This LoRA was trained on a dataset generated with Illustrious-based models. For optimal performance, tag-style prompts are recommended.
Flame colors can be controlled using trigger words. The following color variants are currently supported:
blue fire
purple fire
green fire