IRSTLM is no more supported on SourceForge.
Please visit the new IRSTLM official website:
IRSTLM: A tool for the estimation, representation, and computation of statistical language models.
IRSTLM, the IRST Language Modeling Toolkit, features algorithms and data structures suitable to estimate, store, and access very large n-gram language models. Our software has been integrated into a popular open source Statistical Machine Translation decoder called Moses, and is compatible with language models created with other tools, such as the SRILM Tooolkit.
Very useful, but recently all of the documentation has disappeared from the page linked and basically everywhere else. Would love to have the documentation and some instructions added in the package.
I have a question and would like to ask for your help. Since I checked online, there are few answers about this problem.I don't know why. I follow Moses steps to train irst language model and these are my steps: # language model built on target language. ~/irstlm/bin/add-start-end.sh < /media/Data/Mose/test/raw10k.en > raw10k.sb.en export IRSTLM=$HOME/irstlm; ~/irstlm/bin/build-lm.sh -i raw10k.sb.en -o raw10k.lm.en -n 3 ~/irstlm/bin/compile-lm --text yes raw10k.lm.en.gz raw10k.fr-en.arpa.en Then, the last line command told me that too many arguments. I checked compile-lm command and there is --text options, and no format problem I think. I also tried search online, but few people met with this problem. The default of --text is no, but if I omit --text yes, I can not run the last step "~/mosesdecoder/bin/build_binary raw10k.fr-en.arpa.en raw10k.fr-en.blm.en". Did I do something wrong? I am really confusing. Thank you very much for the help!
Very useful project