Based on the same 22 image dataset as the SDXL version. It is more versatile but reflects less of the style of BG.
Based on the same 22 image dataset as the SDXL version. It is more versatile but reflects less of the style of BG.