Hello,
I tried to train the model on 100 hours of Librispeech audio. I am using some part of the data for testing(around 3-4 hours approx.)
Link for data: http://www.openslr.org/resources/12/train-clean-100.tar.gz
Though the training modules passed, got a lot of errors in the forward-backward iterations, but i couldnt see much in the logs (actually couldnt understand much.)
I am getting the following error rate:
SENTENCE ERROR: 91.5% (162/177) WORD ERROR RATE: 19.5% (1061/5447)
Hello,
I tried to train the model on 100 hours of Librispeech audio. I am using some part of the data for testing(around 3-4 hours approx.)
Link for data: http://www.openslr.org/resources/12/train-clean-100.tar.gz
Though the training modules passed, got a lot of errors in the forward-backward iterations, but i couldnt see much in the logs (actually couldnt understand much.)
I am getting the following error rate:
SENTENCE ERROR: 91.5% (162/177) WORD ERROR RATE: 19.5% (1061/5447)
Link for etc folder and log files: https://drive.google.com/open?id=1ZkfaCBdInk6KPpB1rq0fimqlP7zAJAjC
Please let me know if there's anything wrong with data preparation, or i need to tweak the training parameters.
Thanks and Best Regards,
Nikhil Shirwandkar