search

gooofy / zamia-speech

Open tools and data for cloudless automatic speech recognition
GNU Lesser General Public License v3.0
443 stars 84 forks source link

German STT dont working #32

Closed ercash closed 5 years ago

ercash commented 6 years ago

Hello,

i am trying to test pre-trained Kaldi models with german language. It shows errors:

debian-admin@debian:~$ python kaldi_decode_wav.py -v -m /opt/kaldi/model/kaldi-generic-de-tdnn_sp Wie_geht_es_dir.wav DEBUG:root:/opt/kaldi/model/kaldi-generic-de-tdnn_sp loading model... DEBUG:root:/opt/kaldi/model/kaldi-generic-de-tdnn_sp loading model... done, took 52.890310s. DEBUG:root:/opt/kaldi/model/kaldi-generic-de-tdnn_sp creating decoder... DEBUG:root:/opt/kaldi/model/kaldi-generic-de-tdnn_sp creating decoder... done, took 0.695156s. Traceback (most recent call last): File "kaldi_decode_wav.py", line 72, in if decoder.decode_wav_file(wavfile): File "kaldiasr/nnet3.pyx", line 195, in kaldiasr.nnet3.KaldiNNet3OnlineDecoder.decode_wav_file (kaldiasr/nnet3.cpp:4362) AssertionError

I had installed with this guide https://github.com/gooofy/zamia-speech#debian-9-stretch-amd64

Can you tell me, why is not working :(. Thank you.

svenha commented 6 years ago

Please check the reported AssertionError. Line 195 is: assert wavf.getnchannels()==1 So, I think you recorded a stereo file, but it should be mono.

ercash commented 6 years ago

Thanks for answer. I think, wav-file is mono:

debian-admin@debian:~$ file Wie_geht_es_dir.wav Wie_geht_es_dir.wav: RIFF (little-endian) data, WAVE audio, Microsoft PCM, 16 bit, stereo 16000 Hz

Hier is file: Wie_geht_es_dir.zip

ercash commented 6 years ago

Sorry, you are right, i did't see it "stereo 16000 Hz"

gooofy commented 5 years ago

assuming the issue is solved, I'll close this one.