kardosdrur/opensubtitles-no-da

该数据集包含从OpenSubtitles中提取的丹麦语和挪威语对齐句子，使用了基于启发式的过滤器。数据集旨在帮助丹麦基础模型项目中的句子转换器训练。

The OpenSubtitles Danish-Norwegian dataset contains aligned sentences in Danish and Norwegian, extracted from OpenSubtitles using heuristic-based filters. The dataset is intended to aid in training sentence transformers in the Danish Foundation Models project. It includes train and test splits with 1,772,983 and 443,246 examples respectively. The dataset features include link_id, da, no, and overlap, where da and no are string types and overlap is a float64 type.

原始信息汇总

数据集概述

数据集名称

OpenSubtitles Danish-Norwegian

数据集描述

对齐的丹麦语和挪威语句子，来自OpenSubtitles，通过基于启发式的过滤器筛选。

数据集配置

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

数据集信息

• 特征:

  • link_id: 字符串类型
  • da: 字符串类型
  • no: 字符串类型
  • overlap: float64类型

• 分割:

  • 训练集:
    • 字节数: 270499727.08648384
    • 样本数: 1772983
  • 测试集:
    • 字节数: 67624969.91351616
    • 样本数: 443246

• 下载大小: 201396375

• 数据集大小: 338124697.0