>>55204347Thanks, I love the first one, even if it's the BW1 Cheren its peak reaction image fodder.
Brings me back to when I used Stable Diffusion to make random pictures of characters I liked for fun, but I stopped at one point for some reason.
I think it was because I never could get SDXL LoRAs to get completely like the character due to me only having a palty 8GB of VRAM (which meant training the LoRA was basically something that would take all day) and that I was unable to go back to 1.5 after trying out XL.
...maybe this is proof I should get back to it.
