2

I've been doing some KALDI learning these days and I follow the tutorial and I complete some examples like yesno, voxforge, ynstadial, and a custom digits ASR.

But after all of the above completed I only got something like WER 5% and some log.

How can I use them like a microphone or something?

Also, I am trying to do some incremental learning or transfer training on the KALDI dummy tutorial's example but I don't know how to start. Can anyone help me with what to learn or what to do?

Thanks for helping!

Vishal Parmar
  • 524
  • 7
  • 27
Eric WU
  • 21
  • 3

1 Answers1

0

Kaldi allows online and offline decoding. [https://kaldi-asr.org/doc/online_decoding.html][1]

Demo is available in kaldi/egs/vox directory. See gst and online demo folders . [https://github.com/kaldi-asr/kaldi/tree/master/egs/voxforge][2]

Details to follow in the comments

SAGE
  • 9
  • 3
  • A good way to build up basics before going deep into Kaldi is 1. Learn speech processing fundamentals on [www.speech.zone][3] 2. Learn Linux Hands on. Check [cbtnuggets.com][4] 3. Learn Natural Language Processing and basics of Data science and ML Ops. Check Andrew NG courses on [deeplearning.ai][5] [1]: https://kaldi-asr.org/doc/online_decoding.html [2]: https://github.com/kaldi-asr/kaldi/tree/master/egs/voxforge [3]: https://www.speech.zone [4]: https://cbtnuggets.com [5]: https://deeplearning.ai – SAGE Jul 06 '22 at 18:54
  • Some beginner kaldi help is available on https://jrmeyer.github.io/ https://chrisearch.wordpress.com/2017/03/11/speech-recognition-using-kaldi-extending-and-using-the-aspire-model/# https://www.assemblyai.com/blog/kaldi-speech-recognition-for-beginners-a-simple-tutorial/ https://kunal-dhawan.weebly.com/asr-system-for-hindi-language-from-scratch.html Learn how to integrate your own Language Model built on kaldi with vosk for online and offline Speech to text on https://medium.com/@andimid/offline-foreign-speech-recognition-32d8d63de2dc – SAGE Jul 06 '22 at 18:54