yes, the preprocessing and postprocessing steps can be quite complicated.
if you use the experiment management system (EMS), it will set up the
steps for you. The EMS can be found in
scripts/ems
there is no process than replaces tags, entities etc with placeholders.
You have to do that yourself, perhaps by annotating the input sentence
with xml
http://www.statmt.org/moses/?n=Moses.AdvancedFeatures#ntoc5
On 20/06/2012 04:58, Henry Hu wrote:
> Hi all,
>
> I was wondering where the detailed steps of decoding are? I've look
> through at manual and document on http://www.statmt.org and not found
> them.
>
> Now, I've got some materials to translate. I suppose that I should
> follow these steps blew:
>
> . lowercase
> . replace tags, entities, etc with placeholders
> . tokenize
> . decode
> . detokenize
> . replace placeholders with tags, entities, etc
> . recase
>
> There are so many processes to handle. Where can I get a guide for
> decoding? Thanks for any suggestions.
>
> Best Regards,
> Henry
> _______________________________________________
> Moses-support mailing list
> [email protected]
> http://mailman.mit.edu/mailman/listinfo/moses-support
>
_______________________________________________
Moses-support mailing list
[email protected]
http://mailman.mit.edu/mailman/listinfo/moses-support
