Beam Search

Überblick

Beam search expands several partial sequences at once and keeps the highest-scoring ones, rather than committing to the token chosen at each step. It solves the problem that greedy decoding can be locally optimal and globally poor. It exists as the default for translation and summarisation, and it tends to produce bland text for open-ended generation.

Technologiegraph

Wie Beam Search mit dem übrigen Stack zusammenhängt.

Alternativen

• Top-p Sampling technique

  Eine Alternative zu Beam Search · source

Vergleiche

Detaillierte Gegenüberstellungen mit den Alternativen oben.

• Top-p Sampling vs Beam Search