This is a tool for training language model. The smoothing method is kneser-ney smoothing, to adapt the arpa output format, I use the backoff version of kneser-ney. You need python environment to run the program, the command is 'python LM_train.py dictionary data output', and you should use absolute path.
lzm0706 / language-model Goto Github PK
View Code? Open in Web Editor NEWLicense: GNU General Public License v3.0