ModelsLab/Freedive Girls, Underwater Style - Freediver FLUX V02

This LoRA is meant to expand the range of what the typical FLUX checkpoint can do to generate underwater images. Typically they have a problem doing anything more than a snorkeling selfie, and are nearly incapable of generating freediving gear, masks, fins, etc..For the gear I focused on sticking to specific wording, you should use "color + dive mask", "color + diving fins", specific items are "oval dive mask", "weight belt", "freediving fins", "cressif1 dive mask", "cressiGara diving fins".You should be able to prompt water conditions, like deep, shallow, good visibility, poor visibility, etc...This LoRA was trained on a 1000 image, handpicked dataset; the images have been edited to remove things like logos or texts on clothing. All images have been prompted with a combination of AI and manual editing to correct errors or omissions.