Skip to main content

Vocabulary of natural language processing

Search from vocabulary

Concept information

Preferred term

random seed  

Definition

  • A number that is used to initialize a pseudorandom number generator, which is typically used to determine the initial state of model parameters. (Based on Bethard, We need to talk about random seeds, 2022)

Broader concept

Example

  • All experiments were repeated ten times with varying random seeds but with the same data (subsets). (Hedderich, Adelani, Zhu, Alabi, Markus & Klakow, 2020)
  • Since we run each case using 5 random seeds and report the average performance the final tokens are 25990*5=129950 tokens so we need to spend $0.38985. (Yuan, Chen, Zhang & Jiang, 2024)
  • We ran 2 random seeds for each and report the mean. (Portelance, Frank, Jurafsky, Sordoni & Laroche, 2021)
  • We ran each setup 5 times with different random seeds. (Portelance, Frank, Jurafsky, Sordoni & Laroche, 2021)
  • We train each model 5 times with different random seed and when evaluating we sample 4 different support sets. (Chen, Lin, Han & Sun, 2021)

In other languages

URI

http://data.loterre.fr/ark:/67375/8LP-Q8D7GFWP-7

Download this concept:

RDF/XML TURTLE JSON-LD Last modified 6/28/24