A simple PyTorch Implementation of Generative Adversarial Networks, focusing on anime face drawing. This project use 3 type of GAN to experiment including DCGAN, WGAN_GP and DRAGAN.
Dataset I use is from this paper Towards the Automatic Anime Characters Creation with Generative Adversarial Networks
This dataset is collected by the authors which content 42000 face image of anime character.
You can download the training data from: https://drive.google.com/open?id=1bXXeEzARYWsvUwbW3SA0meulCR3nIhDb
| Name | Paper Link | Value Function |
|---|---|---|
| DCGAN | Arxiv |  |
| WGAN_GP | Arxiv |  |
| DRAGAN | Arxiv |  |
All results are generated from the fixed noise vector.
| Name | Epoch 1 | Epoch 50 | Epoch 100 | GIF |
|---|---|---|---|---|
| DCGAN |  |
 |
 |
 |
| WGAN_GP |  |
 |
 |
 |
| DRAGAN |  |
 |
 |
 |
Google Colab
To reproduce the result of this project you download dataset from above link and place it in 'dataset/raw'.
Then for each of the model, run the bash file in 'script' folder.
Or, you can download the pretrained models from this link https://drive.google.com/drive/folders/1-3cKKUlq_jTdUUIhpHO4fzKBuGKxi03s?usp=sharing
You can try to generate some faces using this app
https://anime-face-app.herokuapp.com/