Fetching the paper…

Emformer: Efficient Memory Transformer Based Acoustic Model For Low Latency Streaming Speech Recognition · Around