pelatihpokemongo/datatalk

datatalk是一个印尼语语音数据集，包含两个列：音频和文本。该数据集适用于自动语音识别和文本到语音的快速原型设计，由本地音频和转录本构建而成，仅包含一个训练集split。数据集的结构参考了Elise数据集，并使用Parquet格式进行快速预览。datatalk is an Indonesian speech dataset consisting of two columns: audio and text. It is designed for rapid prototyping in automatic speech recognition and text-to-speech tasks, built from local audio and transcripts with a single train split. The structure of the dataset is based on the Elise dataset and is auto-converted to Parquet for fast preview on the Hub.