rebnej/textcaps-git-large-5best

该数据集包含图像ID、假设和参考三个主要特征。数据集分为训练集和验证集，训练集包含109,765个样本，验证集包含15,830个样本。训练集的大小为72,839,425字节，验证集的大小为10,336,749字节。数据集的下载大小为32,327,020字节，总大小为83,176,174字节。

The dataset contains three main features: image_id, hypothesis, and reference. The dataset is divided into a training set and a validation set, with the training set containing 109,765 samples and the validation set containing 15,830 samples. The training set size is 72,839,425 bytes, and the validation set size is 10,336,749 bytes. The download size of the dataset is 32,327,020 bytes, and the total size is 83,176,174 bytes.

原始信息汇总

数据集概述

特征

• image_id: 字符串类型
• hypothesis: 字符串序列
• reference: 字符串序列

数据分割

• train:
  • 样本数量: 109765
  • 数据大小: 72839425 字节
• val:
  • 样本数量: 15830
  • 数据大小: 10336749 字节

数据集大小

• 下载大小: 32327020 字节
• 总数据大小: 83176174 字节

配置

• default:
  • train: 数据路径为 data/train-*
  • val: 数据路径为 data/val-*