My first ever lora, improving Z-Image's understanding of this particular style of leggings. Works from the front or back.
Trained with natural-language captions, so no trigger word. Just mention in your prompt that your subject is wearing v-cut leggings.
Recommended weight: 0.5-2 depending on the style. Side-by-side comparison images are there in the samples.
Trained on Base, so should play nice with Turbo finetunes.
Description
Trained on 31 curated images, mostly photorealistic with some anime-style. Captioned using natural language (~60% by hand, 40% by JoyCaption Beta One).