metadata

language:
  - en
tags:
  - cyberpunk
  - anime
  - stable-diffusion
  - 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.
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.

HYPERNETWORKS

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