youssefkhalil320/MedSynth-Combined

该数据集是一个包含对话、笔记和音频的多模态数据集，专门用于训练目的。数据集包括2551个训练样本，总大小约为47.17GB。每个样本包含以下特征：Dialogue（对话文本）、Note（笔记文本）、audio（音频数据，采样率为24000Hz，未解码）和row_idx（整数行索引）。数据集仅提供训练分割，没有验证或测试集。音频特征以原始格式存储，适用于需要处理语音和文本结合的任务，如语音识别、对话生成或多模态学习。

This dataset is a multimodal dataset containing dialogue, notes, and audio, specifically designed for training purposes. It includes 2551 training examples with a total size of approximately 47.17GB. Each example consists of the following features: Dialogue (text of conversations), Note (text of notes), audio (audio data with a sampling rate of 24000Hz, undecoded), and row_idx (integer row index). The dataset only provides a training split, with no validation or test sets. The audio feature is stored in raw format, making it suitable for tasks that involve combining speech and text, such as speech recognition, dialogue generation, or multimodal learning.