Thanks for the info Jörn.

On Thu, Jun 8, 2017 at 8:03 PM, Joern Kottmann <[email protected]> wrote:

> 1) Yes, you can use the API to load multiple models, and afterwards you can
> merge the spans. In that case you have to deal with overlapping spans
> somehow. Today we also support training models on multiple types, that is
> probably better suited for you use case.
>
> 2) There is some code which does that, but it just drops overlapping spans
> without any particular strategy, e.g. there might be some ways for you to
> validate which names are of which type
>
> Jörn
>
> On Thu, Jun 8, 2017 at 7:00 AM, Manoj B. Narayanan <
> [email protected]> wrote:
>
> > Hi,
> >
> > This is with respect to the following bug posted on Source Forge a long
> > time back Multiple Models in CMD
> > <https://sourceforge.net/p/opennlp/bugs/41/>. I have 2 questions
> regarding
> > this.
> >
> > 1) Currently there is support for loading multiple models in CMD tool and
> > it gives a single result. Is is possible with the API ?
> >
> > 2) How does the system resolve conflicts and produces a single result?
> >
> > Thanks,
> > Manoj
> >
> > On Mon, Jun 5, 2017 at 12:41 PM, Manoj B. Narayanan <
> > [email protected]> wrote:
> >
> > > Hi,
> > >
> > > This is with respect to the following bug posted on Source Forge a long
> > > time back Multiple Models in CMD
> > > <https://sourceforge.net/p/opennlp/bugs/41/>. I have 2 questions
> > > regarding this.
> > >
> > > 1) Currently there is support for loading multiple models in CMD tool
> and
> > > it gives a single result. Is is possible with the API ?
> > >
> > > 2) How does the system resolve conflicts and produces a single result?
> > >
> > > Thanks,
> > > Manoj
> > >
> >
>

Reply via email to