MegaBNSpeech

该数据集是一个大规模的、与领域无关的自动语音识别（ASR）数据集，它由YouTube上的内容开发而来，涵盖了广泛的主题、说话风格、方言、噪声环境和对话场景。该数据集包含了从新闻频道、脱口秀和旅行视频博客中提取的多种音频，并已转换为16千赫兹采样率的WAV格式。该数据集规模宏大，总计53,000小时，包含42,000个视频资源，其任务是自动语音识别。This dataset is a large-scale, domain-agnostic Automatic Speech Recognition (ASR) dataset developed from content on YouTube, covering a wide range of topics, speaking styles, dialects, noise environments, and conversational scenarios. It contains various audio clips extracted from news channels, talk shows, and travel video blogs, and has been converted to WAV format with a sampling rate of 16 kHz. This dataset has a massive scale, with a total duration of 53,000 hours and 42,000 video resources, and its targeted task is automatic speech recognition.