お洋服お借りしましたー。↓
背景はAIに描いてもらった

1 7

So I used Dreambooth/Stable Diffusion (thanks to the 's training code) to train on my own drawings, but I think it's still easy to guess which results come from the AI.
Full post: https://t.co/gVbtSR7Rmq

0 1

I got all of my really good results by training a custom model using Dreambooth and then using it on a local installation of Stable Diffusion.

It took like a week to figure it out, but I'm super happy with the results :3

0 8

姉妹二人分まとめてどりぶったせいか、表情を上手く操れない。髪や色、左右が混ざる。描き分けむつかしい。別個にやったほうがいいのかな。

0 2

New finetune just out.

Instant dopamine hit 🚀

I'm going to have some fun with this one.

11 103



Seriously, this is the most Ashita no Joe looking image my custom Dreambooth model has spat out, so far. Those lines, the coloring, the shading, is all bang on. All that's missing is the precise character design. Free PFP, to anybody who wants it.

0 0



Here's the same prompt as the last image, but rando seed, and the NovelAI pre-prompt and negative prompts, at CFG 12. My Ashita no Joe 2 Dreambooth model might be turning into something that's a keeper.

0 1