Skip to content

ryanrussell/HomeDiffusion

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Home Diffusion

user/pass: demo / demo

The demo relies on community resources and may be overloaded at times. Please DO NOT use this for automated or production workloads; it is subject to change.

Disrupting Home Design with Stable Diffusion and ControlNet

Home Diffusion is an open source project that uses Stable Diffusion and ControlNet to revolutionize the way home design is done. With Home Diffusion, you can run your own design studio and craft beautiful, unique designs with ease.

Examples

Tropical Home Remodel Side A Before Tropical Home Remodel Side A After

Tropical Home Remodel Side B Before Tropical Home Remodel Side BAfter

Tropical Outside Courtyard Pool Remodel Before Tropical Outside Courtyard Pool Remodel After

Get Started with Home Diffusion

Home Diffusion is easy to use and requires only a GPU to run. Here's what you need to get started:

  • A GPU with suggested minimum 12GB of VRAM
  • An internet connection
  • Home Diffusion

Once you have the necessary hardware and software, you can start designing your dream home with Home Diffusion.

What is Home Diffusion?

Home Diffusion is a revolutionary open source project that utilizes Stable Diffusion and ControlNet to create highly accurate and customizable home designs. Home Diffusion is a powerful tool that allows users to design their own homes with unprecedented accuracy and precision.

Users can customize every aspect of their home design, from the overall layout and structure to the smallest details. Home Diffusion also provides users with access to a library of pre-made designs and templates, so they can quickly find the perfect home design for their needs.

How Does It Work?

Home Diffusion is an innovative open source project that enables users to create design concepts for their homes. Through a fine-tuned implementation of Stable Diffusion and ControlNet, Home Diffusion provides users with an intuitive and powerful tool to create dream homes that fit their exact specifications.

Run locally

You can run your own design studio with Home Diffusion if you have your own GPU.

Either way, the first step is to download the file and place it in ./models.

Step 1 Clone repo and download model

git clone git@github.com:HomeDiffusion/HomeDiffusion.git
cd HomeDiffusion
# Download the model, this is ~5.4 GB, so grab a coffee. Thanks to lllyasviel for ControlNet!
wget https://huggingface.co/lllyasviel/ControlNet/resolve/main/models/control_sd15_mlsd.pth  -P models/

Launch with conda

conda env create -f environment.yml
conda activate homeDiffusion
# Get another coffee. This be slow the first time, but faster in subsequent uses. Varies based on your internet and machine

Launch with docker

docker-compose up --build
# Get another coffee. This be slow the first time, but faster in subsequent uses. Varies based on your internet and machine

Confirm Home Diffusion is running

Once the build is done, with either conda or docker-compose, you should see a message like this at the end of the output:

homediffusion    | Loaded model config from [./models/cldm_v15.yaml]
homediffusion    | Loaded state_dict from [./models/control_sd15_mlsd.pth]
homediffusion    | Running on local URL:  http://0.0.0.0:7860
homediffusion    | 
homediffusion    | To create a public link, set `share=True` in `launch()`.

Open a browser and load http://0.0.0.0:7860. Your local instance of Home Diffusion should be running and look like this: Home Diffusion Up and Running

Advanced Controls

After running thousands of iterations locally, we have set some sane defaults for you. An advanced user guide is coming soon!

Enjoy! Please give us a star if you like Home Diffusion :)

Acknowledgements

Thank you to many open source contributors who have made this possible, but especially:

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%