sphinx_lm_eval - Man Page

Evaluate perplexity of a transcription


sphinx_lm_eval [ options ]...


This program evaluates the perplexity of a text file according to a given language model.  The text file is assumed to be in transcript format, i.e. one utterance per line, delimited by <s> and </s>.


Shows the usage of the tool


model file


file listing a set of language models


of language model in -lmctlfn to use for all utterances


Base in which all log-likelihoods calculated


file to evaluate


Language model weight


Use memory-mapped I/O for reading binary LM files


definition file for classes in LM


string to evaluate


Unigram probability weight (interpolated with uniform distribution)


Print details of perplexity calculation


Word insertion probability


David Huggins-Daines <dhdaines@gmail.com>