Dear devs,
Currently, I'm working on my thesis on the topic of creating a system that is able to translate English to Bahasa Indonesia (Indonesian language).
Joshua is chosen because of the both phrase-based and hierarchical feature, since the translation engine system in the backend will be
using hierarchical phrased-based decoding technique.
But, I don't see in the documentation how to use both of these technique in the pipeline. Is this by default? Or, should I use some options in the pipeline?
If yes, what would those options be?
It's said in the documentation (Installation),
"You will need to install Moses if either of the following applies to you:
- You wish to build phrase-based models ....
..............................
Follow the instructions for installing Moses
here, and then define the $MOSES
environment variable to point to the root of the Moses installation."So, I'm currently installing Moses to be able to build the phrase-based models. However, I'm not seeing how to enable it on the pipeline documentation.
Or did I miss it? Or is it going to build the phrase-based models after defining the $MOSES environment variable?
Thanks a lot for your help. It means so much for me. Waiting for the answers while preparing my training data. I'm sorry that I ask a lot of questions.
Hopefully, after it is finished, I can contribute the language packs at least. :)
Regards,
Reza Lesmana