Creating Fake Images using DCGAN

*** Required Python version is 3.10 as PyTorch doesn't support 3.11 yet. ***

Please follow the below steps to download the CIFAR-10 data for training the GAN

Step 1: Open the URL http://www.cs.toronto.edu/~kriz/cifar.html

Step 2: Click on the highlighted text to download the training dataset image

Step 3: Extract the files and place them in the "data" folder image

Step 4: Run the 'GANs.py' file

Fake Images generated by Generator

We have set the Epochs to 35, which can be tweaked to get more meaningful images. Below is the fake image generated at 25th Epoch

fake_samples_epoch_024

Real Sample image

real_samples

Related Works