WhissleAI/Meta_STT_HI_Set1

Meta Speech Recognition Hindi Dataset (第一集)是一个包含印地语语音识别样本的元数据和音频文件的数据集，适用于语音识别、实体标注、意图预测、情感分类、年龄预测和性别预测等多种NLP任务。该数据集由AI4Bharat Indic Speech Dataset、Common Voice Hindi和Shrutilipi等多个来源合并而成，遵循不同的开放许可证。数据集分为训练集、验证集和测试集，总共包含超过800,000个样本。

The Meta Speech Recognition Hindi Dataset (Set 1) is a collection of metadata and audio files for Hindi speech recognition samples, suitable for tasks such as speech recognition, entity tagging, intent prediction, emotion classification, age prediction, and gender prediction. The dataset is compiled from multiple sources including AI4Bharat Indic Speech Dataset, Common Voice Hindi, and Shrutilipi, and is licensed under different open licenses. It is divided into training, validation, and test sets, totaling more than 800,000 samples.