Train Acoustic Model Problem

Speech Recognition Toolkit

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

This project can now be found here.

#488 Train Acoustic Model Problem

Milestone: next release

Status: open

Owner: nobody

Labels: None

Priority: 1

Updated: 2019-02-13

Created: 2019-02-13

Creator: raymond

Private: No

I trained my own Mandarin acoustic model, when it finish all the stages, with decoder showing:
SENTENCE ERROR: 0.0% (0/10) WORD ERROR RATE: 0.0% (0/10).(I only trained 10 vocabulary for commanding)
It all seen good, but when I use the lm in my demo, the recognizer response nothing (empty),
do you know how to solve this problem? Thanks.

1 Attachments

etc.rar

Discussion

raymond - 2019-02-13

Sorry,I mean when I use the acoustic model(AM) in my demo.

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

raymond - 2019-02-13

Also,I just saw some guy successfully trained an AM with a small data set,I don't know how he made it,but it works perfectlly. With this,I just want to know how to increase recognization accuracy with a small amount of vocabulary .

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

Train Acoustic Model Problem

Speech Recognition Toolkit

Group

Searches

Help

#488 Train Acoustic Model Problem

Discussion