# LoRA

Freezes the model and learns small low-rank update matrices for selected layers, so only a small fraction of parameters change.

## What it is used for

Cheap fine-tuning of large models, and keeping many task adapters over a single base.