JAV yuna_shiina

Imitation model. Modelled after retired Japanese AV star Yuna Shiina (椎名ゆな).

She retired back around 2015 (from year 2009 since her first debut), this actress was quite popular back then. She recently resurfaced earlier last year after a long break. And just nice, now she probably transition into slightly more mature-themed milf JAV. Also classified within the "Bijin" (beauty class) category, but her refining feature is always her rather high cheekbone.

~~~~~~~~~~~~~~~~~~~~~~~~~~
Recommended LoRA strength: 0.8 (0.6-1.0 are plausible range dependent on checkpoints; higher values >1 lead to sharp facial features, but can bring out unintended masculinity or even distort faces in certain checkpoint)

Main Prompt:

yunashiina_jav, breasts, long hair, brown hair

variations: tied hair, hair covering one eye

facial variations: natural makeup, mature makeup, parted lips, lipstick, eyeshadow, high cheekbone

*Use of facial prompts depend on checkpoints, use it if your checkpoint have a bias towards minimal makeup, OR dont use it if your checkpoint over-react with "mature makeup" tags

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Checkpoints that work well for this LoRA:

1) Realisian. Best when it comes to resemblance (see for youself). ALMOST perfect.

2) RealisticMix. 2nd best, no doubt.

3) ClearRealismMix-Eur, or asian-based/eurasian checkpoints. Not perfect, but still better than pure-European checkpoints.

10) PerfectWorld, UltimateHentai Realistic or other photoRealistic Hentai models (European model works well for this LoRA). This time, PerfectWorld does very well, again.

2.5D Hentai checkpoints dont seem to perform well with this LoRA, very often you get distorted faces - perhaps can try reducing the weight, but it loses some aspect of resemblance : /

If you find one that works well, let me know!

~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Note for achieving good resemblance:

This LoRA actually require very minimal tweaking. Somehow, her facial features and body tyle is unique enough for SD to pick this up fairly quickly in the training session.

Use Adetailer for baseline generation (at 512, 768 etc) only, try to control LoRA strength at 0.8, adjust your denoise accordingly. For this LoRA, recommended to use Adetailer at ~0.4 denoise strength for close-up, and 0.45 to 0.50 for full body.

For this LoRA, you can do without Adetailer prompt, (just use the same prompt). But if you wish, you can put the makeup prompts there.

The trick is to pay attention to:
1) her defined cheekbone

2) her middle parting (hair)

3) her rather serious glance (this is why 'mature' tags help)

For Realisian, chilloutM, MagicMix etc, of the pretty asian checkpoints, you can try to use "mature woman" in the prompt to get some veyr minor wrinkles, and age her accordingly. But do not use those prompt for 2.5D Hentai checkpoints, you will get her to look like a ladyboy....

Recommended approach: is to not use HiRes, but upscale it with img2img instead.

Use a denoise strength of 0.25 -0.3 (anything higher will overcorrect face during the process), and turn off Adetailer.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Issues:

One of the easiest plug-and-play LoRA. No issues so far, apart from some checkpoint simply dont work as well.

in img2img upscaling, sometimes the shape of her eyes can be alittle off - I guess you can try running Adetailer again with very low denoise to try and fix that.

Post-production edits/Reproducibility:

Actually none. No inpainting was needed. Everything fresh from oven this time.

Happy prompting. And do share [SFW] creations~