Package Galaxy

pypi package 'faiss-gpu'

Popularity: High (more popular than 99% of all packages)
Description: A library for efficient similarity search and clustering of dense vectors.
Installation: pip install faiss-gpu
Last version: 1.7.2 (Download)
Homepage: https://github.com/kyamagu/faiss-wheels
Size: 83,482.84 kB
License: MIT
Keywords: search, nearest, neighbors

Activity

Last modified: January 11, 2022 7:09 AM (a year ago)
Versions released in one year: 0
Weekly downloads: 25,761
04/03/202206/19/202209/04/202211/20/202202/05/2023050,000100,000150,000200,000released versions / week
  • Versions released
  • Weekly downloads

What's new in version 1.7.2

Delta between version 1.7.1.post3 and version 1.7.2

Source: Github
Commits:
  • 19cfce357742b671b4651e3ac0fe7a40e4702d37, January 11, 2022 4:54 AM:
    Update faiss to v1.7.2 release (#51)

Update faiss to v1.7.2 release
Files changed:
.github/workflows/build.yml CHANGED
@@ -64,7 +64,7 @@ jobs:
64
  - name: Build wheels
65
  uses: pypa/cibuildwheel@v2.1.1
66
  env:
67
- CIBW_BUILD: cp36-* cp37-* cp38-* cp39-* cp310-*
68
  CIBW_ARCHS: ${{ matrix.arch }}
69
  CIBW_MANYLINUX_X86_64_IMAGE: manylinux2014
70
  CIBW_ENVIRONMENT: >
@@ -80,7 +80,7 @@ jobs:
80
  LIB="${LIB};${CMAKE_PREFIX_PATH}\\lib;${CONDA}\\Library\\lib"
81
  CPATH="${CPATH};${CMAKE_PREFIX_PATH}\\include;${CONDA}\\Library\\include"
82
  CIBW_BEFORE_ALL: bash scripts/build_${{ runner.os }}.sh
83
- CIBW_BEFORE_BUILD_WINDOWS: pip install delvewheel pefile==2021.5.24
84
  CIBW_REPAIR_WHEEL_COMMAND_WINDOWS: delvewheel repair -v -w {dest_dir} {wheel}
85
  CIBW_TEST_REQUIRES: pytest scipy
86
  CIBW_TEST_COMMAND: >
64
  - name: Build wheels
65
  uses: pypa/cibuildwheel@v2.1.1
66
  env:
67
+ CIBW_SKIP: pp*
68
  CIBW_ARCHS: ${{ matrix.arch }}
69
  CIBW_MANYLINUX_X86_64_IMAGE: manylinux2014
70
  CIBW_ENVIRONMENT: >
80
  LIB="${LIB};${CMAKE_PREFIX_PATH}\\lib;${CONDA}\\Library\\lib"
81
  CPATH="${CPATH};${CMAKE_PREFIX_PATH}\\include;${CONDA}\\Library\\include"
82
  CIBW_BEFORE_ALL: bash scripts/build_${{ runner.os }}.sh
83
+ CIBW_BEFORE_BUILD_WINDOWS: pip install delvewheel
84
  CIBW_REPAIR_WHEEL_COMMAND_WINDOWS: delvewheel repair -v -w {dest_dir} {wheel}
85
  CIBW_TEST_REQUIRES: pytest scipy
86
  CIBW_TEST_COMMAND: >
faiss CHANGED
@@ -1 +1 @@
1
- Subproject commit b4eb51dae81084b29ca77834fd9b0537045853e5
1
+ Subproject commit c08cbff1a4d6c9afb6b8f69004c5530aaf80237a
scripts/build_Linux.sh CHANGED
@@ -7,7 +7,8 @@ FAISS_OPT_LEVEL=${FAISS_OPT_LEVEL:-"generic"}
7
 
8
  CUDA_VERSION="10.0"
9
  CUDA_PKG_VERSION="10-0-10.0.130-1"
10
- CUBLAS_PKG_VERSION="10-0-10.0.130-1"
 
11
  NVIDIA_REPO_URL="http://developer.download.nvidia.com/compute/cuda/repos/rhel7/x86_64/cuda-rhel7.repo"
12
  CMAKE_CUDA_ARCHITECTURES="35-real;50-real;60-real;70-real;75"
13
 
@@ -25,6 +26,7 @@ if [[ ${FAISS_ENABLE_GPU} == "ON" ]]; then
25
  yum -y install \
26
  cuda-command-line-tools-${CUDA_PKG_VERSION} \
27
  cuda-cublas-dev-${CUBLAS_PKG_VERSION} \
 
28
  devtoolset-7-gcc \
29
  devtoolset-7-gcc-c++ \
30
  devtoolset-7-gcc-gfortran \
7
 
8
  CUDA_VERSION="10.0"
9
  CUDA_PKG_VERSION="10-0-10.0.130-1"
10
+ CUBLAS_PKG_VERSION=${CUDA_PKG_VERSION}
11
+ CURAND_PKG_VERSION=${CUDA_PKG_VERSION}
12
  NVIDIA_REPO_URL="http://developer.download.nvidia.com/compute/cuda/repos/rhel7/x86_64/cuda-rhel7.repo"
13
  CMAKE_CUDA_ARCHITECTURES="35-real;50-real;60-real;70-real;75"
14
 
26
  yum -y install \
27
  cuda-command-line-tools-${CUDA_PKG_VERSION} \
28
  cuda-cublas-dev-${CUBLAS_PKG_VERSION} \
29
+ cuda-curand-dev-${CURAND_PKG_VERSION} \
30
  devtoolset-7-gcc \
31
  devtoolset-7-gcc-c++ \
32
  devtoolset-7-gcc-gfortran \
setup.py CHANGED
@@ -5,7 +5,7 @@
5
  import os
6
 
7
  NAME = 'faiss-cpu'
8
- VERSION = '1.7.1.post3'
9
 
10
  LONG_DESCRIPTION = """
11
  Faiss is a library for efficient similarity search and clustering of dense
@@ -55,6 +55,7 @@ def __str__(self):
55
  '-Doverride=',
56
  '-I' + FAISS_INCLUDE,
57
  '-I' + FAISS_ROOT,
 
58
  ]
59
 
60
  if sys.platform == 'win32':
5
  import os
6
 
7
  NAME = 'faiss-cpu'
8
+ VERSION = '1.7.2'
9
 
10
  LONG_DESCRIPTION = """
11
  Faiss is a library for efficient similarity search and clustering of dense
55
  '-Doverride=',
56
  '-I' + FAISS_INCLUDE,
57
  '-I' + FAISS_ROOT,
58
+ '-doxygen',
59
  ]
60
 
61
  if sys.platform == 'win32':

Readme

Faiss is a library for efficient similarity search and clustering of dense vectors. It contains algorithms that search in sets of vectors of any size, up to ones that possibly do not fit in RAM. It also contains supporting code for evaluation and parameter tuning. Faiss is written in C++ with complete wrappers for Python/numpy. It is developed by Facebook AI Research.