diff --git a/databases/vectordb/diskann.html b/databases/vectordb/diskann.html index f320983..6b46dd1 100644 --- a/databases/vectordb/diskann.html +++ b/databases/vectordb/diskann.html @@ -183,6 +183,10 @@

DiskANN 通过 BeamSearch(设置 beamwidth 一次读多个数据块) 和缓存最常访问的节点(eg. by caching all vertices that are C = 3 or 4 hops from the starting point s)来加速查询。 另外,DiskANN 将邻居节点的向量保存在磁盘索引文件中,来提高搜索的精度(Implicit Re-Ranking Using Full-Precision Vectors)。

+

Code

+

Further readings

diff --git a/print.html b/print.html index 4f909f6..7400c94 100644 --- a/print.html +++ b/print.html @@ -1167,6 +1167,10 @@

Misc

  • Accelerating Vector Search: Using GPU-Powered Indexes with RAPIDS RAFT
  • Accelerating Vector Search: Fine-Tuning GPU Index Algorithms
  • Accelerated Vector Search: Approximating with RAPIDS RAFT IVF-Flat
  • +
  • Semantic Search: Measuring Meaning From Jaccard to Bert
  • +
  • SPLADE for Sparse Vector Search Explain
  • +
  • Sparse embedding or BM25?
  • +
  • Unlock Advanced Recommendation Engines with Milvus' New Range Search
  • Efficient and robust approximate nearest neighbor search using Hierarchical Navigable Small World graphs

    @@ -1254,6 +1258,10 @@

    DiskANN 通过 BeamSearch(设置 beamwidth 一次读多个数据块) 和缓存最常访问的节点(eg. by caching all vertices that are C = 3 or 4 hops from the starting point s)来加速查询。 另外,DiskANN 将邻居节点的向量保存在磁盘索引文件中,来提高搜索的精度(Implicit Re-Ranking Using Full-Precision Vectors)。

    +

    Code

    +

    Further readings