ModelsLab/Amateur Snapshot Photo - [STYLE] [LORA] [FLUX] - V9.0

dataset further increased from 30 images to 45

• improved facial variety

• more flexibility regarding scenes/better prompt following (I think?)

• more coherent images (I think?)

• further slightly improved FLUX face (I think?)

• further slightly improved FLUX skin (I think?)

• higher "image quality" as in less "low-quality" looking

• improved lighting (I would say)

• smaller model size again