Evaluation Task and Evaluation Data Introduction

Audio Domain Evaluation (Audio)

Currently, it includes the following evaluation tasks:

• Speech Recognition：including evaluation datasets of AISHELL-1, KeSpeech, Librispeech, etc.
• Speech Emotion Recognition：including evaluation datasets of IEMOCAP, etc.
• Intent Classification：including Fluent Speech Commands dataset.
• Language Identification：including Common Language dataset.
• Speech Separation：including LibriMix dataset.
• Slot Filling：including SNIPS dataset.
• Speech Enhancement：including Voicebank-DEMAND dataset.
• Speech Translation：including Covost-en-de dataset.
• Speaker Recognition：including CN-Celeb dataset.