search

srvk / eesen

The official repository of the Eesen project
http://arxiv.org/abs/1507.08240
Apache License 2.0
822 stars 342 forks source link

live mode recognition #84

Closed Bhee closed 8 years ago

Bhee commented 8 years ago

I would like to test the model with live utterance (by live recording ) instead of batch mode,

  1. whether eesen support live mode?
  2. if not how to load a model once and decode each utterance
fmetze commented 8 years ago

Please look here, the transcriber can load multiple utterances easily.

Online/ live recognition would be great to have, but it is a lot of engineering - anyone up for it?