ibm-research/REAL-MM-RAG_FinTabTrainSet_rephrased

REAL-MM-RAG_FinTabTrainSet_rephrased是一个针对金融表格密集型文档的检索模型训练数据集，包含48000个自然语言三元组，通过自动化流程构建，用于提高表格文档的检索性能。REAL-MM-RAG_FinTabTrainSet_rephrased is a retrieval model training dataset for table-intensive financial documents, containing 48,000 natural language triplets constructed through an automated pipeline to enhance the retrieval performance for tabular documents.