Difference between revisions of "Speech:Exps 0212"

From Openitware
Jump to: navigation, search
(Description)
(Description)
 
Line 16: Line 16:
 
Training: 32 Min
 
Training: 32 Min
 
Decode: xRT = 1.08
 
Decode: xRT = 1.08
 +
WER: 33.8

Latest revision as of 18:10, 18 March 2014

Description

Author:
Colby Johnson & Colby Chenard

Date:
3/16/14

Purpose:
Find the difference in performance using genTrans5.pl vs genTrans8.pl

Details:
Mimicking Experiment 0209 we will generate a new acoustic model using the smae dictionary and parameters except using genTrans5.pl to compare the results.

Results:
Training: 32 Min Decode: xRT = 1.08 WER: 33.8