Still in experimental workflow with SDXL_Niji_Seven's text encoder 1 clips, Flux's t5xxl_fp16 and the Katayama_Mix_XL_v2 model and some LoRa's
