Microsoft · DP-800
Models and embeddings
DP-800 · Lesson 17 of 20 · 2:13:34
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 DP-800 exam (Microsoft Certified: SQL AI Developer Associate). In this lesson: evaluating models, CREATE EXTERNAL MODEL with credentials and permissions, choosing columns and chunking text with AI_GENERATE_CHUNKS, generating embeddings, and keeping them current. Includes three original practice questions.
Preview the topics
- The embedding deployment, on the same Azure OpenAI resource, gets a new name. Dozens of procedures call AI_GENERATE_EMBEDDINGS through an external model. What should the team change?
- Product reviews average twenty thousand characters, beyond the input limit of the embedding model, and one embedding per review returns poor matches. What should the team do?
- Product descriptions change a few hundred times a day, and the team wants embeddings refreshed within minutes, outside the database engine, without adding work to each write. Which method fits?