Survival report

I've been diligently studying the Anima model since it came out. Version 1.0 has officially been released, but as expected, it's producing a lot of Western-looking images (and the art style isn't consistent) without a masterpiece face, so I spent the whole day today creating a LoRA art style (;'∀') For training images, I used 700 images of the model I normally use. This is my first time creating a LoRA art style, but I feel like it went better than I expected. Num Repeats=1, Keep Tokens=1 TE=0 seemed to be recommended, so I followed that. Epoch 2 started to burn, so I slightly diluted 2 and generated it. It seems the base model really goes wild without a LoRA art style 💦