Yavzan/Unet-Stable-Diffusion

Diffusion Model Sampler

An implementation of a diffusion model sampler using a UNet transformer to generate handwritten digit samples.
Explore the docs »

View Demo · Report Bug · Request Feature

Table of Contents

1. About The Project
   • Built With
2. Getting Started
   • Prerequisites
   • Installation
3. Usage
4. Results
5. Roadmap
6. Contributing
7. License
8. Contact
9. Acknowledgments

About The Project

Diffusion models have shown great promise in generating high-quality samples in various domains. In this project, we utilize a UNet transformer-based diffusion model to generate samples of handwritten digits. The process involves:

1. Setting up the model and loading pre-trained weights.
2. Generating samples for each digit.
3. Creating a GIF to visualize the generated samples.

(back to top)

Built With

AI and Machine Learning Libraries

(back to top)

Getting Started

To get a local copy up and running follow these simple example steps.

Prerequisites

Ensure you have the following prerequisites installed:

• Python 3.8 or higher
• CUDA-enabled GPU (optional but recommended)
• The following Python libraries:
  • torch
  • torchvision
  • numpy
  • Pillow
  • matplotlib

Installation

1. Clone the repository:

   git clone https://github.com/Yavuzhan-Baykara/Stable-Diffusion.git
   cd Stable-Diffusion
   

2. Install the required Python libraries:

   pip install torch torchvision numpy Pillow matplotlib
   

(back to top)

Usage

To train the UNet transformer with different datasets and samplers, use the following command:

python train.py <dataset> <sampler> <epoch> <batch_size>