Skip to content

mshaikh2/GANs_Comparison

Repository files navigation

MNIST reconstruction using Convnet, Neuralnet and CapsuleNets

Deep Convolutional GAN

The below GIF displays the sample of images generated from epoch 1 to 50 at every 5 epochs.

Conv layers enable GANs to generate better images much faster than neural net.

Each epoch takes around 60 seconds

Images_generated_using_conv_net

Graph of Loss over 50 epochs

Graph1

Deep Neural GAN

The below GIF displays the sample of images generated from epoch 1 to 200 at every 20 epochs.

Neural net enables GANs to generate decent images but after much longer training epochs.

Each epoch takes around 15 seconds.

Images_generated_using_conv_net

Capsule Nets

The below GIF displays the sample of images generated from epoch 1 to 9 at every epoch.

At the decoder end a 28x28 image is reconstructed by passing the latent vector along with its true class variable through two fully connected layers

Each epoch takes around 55 mins seconds.

Images_generated_using_caps_net

Graph of Loss over 9 epochs

Graph3

Libraries

  • Tensorflow
  • Keras
  • openCV
  • PIL
  • numpy

Refrences

About

A basic comparision between conv GANs and neural net GANs

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published