HachiML/timeseries_QA_single_for_projector

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。这些特征分别代表了不同的数据类型和结构，如整数序列、浮点数序列等。数据集分为训练集和评估集，训练集包含152,215个示例，评估集包含8,012个示例。数据集的下载大小为160,458,881字节，总大小为1,676,134,647字节。数据文件根据配置被存储在指定的路径中。

The dataset includes multiple features such as input_ids, attention_mask, time_series_values, time_series_input_mask, and labels. These features represent different data types and structures, such as sequences of integers and floating-point numbers. The dataset is divided into a training set and an evaluation set, with the training set containing 152,215 examples and the evaluation set containing 8,012 examples. The download size of the dataset is 160,458,881 bytes, and the total size is 1,676,134,647 bytes. Data files are stored in specified paths according to the configuration.

原始信息汇总

数据集概述

特征信息

• input_ids: 整数序列，类型为 int32
• attention_mask: 整数序列，类型为 int8
• time_series_values: 浮点数序列，类型为 float32
• time_series_input_mask: 浮点数序列，类型为 float32
• labels: 整数序列，类型为 int64

数据分割

• train: 包含 152215 个样本，大小为 1592321115.0 字节
• eval: 包含 8012 个样本，大小为 83813532.0 字节

数据集大小

• 下载大小: 160458881 字节
• 总大小: 1676134647.0 字节

配置信息

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