randlab / pytorch-wgan Goto Github PK

This repository was forked from https://github.com/Zeleni9/pytorch-wgan for the purpose of using WGAN in the context of training image based inversion.

This is the pytorch implementation of 3 different GAN models using same convolutional architecture.

• DCGAN (Deep convolutional GAN)
• WGAN-CP (Wasserstein GAN using weight clipping)
• WGAN-GP (Wasserstein GAN using gradient penalty)

The prominent packages are:

• numpy
• scikit-learn
• tensorflow 2.5.0
• pytorch 1.8.1
• torchvision 0.9.1

To install all the dependencies quickly and easily you should use pip

pip install -r requirements.txt

Running training of WGAN-GP model on 128x128 extractions from ti/zahner.png :

python main.py --model WGAN-GP-128 \
               --is_train True \
               --dataset ti_sampler \
               --ti_file zahner.png \
               --cuda True \
               --batch_size 64 \
               --dataroot ti

This training took around 4h 21 min on a single GPU.

• WGAN reddit thread
• Blogpost
• Deconvolution and checkboard Artifacts
• WGAN-CP paper
• WGAN-GP paper
• DCGAN paper
• Working remotely with PyCharm and SSH