Bidur-Khanal / BigBiGAN-PyTorch

Unofficail pytorch implementation of BigBiGAN

BigBiGAN-PyTorch

An unofficial small scale implementation of BigBiGAN

Architecture

Requirements

Python 3.6.9
PyTorch 1.6.0
Numpy 1.18.5

Implementation

Generator & Discriminator

Both generator and discriminator follow architecture presented in the BigGAN paper

Encoder

The encoder is implemented as RevNet with few fully connected layers on top.

How to use

python train_gan.py --dataset {supported_dataset} --data_path {path/to/dataset/folder}

Supported Datasets

• MNIST 32x32
• FMNIST 32x32
• CIFAR10 32x32
• CIFAR100 32x32
• Imagewoof 64x64
• Imagenette 64x64

Samples

CIFAR10

Imagewoof

Acknowledgments

BigGAN https://github.com/taki0112/BigGAN-Tensorflow

RevNet https://github.com/google/revisiting-self-supervised

Tensorflow implementation of BigBiGAN https://github.com/LEGO999/BigBiGAN-TensorFlow2.0