← LectureMoment
Inference & Optimization · Intermediate
Inference & Optimization — Intermediate
20 practice questions on Inference & Optimization. Every question is written from a specific moment in a real lecture, and after you answer it links to that exact timestamp so you can check it yourself.
What this pack asks
The free questions in this pack. Answer choices and explanations appear as you play.
On modern accelerators, arithmetic throughput far exceeds memory bandwidth. Why does this make operation counts an incomplete measure of an algorithm's speed?
Answer this one →A quantized linear layer runs input ddequantizes its output before passing it on. Why is the dequantization step needed?
Answer this one →
Written from lectures by Umar Jamil. Not affiliated with or endorsed by any university, channel, creator or certification program.