17
1

Improvements to SDXL in NovelAI Diffusion V3

Abstract

In this technical report, we document the changes we made to SDXL in the process of training NovelAI Diffusion V3, our state of the art anime image generation model.

View on arXiv
Comments on this paper