adupatil/Image_Generator

The Image_Generator is created using the concept of Deep Convolutional GANs. The Generator is created using Transpose Convolution layers and it is the one crearting the images. The second part is the Discriminator created using Convolutional layers which discriminates the images as real or fake and feeds the output to the generator to make it better every epoch

Image_Generator

The Image_Generator is created using the concept of Deep Convolutional GANs. The Generator is created using Transpose Convolution layers and it is the one creating the images. The second part is the Discriminator created using Convolutional layers which discriminates the images as real or fake and feeds the output to the generator to make it better every epoch

Input

The generator and discriminator have been trained on CIFAR-10 dataset

Output

The generator tries to recreate the images and these images are available in the folder named results.

Real Image:

Fake Image epoch 5:

Fake Image epoch 10:

Fake Image final epoch: