Unverified50% confidenceFactExact time
MS MARCO数据集约有800万文档，来自Bing搜索日志，每个查询平均约有1.1个标注正样本
1
Sources
50%
Confidence
Long-term
Relevance
7/16/2026
First Seen
Sources
Related Claims
UnverifiedAs of 2024, an estimated over 1 billion license plate scan records are stored in various databases58% similarVerifiedLAION-5B contains approximately 5.8 billion image-text pairs, making it the largest publicly available multimodal dataset at the time.57% similarUnverifiedNeurIPS 2023收到超过12000篇投稿，每位审稿人平均需要处理6篇以上的论文56% similarUnverifiedThe National Data Administration disclosed that high-quality datasets have reached 120,000, with a total volume exceeding 1,565 petabytes (PB).56% similarUnverifiedCOCO 数据集包含约12万张训练图像、80个类别、超过150万个目标实例，由微软研究院发布55% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/524281API
curl https://kongchang.com/api/v1/knowledge/claims/524281MCP
get_claim(id=524281)