IVF Index: Clustering-Based Vector Search & Partitioning › Key Parameters
Michael BrenndoerferJanuary 27, 20261 min read
Language AI HandbookAll chapters Part 44 of 67Retrieval-Augmented Generation
IVF Index: Clustering-Based Vector Search & Partitioning › Key ParametersOne section · Sections in this chapter11 / 12 ↓
Sections in this chapter12
Reading tools Definitions & inline links
Choose your expertise level to adjust how many terms are explained. Beginners see more tooltips, experts see fewer to maintain reading flow. Hover over underlined terms for instant definitions.
Article links
Make inline references clickable
The key parameters for IVF are:
- k (number of centroids): Controls the granularity of space partitioning. Typical values range from to where is the dataset size. More centroids create smaller cells but increase centroid comparison cost.
- nprobe: Number of cells to search at query time. Controls the accuracy-speed tradeoff. Higher values improve recall but increase query latency linearly.
- n_init: Number of k-means initializations during training. Higher values improve clustering quality but increase training time.
- max_iter: Maximum iterations for k-means convergence. Typical values are 20-300 depending on dataset size and convergence requirements.