metarank/esci

Amazon ESCI/ESCI-S数据集是Amazon ESCI和ESCI-S数据集的结合，采用JSON格式。该数据集用于在Metarank项目中微调双编码器和交叉编码器模型。数据集的格式为JSON-line格式，每个行代表一个排名事件，并附有所有项目元数据。

原始信息汇总

Amazon ESCI/ESCI-S 数据集概述

数据集基本信息

• 许可证: Apache 2.0
• 语言: 英语
• 标签: 购物, 排名, 亚马逊
• 数据集名称: ESCI
• 数据集大小: 10K<n<100K

数据集描述

Amazon ESCI/ESCI-S 数据集是 Amazon ESCI 和 ESCI-S 数据集的组合，采用JSON格式。该数据集主要用于 Metarank 项目中对双编码器和交叉编码器模型进行微调。

数据集格式

数据集采用JSON行格式，每一行代表一个单独的排名事件，所有商品元数据已预先合并。示例数据如下： json { "query": "!qscreen fence without holes", "e": [ { "title": "Zippity Outdoor Products ZP19026 Lightweight Portable Vinyl Picket Fence Kit w/Metal Base(42" H x 92" W), White", "desc": "..." }, { "title": "Sunnyglade 6 feet x 50 feet Privacy Screen Fence Heavy Duty Fencing Mesh Shade Net Cover for Wall Garden Yard Backyard (6 ft X 50 ft, Green)", "desc": "..." }, { "title": "Amgo 6 x 50 Black Fence Privacy Screen Windscreen,with Bindings & Grommets, Heavy Duty for Commercial and Residential, 90% Blockage, Cable Zip Ties Included, (Available for Custom Sizes)", "desc": "..." }, { "title": "Amgo 4 x 50 Black Fence Privacy Screen Windscreen,with Bindings & Grommets, Heavy Duty for Commercial and Residential, 90% Blockage, Cable Zip Ties Included, (Available for Custom Sizes)", "desc": "..." } ] }