Fetching the paper…

Sequence-to-Sequence Speech Recognition with Time-Depth Separable Convolutions · Around