TrevorJS/synth-addresses-ner-mk2

该数据集包含多个与地址相关的特征，如名称、州名、州缩写、邮政编码、城市、街道号码、街道名称等。此外，还包括OCR文本、类型、单元号码、单元指示符和跟踪号码等字段。数据集被分为训练集、评估集和测试集，分别包含1350000、150000和500000个示例。

This dataset includes multiple features related to addresses, such as name, state name, state abbreviation, postal code, city, street number, street name, etc. Additionally, it includes fields like OCR text, type, unit number, unit designator, and tracking number. The dataset is divided into training, evaluation, and test sets, containing 1,350,000, 150,000, and 500,000 examples respectively.