There should be an option to supply an already pos-tagged files rather than always tagging the input files. This can be performed as a check to the availability of the pos file similar to the way the language model perplexity files are checked before calculating them.
There should be an option to supply an already pos-tagged files rather than always tagging the input files. This can be performed as a check to the availability of the pos file similar to the way the language model perplexity files are checked before calculating them.