VoxForge
VoxForge
Good morning everybody ,
I am working on language processing and I need to make an alignement using GIZA++ , I compiled GIZA++ and it worked very well . In the documents I read they say that the most important file generated by this tool for "the user" is the actual.ti.final because it includes the alignements proposed by GIZA and the probability of each. When I tried this with the couple of languages "french-english" it worked and the file was there however when I tried this with Arabic it did not : This file and some others are there but with 0 ko ?? is there any parameters to modify when it comes to arabic?? If any one can give me a clue I will be very thankful.
--- (Edited on 3/30/2012 4:09 am [GMT-0500] by sara) ---
Well nobody was able to help me but thank God I find the problem, in case someone faces the same problem : it has nothing to do with the language the problem is in the corpus, it has to be organised a sentence per ligne.
--- (Edited on 4/2/2012 4:50 am [GMT-0500] by Visitor) ---
Hello friend,
I trained the model using giza++ and all went correctly i.e. all required files are generated correctly. ( simple english to french )
But I don't know how to "Test" unobserved data.
I want to test if my trained model work correctly or not.
Please help me telling the "Test" step by step.
Thank you in advance.
--- (Edited on 9/11/2012 2:42 am [GMT-0500] by Visitor) ---