Recommended workflow:

Use model anime_expmixLine_v2 or anime_expmixLine_v3, negative embedding EasyNegative, upscaler 4x-AnimeSharp

Prompt (add your subject):

(lasterk:1.3), (masterpiece,detailed,highres:1.2) <lora:lasterk-v3:1>

or

lasterk, <lora:lasterk-v3:1>

Negative prompt:

easynegative

1.txt2img (find good seed here):

• Sampler: DPM++ SDE Karras

• Steps: 30

• Hires. fix: True

• Hires upscaler: 4x-AnimeSharp

• Hires steps: 20

• Denoising strength: 0.75

• Hires upscale: 1.5

• Size: 512x768

• CFG scale: 3

• Clip skip: 2

2.img2img:

• Use "send to img2img"

• Size: 1152x1728 (Resize by 1.5)

• Denoising strength: 0.4

Alternative workflow (faster):

txt2img:

• Sampler: DPM++ SDE Karras

• Steps: 40

• Hires. fix: True

• Hires upscaler: 4x-AnimeSharp

• Hires steps: 20

• Denoising strength: 0.4

• Hires upscale: 1.5 or 2

• Size: 512x768

• CFG scale: 7

You can also experiment and share your workflow if you would discover something better or interesting.

Also please share your generations and make reviews.

Feel free to ask anything in comments.

I've made new lora for Pony XL: https://civitai.com/models/346719