cmaldona/Generalization-MultiClass-CLINC150-ROSTD

该数据集名为generalization-test，版本为1.0.0，是三个数据集的合并，主要用于多类分类任务的泛化实验。数据集包含两个配置：clinc150和rostd+，每个配置都有训练、验证和测试集的文件路径。数据集的语言为英语，任务类别为文本分类。此外，数据集用于处理ID、near-OOD、covariate-shift和far-OOD的实验设置，涉及的数据集包括CLINC150、ROSTD+OOD、SST2和News Category。This dataset, named generalization-test with version 1.0.0, is a merged compilation of three datasets, primarily designed for generalization experiments in multi-class classification tasks. The dataset includes two configurations: clinc150 and rostd+, each with file paths for training, validation, and test sets. The dataset uses English as its language, and the task category is text classification. In addition, this dataset is intended for experimental settings including ID, near-OOD, covariate-shift, and far-OOD, with the involved datasets being CLINC150, ROSTD+ OOD, SST2, and News Category.

原始信息汇总

数据集概述

数据集名称

• 名称: generalization-test
• 版本: 1.0.0

描述

• 描述: 由三个数据集合并而成。

配置

• 配置名称: clinc150
  • 默认: 是
  • 数据文件:
    • 训练集: train_clinc150.csv
    • 验证集: validation_clinc150.csv
    • 测试集: test_clinc150.csv
• 配置名称: rostd+
  • 数据文件:
    • 训练集: train_rostd+.csv
    • 验证集: val_rostd+.csv
    • 测试集: test_rostd+.csv

许可证

• 许可证: openrail

任务类别

• 任务类别: 文本分类

语言

• 语言: 英语