lmms-lab/openhermes_instruction

这是一个包含音频和文本字段的数据集，主要用于测试。数据集由四个主要字段组成：音频上下文字段（context），指令字段（instruction），语音指令字段（speech_instruction）和答案字段（answer）。测试集包含100个示例，文件大小约为19092915字节。

This dataset consists of audio and text fields, primarily for testing purposes. It is composed of four main fields: an audio context field (context), an instruction field (instruction), a spoken instruction field (speech_instruction), and an answer field (answer). The test split contains 100 examples, with a total file size of approximately 19092915 bytes.