Overview
- Adaptation ladder — prompt → RAG → fine-tune → align
- Five families — full SFT, soft prompts, PEFT, alignment, federated
- LoRA & QLoRA — math intuition, NF4, merge vs hot-swap
- RLHF, DPO, GRPO — when each fits
- Runnable PEFT + TRL examples
Blog mega-GIF: mega-finetune-everything.gif
- Full tutorial — Parts 1–17 with GIFs
- Examples — LoRA, QLoRA shell, DPO scripts
- Assets — diagram + terminal GIFs, blog poster