My hero! Have to test, but this is one of MANY classic ladies I'd love to see brought back, in more ways than one. Pamela Peaks, Domonique Simone, Sakura Sena, Jasmin Aloha, Mimi Miyagi, Wendy Whoppers just to name a few.

Your LoRAs are truly some of the BEST i've ever seen, but if it's not too much trouble, could you give some more detail on HOW you make such amazing / accurate loras? I know you said you used 102 training images, but what KIND of images (portrait, landscape, standing, sitting, nude, clothed, etc.)??? Also, how detailed are the "captions" you put for each image? (Do you just include the trigger word itself for each image, or do you spell out each individual image's details?)

I know the LoRA itself reveals many "settings" (e.g. "fp16" vs "bf16" / etc.), but any details you can give on your techniques would be INCREDIBLE.

Seriously, you're at a whole different level of competency with these LoRAs! Keep up the amazing work!!!

Another question (if you're reading these), but are you meticulously adding "captions" for each photo in your dataset, or do you ignore descriptive captions when creating these??? Any help you can give would be AWESOME!!!