gruhit-patel/alpaca_speech_instruct

该数据集包含四个主要特征：output、input、instruction和filename，其中filename特征包含音频数据，采样率为16000Hz。数据集分为一个训练集，包含51760个样本，总大小为12494864368.04字节。数据集的下载大小为12288466427字节，许可证为cc，任务类别为文本生成，语言为英语。数据集的pretty_name为Alpaca Speech Inst，规模类别为10K<n<100K。

The dataset includes four features: output (string type), input (string type), instruction (string type), and filename (audio type with a sampling rate of 16000). The dataset is divided into a training set (train) containing 51760 samples, with a total size of 12494864368.04 bytes. The download size of the dataset is 12288466427 bytes. The dataset is suitable for text generation tasks, in English, named Alpaca Speech Inst, with the number of samples between 10K and 100K.