The original weights are frozen, and small adapter matrices are trained alongside them.

In depth: 10 · Fine-tuning: when and how → “LoRA and PEFT: training only a part”

Related: SFT · model

Appears in

• 10 · Model Training: Practice: fine-tune a small model

The course is free — start on the site →

Write to the author: Telegram · LinkedIn

checked against the course: 2026-11-02 · 664b147