About this version

Apparently there is still room for improvement with SDXL models :)

Based on Fv5, merged with UnstableIllusion and fameGridSDXL and some personal Loras trained on my own dataset, this Fv6 version brings better realism overall with more details, better skin texture and better contrasts (see Fv5-Fv6 comparison images).

Credit:

https://civitai.com/models/463163/the-araminta-experiment-sdxlflux