NVIDIA / OpenSeq2Seq

Toolkit for efficient experimentation with Speech Recognition, Text2Speech and NLP
https://nvidia.github.io/OpenSeq2Seq
Apache License 2.0
1.54k stars 369 forks source link

Alignment #391

Closed arnav1993k closed 5 years ago

arnav1993k commented 5 years ago

The sample rate change was causing distortions in audio files generated due to inconsistencies in librosa vs soundfile. Fixed it using librosa at all read and write commands