The latest in pursuit of extreme realism using Flux Dev, this Lora is trained on almost 1000 hand-selected high-resolution photographs of ordinary people, and enables Flux to produce a massive variety of faces of everyday people of all races, ages and genders, without the stereotypical Flux 'look' of highly sylized models.

Usage: Weight 0.5 - 1.5, default is 1.0. Keyword 'rlppl' is helpful.

Prompting: use ages, ethnicities, hair colours, styles and body types (eg 'underweight', 'overweight', 'obese', 'muscular', 'skinny') to push the model into even more variety.

The grid below shows the effect at different weights.