This repository contains the hands-on notebooks used in the D200 course for fine-tuning LLMs.
- Slides:
docs/slides/ - Notebooks:
notebooks/ - Solutions:
notebooks/solutions/
The exercises focus on fine-tuning techniques (LoRA, DPO) using lightweight models that run on standard GPUs.
All exercises are designed to run on ≥16GB NVIDIA GPUs, with constrained configurations to ensure reliability.
Work through the notebooks in order.
Try implementing each section yourself before checking the solutions.