Info
P2HNNS-benchmarks is a benchmarking environment for approximate nearest neighbor search algorithms. This website contains the current benchmarking results. Please visit P2HNNS-benchmarks to get an overview of the evaluated datasets and algorithms. Make a pull request on GitHub to add your own code or improvements to the benchmarking system. We acknowledge and give full credit to the original ANN-BENCHMARKS repository, developed by Martin Aumueller, Erik Bernhardsson, and Alec Faitfull, from which this project is forked.
Benchmarking Results
Results are split by distance measure and dataset. In the bottom, you can find an overview of an algorithm's performance on all datasets. Each dataset is annoted by (k = ...), the number of nearest neighbors an algorithm was supposed to return. The plot shown depicts Recall (the fraction of true nearest neighbors found, on average over all queries) against Queries per second. Clicking on a plot reveils detailled interactive plots, including approximate recall, index size, and build time.
Benchmarks for Single Queries
Results by Dataset
Distance: Euclidean
cifar10-512-euclidean-med (k = 10)
gist-960-euclidean-med (k = 10)
glove-100-euclidean-med (k = 10)
glove-25-euclidean (k = 10)
glove-25-euclidean-med (k = 10)
glove-25-euclidean-med-bctree (k = 10)
glove-25-euclidean-med-rpsd (k = 10)
trevi-4096-euclidean-med (k = 10)
Results by Algorithm
mh
mqh-kjl
balltree
bt-mqh
mh-mqh
mqh
nh
fh
bctree
Contact
P2HNNS-benchmarks is a fork of the original ANN-BENCHMARKS repository, which was developed by Martin Aumueller (maau@itu.dk), Erik Bernhardsson (mail@erikbern.com), and Alec Faitfull (alef@itu.dk). This fork, reconfigured by Søren Majlund Jensen and Christian Porsmose Stender, includes minor modifications to address the P2HNNS problem. We acknowledge and give full credit to the original developers for their work. Please use GitHub to submit your implementations or improvements.