Paella

lastmod 2023-04-19created 2023-04-191 edits

A simple and efficient Text-to-Image generative model

Features

• Aims to build an image synthesis model anyone can understand
  • The sampling code is only 12 lines, so anyone can understand it in a few minutes
• Operates in a Quantized Latent Space like StableDiffusion to reduce compute

See also

• Text-to-Image
• StableDiffusion

Favorite site

• A new Paella: Simple & Efficient Text-To-Image generation | LAION