Just for reference, The issue was that I didn't filter the phrase table
beforehand. based on how the model was trained, I knew this wouldn't reduce
the size of the phrase table enough to make it worthwhile (or so I
thought), but as soon as I filtered the phrase table it stopped segFaulting.
On Fri
there's no tools to do this but you can write it yourself. You need to make
up some scores to give each phrase.
The other methods to use your phrases are:
1. Add it to the training data and retrain your model.
2. Create a 2nd phrase-table with just your phrases and get the decoder
to use it,
The moses.ini file looks ok.
The segfault is difficult to debug by inspection. It's best sprinkle the
decoder code with debugging message and run 1 instance of the decoder to
see exactly where it goes wrong. And you should binarise the pt and
reordering models to save yourself time
* Looking for
