LoRA

image

Low-Rank Adaptation — a small add-on file that teaches an existing model a new style, character, or concept without retraining the whole thing. Instead of adjusting billions of parameters, a LoRA trains a few million and layers on top, so files are tiny (megabytes) and training takes hours, not months. The open-weights community trades thousands of LoRAs — specific art styles, likenesses, product types. For creators it is the practical path to a custom brand look: train a LoRA on 20-30 brand images, and every generation comes out on-brand.