For best outputs:

Sampler: DPM++ 2S a Karras

Steps: 30, cfg 7, CLIP skip 2

VAE: not needed, but yes Adetailer with a strenght of 0.35.

Neg prompt: "SFW, children, loli, child, kids, teens, ((Watermarks)), ((website)), ((logo)), ((artist name)), ((web page)), ((letterbox)), ((signature)), ((webpage)), ((text at bottom)), ((text in corner)), ((company name)), ((names)), ((labels)), ((captions)), ((websites)), ((comments))), ((words)), ((text))"

This model tends to create watermarks to websites that don't actually exist, make sure u put that in ur neg.

For the pos: no needed, but this may increase the quality a little -> "high resolution, best quality, 4k, highly detailed, realism, real photo, realistic skin texture, skin imperfections, skin pores, sharp details, bright light, warm colors, 8k, HDR, f.1/2, 70mm lens, Kodak, Analog style"

This model is mainly intended for basic positions having POV as core. On its own cant do much more, but accepts and understands LoRa's quite well. With the proper LoRa it will create the output you need.