Image-to-LoRA V2 converts images to style LoRA in one pass

View on Reddit

The updated i2L V2 turns one or more reference images into a style LoRA without explicit training, requiring only a single forward pass. It's built on a model collection shared by the developer.