You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@opennlp.apache.org by "Russ, Daniel (NIH/CIT) [E]" <dr...@mail.nih.gov> on 2017/01/10 14:53:56 UTC
merge TrainingParameters and PluggableParameters
Hi,
The point of PluggableParameters was to move the “get(Int/String/Boolean)Parmeters” out of AbstractTrainer and into a parameters. I would like to merge the functionality of PluggableParameters into trainingParameters. Before I start, is there any reason why the GISTrainer and the indexer have separate cutoffs? I assume it is just historic (maybe the trainer used to index), only the larger cutoff is in effect.
Daniel
Daniel Russ, Ph.D.
Staff Scientist, Office of Intramural Research
Center for Information Technology
National Institutes of Health
U.S. Department of Health and Human Services
12 South Drive
Bethesda, MD 20892-5624
Re: merge TrainingParameters and PluggableParameters
Posted by "Russ, Daniel (NIH/CIT) [E]" <dr...@mail.nih.gov>.
Sounds goods. I’ll take care of both of them.
Daniel
On 1/10/17, 10:12 AM, "Joern Kottmann" <ko...@gmail.com> wrote:
1. Yes, there are historic reasons, opennlp-tools used to depend on maxent,
but maxent didn't depend on opennlp-tools. Therefore maxent couldn't use
any opennlp-tools classes. It would be good to open a jira for this and get
this re-factored.
2. The cutoff code from GISTrainer should be removed. See issue OPENNLP-123
(half a decade old). PR for that would be very welcome.
3. And when you are it you might be interested to also help us resolve
OPENNLP-675.
Jörn
On Tue, Jan 10, 2017 at 3:53 PM, Russ, Daniel (NIH/CIT) [E] <
druss@mail.nih.gov> wrote:
> Hi,
>
> The point of PluggableParameters was to move the
> “get(Int/String/Boolean)Parmeters” out of AbstractTrainer and into a
> parameters. I would like to merge the functionality of PluggableParameters
> into trainingParameters. Before I start, is there any reason why the
> GISTrainer and the indexer have separate cutoffs? I assume it is just
> historic (maybe the trainer used to index), only the larger cutoff is in
> effect.
>
> Daniel
>
> Daniel Russ, Ph.D.
> Staff Scientist, Office of Intramural Research
> Center for Information Technology
> National Institutes of Health
> U.S. Department of Health and Human Services
> 12 South Drive
> Bethesda, MD 20892-5624
>
>
Re: merge TrainingParameters and PluggableParameters
Posted by Joern Kottmann <ko...@gmail.com>.
1. Yes, there are historic reasons, opennlp-tools used to depend on maxent,
but maxent didn't depend on opennlp-tools. Therefore maxent couldn't use
any opennlp-tools classes. It would be good to open a jira for this and get
this re-factored.
2. The cutoff code from GISTrainer should be removed. See issue OPENNLP-123
(half a decade old). PR for that would be very welcome.
3. And when you are it you might be interested to also help us resolve
OPENNLP-675.
Jörn
On Tue, Jan 10, 2017 at 3:53 PM, Russ, Daniel (NIH/CIT) [E] <
druss@mail.nih.gov> wrote:
> Hi,
>
> The point of PluggableParameters was to move the
> “get(Int/String/Boolean)Parmeters” out of AbstractTrainer and into a
> parameters. I would like to merge the functionality of PluggableParameters
> into trainingParameters. Before I start, is there any reason why the
> GISTrainer and the indexer have separate cutoffs? I assume it is just
> historic (maybe the trainer used to index), only the larger cutoff is in
> effect.
>
> Daniel
>
> Daniel Russ, Ph.D.
> Staff Scientist, Office of Intramural Research
> Center for Information Technology
> National Institutes of Health
> U.S. Department of Health and Human Services
> 12 South Drive
> Bethesda, MD 20892-5624
>
>