Fetching the paper…

Simplified Self-Attention for Transformer-based End-to-End Speech Recognition · Around