Fetching the paper…

Conv-Transformer Transducer: Low Latency, Low Frame Rate, Streamable End-to-End Speech Recognition · Around