Train on Last_5hr with forced minimum iteration
Author: Eric Beikman
Purpose: To determine the effect of forcing the Sphinx trainer to iterate more than the default.
The Sphinx Trainer utilizes multiple iterations of the Baum-Welch algorithm to refine the Acoustic model. The Sphinx trainer will automatically know how many iterations it needs to create a good model; as such, instead of determining a definitive number of iterations, the configuration file is given a minimum and maximum amounts of iterations. CMU does not recommend that you set the maximum amounts of iterations past 10.
By default, the minimum amount of iterations is set to 1. For this experiment, this was increased to 4.
The remaining aspects of the experiment was taken directly from experiment 0089. We are using its dictionaries, phone list, and using genTrans5.pl to generate the experiment's transcript.
Results The train was able to run successfully. This experiment is continued in Experiment 0103.