This is a LORA model of ITZY - Yeji v 2.1 , trained with the 1.5 base model, It should work well with any photorealistic models, You'll get best results with a LORA weight 0.7-0.9, Steps: 30, Sampler: DPM++ SDE Karras, CFG scale: 4.
The version 2.1 lora is being fine tune that the last version if there's a problem u can check me i think the most possible improvement to look for more high res data set.