I was recently experimenting again with training in Google Colab with using a small file size. honestly The results are not that bad, althought it seems lora struggles generating glass sometimes and hair can inconsistent when generating og outfit, or without ,not sure why ? all images generated with AutismMix SDXL weight 1 There has had LORA of SD1.5 for this character, I have pinned below

Tulip Olsen -Infinity Train

Tulip_Olsen-10.safetensors

v1.0

ill removed this if the owner asked

Files with SHA256: d78cee5c2b156c8bd39fb7444e7bed422965366e830cbffd3a2d88129d5d208c