rebecca

REBECCA

Shameless plug: If you want a free open source fully local booru alternative, I have created one. It automatically tags your images and let's you easily filter by generation parameters. The wiki has some pictures of how it looks like.

DATASET

• 71a: 71 images of Rebecca, with dimensions of 640x640.

VERSION

• v1: Trained with clip skip 1 (off) (only applicable for embedding/hypernetwork, as I don't know the status for dreambooth). NAI-analog base model.

DREAMBOOTH

Trained via this GUI, but note that it's quite restrictive in training options.

• rebecca-71a-v1: Trained with "very high quality" preset and "normal" learning rate.
  • Examples

EMBEDDINGS

Note: You'll have to rename your chosen embedding to "rebecca"

• rebecca-71a-v1a-embeddings: Embedding trained with voldy's web-ui with 8 tokens per vector, using a learning rate of 0.02:200, 0.01:1000, 0.005:2000, 0.002:3000, 0.0005:4000, 0.00005, for 20k steps.
  • Examples with the 20k embed
• rebecca-71a-v1b-embeddings: Embedding trained with voldy's web-ui with 8 tokens per vector, using a learning rate of 0.005, for 30k steps.
  • Examples with the 5k embed
  • Examples with the 10k embed
  • Examples with the 15k embed
  • Examples with the 20k embed
  • Examples with the 30k embed
• rebecca-71a-v1c-embeddings: Embedding trained with voldy's web-ui with 12 tokens per vector, using a learning rate of 0.005, for 30k steps.
  • Examples with the 5k embed
  • Examples with the 10k embed
  • Examples with the 15k embed
  • Examples with the 20k embed
  • Examples with the 30k embed

HYPERNETWORKS

Apparently voldy's hypernetwork training is broken at the moment. Will either rollback, or wait for it to get fixed.