For best outputs:

Sampler: DPM++ 2M Karras.

Steps: 50, cfg 7.5, CLIP skip 2.

VAE: clearvae_v2.3_fp16.safetensors

Pos prompt: add the LoRa "more_details" with a weight of 0.7 (found in my example runs). Aside of this, as many realism keywords as you want. For example, wrap your text between "(Highest quality:1.3), cinematic shot, masterpiece, (sharp focus:1.5), (photorealistic:1.3), (full body view), (solo:1.5)" and "highly detailed skin), (detailed face), detailed background, volumetric lighting, intricate details, UHD"

Neg prompt: keywords that enhance the realism, for example, "logo, cartoon, painting, illustration, (worst quality, low quality, normal quality:2), logo, chinese, cartoon, 3d, (disfigured), (bad art), (deformed), (poorly drawn), (extra limbs), strange colours, blurry, boring, sketch, lacklustre, repetitive, cropped".

Note: the model is able to create other art styles, like vector illustrations, ect, but it's strong feature is the realism.

What changes from v5.0 to v5.5: input_image feature.

Checkpoint Merge.