uima-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Tommaso Teofili <tommaso.teof...@gmail.com>
Subject Re: Guidelines for a mutual contribution
Date Fri, 02 Sep 2011 07:00:47 GMT
2011/9/1 Jörn Kottmann <kottmann@gmail.com>

> On 9/1/11 5:38 PM, Tommaso Teofili wrote:
>
>> 2011/9/1 Jörn Kottmann<kottmann@gmail.com>
>>
>>  >  On 9/1/11 2:39 PM, Tommaso Teofili wrote:
>>> >
>>>
>>>> >>  I am reviewing the legal stuff for this; if no one objects, once
I'm
>>>> >>  finished I'll proceed with the vote for the acceptance for HMM
>>>> Tagger
>>>> >>  French
>>>> >>  Models.
>>>> >>
>>>>
>>> >
>>> >  Will it be possible for us to retrain these models? And then also
>>> release
>>> >  the retrained models?
>>> >
>>>
>> As long as one can read French (that is a false sentence for me at the
>> moment :P) Nicolas wrote something here:
>> http://enicolashernandez.**blogspot.com/2011/05/**
>> construire-des-modelisations-**du-french.html<http://enicolashernandez.blogspot.com/2011/05/construire-des-modelisations-du-french.html>
>>
>> The models were built on French Treebank corpus [1].
>> It would be nice if it could be translated to English, and hopefully added
>> to the Tagger documentation.
>>
>>
>>  >
>>> >  Otherwise it will be hard to change to code, since strict backward
>>> >  compatibility
>>> >  must be maintained.
>>> >
>>>
>> As far as I know, not at the moment as the legal stuff and this
>> contribution
>> regard only the models as they are without the data used to train them.
>> Asking the French Treebank corpus rights owner to grant ASF a SGA for such
>> data would be another piece of work I think.
>>
>
> I guess it will be possible to follow the steps to retrain the model.
> Then it would not be allowed to distribute this new model, right?
>
>
I think so.
Tommaso

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message