You know, during this entire ordeal I've been going through I think I learned a lot something, it is possible for LORAs to have TOO MANY images in a dataset...crazy I know, it can cause the thing to barely function like the ai got overloaded or something, might be a good idea to slow down the training process if I try making a lora with 1000 images; anyway-

Ko-fi: https://ko-fi.com/keyennorman

The models I make are tested for resolutions under 512 (Not by choice), results may be different (Probably better) with higher resolutions. Ref images are in the lower quality as well.

Unirequisets
1. 1girl, filo, cleavage, pink hair, thigh highs, sword, cape, gloves, blue skirt, short hair, blue eyes,

2. 0.6-0.7

3. Trained with 6 images, do not expect too much flexibility

On to the last one and then I can finally sleep, where the buzz at? I wanna be like a bee