Chiharu Yamada - Bounty Request版本Chiharu Red (ID: 301970)

This one was made for a bounty right here on CivitAI.

Versions (Important!)

• Chiharu - Very accurate in reproducing the given character's face, but may need reinforcement for hair color occasionally. If you're going for consistency of look, this is the one you want.

• Chiharuyamada - The final version of the Embed. It's the most flexible/versatile version, but its occasionally a little less accurate than the plain Chiharu variant.

• Chiharu_red - An slightly overtrained version of Chiharu. Why include it? Because not only is it incredibly faithful in reproducing the character, it developed an incredibly cool style that acts a heavy dose of artistry to image made with it. Check out the gallery images! It can mostly be prompted out, but it'll tend to stick to the red color palette for the background

Important Prompt Notes

All versions, if prompted strictly with only the trigger word, lean towards monochrome and manga style with light color. It looks fantastic, but, well, you might not want that. Therefore its highly recommended to use the following neg prompt:

(low quality, worst quality:1.3), monochrome

---

As per the bounty text, Chiharu is a mascot for Oobabooga Text Generation UI - Essentially A1111 for textual AI (it runs on Gradio, just like A1111). Check it out if you're interested in that!

"Chiharu Yamada is a youthful adult, computer engineer-nerd with a knack for problem solving and a passion for technology"

I worked strictly off the dataset provided, which was 100% headshots, so this one was an interesting challenge (did about 7-8 separate trainings adjusting tags and learn rates), but I have 3 results I'm happy with I've decided to upload here.

描述:

Like Chiharu, but even more accurate, and red.

The showcase will show you what I mean!

训练词语: chiharu_red