LoRA: Easy Training of Stable Diffusion Models

Low-Ranking Adaptation (LoRA) models are small models that are trained on a certain specific concept, character, or style.

In order to use a LoRA model, you can put a LoRA file from Civitai (opens in a new tab) into the stable-diffusion-webui/models/lora folder. This will have them show up in your UI under the Loras tab.

If you would like to get into training LoRA, I highly recommend checking out Laura Carnevali's video (opens in a new tab).

P.S.

• to change the image on the model card in your UI, you need just save a png/jpg and a .txt with the same name into the folder the model is in. This works for checkpoints and embeddings, too.