Skip to content
AI Engineer Guide.

Track

Model adaptation

SFT, LoRA, preference optimisation, and when tuning is the wrong tool.

  1. Adaptation decision treeconcept · T1
  2. SFTconcept · T1
  3. Instruction tuningconcept · T1
  4. Dataset constructionconcept · T1
  5. Data qualityconcept · T2
  6. PEFTconcept · T1
  7. LoRAconcept · T1
  8. QLoRAconcept · T1
  9. Adaptersconcept · T2
  10. Full fine-tuningconcept · T2
  11. Preference dataconcept · T2
  12. RLHFconcept · T2
  13. Reward modelsconcept · T2
  14. PPO/DPO/GRPOconcept · T2
  15. Distillationconcept · T2
  16. Synthetic dataconcept · T1
  17. Fine-tuning evaluationconcept · T1
  18. Catastrophic forgettingconcept · T2
  19. Model mergingconcept · T3
  20. When not to fine-tuneconcept · T1
  21. Adapter servingconcept · T2
  22. Training costconcept · T2