Search

Word Search

Information System News

Decoding Strategies and Output Control
Rick W

Decoding Strategies and Output Control

This chapter is divided into nine parts; they are: • Reading Logits from a Model • Greedy Decoding • Temperature Sampling • Top-$k$ Sampling • Nucleus Sampling • Repetition Penalties • Beam Search • Stop Conditions • Structured Output Constraints The model returns a vector of logits for every position in the input sequence.
Previous Article Using a Transformer Model: From Training to Inference
Next Article Measuring Performance of Transformer Inference
Print
6