Wuming/reactome

该数据集包含两个配置：pair_classification和reranking。pair_classification配置用于基因对的分类任务，包含基因名称、Reactome名称、类别、基因摘要、维基百科信息和Wikicrow信息等特征。reranking配置用于基因对的重排序任务，包含基因名称、正样本、负样本、基因摘要、维基百科信息和Wikicrow信息等特征。两个配置都包含训练、验证和测试集，分别有不同的字节大小和样本数量。

The dataset includes two configurations: pair_classification and reranking. The pair_classification configuration is for pair classification tasks and includes features such as external gene name, reactome name, class label, gene summary, Wikipedia, and Wikipedia summary. The reranking configuration is for reranking tasks and includes features such as external gene name, positive, negative, gene summary, Wikipedia, and Wikipedia summary. The dataset is divided into training, validation, and test sets, each with corresponding sizes and number of examples.