README.md

Generative Adversarial Networks

1. GANs

Architecture (Tensorflow)

Play with shallow gan the min-max game

tensorboard --logdir=summary_directory/

localhost:6006

2. DCGANs

Architecture (PyTorch)

Discriminator (
(sigmoid): Sigmoid ()
(batch_norm_2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True)
(leaky_relu): LeakyReLU (0.2)
(reduce): Linear (512 -> 1)
(d_layer_1): Conv2d(3, 128, kernel_size=(5, 5), stride=(2, 2))
(d_layer_2): Conv2d(128, 256, kernel_size=(5, 5), stride=(2, 2))
(d_layer_3): Conv2d(256, 512, kernel_size=(5, 5), stride=(2, 2))
)
Generator (
(project): Linear (100 -> 4608)
(batch_norm_2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True)
(relu): ReLU ()
(tanh): Tanh ()
(g_layer_1): ConvTranspose2d(512, 256, kernel_size=(3, 3), stride=(2, 2))
(g_layer_2): ConvTranspose2d(256, 128, kernel_size=(3, 3), stride=(2, 2))
(g_layer_3): ConvTranspose2d(128, 3, kernel_size=(4, 4), stride=(2, 2))
)

epoch 1

epoch 2

epoch 3

epoch 4

epoch 5

epoch 6

epoch 7

epoch 8

epoch 9

epoch 10

epoch 11

epoch 12

epoch 13

epoch 14

epoch 15

3. Conditional GANs

4. InfoGANs

5. Wasserstein GANs