Skip to content
A Deep Convolutional Generative Adversarial Network to generate Faces of Anime Characters
Jupyter Notebook
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
images
README.md
animegan.ipynb

README.md

AnimeGAN

A Deep Convolutional Generative Adversarial Network to generate Faces of Anime Characters

Kaggle Notebook Link: https://www.kaggle.com/soumikrakshit/animegan

Dataset Link: https://www.kaggle.com/soumikrakshit/anime-faces

Dataset Sample

Sample From the Dataset

AnimeGAN Architecture

AnimeGAN is based on the standard DCGAN architecture by Alec Radford, Luke Metz, Soumith Chintala. It mainly comprises of convolution layers without max pooling and fully connected layers. It uses convolutional stride and transposed convolution for the downsampling and the upsampling.

Generator Architecture

Generator Architecture

Discriminator Architecture

Discriminator Architecture

Final Model

Final Model Architecture

Training AnimeGAN

The AnimeGAN was trained on 21551 anime face images from size (64, 64) and was trained for 15000 epochs for a batch size of 32 images using Nvidia Tesla K80 GPU in the Kaggle Kernel environment.

Learning Faces

Training Results

Learning Faces

Learning Faces

Learning Faces

You can’t perform that action at this time.