# Transfer learning

Starts from weights learned on a large source task and adapts them to a smaller target task.

## What it is used for

Fine-tuning vision backbones for defect detection, and adapting language models to a domain with little data.