AI Glossary

Fine-tuning

What does “Fine-tuning” mean?

Fine-tuning: Training an already pre-trained model further on a smaller, targeted dataset so it gets better at a specific task, style or domain. It adjusts existing weights instead of starting from scratch, so it needs far less data and compute than pre-training.

Fine-tuning visual guide: Training an already pre-trained model further on a smaller, targeted dataset so it gets better at a specific task, style or domain. It adjusts existing weights instead of starting from scratch, so it needs far less data and compute than pre-training.
Visual guide Open full size (new tab)

What is an example of “Fine-tuning”?

A company might fine-tune a general model on its past support tickets so replies match its products and tone.

Keep learning

See also

Pre-trainingTermTransfer learningTermInstruction tuningTermLoRATerm

Same topic

CheckpointTermCurriculum learningTermDirect preference optimizationTermFoundation modelTermIn-context learningTerm