beats3
收藏资源简介:
该数据集包含10,000个训练样本,总大小约806MB。每个样本包含多种特征:1) JSON格式的结构化数据(含分类标签字段);2) 文本内容(txt字段)及标注文本(annotated_txt);3) 音频数据(采样率16kHz);4) 事件相关统计字段(事件总数、独立事件数及事件序列)。数据以训练集单一分割形式组织,原始文件采用分块存储格式(data/train-*)。
This dataset contains 10,000 training samples with a total size of approximately 806 MB. Each sample includes multiple types of features: 1) Structured data in JSON format with a classification label field; 2) Text content (txt field) and annotated text (annotated_txt); 3) Audio data with a sampling rate of 16 kHz; 4) Event-related statistical fields covering the total number of events, the number of independent events, and the event sequence. The dataset is organized as a single training set split, with its original files stored in a chunked storage format under the path pattern `data/train-*`.




