thinQmaster Virtual testing partner Adaptive practice exams that find your weak spots and drill them. Every answer explained. From $69
Enroll

thinQmaster

Find your exam

Five questions free, before you decide.

Step 1 · choose a vendor

In one of our camps? Your practice exams come with the tuition — they are on your study page already.

AI & Machine Learning · Advanced

Speculative Decoding

Speeding up generation by having a small draft model propose tokens that the large model verifies in parallel.

Definition

The large model checks several drafted tokens in one forward pass and accepts the correct prefix, yielding identical output distribution with lower latency. Variants include Medusa heads and EAGLE.

Exam tip

Speculative decoding reduces latency without changing the large model's output quality.

Related terms