[
https://issues.apache.org/jira/browse/OPENNLP-751?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14483135#comment-14483135
]
Joern Kottmann commented on OPENNLP-751:
----------------------------------------
The way the beam size is handled will result in the wrong beam size when the
SourceForge models are used.
> Chunker trainer doesn't write beam size into the model
> ------------------------------------------------------
>
> Key: OPENNLP-751
> URL: https://issues.apache.org/jira/browse/OPENNLP-751
> Project: OpenNLP
> Issue Type: Bug
> Components: Chunker
> Reporter: Joern Kottmann
> Assignee: Joern Kottmann
> Priority: Minor
> Fix For: 1.6.0
>
>
> The beam size is retrieved from the parameter file but not passed as an
> argument to the model in the ChunkerME.train method.
> Additionally it uses the NameFinderME default beam size instead of the
> ChunkerME default beam size.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
