I'm trying to use the Aligner (unsupervised) to align Korean and English sentences. I see that when I run the aligner on the example data, it does the "test" step on the example test corpus. Based on this, I'm assuming the Aligner can align unknown pairs of sentences using the model obtained through the alignment process with the train corpus. However, I couldn't find a way to use the Aligner to align unknown parallel sentences. How can I perform this task? Poking around the code, I discovered an abstract class called WordAligner, but I'm not too sure how to load a model on to a class that implements it.
[link][comment]