TomoroAI/temporal_cookbook_db

Temporal Cookbook DB是一个设计用于事件抽取、时间推理和事实表示的的多表结构数据集。该数据集最初构建为SQLite数据库，并转换为CSV文件以供在Hugging Face Hub上托管。数据集是从AMD和Nvidia的盈利电话会议数据子集处理而来。数据集组织为多个配置，每个配置对应于原始数据库中的一个表，包括转录文本、文本块、带时间元数据的事件、命名或链接实体、主谓宾三元组和事件实体关联表。每个配置都是一个独立的CSV支持的表，可以独立加载。

Temporal Cookbook DB is a multi-table dataset designed for event extraction, temporal reasoning, and fact representation. The dataset was originally built as an SQLite database and converted into CSV files for hosting on the Hugging Face Hub. It is derived from a subset of data from the earnings calls of AMD and Nvidia. The dataset is organized into multiple configurations, each corresponding to a table in the original database, including transcripts, chunks, events, entities, triplets, and event_entities. Each configuration is a standalone CSV-backed table that can be loaded independently.