Style LoRA trained for CircleStone Labs Anima.
This is a triggerless style LoRA. No specific trigger word is required; load the LoRA and adjust the weight to taste.
Recommended weight: 0.8-1.0
Base model: Anima
Captions: Danbooru-style tags
Training resolution: 1024
Training: 6 epochs / 1080 steps
Trigger word: none
Works well with Danbooru-style tag prompts and mixed tag + natural language prompting. Lower the weight to around 0.6-0.8 if the style becomes too strong or starts overriding prompt details.
Description
First release
Trained on Anima Base V1.0