Latest version
Dec 30, 2017 SpaceX Drone Ship Landing videos are a favourite for all of us. So what better way to celebrate Elon Musk than to try to badly mirror SpaceX's real life achievements. Of course we are doing this. Build spacecraft, fly them, and try to help the Kerbals fulfill their ultimate mission of conquering space. Sep 04, 2017 UVI Workstation updated to 2.6.12 and Falcon to 1.3.1 UVI has updated the free player UVI Workstation to version 2.6.12. Changes: Fix Issue when two different MIDI devices have the same name. UVI has also updated Falcon to version 1.3.1. Changes: Fix Wavetable SymForm mode when phase distortion is at 0. This is a mod that adds parts like the Dragon V2, Falcon 9 and more!Homepage. Sport S/T1 Suspension System (1.5' Leveling) w/ Falcon 3.1 Shocks for Wrangler JK The all-inclusive TeraFlex Sport S/T (Street & Trail) Suspensions Systems deliver performance and capability in a dependable and maintenance-free package.
Released:
A library for similarity search over high-dimensional data based on Locality-Sensitive Hashing (LSH)
Project description
FALCONN - FAst Lookups of Cosine and Other Nearest Neighbors
FALCONN is a library with algorithms for the nearest neighbor searchproblem. The algorithms in FALCONN are based on Locality-SensitiveHashing(LSH), which is a popular class of methods for nearest neighbor searchin high-dimensional spaces. The goal of FALCONN is to provide veryefficient and well-tested implementations of LSH-based data structures.
Currently, FALCONN supports two LSH families for the cosinesimilarity:hyperplane LSH and cross polytope LSH. Both hash families areimplemented with multi-probe LSH in order to minimize memory usage.Moreover, FALCONN is optimized for both dense and sparse data. Despitebeing designed for the cosine similarity, FALCONN can often be used fornearest neighbor search under the Euclidean distance or a maximum innerproduct search.
FALCONN is written in C++ and consists of several modular core classeswith a convenient wrapper around them. Many mathematical operations inFALCONN are vectorized through theEigen andFFHT libraries. The coreclasses of FALCONN rely ontemplates inorder to avoid runtime overhead.
How to use FALCONN
We provide a C++ interface for FALCONN as well as aPython wrapper (that usesNumPy). In the future, we plan to supportmore programming languages such as Julia.For C++, FALCONN is a header-only library and has no dependenciesbesides Eigen (which is also header-only), so FALCONN is easy to set up.For further details, please see ourdocumentation.
How fast is FALCONN?
On data sets with about 1 million points in around 100 dimensions,FALCONN typically requires a few milliseconds per query (running on areasonably modern desktop CPU).
For more detailed results, seeann-benchmarks of ErikBernhardsson. Let us point out that FALCONNis especially competitive, when the RAM budget is quite restrictive,which is not the regime the above benchmarks use.
Questions
Maybe your question is already answered in our Frequently AskedQuestions. If youhave additional questions about using FALCONN, we would be happy tohelp. Please send an email to falconn.lib@gmail.com.
Authors
FALCONN is mainly developed by IlyaRazenshteyn and LudwigSchmidt. FALCONN has grown outof a researchprojectwith our collaborators AlexandrAndoni, PiotrIndyk, and ThijsLaarhoven.
Many of the ideas used in FALCONN were proposed in research papers overthe past 20 years (see thedocumentation).
If you want to cite FALCONN in a publication, here is the bibliographicinformation of our research paper(bibtex):
Falcon 130 Motor
Practical and Optimal LSH for AngularDistanceAlexandr Andoni, Piotr Indyk, Thijs Laarhoven, Ilya Razenshteyn,Ludwig Schmidt NIPS 2015
License
FALCONN is available under the MITLicense (see LICENSE.txt). Notethat the third-party libraries in the external/ folder aredistributed under other open source licenses. The Eigen library islicensed under the MPL2.The googletest and googlemock libraries are licensed under the BSD3-Clause License. Thepybind11 library is licensed under a BSD-stylelicense.
Release historyRelease notifications
1.3.1
1.3.0
1.2.2
Falcon 1390 Parts List
1.2.1
1.2
Falcon 1390
1.1
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
| Filename, size | File type | Python version | Upload date | Hashes |
|---|---|---|---|---|
| Filename, size FALCONN-1.3.1.tar.gz (1.4 MB) | File type Source | Python version None | Upload date | Hashes |
Hashes for FALCONN-1.3.1.tar.gz
| Algorithm | Hash digest |
|---|---|
| SHA256 | 00cf778a97e9f2d9b501e86c16ff6f39528d70142302028675b3947e4a565eef |
| MD5 | ec7c6687d2c54e56b3ec78a4ebc9cc46 |
| BLAKE2-256 | 96b80d2c629d59398a7b3ed8726ce049abf6746bbf09d1ad15878d4fcf8048a6 |