LightsushiReal

This a merge between Darksushi and Maijikmix realistic.

you can make anime to realistic portraits and much more

Original Models credits goes to https://civitai.com/models/24779/dark-sushi-mix-mix

https://civitai.com/models/43331?modelVersionId=176425

you can check images for generation prompts.

Recommended settings

Sampling steps: 15-30 (Only use more steps as a last resort for very complex images, as an excessive number of steps tends to ruin the image)

CFG Scale: 5-12

Optional settings

Hires. Fix

Upscaler: 4x-UltraSharp

Hires Steps: 15-20

Denoising Strength: 0,20-0,40 (The more force, the cleaner the image, but the more the original quality is lost.)

Upscale By: 1,5 - 2,0

Must use adetailer for better faces

Adetailer:

Adetailer Model: face_yolov8n.pt

Inpaint denoising strength: 0,25 - 0,40

ADetailer steps: 15 - 20

it works with most of the portraits & dress Lora's. just need to tweak Lora's settings and you're good to go.

some notes about myself:

I am new to stable diffusion and Civitai so I don't know much about it

I am just using my mid-end PC to generate images. which has an AMD card and it did not support stable diffusion so I had to make it with CPU only. which took 45 min to 1 hour for a single image. I want to create my own checkpoint in the future. So I need support from the community. it's all on you if you want to help. it will be a huge support.

thank you guys.

and enjoy.

support here if you want: https://ko-fi.com/showywharf