DCGAN implementation using Pytorch

Paper link

Implementation of Deep Convolutional GAN (DCGAN) from scratch using Pytorch.

Libraries required

• pytorch
• matplotlib
• tqdm

Dataset

MNIST (dataset will be automatically dowloaded when using this repository)

Training

To train DCGAN model just run

For CPU

python train.py --epochs=10 --batch_size=32 --device="cpu"

For GPU

python train.py --epochs=10 --batch_size=64 --device="cuda"

Output Images during Training DCGAN

When step = 500

When step = 1000

When step = 2000

When steps = 23000