VoltaML - Fast Stable Diffusion

Stable Diffusion WebUI and API accelerated by AITemplate

Documentation · Report Bug · Request Feature

---

Made with ❤️ by Stax124

---

Table of Contents

• About the Project
  • Screenshots
  • Tech Stack
  • Main features
  • Speed comparison
  • Installation
• Contributing
  • Code of Conduct
• License
• Contact

About the Project

Screenshots

Tech Stack

Client

• Typescript
• Vue.js
• NaiveUI
• Ionicons

API

• Python
• FastAPI
• PyTorch
• AITemplate
• xFormers
• WebSockets

Discord Bot

• Discord.py

DevOps

• Docker
• GitHub Actions
• GitHub Pages
• VitePress

Main features

• Easy install with Docker
• Clean and simple Web UI
• Supports PyTorch as well as AITemplat for inference
• Support for Windows and Linux
• xFormers supported out of the box
• GPU cluster load balancing
• Discord bot
• Documented API
• Clean source code that should be easy to understand

Speed comparison

The below benchmarks have been done for generating a 512x512 image, batch size 1 for 50 iterations.

GPU (it/s)	T4	A10	A100	4090	4080	3090	2080Ti	3050
PyTorch	4.3	8.8	15.1	19	15.5	11	8	4.1
xFormers	5.5	15.6	27.5	28	20.2	15.7	N/A	5.1
AITemplate	N/A	23	N/A	N/A	40.5	N/A	N/A	10.2

Installation

Please see the documentation for installation instructions.

Contributing

Contributions are always welcome!

See contributing.md for ways to get started.

Code of Conduct

Please read the Code of Conduct

License

Distributed under the GPL v3. See License for more information.

Contact

Feel free to contact us on discord: https://discord.gg/pY5SVyHmWm

Project Link: https://github.com/VoltaML/voltaML-fast-stable-diffusion