Track 10: Training LLMs, How Training Works
GPUs for Training: Colab, Kaggle, Cloud GPUs Safely
Pick hardware and shut it down after the job.
Part of Train and Fine-Tune Models (19 of 45)
Course
Part of the free syllabus, not yet in a certificate course.
Udemy: coming soon
Tools
- Google Colab (freemium)
- Kaggle Notebooks (free)
- GPU clouds (RunPod, Lambda, Vast.ai, E2E Networks) (paid)
Watch first
- How LLM Training Works: Tokens, Loss, Pre-training vs Fine-tuning
- Colab GPUs: T4, L4, A100 and What Costs Units
In this module
- Fine-Tune vs Prompt vs RAG
- How LLM Training Works: Tokens, Loss, Pre-training vs Fine-tuning
- GPUs for Training: Colab, Kaggle, Cloud GPUs Safely
- Pre-training at Scale, Explained
- Scaling Laws and Compute-Optimal Training