Tanel Alumäe
Tanel Alumäe
Most likely, your MFCC conf file (decoder->mfcc-config in YAML) is not compatible with your acoustic model.
You have to be careful with piping long audio files. You should pipe them using a data rate that is similar to real-time. Otherwise, the whole file is sent to...
Our nnet3-based decoder is implemented based on nnet3 decoder in Kaldi. I have seen claims that our decoder is not as accurate as Kaldi's. I believe there could be a...
Yes, "nnet3 mode" is the only mode to use nnet3 models. It has historical reasons. Originally kaldi-gstreamer-server supported only GMM models (because DNN models were not used at time --...
The e2e model should be usable as other Kaldi chain models. Just don't specify the i-vetor extractor. Make sure that you are linking to recent Kaldi.
I think it's because you need to run the server using Python 2.7, but you are probably using Python 3+
Does this happen when we use client.py? Make sure that the terminal you use client.py in, is configured for UTF-8. This can be usually accomplished with `export LC_ALL=en_US.UTF-8` or something...
[3] should be fixed now, if you recompile gst-kaldi-nnet2-online.
Please clarify your issue. It's difficult to understand where and when this problem occurs.
I think you have to modify the code on the Kaldi Gstreamer component.