Hi,
I used Kaldi Speech Recognition Toolkit to decode a test utterance. Can we achived a probability which shows the assurance of kaldi's prediction for this test utterance?
Dose kaldi have a script which split a long wave into small waves? I have used "segment_long_utterances.sh" script and it gives segment file as the output, which I don't know how to use this "segment" file I want to have some short ".wav" files?
I would appreciate you if you answer my questions
Best regards--
Go to http://kaldi-asr.org/forums.html to find out how to join the kaldi-help group
---
You received this message because you are subscribed to the Google Groups "kaldi-help" group.
To unsubscribe from this group and stop receiving emails from it, send an email to kaldi-help+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/kaldi-help/a7fed2a9-906b-4a07-bb42-08dba4a1bc05n%40googlegroups.com.