uoh cunny
https://files.catbox.moe/cj1okn.webp
decided to train a new version of this lora
lora and toml is here
https://mega.nz/folder/47Yj3ZIS#klaoBwVZI_u5DbjmCjkqRQ/folder/Fm4RUbiT
alk didn't train at all i'll probably need to inspect my dataset for that artist
i was thinking of trying out that lora finetune extract but this lora took six hours on my normal settings for 1 epoch with the amount of images and i cant imagine how slow it would be with everything i need to do to make finetuning work on 8gb of vram without crashing immediately