https://github.com/mcaimi/sd-webapp
Example Stable Diffusion Web Application based on FastAPI, pyTORCH and Gradio
https://github.com/mcaimi/sd-webapp
Last synced: 3 months ago
JSON representation
Example Stable Diffusion Web Application based on FastAPI, pyTORCH and Gradio
- Host: GitHub
- URL: https://github.com/mcaimi/sd-webapp
- Owner: mcaimi
- License: gpl-3.0
- Created: 2024-09-04T14:59:08.000Z (almost 2 years ago)
- Default Branch: main
- Last Pushed: 2024-11-27T16:08:49.000Z (over 1 year ago)
- Last Synced: 2025-01-16T21:15:55.298Z (over 1 year ago)
- Language: Python
- Size: 2.17 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Stable Diffusion Web App
This is a Stable-Diffusion Web Application written using Streamlit, Pytorch/Torchvision and HuggingFace Diffusers
With proper hardware, it can generate images locally from pre-trained Stable Diffusion Models.
To download model checkpoints, have a look at [HuggingFace](https://huggingface.co/models?pipeline_tag=text-to-image&library=safetensors&sort=trending).
## Run Locally
The application expects to find model checkpoints files in the `models/stable-diffusion` folder, and loras in `models/loras` folder.
```
$ uv run streamlit run main.py
```
It tries to use any discovered GPU that is supported by pytorch, but it can run on CPU (*very* slowly)
Tested:
- Apple Metal Performance Shaders
- NVIDIA CUDA
- AMD CPU
## Current supported features
- SD1.5 Inference with optional LoRA support
- SDXL Inference with optional LoRA support
- SD1.5 Model Merging Capabilities
- Support for custom VAEs in both SDXL and SD1.5 inference phases
- SD1.5 and SDXL Inpainting