Anime-GAN: Exploring Generative Adversarial Networks

Overview

Anime-GAN is a pioneering project that leverages the capabilities of Deep Convolutional Generative Adversarial Networks (DCGAN) to generate high-quality anime faces. Inspired by the website "thispersondoesnotexist.com," this project aims to provide a similar experience but with a focus on anime characters. Each refresh on the project's web page presents a new set of 16 anime faces arranged in a 4x4 grid, showcasing the diversity and artistic potential of the generated characters.

Installation

Prerequisites

Anaconda or Miniconda installed on your machine.

Setting Up the Environment

To set up the project environment and run the application, follow these steps:

Create and activate a new Conda environment:

conda create -n anime-gan python=3.10
conda activate anime-gan

Install required packages:
```
pip install -r requirements.txt
```

Running the Application

Start the Flask app: Open a terminal and run the following command:
```
python app.py
```
Expose the app to the internet using Ngrok: Open another terminal and run:
```
ngrok http 5000
```
This will generate a public URL that can be used to access the Flask app from anywhere.

Usage

Navigate to the Ngrok URL or http://localhost:5000 provided by the Ngrok output. The webpage will display 16 anime faces in a 4x4 grid. Each refresh of the page generates a new set of faces, demonstrating the model's ability to create diverse and visually appealing anime characters.

Training the Model

For those interested in training the Anime-GAN model themselves or making modifications, a Jupyter Notebook (training.ipynb) is provided. You can use this notebook to train the model on Kaggle's kernel with the following dataset:

Anime Face Dataset on Kaggle

This allows for customization and experimentation with different model parameters or datasets.

Screenshots

Gradio Integration

Setting Up the Gradio App

Navigate to the Gradio folder:

git clone https://github.com/TanmayDoesAI/anime-gan.git
cd gradio-app

Install the required packages:
```
pip install -r requirements.txt
```
Run the Gradio app:
```
python app.py
```

Accessing the Gradio App

Access the app locally at http://localhost:7860. The Gradio interface allows you to generate and view unique anime characters.

Hugging Face Space

Explore AnimeGAN on Hugging Face Spaces: AnimeGAN on Hugging Face.

MacBook Users

One-Click Setup Script

For MacBook users, you can use the following shell script to simplify the setup and hosting process. Ensure that jq is installed on your machine.

Install jq:
```
brew install jq
```
Make the run_app.sh executable:
```
chmod +x run_app.sh
```
Run the script:
```
./run_app.sh
```

License

This project is licensed under the MIT License - see the LICENSE file for details. Feel free to contribute to Anime-GAN! We welcome contributions from the community to help improve and expand the project. Whether you're interested in fixing bugs, adding new features, or enhancing documentation, your contributions are highly valued.

Enjoy exploring the fascinating intersection of AI and anime art with Anime-GAN!

TanmayDoesAI / anime-gan