hnswlib
hnswlib copied to clipboard
Header-only C++/python library for fast approximate nearest neighbors
I'm integrating hnswlib into my own code, I have the following suggestions: 1. I think that the data passed to `addPoint` should be `const void *` . https://github.com/nmslib/hnswlib/blob/44f20f0b2257af1b3251ac9ed976a6d6583967c0/hnswlib/hnswalg.h#L773 2. I...
https://github.com/nmslib/hnswlib/blob/68c5531caa8be522b19ac20088ffca39d1474520/hnswlib/hnswalg.h#L842-L854 This is different from the algorithm in the paper  I think that `currObj` should be set to `top_candidates.top().second`.
My application occasionally met a case that when all neighbors of entorpoint are marked delete and are further to the query point. The search process will terminate and only return...
A typo
I think it maybe a typo in file "space_l2.h" at line 225: ``` data_size_ = dim * sizeof(unsigned char); ``` why not: ``` data_size_ = dim * sizeof(int); ```
I want to know how many distance calculations are needed to complete a query with fixed parameters
Hi, I want to know how many distance calculations are needed to complete a query with fixed parameters. How can I get this value? Do I need to change the...
When specifying the option space='ip', the same algorithm that is being used for 'l2' distance is used? Or that there is a reduction to a metric space and then HNSW...
I have tried this algorithm and flann on low dimensional dataset(3-d point cloud). Here are 800000+ map points, and 300 query points, and find the query points' k neighbor with...
I build an index with hnswlib in inner product space(normalized data),the data dimension is 128,m= 60, ef_construction = 400。I use random vec(normalized) to test recall performance,I get the result below:...
Hi @yurymalkov & @searchivarius, I have a dataset want to use hnswlib in inner product space(not cosine), it's dimension is 100. I build a index(M = 64, efc = 5000,...
In cosine space, `get_items` returns some irrelevant vectors instead of original ones. ``` import hnswlib import numpy as np dim = 128 num_elements = 10000 # Generating sample data data...