Collection of GAN models and implementations
Basic GAN implementation using MNIST
Using a relatively simple GAN model.
MNIST data can be found here or in this case from torchvision
GAN model trained on 200 epochs.
1 epoch | 50 epochs | 100 epochs | 150 epochs | 200 epochs |
---|---|---|---|---|
![]() |
![]() |
![]() |
![]() |
![]() |
Deep-Convolutional GAN implementation using CelebA
Using a Deep-Convolutional GAN model.
CelebA data can be found here
GAN model trained on 5 epochs.
Result after 5 epochs
- WGAN
- Semi-supervised GAN
- CGAN
- InfoGAN
- Cycle-GAN