The model has two different subsets. 1) trigger "pkgta" - trained on art from GTA V and a few GTA IV images (official artwork)2) trigger "pkgta6" - trained on the very few GTA VI official artwork images because of this, some facial features and/or the hairstyle might be repetitive, but the shading, texturing and colours are also quite faithful to the dataset.

Files with SHA256: 664f0489e6c9aebb737cfe5c021a82697e26215276d7ca9ac86f20918fd09762