tips.
Up: input, Down: output
If you use this code for your research, please cite our paper.
@inproceedings{YiLLR19,
title = {{APDrawingGAN}: Generating Artistic Portrait Drawings from Face Photos with Hierarchical GANs},
author = {Yi, Ran and Liu, Yong-Jin and Lai, Yu-Kun and Rosin, Paul L},
booktitle = {{IEEE} Conference on Computer Vision and Pattern Recognition (CVPR '19)},
pages = {10743--10752},
year = {2019}
}
pip install -r requirements.txt
Download a pre-trained model (using 70 pairs in training set and augmented data) from https://cg.cs.tsinghua.edu.cn/people/~Yongjin/APDrawingGAN-Models1.zip (Model1) and put it in checkpoints/formal_author
.
Then generate artistic portrait drawings for example photos in dataset/data/test_single
using
python test.py --dataroot dataset/data/test_single --name formal_author --model test --dataset_mode single --norm batch --use_local --which_epoch 300
The test results will be saved to a html file here: ./results/formal_author/test_300/index.html
.
python test.py --dataroot {path_to_aligned_photos} --name formal_author --model test --dataset_mode single --norm batch --use_local --which_epoch 300
dataset
folderpython -m visdom.server
checkpoints/[name]
, e.g. checkpoints/formal
), and copy "auxiliary" models into checkpoints/auxiliary
.python train.py --dataroot dataset/data --name formal --continue_train --use_local --discriminator_local --niter 300 --niter_decay 0 --save_epoch_freq 25
checkpoints/auxiliary
.python train.py --dataroot dataset/data --name formal_noinit --use_local --discriminator_local --niter 300 --niter_decay 0 --save_epoch_freq 25
./checkpoints/formal/web/index.html
python test.py --dataroot dataset/data --name formal --use_local --which_epoch 250
The test results will be saved to a html file here: ./results/formal/test_250/index.html
.
--model test
, --dataset_mode single
and --norm batch
):python test.py --dataroot dataset/data/test_single --name formal --model test --dataset_mode single --norm batch --use_local --which_epoch 250
You can find these scripts at scripts
directory.
Preprocessing steps for your own data (either for testing or training).
Best practice for training and testing your models.
You can contact email [email protected] for any questions.
Our code is inspired by pytorch-CycleGAN-and-pix2pix.