For best outputs:

Version Pro:

Sampler: DPM++ 2M Karras.

Steps: 35, cfg 7, CLIP skip 2.

VAE: vae-ft-mse-840000-ema-pruned_fp16.safetensors (optional for a little bit more of image's detail).

Pos prompt: add this at the end of your pos: "realistic, real photo, realistic skin texture, skin imperfections, skin pores, sharp details, bright light, 8k, HDR, f.1/2, 70mm lens, Kodak, Analog style, epiCPhoto, epiC35mm, ", being the LoRa "epiC35mm" (which trigger is "epiC35mm" as well) and the "epiCPhoto" embedding both included in my vPro example runs.

Neg prompt: "(cleavage:1.5), (NSFW:1.5), (deformed, distorted, disfigured:1.3), poorly drawn, bad anatomy, wrong anatomy, extra limb, missing limb, floating limbs, (mutated hands and fingers:1.4), disconnected limbs, mutation, mutated, ugly, disgusting, blurry, amputation, (piercing:1.3) (hands:1.35), epiCNegative, realisticvision-negative-embedding", being the embeddings "epiCNegative" and "realisticvision-negative-embedding" both included in my vPro example runs.

Version 3.0:

Sampler: DPM++ 2M SDE Heun Karras.

Steps: 35, cfg 4.5, CLIP skip 2.

VAE: clearvae_v2.3_fp16.safetensors

Pos prompt: use the LoRa "FaceBeauty_qinglong_V3" with a weight of 1 ~ 1.5 for small faces (included in my example runs for both versions).

Neg prompt: "(deformed, distorted, disfigured:1.3), poorly drawn, bad anatomy, wrong anatomy, extra limb, missing limb, floating limbs, (mutated hands and fingers:1.4), disconnected limbs, mutation, mutated, ugly, disgusting, blurry, amputation".

What changes from v3.0 to vPro: input_image feature, overall quality and realism increased.

Checkpoint Merge.