Speech:Exps 0247

From Openitware
Jump to: navigation, search

Description

Author: Colby Johnson

Date: 3/26/14

Purpose: using what we know now, we will run a train and decode it to get the best percentage we can at this point in time.

Details: Decode on last_5hr/train acoustic model using the last_5hr/train data set. LM was build using last_5hr/train data set.


Results: Decode: xRT = 1.19

                    SYSTEM SUMMARY PERCENTAGES by SPEAKER
     ,-----------------------------------------------------------------.
     |                            hyp.trans                            |
     |-----------------------------------------------------------------|
     | SPKR    | # Snt # Wrd | Corr    Sub    Del    Ins    Err  S.Err |
     |=================================================================|
     | Sum/Avg | 4201  65836 | 92.4    4.4    3.1    9.5   17.1   80.8 |
     |=================================================================|
     |  Mean   | 34.4  539.6 | 92.3    4.6    3.1   10.5   18.2   81.2 |
     |  S.D.   |  9.1  157.3 |  3.5    2.3    1.7    6.2    8.2   11.1 |
     | Median  | 33.0  518.0 | 92.7    4.0    2.8    8.8   16.5   81.9 |
     `-----------------------------------------------------------------'

Successful Completion