Anima

Trained at 768x768

Best as a 1:1 ratio. (Examples are 1024x1024)

Weight best at 1.0

CFG best at 1 - 1.5 (using 8-step Flash speedup)

I used https://civitai.red/models/2632277/anima-v1-turbo-aio as my AIO base model.

Illustrious

Trained at 1024x1024 or 1:1 ratio.

Weight best at 1.0

CFG best at 3.5 - 5

SD15

Trained at 512x512, 768x768 and 1024x1024. Just about any reasonable resolution should work fine.

Weights are best at .9 - 1.0.

Some character LoRA weights need to be reduced.

As with most pixel based and lo-fi LoRAs - easy on the negatives.