aaekay/pubmed-embeddings

该数据集包含从本地PubMed标题和摘要文本生成的嵌入向量，专为生物医学检索和最近邻研究设计。公共文件故意不包括PubMed标题、摘要或全文，行中包含PMID、嵌入向量、哈希值和轻量级元数据，以便研究人员可以将其与自己的授权PubMed镜像或官方NCBI/PubMed服务连接。

This dataset contains embedding vectors generated from local PubMed title and abstract text. It is designed for biomedical retrieval and nearest-neighbor research. The public files intentionally do not include PubMed titles, abstracts, or full text. Rows contain PMIDs, embeddings, hashes, and lightweight metadata so researchers can join against their own authorized PubMed mirror or the official NCBI/PubMed services.