title: Stable Diffusion WebUI ControlNet
emoji: 🦄
colorFrom: pink
colorTo: yellow
sdk: docker
app_port: 7860
pinned: true
tags:
- stable-diffusion
- stable-diffusion-diffusers
- text-to-image
models:
- stabilityai/stable-diffusion-2-1
- runwayml/stable-diffusion-v1-5
- lllyasviel/ControlNet
- webui/ControlNet-modules-safetensors
- dreamlike-art/dreamlike-diffusion-1.0
- Anashel/rpg
- Lykon/DreamShaper
Stable Diffusion WebUI + ControlNet
Private image builds with both with Stable Diffusion 2.1 models and Stable Diffusion 1.5 models and bundles several popular extensions to AUTOMATIC1111's WebUI, including the ControlNet WebUI extension. ControlNet models primarily works best with the SD 1.5 models at the time of writing.
Shared UI space would usually load with a model based on Stable Diffusion 1.5.
🐳 🦄 Builds a Docker image to be run as a Space at Hugging Face using A10G or T4 hardware.
Setup on Hugging Face
- Duplicate this space to your Hugging Face account or clone this repo to your account.
- Under the "Settings" tab of your space you can choose which hardware for your space, that you will also be billed for.
- The
on_start.sh
file will be run when the container is started, right before the WebUI is initiated. This is where you can install any additional extensions or models you may need. Make sure the env valueIS_SHARED_UI
is set to0
or is unset for your space, or else only the lightweight model installation will run and some features will be disabled.
Relevant links for more information
Repo for this builder
This repo, containing the Dockerfile
, etc. for building the image can originally be found on both 🤗 Hugging Face ➔ carloscar/stable-diffusion-webui-controlnet-docker
and 🐙 GitHub ➔ kalaspuff/stable-diffusion-webui-controlnet-docker
.
Stable Diffusion Web UI
- Source Code: https://github.com/AUTOMATIC1111/stable-diffusion-webui
- Documentation: https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki
WebUI extension for ControlNet
- Source Code: https://github.com/Mikubill/sd-webui-controlnet
ControlNet models
- Trained models: https://github.com/lllyasviel/ControlNet
- Pre-extracted models: https://huggingface.co/webui/ControlNet-modules-safetensors/tree/main
Licenses for using Stable Diffusion models and ControlNet models
- https://huggingface.co/stabilityai/stable-diffusion-2/blob/main/LICENSE-MODEL
- https://huggingface.co/spaces/CompVis/stable-diffusion-license
- https://github.com/lllyasviel/ControlNet/blob/main/LICENSE
Enable additional models (checkpoints, LoRA, VAE, etc.)
Enable the models you want to use on the bottom of the on_start.sh
file. This is also the place to add any additional models you may want to install when starting your space.
## Checkpoint · Example:
download-model --checkpoint "FILENAME" "URL"
## LORA (low-rank adaptation) · Example:
download-model --lora "FILENAME" "URL"
## VAE (variational autoencoder) · Example:
download-model --vae "FILENAME" "URL"
Some examples of additional (optional) models
Some models such as additional checkpoints, VAE, LoRA, etc. may already be present in the on_start.sh
file. You can enable them by removing the #
in front of their respective line or disable them by removing the line or adding a leading #
before download-model
.
- Checkpoint · Dreamlike Diffusion 1.0 (license)
- Checkpoint · Dreamshaper 3.31
- Checkpoint · The Ally's Mix III: Revolutions
- Checkpoint · Deliberate v2
- Checkpoint · dalcefo_painting
- Checkpoint · RPG v4
- Checkpoint · A to Zovya RPG Artist's Tools (1.5 & 2.1)
- LoRA · epi_noiseoffset v2
- VAE · sd-vae-ft-mse-original
- Embedding · bad_prompt_version2
- See https://huggingface.co/models?filter=stable-diffusion and https://civitai.com/ for more.
Visit the individual model pages for more information on the models and their licenses.
Extensions
- GitHub ➔ deforum-art/deforum-for-automatic1111-webui
- GitHub ➔ yfszzx/stable-diffusion-webui-images-browser
- GitHub ➔ Vetchems/sd-civitai-browser
- GitHub ➔ kohya-ss/sd-webui-additional-networks
- GitHub ➔ Mikubill/sd-webui-controlnet
Additional acknowledgements
A lot of inspiration for this Docker build comes from GitHub ➔ camenduru. Amazing things! 🙏