dianavdavidson/gram-vaani-preprocessed

这是一个包含印地语音频及其转录文本的数据集，带有说话者的人口统计信息和情感标签。数据集包括音频特征（采样率16000Hz）、转录文本、口音、年龄、性别、背景、情感、地区、州等信息，用于语音识别或情感分析等自然语言处理任务。数据分为训练集（22454个样本）、验证集（1220个样本）和测试集（716个样本）。

This is a dataset containing Hindi audio and corresponding transcriptions, with speaker demographic information and sentiment labels. It includes audio features (sampling rate 16000 Hz), transcripts, accent, age, gender, background, sentiment, district, and state, designed for natural language processing tasks such as speech recognition or sentiment analysis. The data is split into train (22454 examples), validation (1220 examples), and test (716 examples) sets.