Hi

 
> Questions:
> 
> 1. Where can I get the new TAUS paralleled corpus from, such as English-
> Vietnamese language pair?
> 

This sounds like a question for TAUS. Have you looked through their website?

> 
> 
> 2. Suppose that I have trained paralleled corpus and generate the configure
> file as moses1.ini and I will train new corpus generating configure file as
> moses2.ini. How can I improve the translation quality based on moses1.ini
> with moses2.ini?
> 
> I mean, how can I merge the moses1.ini and moses2.ini together, which will
> enlarge the vocabulary of MT?
>

You have to concatenate the corpora, then retrain a new model with the new 
large corpus. 

There is support for incremental retraining, but it's not straightforward to 
use.
http://www.statmt.org/moses/?n=Moses.AdvancedFeatures#ntoc27
 
> 
> 
> 3. Is there any method allowing me to alter the translation between source
> language & target language? How can I do it?
> 

If you want to inject your own translations, look at the xml-input 
functionality

http://www.statmt.org/moses/?n=Moses.AdvancedFeatures#ntoc4

cheers - Barry
_______________________________________________
Moses-support mailing list
[email protected]
http://mailman.mit.edu/mailman/listinfo/moses-support

Reply via email to