LibriSpeech - Githubissues

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

Closed blisc closed 5 years ago

blisc commented 5 years ago

moved pad_to from get_speech_features to data layer
added flag to precompute mel_basis
added librispeech config
- conv masking
- hamming window
- norm per feature
- dithering