mirror of
https://github.com/facebookresearch/faiss.git
synced 2026-10-12 02:13:36 +00:00
Saim-naushad
506600962f
Summary: This PR adds a Python HNSW tutorial corresponding to the existing C++ HNSW tutorial. The example creates synthetic vectors, builds an IndexHNSWFlat index, runs a search, and prints the nearest-neighbor IDs and distances for the final queries. Pull Request resolved: https://github.com/facebookresearch/faiss/pull/5260 Reviewed By: mnorris11 Differential Revision: D107272616 Pulled By: alibeklfc fbshipit-source-id: 721d7a1a95aa79150d6cb9653bd803065c82dcd7
42 lines
860 B
Python
42 lines
860 B
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 numpy as np
|
|
|
|
import faiss
|
|
|
|
|
|
d = 64 # dimension
|
|
nb = 100000 # database size
|
|
nq = 10000 # number of queries
|
|
|
|
np.random.seed(1234)
|
|
|
|
xb = np.random.random((nb, d)).astype("float32")
|
|
xb[:, 0] += np.arange(nb) / 1000.0
|
|
|
|
xq = np.random.random((nq, d)).astype("float32")
|
|
xq[:, 0] += np.arange(nq) / 1000.0
|
|
|
|
k = 4
|
|
|
|
index = faiss.IndexHNSWFlat(d, 32) # M=32: neighbors per node
|
|
index.hnsw.efConstruction = 40 # graph construction quality
|
|
index.add(xb)
|
|
index.hnsw.efSearch = 64 # higher = better recall, slower
|
|
|
|
# sanity check
|
|
D, I = index.search(xb[:5], k)
|
|
print(I)
|
|
print(D)
|
|
|
|
D, I = index.search(xq, k)
|
|
|
|
print("I=")
|
|
print(I[-5:])
|
|
|
|
print("D=")
|
|
print(D[-5:])
|