mirror of
https://github.com/facebookresearch/faiss.git
synced 2025-06-03 21:54:02 +08:00
383b5d908c
Summary: `torch.load` without `weights_only` parameter is unsafe. Explicitly set `weights_only` to False only if you trust the data you load and full pickle functionality is needed, otherwise set `weights_only=True`. If `weights_only=True` doesn't work for some cases, then explicit `weights_only=False` should be used. Found with https://github.com/pytorch-labs/torchfix/ Pull Request resolved: https://github.com/facebookresearch/faiss/pull/3796 Reviewed By: asadoughi Differential Revision: D61824340 Pulled By: kit1980 fbshipit-source-id: bc013d06d4f368f730ffee6898e75fd0b0ff1d40
78 lines
2.3 KiB
Python
78 lines
2.3 KiB
Python
# Copyright (c) Facebook, Inc. and its affiliates.
|
|
#
|
|
# This source code is licensed under the MIT license found in the
|
|
# LICENSE file in the root directory of this source tree.
|
|
|
|
"""
|
|
This demonstrates how to reproduce the QINCo paper results using the Faiss
|
|
QINCo implementation. The code loads the reference model because training
|
|
is not implemented in Faiss.
|
|
|
|
Prepare the data with
|
|
|
|
cd /tmp
|
|
|
|
# get the reference qinco code
|
|
git clone https://github.com/facebookresearch/Qinco.git
|
|
|
|
# get the data
|
|
wget https://dl.fbaipublicfiles.com/QINCo/datasets/bigann/bigann1M.bvecs
|
|
|
|
# get the model
|
|
wget https://dl.fbaipublicfiles.com/QINCo/models/bigann_8x8_L2.pt
|
|
|
|
"""
|
|
|
|
import numpy as np
|
|
from faiss.contrib.vecs_io import bvecs_mmap
|
|
import sys
|
|
import time
|
|
import torch
|
|
import faiss
|
|
|
|
# make sure pickle deserialization will work
|
|
sys.path.append("/tmp/Qinco")
|
|
import model_qinco
|
|
|
|
with torch.no_grad():
|
|
|
|
qinco = torch.load("/tmp/bigann_8x8_L2.pt", weights_only=False)
|
|
qinco.eval()
|
|
# print(qinco)
|
|
if True:
|
|
torch.set_num_threads(1)
|
|
faiss.omp_set_num_threads(1)
|
|
|
|
x_base = bvecs_mmap("/tmp/bigann1M.bvecs")[:1000].astype('float32')
|
|
x_scaled = torch.from_numpy(x_base) / qinco.db_scale
|
|
|
|
t0 = time.time()
|
|
codes, _ = qinco.encode(x_scaled)
|
|
x_decoded_scaled = qinco.decode(codes)
|
|
print(f"Pytorch encode {time.time() - t0:.3f} s")
|
|
# multi-thread: 1.13s, single-thread: 7.744
|
|
|
|
x_decoded = x_decoded_scaled.numpy() * qinco.db_scale
|
|
|
|
err = ((x_decoded - x_base) ** 2).sum(1).mean()
|
|
print("MSE=", err) # = 14211.956, near the L=2 result in Fig 4 of the paper
|
|
|
|
qinco2 = faiss.QINCo(qinco)
|
|
t0 = time.time()
|
|
codes2 = qinco2.encode(faiss.Tensor2D(x_scaled))
|
|
x_decoded2 = qinco2.decode(codes2).numpy() * qinco.db_scale
|
|
print(f"Faiss encode {time.time() - t0:.3f} s")
|
|
# multi-thread: 3.2s, single thread: 7.019
|
|
|
|
# these tests don't work because there are outlier encodings
|
|
# np.testing.assert_array_equal(codes.numpy(), codes2.numpy())
|
|
# np.testing.assert_allclose(x_decoded, x_decoded2)
|
|
|
|
ndiff = (codes.numpy() != codes2.numpy()).sum() / codes.numel()
|
|
assert ndiff < 0.01
|
|
ndiff = (((x_decoded - x_decoded2) ** 2).sum(1) > 1e-5).sum()
|
|
assert ndiff / len(x_base) < 0.01
|
|
|
|
err = ((x_decoded2 - x_base) ** 2).sum(1).mean()
|
|
print("MSE=", err) # = 14213.551
|