Skip to content

Latest commit

 

History

History
113 lines (83 loc) · 2.2 KB

README.md

File metadata and controls

113 lines (83 loc) · 2.2 KB
Raw

ISF-GAN with StyleGAN v2

Configuration

  • Python: 3.6.8
  • Pytorch: 1.6.0+cu101
  • torchvision: 0.7.0a0+78ed10c
  • timm: 0.4.5
  • ffmpeg
  • munch

Download pretrained models:

Download pretrained models from here.

Preparing dataset

  • Generate data through stylegan2 (with the released data from StyleFlow):
# stylegan
$ python3 generate_data.py --save_dir /path/to/save_dir
  • Download pretrained classifiers:
$ python3 ../styleganv1/preparing/download.py
  • Collect attributes of generated images:
$ python3 ../styleganv1/preparing/collect_attributes.py
$ python3 ../styleganv1/preparing/merge.py

Finally, we can organize the dataset in this format:

../datasets/stylegan2-ffhq
  |__train
  |   |__xxx.jpg
  |   |__log.txt
  |   |__w.npy
  |   |__wp.npy
  |   |__z.npy
  |
  |__test
  |   |__xxx.jpg
  |   |__log.txt
  |   |__w.npy
  |   |__wp.npy
  |   |__z.npy
  |
  |__list_attr_ffhq-test.txt
  |__list_attr_ffhq-train.txt

Training

$ sh ./scripts/train.sh

Testing

  • Image Editing
$ python3 isfgan_edit.py \
  --checkpoint_dir /path/to/checkpoint_dir \
  --use_post 1 \
  --save_dir /path/to/save_dir
  • Fast testing:

The pretrained models can be downloaded here.

$ python3 isfgan_edit_single.py \
  --checkpoint_dir /path/to/checkpoint_dir \
  --use_post 1 \
  --save_dir /path/to/save_dir
Input Gender Eyeglasses Age Expression
  • Image interpolation
$ python3 isfgan_interp.py \
  --checkpoint_dir /path/to/checkpoint_dir \
  --use_post 1 \
  --save_dir /path/to/save_dir
  • Image sampling
$ python3 isfgan_sample.py \
  --checkpoint_dir /path/to/checkpoint_dir \
  --use_post 1 \
  --save_dir /path/to/save_dir