For previous version
V1 trained with 6400*3+6400*2 steps, network rank is 128,no flipped img in the dataset.
V2(my preference for now) trained with 6000*2+6000*2 steps, network rank is 128, flipped img in the dataset.
V2-lite is a lite version of v2,6000*2 steps. how lite actually？Just lose 93.75% of the "weight".
New version : v3n
Set network alpha much larger, which largely effect learning rate.
Outputs result better and more stable , which could means less stylized.
Also perform well with simple prompts (last 1 post for example)
Without trigger words the outputs result slightly different, maybe worthwhile to try.
Perhaps you noticed that the negative prompt I used in the posts seems needless,
well,that's because it's designed for AOM3(nsfw yes)
By the way,you can easily call out a big cat to accompany her.