Second try in training this LoRA, with bigger and more diverse set of data
First try in cooking LoRA for OC character
