HKUST

The HKUST dataset is a benchmark for evaluating speech recognition systems, containing Mandarin Chinese speech data used to assess performance metrics such as Character Error Rate (CER).

Papers using HKUST (5)

EfficientASR: Speech Recognition Network Compression via Attention Redundancy and Chunk-Level FFN Optimization2024 · 2 citesDual Causal/Non-Causal Self-Attention for Streaming End-to-End Speech Recognition2021GNCformer Enhanced Self-attention for Automatic Speech Recognition2023Conversational Speech Recognition by Learning Audio-textual Cross-modal Contextual Representation2023Online Hybrid CTC/Attention End-to-End Automatic Speech Recognition Architecture2023