CGIAR/AgricultureVideosTranscript

该数据集包含印地语和奥里亚语的农业视频。数据集由多个xls文件组成，每个xls文件包含视频网址（YouTube视频链接）列和相应的转录文本。转录文本包括：1）由ASR模型生成的转录（用于基准测试）；2）人工转录；3）时间戳；4）人工翻译。该数据集可用于训练和基准测试针对ASR和翻译的特定领域模型。时间戳作为音频的来源，相应的人工转录可以用于对照。

This dataset consists of agriculture videos in Hindi and Oriya. The dataset comprises multiple xls files, each containing a column of video URLs (YouTube video links) and corresponding transcripts. The transcripts include: 1) Transcripts generated by ASR models (for benchmarking purposes); 2) Manual transcripts; 3) Time stamps; 4) Manual translations. The dataset can be used for training and benchmarking domain-specific models for ASR and translation, with time stamps serving as the source of audio and the corresponding human transcripts for reference.