Fetching the paper…

Serialized Output Training for End-to-End Overlapped Speech Recognition · Around