Benchmarking nearest neighbors

Doing fast searching of nearest neighbors in high dimensional spaces is an increasingly important problem, but so far there has not been a lot of empirical attempts at comparing approaches in an objective way.

This project contains some tools to benchmark various implementations of approximate nearest neighbor (ANN) search for different metrics. We have pregenerated datasets (in HDF5) formats and we also have Docker containers for each algorithm. There's a test suite that makes sure every algorithm works.

Evaluated

Annoy
FLANN
scikit-learn: LSHForest, KDTree, BallTree
PANNS
NearPy
KGraph
NMSLIB (Non-Metric Space Library): SWGraph, HNSW, BallTree, MPLSH
hnswlib (a part of nmslib project)
RPForest
FAISS
DolphinnPy
Datasketch
PyNNDescent
MRPT
NGT: ONNG, PANNG
SPTAG
PUFFINN
N2
ScaNN
Elastiknn

Data sets

We have a number of precomputed data sets for this. All data sets are pre-split into train/test and come with ground truth data in the form of the top 100 neighbors. We store them in a HDF5 format:

Dataset	Dimensions	Train size	Test size	Neighbors	Distance	Download
DEEP1B	96	9,990,000	10,000	100	Angular	HDF5 (3.6GB)
Fashion-MNIST	784	60,000	10,000	100	Euclidean	HDF5 (217MB)
GIST	960	1,000,000	1,000	100	Euclidean	HDF5 (3.6GB)
GloVe	25	1,183,514	10,000	100	Angular	HDF5 (121MB)
GloVe	50	1,183,514	10,000	100	Angular	HDF5 (235MB)
GloVe	100	1,183,514	10,000	100	Angular	HDF5 (463MB)
GloVe	200	1,183,514	10,000	100	Angular	HDF5 (918MB)
Kosarak	27983	74,962	500	100	Jaccard	HDF5 (2.0GB)
MNIST	784	60,000	10,000	100	Euclidean	HDF5 (217MB)
NYTimes	256	290,000	10,000	100	Angular	HDF5 (301MB)
SIFT	128	1,000,000	10,000	100	Euclidean	HDF5 (501MB)
Last.fm	65	292,385	50,000	100	Angular	HDF5 (135MB)

Results

These are all as of 2020-07-12, running all benchmarks on a c5.4xlarge machine on AWS with --parallelism set to 3:

glove-100-angular

sift-128-euclidean

fashion-mnist-784-euclidean

lastfm-64-dot

nytimes-256-angular

glove-25-angular

Install

The only prerequisite is Python (tested with 3.6) and Docker.

Clone the repo.
Run pip install -r requirements.txt.
Run python install.py to build all the libraries inside Docker containers (this can take a while, like 10-30 minutes).

Running

Run python run.py (this can take an extremely long time, potentially days)
Run python plot.py or python create_website.py to plot results.

You can customize the algorithms and datasets if you want to:

Check that algos.yaml contains the parameter settings that you want to test
To run experiments on SIFT, invoke python run.py --dataset glove-100-angular. See python run.py --help for more information on possible settings. Note that experiments can take a long time.
To process the results, either use python plot.py --dataset glove-100-angular or python create_website.py. An example call: python create_website.py --plottype recall/time --latex --scatter --outputdir website/.

Including your algorithm

Add your algorithm into ann_benchmarks/algorithms by providing a small Python wrapper.
Add a Dockerfile in install/ for it
Add it to algos.yaml
Add it to .travis.yml

Principles

Everyone is welcome to submit pull requests with tweaks and changes to how each library is being used.
In particular: if you are the author of any of these libraries, and you think the benchmark can be improved, consider making the improvement and submitting a pull request.
This is meant to be an ongoing project and represent the current state.
Make everything easy to replicate, including installing and preparing the datasets.
Try many different values of parameters for each library and ignore the points that are not on the precision-performance frontier.
High-dimensional datasets with approximately 100-1000 dimensions. This is challenging but also realistic. Not more than 1000 dimensions because those problems should probably be solved by doing dimensionality reduction separately.
Single queries are used by default. ANN-Benchmarks enforces that only one CPU is saturated during experimentation, i.e., no multi-threading. A batch mode is available that provides all queries to the implementations at once. Add the flag --batch to run.py and plot.py to enable batch mode.
Avoid extremely costly index building (more than several hours).
Focus on datasets that fit in RAM. Out of core ANN could be the topic of a later comparison.
We mainly support CPU-based ANN algorithms. GPU support exists for FAISS, but it has to be compiled with GPU support locally and experiments must be run using the flags --local --batch.
Do proper train/test set of index data and query points.
Note that set similarity was supported in the past. This might hopefully be added back soon.

Authors

Built by Erik Bernhardsson with significant contributions from Martin Aumüller and Alexander Faithfull.

Related Publication

The following publication details design principles behind the benchmarking framework:

M. Aumüller, E. Bernhardsson, A. Faithfull: ANN-Benchmarks: A Benchmarking Tool for Approximate Nearest Neighbor Algorithms. Information Systems 2019. DOI: 10.1016/j.is.2019.02.006

chunelfeng / ann-benchmarks Goto Github PK

ann-benchmarks's Introduction