LoCon trained on human characters and backgrounds from the Keroro anime. Works better with the Anything V5, AbyssOrangeMixV3 or Daydream models. This LoCon uses a smaller sample than the previous one, but with higher quality control on samples and tags.
