mirror of
https://github.com/facebookresearch/faiss.git
synced 2026-10-12 02:13:36 +00:00
Subhadeep Karan
451ca8429d
Summary: Pull Request resolved: https://github.com/facebookresearch/faiss/pull/4666 * `for OpenMP support support` → `for OpenMP support` * `neccesary` → `necessary` * `journal= {arXiv:1702.08734},,` → `journal = {arXiv:1702.08734},` * `verinsion` → `version` * `consistency_check--config` → `consistency_check --config` * `etracted` → `extracted` (2 occurrences) * `sb` → `xb` (variable reference) * `statists` → `statistics` * `elemnts` → `elements` Reviewed By: alibeklfc Differential Revision: D86474214 fbshipit-source-id: f63272423b0a767f7576cd433004c87d85e2d740
39 lines
1.3 KiB
Python
39 lines
1.3 KiB
Python
# Copyright (c) Meta Platforms, Inc. and affiliates.
|
|
#
|
|
# This source code is licensed under the MIT license found in the
|
|
# LICENSE file in the root directory of this source tree.
|
|
|
|
import faiss
|
|
import numpy as np
|
|
|
|
d = 64 # dimension
|
|
nb = 100000 # database size
|
|
nq = 10000 # nb of queries
|
|
np.random.seed(1234) # make reproducible
|
|
xb = np.random.random((nb, d)).astype('float32') # 64-dim *nb queries
|
|
xb[:, 0] += np.arange(nb) / 1000.
|
|
xq = np.random.random((nq, d)).astype('float32')
|
|
xq[:, 0] += np.arange(nq) / 1000.
|
|
|
|
m = 8 # 8 specifies that the number of sub-vector is 8
|
|
k = 4 # number of dimension in extracted vector
|
|
n_bit = 4 # 4 specifies that each sub-vector is encoded as 4 bits
|
|
bbs = 32 # build block size ( bbs % 32 == 0 ) for PQ
|
|
|
|
index = faiss.IndexPQFastScan(d, m, n_bit, faiss.METRIC_L2)
|
|
index_refine = faiss.IndexRefineFlat(index)
|
|
# construct FastScan and run index refinement
|
|
|
|
assert not index_refine.is_trained
|
|
index_refine.train(xb) # Train vectors data index within mockup database
|
|
assert index_refine.is_trained
|
|
|
|
index_refine.add(xb)
|
|
params = faiss.IndexRefineSearchParameters(k_factor=3)
|
|
D, I = index_refine.search(xq[:5], 10, params=params)
|
|
print(I)
|
|
print(D)
|
|
index.nprobe = 10 # make comparable with experiment above
|
|
D, I = index.search(xq[:5], k) # search
|
|
print(I[-5:])
|