"Please read this before using, thank you."

Generate simple anime style portraits for Flux Dev.

(Currently Works for Simple Prompts Only!)

This is v1 release. This changes pretty much everything including dataset, training process, and trigger word. The trigger word is inspired by lots of amazing creators when I was looking for a solution to force Flux into anime even when using natural language and without mentioning anime at all in the prompt. You can still integrate "anime" in prompt or even use tag prompting if you want to fully lock this into 2D (especially if you don't like full lips). This version has a bit different style from B3 (or Beta 3 in the model info), but I believe it's better in terms of flexibility (thus requires higher weight). More of this in the changelog.

Use settings:

> Put trigger words at the beginning or the end of prompt.

> Weight from 0.7 to 1. Using 0.6 is okay if you put "anime" in prompt but otherwise start from 0.7.

> Use Euler. If you're generating in comfy, use Euler + Normal, Euler + Simple, and/or Euler + Beta. I never tested others, but if you're comfortable with certain samplers, you can definitely try with that.

> 20 steps is fine. 25 is also fine. 30 is fine. I never tested more than 30 since it's expensive.

> You don't really need upscaler but you can use it if you want higher fidelity.

> Clip skip... I don't really know. I usually use 1, but 2 also works.

> Recommended for portrait images. I'm not talking about image orientation, but like, that portraits. I hope you understand this. But if you want to create something else then feel free to do it since I'm not really a good prompter and can't test it with good prompt.

Changelog (v1):

> Smaller size! Finally, I can get it to half the size of B3 but still performs quite well.

> Style changed to be quite similar to Beta 2 (very noticeable in close-ups).

> Trigger word changed into "artwork by ry012".

> Improved flexibility compared to B3.

> Changes to the training parameter resulting in quite a difference in behavior (compared to B3).

Known issue:

> Feet still somewhat sucks

> Sometimes there is a watermark "ry012" since I use the trigger word to trick Flux that it's an art by an artist.

> You tell me

Thank you for using this model for your image generation.

====================================================