---
title: README
emoji: 📉
colorFrom: green
colorTo: purple
sdk: static
pinned: false
---

# Welcome to the 🧨 `diffusers` organization!

`diffusers` is the go-to library for state-of-the-art pretrained diffusion models for multi-modal generative AI.

[GitHub](https://github.com/huggingface/diffusers), [Docs](https://huggingface.co/docs/diffusers/index)

In this organization, you can find some utilities and models we have made for you 🫶

## 🆕 SDXL Inpainting 🎨
We are training Inpainting specialist models for Stable Diffusion XL ✨. They are way better at inpainting and outpainting than the base-model, but are still experimental. Try it out and consider giving feedback! 🤗
| Inpanting model | Demo |
|-----------------------------------|---------------------|
| [`inpainting-sdxl-0.1`](https://huggingface.co/diffusers/inpainting-sdxl-0.1) | [Spaces demo 🕹️](https://huggingface.co/spaces/diffusers/stable-diffusion-xl-inpainting) |

## SDXL ControlNets

We are training ControlNets for Stable Diffusion XL ✨

### Small sized ControlNets 🤏

We are experimenting with training smaller ControlNets that can still retain their control capabilities! This is the [training script used](https://github.com/huggingface/diffusers/blob/7b93c2a882d8e12209fbaeffa51ee2b599ab5349/examples/research_projects/controlnet/train_controlnet_webdataset.py).
Please check the models released, and consider giving us feedback on our [Discord server](https://t.co/1n75wi976V?amp=1):

| ControlNet                                                                                            | Size in fp16 |
| ----------------------------------------------------------------------------------------------------- | ------------ |
| [`controlnet-canny-sdxl-1.0-small`](https://huggingface.co/diffusers/controlnet-canny-sdxl-1.0-small) | 320MB        |
| [`controlnet-canny-sdxl-1.0-mid`](https://huggingface.co/diffusers/controlnet-canny-sdxl-1.0-mid)     | 545MB        |
| [`controlnet-depth-sdxl-1.0-small`](https://huggingface.co/diffusers/controlnet-depth-sdxl-1.0-small) | 320MB        |
| [`controlnet-depth-sdxl-1.0-mid`](https://huggingface.co/diffusers/controlnet-depth-sdxl-1.0-mid)     | 545MB        |

### Full sized ControlNets 💪

The team has also trained full sized ControlNets - as big as the SDXL model (5GB in fp32). We have provided a [ControlNet training script and guide here](https://github.com/huggingface/diffusers/blob/main/examples/controlnet/README_sdxl.md)

| ControlNet                                                                                        |
| ------------------------------------------------------------------------------------------------- |
| [`controlnet-canny-sdxl-1.0`](https://huggingface.co/diffusers/controlnet-canny-sdxl-1.0)         |
| [`controlnet-depth-sdxl-1.0`](https://huggingface.co/diffusers/controlnet-depth-sdxl-1.0)         |
| [`controlnet-zoe-depth-sdxl-1.0`](https://huggingface.co/diffusers/controlnet-zoe-depth-sdxl-1.0) |

### Utilities 🔧

- [Convert Stable Diffusion XL weights to `diffusers` format](https://huggingface.co/spaces/diffusers/sdxl-to-diffusers)
- [Convert Stable Diffusion weights to `diffusers` format](https://huggingface.co/spaces/diffusers/sd-to-diffusers)
- [Convert diffusers models to `safetensors` format](https://huggingface.co/spaces/diffusers/convert)

More coming soon ✨