Skip to content
Search

Edit Anything

lastmod 2023-04-18created 2023-04-182 edits

Uses Meta's Segment-Anything Model to recognize every part of an image, and edits and regenerates each one

Features

  • Uses Meta's Segment-Anything Model to recognize every part of an image, and edits and regenerates each one
  • Generates images with fine control through a ControlNet trained on SAM masks
  • Image control through text prompts using BLIP2
  • Segment Anything + ControlNET + BLIP2 + Stable Diffusion

See also

Favorite site

author Zer0 sitemap