This is an improved version of my previous Harukawa style. This was made using more accurate descriptions of sample images.
Harukawa_V_2.0.safetensors