AWS · AIF-C01
Training and fine-tuning FMs
AIF-C01 · Lesson 12 of 17 · 1:28:23
YouTube loads only after you press play. This chapter opens at its timestamp in the full course. Watch on YouTube ↗
In this lesson
Prepare for the AIF-C01 exam (AWS Certified AI Practitioner). In this lesson: pre-training, continued pre-training, fine-tuning, and distillation, instruction tuning, domain adaptation, and transfer learning, how to prepare data for fine-tuning, and reinforcement learning from human feedback. Includes three original practice questions.
Preview the topics
- AnyCompany has two million unlabeled insurance documents, and wants a model that understands insurance vocabulary better, before any task-specific tuning. Which method fits?
- AnyCompany’s fine-tuning examples all come from car claims in one region, but the model will handle every claim type, nationwide. Which data property is missing?
- AnyCompany’s reviewers compare pairs of model answers and pick the better one, and these preferences then train the model. What is this?