My usual prompt:

Positive:
instagram photo of a thin 18 y.o. indonesian woman with (input whatever you want after)

Negative:
busty, cartoon, painting, illustration, (worst quality, low quality, normal quality:2)

If your face comes out badly, use after detailer instead.

My usual Settings:

CFG: 5

Clip Skip: 1 or 2

Hi-res steps: 11 - 15

Denoise Strenght: 0.35 ~ 0.5

Sampling Method: DPM++ 2M Karras or 2M SDE KARRAS
Sampling Steps: 20 - 40

Another Tips:

My workflow usually I generate with euler a, and then send it to img2img and rezize it by two and change the sampler method to DPM++ 2M Karras

NOTES:

Please kindly post your result on my model page, that's all i ask


MY WORKFLOW:

1. Use your Positive and negative prompt

2. if the image not have a good face, i use inpaint with denoise strengh 0.35 ~ 0.5

3. Send the generated image to img2img and resize it by 2.
   
   Notes:

   1. if you got an error on stable diffusion please add disable safe unpickle line on the webui-user.bat
   
   set COMMANDLINE_ARGS= --autolaunch --update-check --xformers --disable-safe-unpickle
   

   2. For a better face just use inpaint with denoising strengh from 0.35 - 0.5