I originally did a kofi goal for creating this LORA, but no one supports me. So you know what? Suck it I will ball.
我原本是要等到达到Kofi预定金额的时候还开始制作这个LORA。不过我的人气太少了或没有人要捐，到现在为止Kofi里面还是0。所以。。。算了我干脆就做吧。

This model is basically Tatsuki Fujimoto's another cooking style other than the KINO and Best Boy (Togata). This model take much less cautions than my last one, where you just need to know which style are you going to aim to.

I only give 36 image dataset variant and 51 image dataset variant models only due to 23iv and 14iv being far too unstable than 36 and 51, especially on eyes. Even with 51, you can still see AI struggled with the eyes, I am still trying to find more times to fill in and upload even more image dataset's model.

I had experimented with those 2 models in InvokeAI, ComfyUI (primary), and AUTOMATIC1111, and also style merging stuff. And I found some interesting stuff during the experiments which you can see in the sample pic below ( It is too much for me to say... )

This model still use AnyLORA as always. Oh! and also the dataset is only using Chapter 139-142 images for 51iv variant. So for now it can't emulate anything other than Denji, Fumiko, Barem, Nayuta and Miri. ( spoiler :skull: )

If you have any other questions or recommendation, just ask in the comment.

明天再来做中文的吧，你们先看英文的先。