>>84154451it barely works. it doesn't want to learn details like correct clothing or hair ornaments
Maybe it will work better if i only train clip and proj_out/linear2 like this guy
https://www.reddit.com/r/StableDiffusion/comments/1f523bd/good_flux_loras_can_be_less_than_45mb_128_dim/