V3

New merge. Similar vibe and texture, but better lighting, hair, and my preferred facial characteristics. Sample gallery images were not facefixed or processed in any way, I uploaded straight out of my Comfyui folder to Civitai.

To get results similar to the gallery sample images, use Hassaku 1.3a, the SssslipperyP0s and SmoothN3g embeddings, along with the following chunks in your prompt:

Positive prompt:

realistic hair, detailed hair, individual hair strands,

thick eyelashes, gorgeous face, cute face, aegyo sal, lips, nose, perfect eyes,

Negative prompt:

anime, flat color, anime coloring, cel rendering,

Uploaded two different files, sv4_64dim_fp16 is my recommendation for the best balance between file size and quality for local generation.

V2

Use Hassaku 1.3a and the following tags in your positive prompt:

lineart, thick lineart, painterly, painting /(medium/), semi realistic, cartoon, western comics,

V2 was my first merge ever, essentially re-doing V1 with some slight adjustments. V1 is strictly better, in my estimation, but I'll leave V2 up here for posterity.

V1

Use Hassaku 1.3a and the following tags in your positive prompt:

lineart, thick lineart, painterly, painting /(medium/), semi realistic, cartoon, western comics,

V1 was very kindly created for me by n_Arno, it's a merge of one of their LoRAs and another LoRA that I use frequently. All credit to the original creators of both LoRAs.

Best used with this embedding n_Arno also created for me.

Details in this post.