Problem in voice corpus tool augment when converting stereo to mono

[nmstoker (Neil Stoker)]

As per comments here it’s not bidirectional, just a unidirectional LSTM