Track
Fine-Tuning
Goal: Adapt a pretrained model with LoRA/QLoRA, instruction tuning, DPO, and the data work behind good adapters.
Prereqs: LLMs.
Status: done
Work through the steps in order. Bold links open YouTube.
| Step | Concept | YouTube | Read |
|---|---|---|---|
| 1 | Full fine-tuning vs. parameter-efficient fine-tuning | LoRA & QLoRA Explained Simply — Full Fine-Tuning vs PEFT + Intuition + Practical | |
| 2 | LoRA | Hu et al. 2021 — LoRA: Low-Rank Adaptation of Large Language Models | |
| 3 | QLoRA | Dettmers et al. 2023 — QLoRA: Efficient Finetuning of Quantized LLMs | |
| 4 | Fine-tuning with LoRA in practice | Hugging Face PEFT — LoRA conceptual guide | |
| 5 | Data for fine-tuning | Meta AI — How to fine-tune: Focus on effective datasets | |
| 6 | Direct Preference Optimization | Direct Preference Optimization (DPO) — How to fine-tune LLMs directly without reinforcement learning | Rafailov et al. 2023 — DPO: Your Language Model is Secretly a Reward Model |
Found a broken link or an unclear step? Report a problem with this track.