First attempt at training from manga/doujins (V1).
Best at 0.8 weight. No trigger words. Best for monochrome.
Description
Slight changes to the dataset, trained for more steps. Overall, I think a minor improvement over V1
First attempt at training from manga/doujins (V1).
Best at 0.8 weight. No trigger words. Best for monochrome.
Slight changes to the dataset, trained for more steps. Overall, I think a minor improvement over V1