peacasso 0.0.3a0

Create with AI models

Peacasso

Peacasso is an experimental UI interface that makes it easy to generate interesting images with multimodal (text, image) models (stable diffusion).

Installation

Acccess to the model weights needs a HuggingFace model account and access token. Please create an account at huggingface.co, get an access token and agree to the model terms here. Finally, create a HF_API_TOKEN environment variable containing your token.

pip install peacasso

once installed, run the ui with the following command:

peacasso ui   # or peacasso ui --port=8080

Note that you will need to run this on a machine that meets the requirements for the stable diffusion model (cuda).

Implementation Plan

• Command line interface
• UI interface with controls for changing the model parameters
• Enable curation of galleries
• Image inpainting
• Latent space visualization

Acknowledgement

This work builds on the stable diffusion model and code is adapted from the HuggingFace implementation.