Commit Graph

14 Commits (8b682606e1564b659f56afae5e16e672ac6ab564)

Author SHA1 Message Date
Lucas Hosseini 3896b12c65
Facebook sync (Jun 2019) (#862)
Bugfixes:
- slow scanning of inverted lists (#836).

Features:
- add basic support for 6 new metrics in CPU `IndexFlat` and `IndexHNSW` (#848);
- add support for `IndexIDMap`/`IndexIDMap2` with binary indexes (#780).

Misc:
- throw python exception for OOM (#758);
- make `DistanceComputer` available for all random access indexes;
- gradually moving from `long` to `int64_t` for portability.
2019-06-19 15:59:06 +02:00
Lucas Hosseini a8118acbc5
Facebook sync (May 2019) + relicense (#838)
Changelog:

- changed license: BSD+Patents -> MIT
- propagates exceptions raised in sub-indexes of IndexShards and IndexReplicas
- support for searching several inverted lists in parallel (parallel_mode != 0)
- better support for PQ codes where nbit != 8 or 16
- IVFSpectralHash implementation: spectral hash codes inside an IVF
- 6-bit per component scalar quantizer (4 and 8 bit were already supported)
- combinations of inverted lists: HStackInvertedLists and VStackInvertedLists
- configurable number of threads for OnDiskInvertedLists prefetching (including 0=no prefetch)
- more test and demo code compatible with Python 3 (print with parentheses)
- refactored benchmark code: data loading is now in a single file
2019-05-28 16:17:22 +02:00
Lucas Hosseini afe0fdc161
Facebook sync (Mar 2019) (#756)
Facebook sync (Mar 2019)

- MatrixStats object
- option to round coordinates during k-means optimization
- alternative option for search in HNSW
- moved stats and imbalance_factor of IndexIVF to InvertedLists object
- range search for IVFScalarQuantizer
- direct unit8 codec in ScalarQuantizer
- renamed IndexProxy to IndexReplicas and moved to main Faiss
- better support for PQ code assignment with external index
- support for IMI2x16 (4B virtual centroids!)
- support for k = 2048 search on GPU (instead of 1024)
- most CUDA mem alloc failures throw exceptions instead of terminating on an assertion
- support for renaming an ondisk invertedlists
- interrupt computations with ctrl-C in python
2019-03-29 16:32:28 +01:00
Lucas Hosseini 323dbf3be3
Facebook sync (Dec 2018). (#660)
* Add GpuIndexBinaryFlat
* Add IndexBinaryHNSW
2018-12-19 17:48:35 +01:00
Lucas Hosseini 6880286ea0
Facebook sync (#504)
* Facebook sync

* Update swig wrappers.

* Fix comment.
2018-07-06 14:12:11 +02:00
pigghost 53876970f0 Update IndexIDMap reset (#401) 2018-05-02 13:40:58 +02:00
Matthijs Douze 0c482e54eb sync with FB version 2018-02-23 (#347)
- support on-disk IVF
2018-02-23 07:49:45 -08:00
matthijs 250a3d3f18 sync with FB version 2017-11-22
various bugfixes from github issues
kmean with some frozen centroids
GPU better tiling for large flat datasets
default AVX for vector ops
2017-11-22 05:11:28 -08:00
matthijs 8e3dc6f2b0 changed license 2017-07-30 00:18:45 -07:00
matthijs ca2ea5097b Sync with FB version 2017-07-16
- access GPU resources

- all Hamming distances of all sizes supported
2017-07-26 03:11:01 -07:00
matthijs f7aedbdfc0 sync with FB version 2017-07-18
- implemented ScalarQuantizer (without IVF)
- implemented update for IndexIVFFlat
- implemented L2 normalization preproc
2017-07-18 02:51:27 -07:00
matthijs 784e2facd8 Synchronization with FB version 2017-06-21
* moved most FAISS_ASSERT calls to C++ exceptions, and adjusted
  memory allocation to avoid mem leaks

* added an IndexIVFScalarQuantizer type that offers an
  intermediate compression between IVFFlat and IVFPQ

* support removal of indices in IndexIDMap / IndexFlat combination

* various fixes in GPU code
2017-06-21 09:01:06 -07:00
matthijs 00c288c0d9 added benchmarking scripts 2017-02-24 08:37:57 -08:00
Matthijs Douze c670118a4e Initial commit 2017-02-22 23:26:44 +01:00