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.

GTA Art Style

NoobAI v-prediction version 1.0, LoCon 48/48 and 48/48 models, trained on Noob are very weak and hard to pop with Illustrious-based checkpoints, so dims have to be high

Files with SHA256: 43720c959ad47a9cb9cef553cd0dcc6744939d78bcd8fddea3be9af9a6e9e040