/stable-diffusion-colab-ui

Stable Diffusion UI using google colab backend

Primary LanguageShell

This project is inspired by Easy Diffusion

Stable Diffusion Colab UI

This Colab notebook allows you to use the Stable Diffusion software to create visually appealing images using pretrained models. Stable Diffusion is a generative model that can produce high-quality images from text descriptions.

Original notebook: StableDiffusionUI

Instructions

  1. Open the provided Colab notebook link: StableDiffusionUI

  2. Run the notebook cells to set up the environment and download necessary files.

Or Create from scratch:

  1. The following commands are used to download the model checkpoints:

    wget -O ./easy-diffusion/models/stable-diffusion/anythingv45.ckpt https://huggingface.co/andite/anything-v4.0/resolve/main/anything-v4.5.ckpt

  2. Install the required pyngrok package by running:

    !pip install pyngrok

  3. Import the necessary modules and set up an authentication token for ngrok tunneling:

    from pyngrok import ngrok
token = str(input("Please insert your auth token here: "))
ngrok.set_auth_token(token)

  4. Connect ngrok to a port (e.g., port 8888) and print the tunnel details:

    ngrok_tunnel = ngrok.connect(8888)
print(ngrok_tunnel)

  5. Run the Stable Diffusion user interface using the following command:

    SD_UI_BIND_PORT=8888 ./easy-diffusion/start.sh

  6. Access the Stable Diffusion user interface by clicking on the ngrok tunnel link displayed in the output. This will open a web interface where you can provide text descriptions and generate images using the pretrained models.

  7. Experiment with different text descriptions to see how the model interprets and generates corresponding images.

Remember to terminate the ngrok tunnel and stop the Stable Diffusion user interface when you are done.

For more details, refer to the original Colab notebook: StableDiffusionUI