[
https://issues.apache.org/jira/browse/OPENNLP-831?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15223787#comment-15223787
]
Rodrigo Agerri edited comment on OPENNLP-831 at 4/4/16 7:54 AM:
----------------------------------------------------------------
Do you train MAXENT or PERCEPTRON?
Sentences you mean 5K sentences or 500K sentences? I wrote it wrong in my
previous message, I meant 2M words, not sentences.
If you used MAXENT, how many interations? If you used PERCEPTRON, how many
iterations until it converged?
How many entities in the training data? How many classes/types?
R
was (Author: ragerri):
Do you train MAXENT or PERCEPTRON?
Sentences you mean 5K sentences or 500K sentences? I wrote it wrong in my
previous message, I meant 2M words, not sentences.
If you used MAXENT, how many interations? If you used PERCEPTRON, how many
iterations until it converged?
R
> NER model creation is taking much time for large files
> ------------------------------------------------------
>
> Key: OPENNLP-831
> URL: https://issues.apache.org/jira/browse/OPENNLP-831
> Project: OpenNLP
> Issue Type: Improvement
> Components: Name Finder
> Affects Versions: 1.6.0
> Environment: linux
> Reporter: Rakesh
> Priority: Blocker
>
> opennlp 1.6 is taking huge amount of time (nearly 6 hours ) to train a set of
> sentences (sentences count = 20000) . Each sentence may vary in length .You
> must try to decrease the training time. Is there any way to decrease the
> training time.If yes please let me know.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
