jobs-demo
收藏资源简介:
该数据集名为“Bag-of-Documents — Unified Jobs Artifacts”,是用于支持统一职位搜索空间的配套资源。它包含来自四个公开语料库(Open-Apply、LinkedIn、JobStreet和USAJobs)的347,900个职位发布数据。数据集内容涵盖职位标题、精简元数据、完整元数据(以JSONL格式存储),以及用于增强搜索能力的预计算向量表示,包括bge-small向量、te3-large @ 1024目录向量和一个预编码的te3查询缓存(约196,000个热门查询)。该数据集适用于信息检索、语义搜索、职位推荐系统等自然语言处理任务,旨在为职位搜索相关的研究和应用提供结构化和向量化的数据支持。数据以CC BY 4.0许可证发布。
The dataset is named Bag-of-Documents — Unified Jobs Artifacts and serves as a supporting resource for a unified job search space. It contains 347,900 job postings from four public corpora (Open-Apply, LinkedIn, JobStreet, and USAJobs). The dataset includes job titles, simplified metadata, complete metadata (stored in JSONL format), and pre-computed vector representations to enhance search capabilities, such as bge-small vectors, te3-large @ 1024 catalog vectors, and a pre-encoded te3 query cache (approximately 196,000 popular queries). It is suitable for natural language processing tasks like information retrieval, semantic search, and job recommendation systems, aiming to provide structured and vectorized data support for job search-related research and applications. The data is released under the CC BY 4.0 license.
数据集概述
数据集名称:Bag-of-Documents — Unified Jobs Artifacts
许可证:CC-BY-4.0
数据内容:
- 包含347,900条职位发布信息
- 覆盖4个语料库:Open-Apply、LinkedIn、JobStreet、USAJobs
- 提供的数据类型包括:
- 职位标题
- 精简元数据
- 完整元数据(JSONL格式)
- bge-small向量表示
- te3-large @ 1024 目录向量
- 预编码的te3查询缓存(约196k常见查询)
相关资源:
- 配套的职位搜索Space:https://huggingface.co/spaces/dtunkelang/bag-of-documents-jobs
- 源代码仓库:https://github.com/dtunkelang/bag-of-documents




