Fetching the paper…

Mockingjay: Unsupervised Speech Representation Learning with Deep Bidirectional Transformer Encoders · Around