tasksource/doc-nli

--- configs: - config_name: default data_files: - split: train path: data/train-* - split: validation path: data/validation-* - split: test path: data/test-* dataset_info: features: - name: premise dtype: string - name: label dtype: string - name: hypothesis dtype: string splits: - name: train num_bytes: 1758995420.312401 num_examples: 861708 - name: validation num_bytes: 553908139.6429919 num_examples: 210683 - name: test num_bytes: 637709787.6263076 num_examples: 243460 download_size: 401416382 dataset_size: 2950613347.5817003 license: bsd task_categories: - text-classification language: - en tags: - natural-language-inference --- # Dataset Card for "doc-nli" https://github.com/salesforce/DocNLI/tree/main ``` @inproceedings{yin-etal-2021-docnli, title = "{D}oc{NLI}: A Large-scale Dataset for Document-level Natural Language Inference", author = "Yin, Wenpeng and Radev, Dragomir and Xiong, Caiming", editor = "Zong, Chengqing and Xia, Fei and Li, Wenjie and Navigli, Roberto", booktitle = "Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021", month = aug, year = "2021", address = "Online", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2021.findings-acl.435", doi = "10.18653/v1/2021.findings-acl.435", pages = "4913--4922", } ```

原始信息汇总

数据集概述

数据集配置

• 配置名称: default
• 数据文件:
  • 训练集: data/train-*
  • 验证集: data/validation-*
  • 测试集: data/test-*

数据集信息

• 特征:
  • premise: 数据类型为字符串
  • label: 数据类型为字符串
  • hypothesis: 数据类型为字符串
• 分割详情:
  • 训练集:
    • 字节数: 1758995420.312401
    • 示例数: 861708
  • 验证集:
    • 字节数: 553908139.6429919
    • 示例数: 210683
  • 测试集:
    • 字节数: 637709787.6263076
    • 示例数: 243460
• 下载大小: 401416382
• 数据集大小: 2950613347.5817003

许可证

• 许可证类型: bsd

任务类别

• 任务类别: text-classification

语言

• 语言: en

标签

• 标签: natural-language-inference

搜集汇总

背景与挑战

背景概述

DocNLI是一个大规模文档级自然语言推理数据集，包含超过130万行数据，用于文本分类任务，支持英语。该数据集基于前提和假设对，标注为'entailment'或'not_entailment'，旨在评估模型在文档级别上的逻辑推理能力，并已划分为训练、验证和测试集。