HachiML/timeseries_QA_multi_right

该数据集包含多个特征，包括input_ids、attention_mask、time_series_values、time_series_input_mask和labels，这些特征分别以int32、int8、float32和int64的数据类型表示。数据集分为训练集和评估集，训练集包含152,014个样本，评估集包含8,001个样本。数据集的下载大小为160,925,879字节，总大小为2,083,555,315字节。数据文件的路径在配置部分指定。

The dataset includes multiple features such as input_ids, attention_mask, time_series_values, time_series_input_mask, and labels, represented by data types int32, int8, float32, and int64 respectively. The dataset is divided into a training set and an evaluation set, with the training set containing 152,014 samples and the evaluation set containing 8,001 samples. The download size of the dataset is 160,925,879 bytes, and the total size is 2,083,555,315 bytes. The paths to the data files are specified in the configuration section.

原始信息汇总

数据集概述

特征信息

• input_ids: 序列类型为 int32
• attention_mask: 序列类型为 int8
• time_series_values: 序列类型为 float32 的序列
• time_series_input_mask: 序列类型为 float32
• labels: 序列类型为 int64

数据分割

• train:
  • 字节数: 1979374294.0
  • 样本数: 152014
• eval:
  • 字节数: 104181021.0
  • 样本数: 8001

数据集大小

• 下载大小: 160925879 字节
• 总大小: 2083555315.0 字节

配置信息

• default:
  • train: 路径为 data/train-*
  • eval: 路径为 data/eval-*