ECG_LM_Final_Dataset_8715
收藏资源简介:
该数据集是一个多模态医疗数据集,包含心电图图像及其相关的文本序列。数据集由三个核心字段构成:ecg_id为心电图样本的唯一整数标识符;ecg_image为心电图图像数据;text_sequence为与心电图相关的文本描述或报告序列。数据集总样本量为8,715例,已划分为训练集(6,972例)、验证集(871例)和测试集(872例),适用于需要结合视觉与文本信息的任务,例如基于心电图图像的自动报告生成、医疗图像描述、跨模态检索或医疗人工智能模型的训练与评估。
This dataset is a multimodal medical dataset containing electrocardiogram (ECG) images and their associated text sequences. It consists of three core fields: ecg_id is a unique integer identifier for each ECG sample; ecg_image represents the ECG image data; text_sequence is the text description or report sequence related to the ECG. The total sample size is 8,715 cases, divided into a training set (6,972 cases), a validation set (871 cases), and a test set (872 cases). It is suitable for tasks that require combining visual and textual information, such as automatic report generation based on ECG images, medical image description, cross-modal retrieval, or training and evaluation of medical artificial intelligence models.
- 数据集名称: ECG_LM_Final_Dataset_8715
- 数据集来源: Hugging Face Datasets
- 数据特征: 包含三个字段:
ecg_id: 整数类型 (int64),心电图记录的唯一标识符。ecg_image: 图像类型 (image),对应的心电图图像。text_sequence: 字符串类型 (string),相关的文本序列描述。
- 数据集划分: 共8715个样本,分为三个子集:
- 训练集 (train): 6972 个样本
- 验证集 (validation): 871 个样本
- 测试集 (test): 872 个样本
- 数据集大小:
- 下载大小: 2,575,599,373 字节
- 数据集总大小: 2,581,081,026.744 字节
- 配置与文件: 默认配置 (
config_name: default),数据以data/train-*、data/validation-*、data/test-*模式存储的parquet文件。





