My result training model is word error rate 100 %

Speech Recognition Toolkit

Brought to you by: air, arthchan2003, awb, bhiksha, and 5 others

This project can now be found here.

My result training model is word error rate 100 %

Forum: Help

Creator: patcharapol

Created: 2021-08-01

Updated: 2021-08-01

patcharapol - 2021-08-01

I need a specialized model for a small vocabulary application with thai language.
when I train my model with command "sphinxtrain run" in linux .
I wonder why my word error rate is 100 %

My Result : result.png

Someone can help me improve this error rate pls.

Last edit: patcharapol 2021-08-01

result.png

rs2.png

rs3.png

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

patcharapol - 2021-08-01

*And I have 2 wav file that .

hiwnam.wav
Input File : 'hiwnam.wav'
Channels : 1
Sample Rate : 8000
Precision : 16-bit
Duration : 00:32:29.88 = 15599072 samples ~ 146241 CDDA sectors
File Size : 31.2M
Bit Rate : 128k
Sample Encoding: 16-bit Signed Integer PCM

=========================================
puathau.wav

Input File : 'puathau.wav'
Channels : 1
Sample Rate : 8000
Precision : 16-bit
Duration : 00:31:16.60 = 15012832 samples ~ 140745 CDDA sectors
File Size : 30.0M
Bit Rate : 128k
Sample Encoding: 16-bit Signed Integer PCM

wav.png

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Log in to post a comment.