Important Announcement

#32
by CaptiveDreamer - opened

Hello everyone.

  • The entire dataset is now inside catalog.db
  • It includes all 12 million image links
  • You can use tool.py on it to query specific users or posts or artworks.
  • the tool.py was added because some people were literally searching their names in the links list and then claimed the dataset was incomplete
  • There are now labels like "software: blender" etc that may be useful for some applications
  • more info in README.md

Can we try to make LORAs for image models that uses the cara dataset to get similar artworks based on artist words? It could be a good community exercise. Please let us know how we can co-ordinate and help!

Sign up or log in to comment