Summary:
## Description
This PR added support for LSQ on GPU. Only the encoding part is running on GPU and the others are still running on CPU.
Multi-GPU is also supported.
## Usage
``` python
lsq = faiss.LocalSearchQuantizer(d, M, nbits)
ngpus = faiss.get_num_gpus()
lsq.icm_encoder_factory = faiss.GpuIcmEncoderFactory(ngpus) # we use all gpus
lsq.train(xt)
codes = lsq.compute_codes(xb)
decoded = lsq.decode(codes)
```
## Performance on SIFT1M
On 1 GPU:
```
===== lsq-gpu:
mean square error = 17337.878528
training time: 40.9857234954834 s
encoding time: 27.12640070915222 s
```
On 2 GPUs:
```
===== lsq-gpu:
mean square error = 17364.658176
training time: 25.832106113433838 s
encoding time: 14.879548072814941 s
```
On CPU:
```
===== lsq:
mean square error = 17305.880576
training time: 152.57522344589233 s
encoding time: 110.01779270172119 s
```
Pull Request resolved: https://github.com/facebookresearch/faiss/pull/1978
Test Plan: buck test mode/dev-nosan //faiss/gpu/test/:test_gpu_index_py -- TestLSQIcmEncoder
Reviewed By: wickedfoo
Differential Revision: D29609763
Pulled By: mdouze
fbshipit-source-id: b6ffa2a3c02bf696a4e52348132affa0dd838870
Summary:
## Description
The process of updating the codebook in LSQ may be unstable if the data is not zero-centering. This diff fixed it by using `double` instead of `float` during codebook updating. This would not affect the performance since the update process is quite fast.
Users could switch back to `float` mode by setting `update_codebooks_with_double = False`
## Changes
1. Support `double` during codebook updating.
2. Add a unit test.
3. Add `__init__.py` under `contrib/` to avoid warnings.
Pull Request resolved: https://github.com/facebookresearch/faiss/pull/1975
Reviewed By: wickedfoo
Differential Revision: D29565632
Pulled By: mdouze
fbshipit-source-id: 932d7932ae9725c299cd83f87495542703ad6654
Summary:
Pull Request resolved: https://github.com/facebookresearch/faiss/pull/1906
This PR implemented LSQ/LSQ++, a vector quantization technique described in the following two papers:
1. Revisiting additive quantization
2. LSQ++: Lower running time and higher recall in multi-codebook quantization
Here is a benchmark running on SIFT1M for 64 bits encoding:
```
===== lsq:
mean square error = 17335.390208
training time: 312.729779958725 s
encoding time: 244.6277096271515 s
===== pq:
mean square error = 23743.004672
training time: 1.1610801219940186 s
encoding time: 2.636141061782837 s
===== rq:
mean square error = 20999.737344
training time: 31.813055515289307 s
encoding time: 307.51959800720215 s
```
Changes:
1. Add LocalSearchQuantizer object
2. Fix an out of memory bug in ResidualQuantizer
3. Add a benchmark for evaluating quantizers
4. Add tests for LocalSearchQuantizer
Pull Request resolved: https://github.com/facebookresearch/faiss/pull/1862
Test Plan:
```
buck test //faiss/tests/:test_lsq
buck run mode/opt //faiss/benchs/:bench_quantizer -- lsq pq rq
```
Reviewed By: beauby
Differential Revision: D28376369
Pulled By: mdouze
fbshipit-source-id: 2a394d38bf75b9de0a1c2cd6faddf7dd362a6fa8
Summary:
## Description:
This diff implemented Navigating Spreading-out Graph (NSG) which accepts a KNN graph as input.
Here is the interface of building an NSG graph:
``` c++
void IndexNSG::build(idx_t n, const float *x, idx_t *knn_graph, int GK);
```
where `GK` is the nb of neighbors per node and `knn_graph[i * GK + j]` is the j-th neighbor of node i.
The `add` method is not implemented yet.
The unit tests could be found in `tests/test_nsg.cpp`.
mdouze beauby Maybe I need some advice on how to design the interface and support python.
Pull Request resolved: https://github.com/facebookresearch/faiss/pull/1707
Test Plan: buck test //faiss/tests/:test_index -- TestNSG
Reviewed By: beauby
Differential Revision: D26748498
Pulled By: mdouze
fbshipit-source-id: 3280f705fb1b5f9c8cc5efeba63b904c3b832544
Summary:
IndexPQ and IndexIVFPQ implementations with AVX shuffle instructions.
The training and computing of the codes does not change wrt. the original PQ versions but the code layout is "packed" so that it can be used efficiently by the SIMD computation kernels.
The main changes are:
- new IndexPQFastScan and IndexIVFPQFastScan objects
- simdib.h for an abstraction above the AVX2 intrinsics
- BlockInvertedLists for invlists that are 32-byte aligned and where codes are not sequential
- pq4_fast_scan.h/.cpp: for packing codes and look-up tables + optmized distance comptuation kernels
- simd_result_hander.h: SIMD version of result collection in heaps / reservoirs
Misc changes:
- added contrib.inspect_tools to access fields in C++ objects
- moved .h and .cpp code for inverted lists to an invlists/ subdirectory, and made a .h/.cpp for InvertedListsIOHook
- added a new inverted lists type with 32-byte aligned codes (for consumption by SIMD)
- moved Windows-specific intrinsics to platfrom_macros.h
Pull Request resolved: https://github.com/facebookresearch/faiss/pull/1542
Test Plan:
```
buck test mode/opt -j 4 //faiss/tests/:test_fast_scan_ivf //faiss/tests/:test_fast_scan
buck test mode/opt //faiss/manifold/...
```
Reviewed By: wickedfoo
Differential Revision: D25175439
Pulled By: mdouze
fbshipit-source-id: ad1a40c0df8c10f4b364bdec7172e43d71b56c34
Summary:
This diff streamlines the code that collects results for brute force distance computations for the L2 / IP and range search / knn search combinations.
It introduces a `ResultHandler` template class that abstracts what happens with the computed distances and ids. In addition to the heap result handler and the range search result handler, it introduces a reservoir result handler that improves the search speed for large k (>=100).
Benchmark results (https://fb.quip.com/y0g1ACLEqJXx#OCaACA2Gm45) show that on small datasets (10k) search is 10-50% faster (improvements are larger for small k). There is room for improvement in the reservoir implementation, whose implementation is quite naive currently, but the diff is already useful in its current form.
Experiments on precomputed db vector norms for L2 distance computations were not very concluding performance-wise, so the implementation is removed from IndexFlatL2.
This diff also removes IndexL2BaseShift, which was never used.
Pull Request resolved: https://github.com/facebookresearch/faiss/pull/1502
Test Plan:
```
buck test //faiss/tests/:test_product_quantizer
buck test //faiss/tests/:test_index -- TestIndexFlat
```
Reviewed By: wickedfoo
Differential Revision: D24705464
Pulled By: mdouze
fbshipit-source-id: 270e10b19f3c89ed7b607ec30549aca0ac5027fe
Summary:
Legacy __print__ statements are syntax errors in Python 3 but __print()__ function works as expected in both Python 2 and Python 3.
Pull Request resolved: https://github.com/facebookresearch/faiss/pull/1443
Reviewed By: LowikC
Differential Revision: D24157415
Pulled By: mdouze
fbshipit-source-id: 4ec637aa26b61272e5337d47b7796a330ce25bad
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.
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
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
Features:
- automatic tracking of C++ references in Python
- non-intel platforms supported -- some functions optimized for ARM
- override nprobe for concurrent searches
- support for floating-point quantizers in binary indexes
Bug fixes:
- no more segfaults in python (I know it's the same as the first feature but it's important!)
- fix GpuIndexIVFFlat issues for float32 with 64 / 128 dims
- fix sharding of flat indexes on GPU with index_cpu_to_gpu_multiple
Chengqi Deng
Matthijs Douze
Lucas Hosseini
Lucas Hosseini
shengjun.li
Hap-Hugh
cclauss