This LoRA was trained using anima yume v0.25.

This is a LoRA trained on AI-generated images.
Although it is a style LoRA, the dataset consists entirely of a single character (Tojo Nozomi), so there is some bias in the results.

In particular, if you generate with just “1girl,” there is a high probability that Tojo Nozomi will be produced. Additionally, since the model was trained on AI-generated images, anatomical inaccuracies may occur.

I believe the style reproduction has achieved about 80% of the intended target.

trigger word is @eyei