trained with 288 images (including the 94 images used in v1) in 1024x1024, based on SD1.5, 19200 steps,
trained with 288 images (including the 94 images used in v1) in 1024x1024, based on SD1.5, 19200 steps,