write_text_params.1

有关自然语言理解理解方面的源码

.\" Process this file with 
.\"    groff -man -Tascii write_text_params.1

.TH WRITE_TEXT_PARAMS 1 "FEBRUARY 2004" "Infomap Project" "Infomap NLP Manual"

.SH NAME

.TP
write_text_params \- convert binary model params to text

.SH SYNOPSIS
.B write_text_params
.BR -mdir " <model_data_dir> " 

.SH DESCRIPTION
.B write_text_params
reads in the three binary files expressing various model and corpus
parameters, and writes out the information they contain in a textual
format.  Keeping these data around as text is a Good Idea because it
allows easy inspection of the parameters and could aid portability to
other machines, since the binary files are not portable across
architectures.

.SH OPTIONS

.TP
.BI -mdir \ <model_data_dir>
The directory from which we read input and to which we write output.

.SH INPUT FILES
These files are read from the model data directory, specified
as an argument to the 
.B -mdir 
option.

.I model_params.bin
.RS
Basic parameters describing the model.  Parameters that will need to 
be read by search-side programs (such as
.BR associate (1))
are kept here.  Other parameters are kept in 
.I model_info.bin
to keep this file small and efficient to load.  This design reflects
probable premature optimization.  See
.BR prepare_corpus (1)).
.RE

.I model_info.bin
.RS
Additional parameters describing the model.
See 
.BR prepare_corpus (1)).
.RE

.I corpus_format.bin
.RS
Parameters describing the format of the corpus and how it has 
been processed.  For instance, the tags that mark the beginning
and end of documents, and XML/SGML character entities that have
been stripped.  See
.BR prepare_corpus (1)).
.RE

.SH OUTPUT FILES
All of these files are produced in the model data directory, specified
as an argument to the
.B -mdir
option.

.I model_params.txt
.RS
Textual version of
.IR model_params.bin .
.RE

.I model_info.txt
.RS
Textual version of
.IR model_info.bin .
.RE

.I corpus_format.txt
.RS
Textual version of
.IR corpus_format.bin .
.RE

.SH SEE ALSO
.BR prepare_corpus (1), \ count_wordvec (1), \ svdinterface (1), \ encode_wordvec (1), \ count_artvec (1), \ associate(1).

.SH DIAGNOSTICS
Returns 0 to indicate success, number of errors in case of error.

.SH BUGS
Probably this program should not exist, as we would most likely be
better off using textual formats throughout and doing away with the
binary.  This shouldn't cause much of a performance hit, but would 
require some text-parsing code.
Please report bugs to 
.BR [e-mail] .

.SH AUTHOR
This program and manual page were written by Scott Cederberg.  Please direct
inquiries and bug reports to 
.BR [e-mail] .