V2, better captioned and add some more images
My first LORA made using Kohya, better results using squares images
Trained with AnyV4
