Top Ten Tips for Fine-tuning Large Language Models

Top Ten Tips for Fine-tuning Large Language Models

Trelis Research via YouTube Direct link

Playlists to watch

15 of 23

15 of 23

Playlists to watch

Class Central Classrooms beta

YouTube videos curated by Class Central.

Classroom Contents

Top Ten Tips for Fine-tuning Large Language Models

Automatically move to the next video in the Classroom when playback concludes

  1. 1 Top Ten Fine-tuning Tips
  2. 2 Tip 1: Start with a Small Model
  3. 3 Tip 2: Use LoRA or QLoRA
  4. 4 Tip 3: Create 10 manual questions
  5. 5 Tip 4: Create datasets manually
  6. 6 Tip 5: Start training with just 100 rows
  7. 7 Tip 6: Always create a validation data split
  8. 8 Tip 7: Start by only training on one GPU
  9. 9 Tip 8: Use weights and biases for logging
  10. 10 Scale up rows, tuning type, then model size
  11. 11 Tip 9: Consider unsupervised fine-tuning if you've lots of data
  12. 12 Tip 10: Use preference fine-tuning ORPO
  13. 13 Recap of the ten tips
  14. 14 Ten tips applied to multi-modal fine-tuning
  15. 15 Playlists to watch
  16. 16 Trelis repo overview
  17. 17 ADVANCED Fine-tuning repo Trelis.com/ADVANCED-fine-tuning
  18. 18 Training on completions only
  19. 19 ADVANCED fine-tuning repo CONTINUED
  20. 20 ADVANCED vision Trelis.com/ADVANCED-vision
  21. 21 ADVANCED inference trelis.com/enterprise-server-api-and-inference-guide/
  22. 22 ADVANCED transcription trelis.com/ADVANCED-transcription
  23. 23 Support + Resources Trelis.com/About

Never Stop Learning.

Get personalized course recommendations, track subjects and courses with reminders, and more.

Someone learning on their laptop while sitting on the floor.