Neon G A N

Introduction

NeonGAN is a model designed for unconditional image generation using the CycleGAN architecture. It transforms images into a futuristic neon style. The model is particularly useful for style-transfer tasks and operates using PyTorch.

Architecture

NeonGAN is built upon the CycleGAN architecture, which is known for its capability in style transfer applications. The model takes standard images, including those of people and scenery, and converts them into neon-styled images, utilizing a dataset of high-contrast neon images.

Training

The model was trained using a dataset comprising 256x256 high-contrast neon images for style and regular images as base images. This dataset can be accessed on Kaggle under "Futuristic Images."

Guide: Running Locally

1. Installation: Clone the NeonGAN repository from GitHub: NeonGAN GitHub.
2. Dependencies: Ensure you have Python and PyTorch installed. Install additional dependencies as listed in the repository's requirements.txt.
3. Dataset: Download the training dataset from Kaggle.
4. Run Model: Follow the instructions on GitHub for running the model locally.
5. Cloud GPUs: For improved performance, consider using cloud GPU services such as AWS, Google Cloud, or Azure.

License

NeonGAN is released under the MIT License, allowing for flexible reuse and modification of the code.