LM-Polygraph/wmt14

这是wmt14数据集的预处理版本，主要用于LM-Polygraph的基准测试。数据集包含两个配置：deen和fren，每个配置都有训练和测试两个分割，每个分割包含输入和输出两个字符串列。数据集的结构、用途、创建动机、来源数据、偏见和风险等方面都有详细描述。

This is a preprocessed version of wmt14 dataset for benchmarks in LM-Polygraph. The dataset contains two configurations: deen and fren, each with train and test splits, and each split contains two string columns: input and output. The structure, uses, creation rationale, source data, biases, risks, and limitations of the dataset are described in detail.