| Name | faiss-gpu-cu12 JSON |
| Version |
1.12.0
JSON |
| download |
| home_page | None |
| Summary | A library for efficient similarity search and clustering of dense vectors. |
| upload_time | 2025-08-24 14:25:19 |
| maintainer | None |
| docs_url | None |
| author | None |
| requires_python | <3.13,>=3.9 |
| license | MIT License
Copyright (c) 2024 Di-Is
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
The major design pattern of this repository was based in Kota Yamaguchi's faiss-wheels, which is subject to the same license.
Here is the original copyright notice for faiss-wheels:
MIT License
Copyright (c) 2019 Kota Yamaguchi
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
This repository uses meta's faiss and is subject to the same license.
Here is the original copyright notice for faiss:
MIT License
Copyright (c) Facebook, Inc. and its affiliates.
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE. |
| keywords |
search
nearest
neighbors
|
| VCS |
 |
| bugtrack_url |
|
| requirements |
No requirements were recorded.
|
| Travis-CI |
No Travis.
|
| coveralls test coverage |
No coveralls.
|
# faiss-wheels
This repository is based on [kyamagu/faiss-wheels](https://github.com/kyamagu/faiss-wheels).
[](https://pypi.org/project/faiss-gpu-cu11/)
[](https://pypi.org/project/faiss-gpu-cu12/)
## Overview
This repository provides scripts to build GPU-enabled wheels for the [faiss](https://github.com/facebookresearch/faiss) library.
Distributes `faiss-gpu-cuXX` packages to PyPI using the contents of this repository.
### Key Features
* **No local CUDA installation required** - Dynamically links to CUDA Runtime and cuBLAS libraries from PyPI
* Builds CUDA 11.8+ and CUDA 12.1+ compatible wheels
* Supports Volta to Ada Lovelace architecture GPUs (Compute Capability 7.0–8.9)
* Bundles OpenBLAS in Linux
* Reduces wheel file size through dynamic linking instead of static compilation
## Important Requirements
The published `faiss-gpu-cuXX` packages require proper system setup that cannot be managed by pip. It is your responsibility to prepare a suitable environment:
1. **NVIDIA Driver**: Your host must have a CUDA-compatible NVIDIA driver installed
* The minimum driver version depends on the CUDA version that gets installed
* NVIDIA drivers are backward compatible with older CUDA versions ([See CUDA Compatibility Documentation](https://docs.nvidia.com/deploy/cuda-compatibility/))
2. **GPU Architecture**: Your GPU must be compatible (Compute Capability 7.0–8.9)
* Supported: Volta, Turing, Ampere, Ada Lovelace
3. **Library Compatibility**: If you install multiple CUDA-dependent libraries (e.g., PyTorch) in the same environment, they must link to the same CUDA version
## GPU Architecture Support for PyPI Packages
### Support Policy for `faiss-gpu-cu11` and `faiss-gpu-cu12`
**Note**: This is an **unofficial, personal development project** with limited computational resources. Due to these constraints, comprehensive testing across all NVIDIA GPU architectures is not feasible. The pre-built `faiss-gpu-cu11` and `faiss-gpu-cu12` packages on PyPI aim to support the same GPU architecture range (Compute Capability 7.0–8.9) as the official Faiss repository.
### For Unsupported GPU Architectures
If you have a GPU architecture that is not supported by these pre-built wheels:
1. **Official Faiss**: Follow the [official Faiss repository build instructions](https://github.com/facebookresearch/faiss/blob/main/INSTALL.md)
2. **Build from Source**: Use this repository's code to build wheels for your specific architecture (see [Building from Source](#building-from-source) section)
## Installation
The `faiss-gpu-cu11` and `faiss-gpu-cu12` wheels are available on PyPI. Choose the appropriate version for your CUDA environment.
### For CUDA 12
```bash
# Install with fixed CUDA 12.1 (requires NVIDIA Driver ≥R530)
pip install 'faiss-gpu-cu12[fix-cuda]'
# Install with CUDA 12.X (X≥1) - allows flexibility but driver requirement varies
pip install faiss-gpu-cu12
```
**Details:**
* `faiss-gpu-cu12` is built with CUDA Toolkit 12.1 and maintains minor version compatibility
* With `[fix-cuda]`: Installs exactly CUDA 12.1, requiring NVIDIA Driver ≥R530
* Without `[fix-cuda]`: Allows any CUDA 12.X (X≥1), driver requirement depends on the actual CUDA version installed
* For example: CUDA 12.4 requires Driver ≥R550
* Use without `[fix-cuda]` when integrating with other CUDA-dependent packages (e.g., PyTorch with CUDA 12.4)
**System Requirements:**
* OS: Linux x86_64 (glibc ≥2.17)
* GPU: Compute Capability 7.0–8.9
### For CUDA 11
```bash
# Install with CUDA 11.8 (requires NVIDIA Driver ≥R520)
pip install faiss-gpu-cu11[fix-cuda]
# Same as above (CUDA 11.8 is the final version)
pip install faiss-gpu-cu11
```
**Details:**
* `faiss-gpu-cu11` is built with CUDA Toolkit 11.8
* Both commands install CUDA 11.8 since no newer CUDA 11.X versions exist
* Requires NVIDIA Driver ≥R520
**System Requirements:**
* OS: Linux x86_64 (glibc ≥2.17)
* GPU: Compute Capability 7.0–8.9
### Driver Compatibility Reference
| CUDA Version | Minimum Driver Version |
|--------------|------------------------|
| CUDA 11.8 | ≥R520 (520.61.05) |
| CUDA 12.1 | ≥R530 (530.30.02) |
| CUDA 12.2+ | Check [NVIDIA Documentation](https://docs.nvidia.com/deploy/cuda-compatibility/) |
**Warning**: When installing without `[fix-cuda]`, pip may resolve to a newer CUDA version that requires a newer driver than you have installed. Always verify driver compatibility before installation.
### Advanced: Using System CUDA Libraries
If you need to use system-installed CUDA instead of PyPI CUDA packages, you can bypass the automatic CUDA loading:
1. **Exclude PyPI CUDA dependencies** using your package manager (e.g., [uv](https://github.com/astral-sh/uv/issues/7214), [pdm](https://pdm-project.org/en/latest/usage/config/#exclude-specific-packages-and-their-dependencies-from-the-lock-file))
2. **Set environment variable**: `_FAISS_WHEEL_DISABLE_CUDA_PRELOAD=1`
3. **Ensure CUDA libraries are accessible** via `LD_LIBRARY_PATH`
Example with uv (workaround):
```toml
# In pyproject.toml
[tool.uv]
override-dependencies = [
"nvidia-cuda-runtime-cu11==0.0.0; sys_platform == 'never'",
"nvidia-cublas-cu11==0.0.0; sys_platform == 'never'",
]
```
## Versioning
* Follows the original faiss repository versioning (e.g., `1.11.0`)
* Patches specific to this repository use `postN` suffix (e.g., `1.11.0.post1`)
## Building from Source
Build `faiss-gpu-cu11` and `faiss-gpu-cu12` wheels using [cibuildwheel](https://github.com/pypa/cibuildwheel).
### Build Configuration
```bash
# Configure build parameters
export NJOB="32" # Number of parallel build jobs
export FAISS_OPT_LEVEL="generic" # Options: generic, avx2, avx512
export CUDA_ARCHITECTURES="70-real;80-real" # Target GPU architectures
# For builds without GPU testing
export CIBW_TEST_COMMAND_LINUX=""
# For builds with GPU testing (requires NVIDIA Docker)
export CIBW_CONTAINER_ENGINE='docker; create_args: --gpus all'
# Note: GPU testing requires Docker with NVIDIA Container Toolkit configured
```
### Build Commands
```bash
# Build faiss-gpu-cu11 wheels
uvx cibuildwheel@2.23.2 variant/gpu-cu11 --output-dir wheelhouse/gpu-cu11
# Build faiss-gpu-cu12 wheels
uvx cibuildwheel@2.23.2 variant/gpu-cu12 --output-dir wheelhouse/gpu-cu12
```
Wheels will be created in `{repository_root}/wheelhouse/gpu-cuXX/`.
### Build Requirements
* OS: Linux x86_64
* NVIDIA Container Toolkit (if running tests)
* NVIDIA Driver: ≥R530 (if running tests with CUDA 12)
Raw data
{
"_id": null,
"home_page": null,
"name": "faiss-gpu-cu12",
"maintainer": null,
"docs_url": null,
"requires_python": "<3.13,>=3.9",
"maintainer_email": null,
"keywords": "search nearest neighbors",
"author": null,
"author_email": "Di-Is <rhoxbox@gmail.com>",
"download_url": null,
"platform": null,
"description": "# faiss-wheels\n\nThis repository is based on [kyamagu/faiss-wheels](https://github.com/kyamagu/faiss-wheels).\n\n[](https://pypi.org/project/faiss-gpu-cu11/)\n[](https://pypi.org/project/faiss-gpu-cu12/)\n\n## Overview\n\nThis repository provides scripts to build GPU-enabled wheels for the [faiss](https://github.com/facebookresearch/faiss) library.\nDistributes `faiss-gpu-cuXX` packages to PyPI using the contents of this repository.\n\n### Key Features\n\n* **No local CUDA installation required** - Dynamically links to CUDA Runtime and cuBLAS libraries from PyPI\n* Builds CUDA 11.8+ and CUDA 12.1+ compatible wheels\n* Supports Volta to Ada Lovelace architecture GPUs (Compute Capability 7.0\u20138.9)\n* Bundles OpenBLAS in Linux\n* Reduces wheel file size through dynamic linking instead of static compilation\n\n## Important Requirements\n\nThe published `faiss-gpu-cuXX` packages require proper system setup that cannot be managed by pip. It is your responsibility to prepare a suitable environment:\n\n1. **NVIDIA Driver**: Your host must have a CUDA-compatible NVIDIA driver installed\n * The minimum driver version depends on the CUDA version that gets installed\n * NVIDIA drivers are backward compatible with older CUDA versions ([See CUDA Compatibility Documentation](https://docs.nvidia.com/deploy/cuda-compatibility/))\n\n2. **GPU Architecture**: Your GPU must be compatible (Compute Capability 7.0\u20138.9)\n * Supported: Volta, Turing, Ampere, Ada Lovelace\n\n3. **Library Compatibility**: If you install multiple CUDA-dependent libraries (e.g., PyTorch) in the same environment, they must link to the same CUDA version\n\n## GPU Architecture Support for PyPI Packages\n\n### Support Policy for `faiss-gpu-cu11` and `faiss-gpu-cu12`\n\n**Note**: This is an **unofficial, personal development project** with limited computational resources. Due to these constraints, comprehensive testing across all NVIDIA GPU architectures is not feasible. The pre-built `faiss-gpu-cu11` and `faiss-gpu-cu12` packages on PyPI aim to support the same GPU architecture range (Compute Capability 7.0\u20138.9) as the official Faiss repository.\n\n### For Unsupported GPU Architectures\n\nIf you have a GPU architecture that is not supported by these pre-built wheels:\n\n1. **Official Faiss**: Follow the [official Faiss repository build instructions](https://github.com/facebookresearch/faiss/blob/main/INSTALL.md)\n2. **Build from Source**: Use this repository's code to build wheels for your specific architecture (see [Building from Source](#building-from-source) section)\n\n## Installation\n\nThe `faiss-gpu-cu11` and `faiss-gpu-cu12` wheels are available on PyPI. Choose the appropriate version for your CUDA environment.\n\n### For CUDA 12\n\n```bash\n# Install with fixed CUDA 12.1 (requires NVIDIA Driver \u2265R530)\npip install 'faiss-gpu-cu12[fix-cuda]'\n\n# Install with CUDA 12.X (X\u22651) - allows flexibility but driver requirement varies\npip install faiss-gpu-cu12\n```\n\n**Details:**\n\n* `faiss-gpu-cu12` is built with CUDA Toolkit 12.1 and maintains minor version compatibility\n* With `[fix-cuda]`: Installs exactly CUDA 12.1, requiring NVIDIA Driver \u2265R530\n* Without `[fix-cuda]`: Allows any CUDA 12.X (X\u22651), driver requirement depends on the actual CUDA version installed\n * For example: CUDA 12.4 requires Driver \u2265R550\n* Use without `[fix-cuda]` when integrating with other CUDA-dependent packages (e.g., PyTorch with CUDA 12.4)\n\n**System Requirements:**\n\n* OS: Linux x86_64 (glibc \u22652.17)\n* GPU: Compute Capability 7.0\u20138.9\n\n### For CUDA 11\n\n```bash\n# Install with CUDA 11.8 (requires NVIDIA Driver \u2265R520)\npip install faiss-gpu-cu11[fix-cuda]\n\n# Same as above (CUDA 11.8 is the final version)\npip install faiss-gpu-cu11\n```\n\n**Details:**\n\n* `faiss-gpu-cu11` is built with CUDA Toolkit 11.8\n* Both commands install CUDA 11.8 since no newer CUDA 11.X versions exist\n* Requires NVIDIA Driver \u2265R520\n\n**System Requirements:**\n\n* OS: Linux x86_64 (glibc \u22652.17)\n* GPU: Compute Capability 7.0\u20138.9\n\n### Driver Compatibility Reference\n\n| CUDA Version | Minimum Driver Version |\n|--------------|------------------------|\n| CUDA 11.8 | \u2265R520 (520.61.05) |\n| CUDA 12.1 | \u2265R530 (530.30.02) |\n| CUDA 12.2+ | Check [NVIDIA Documentation](https://docs.nvidia.com/deploy/cuda-compatibility/) |\n\n**Warning**: When installing without `[fix-cuda]`, pip may resolve to a newer CUDA version that requires a newer driver than you have installed. Always verify driver compatibility before installation.\n\n### Advanced: Using System CUDA Libraries\n\nIf you need to use system-installed CUDA instead of PyPI CUDA packages, you can bypass the automatic CUDA loading:\n\n1. **Exclude PyPI CUDA dependencies** using your package manager (e.g., [uv](https://github.com/astral-sh/uv/issues/7214), [pdm](https://pdm-project.org/en/latest/usage/config/#exclude-specific-packages-and-their-dependencies-from-the-lock-file))\n2. **Set environment variable**: `_FAISS_WHEEL_DISABLE_CUDA_PRELOAD=1`\n3. **Ensure CUDA libraries are accessible** via `LD_LIBRARY_PATH`\n\nExample with uv (workaround):\n\n```toml\n# In pyproject.toml\n[tool.uv]\noverride-dependencies = [\n \"nvidia-cuda-runtime-cu11==0.0.0; sys_platform == 'never'\",\n \"nvidia-cublas-cu11==0.0.0; sys_platform == 'never'\",\n]\n```\n\n## Versioning\n\n* Follows the original faiss repository versioning (e.g., `1.11.0`)\n* Patches specific to this repository use `postN` suffix (e.g., `1.11.0.post1`)\n\n## Building from Source\n\nBuild `faiss-gpu-cu11` and `faiss-gpu-cu12` wheels using [cibuildwheel](https://github.com/pypa/cibuildwheel).\n\n### Build Configuration\n\n```bash\n# Configure build parameters\nexport NJOB=\"32\" # Number of parallel build jobs\nexport FAISS_OPT_LEVEL=\"generic\" # Options: generic, avx2, avx512\nexport CUDA_ARCHITECTURES=\"70-real;80-real\" # Target GPU architectures\n\n# For builds without GPU testing\nexport CIBW_TEST_COMMAND_LINUX=\"\"\n\n# For builds with GPU testing (requires NVIDIA Docker)\nexport CIBW_CONTAINER_ENGINE='docker; create_args: --gpus all'\n# Note: GPU testing requires Docker with NVIDIA Container Toolkit configured\n```\n\n### Build Commands\n\n```bash\n# Build faiss-gpu-cu11 wheels\nuvx cibuildwheel@2.23.2 variant/gpu-cu11 --output-dir wheelhouse/gpu-cu11\n\n# Build faiss-gpu-cu12 wheels\nuvx cibuildwheel@2.23.2 variant/gpu-cu12 --output-dir wheelhouse/gpu-cu12\n```\n\nWheels will be created in `{repository_root}/wheelhouse/gpu-cuXX/`.\n\n### Build Requirements\n\n* OS: Linux x86_64\n* NVIDIA Container Toolkit (if running tests)\n* NVIDIA Driver: \u2265R530 (if running tests with CUDA 12)\n",
"bugtrack_url": null,
"license": "MIT License\n \n Copyright (c) 2024 Di-Is\n \n Permission is hereby granted, free of charge, to any person obtaining a copy\n of this software and associated documentation files (the \"Software\"), to deal\n in the Software without restriction, including without limitation the rights\n to use, copy, modify, merge, publish, distribute, sublicense, and/or sell\n copies of the Software, and to permit persons to whom the Software is\n furnished to do so, subject to the following conditions:\n \n The above copyright notice and this permission notice shall be included in all\n copies or substantial portions of the Software.\n \n THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\n IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\n FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE\n AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER\n LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,\n OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE\n SOFTWARE.\n \n The major design pattern of this repository was based in Kota Yamaguchi's faiss-wheels, which is subject to the same license.\n Here is the original copyright notice for faiss-wheels:\n \n MIT License\n \n Copyright (c) 2019 Kota Yamaguchi\n \n Permission is hereby granted, free of charge, to any person obtaining a copy\n of this software and associated documentation files (the \"Software\"), to deal\n in the Software without restriction, including without limitation the rights\n to use, copy, modify, merge, publish, distribute, sublicense, and/or sell\n copies of the Software, and to permit persons to whom the Software is\n furnished to do so, subject to the following conditions:\n \n The above copyright notice and this permission notice shall be included in all\n copies or substantial portions of the Software.\n \n THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\n IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\n FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE\n AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER\n LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,\n OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE\n SOFTWARE.\n \n This repository uses meta's faiss and is subject to the same license.\n Here is the original copyright notice for faiss:\n \n MIT License\n \n Copyright (c) Facebook, Inc. and its affiliates.\n \n Permission is hereby granted, free of charge, to any person obtaining a copy\n of this software and associated documentation files (the \"Software\"), to deal\n in the Software without restriction, including without limitation the rights\n to use, copy, modify, merge, publish, distribute, sublicense, and/or sell\n copies of the Software, and to permit persons to whom the Software is\n furnished to do so, subject to the following conditions:\n \n The above copyright notice and this permission notice shall be included in all\n copies or substantial portions of the Software.\n \n THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\n IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\n FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE\n AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER\n LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,\n OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE\n SOFTWARE.",
"summary": "A library for efficient similarity search and clustering of dense vectors.",
"version": "1.12.0",
"project_urls": {
"faiss-gpu-wheels": "https://github.com/Di-Is/faiss-gpu-wheels"
},
"split_keywords": [
"search",
"nearest",
"neighbors"
],
"urls": [
{
"comment_text": null,
"digests": {
"blake2b_256": "4f6a2b4e10ed0ef29870add4f4861708cb2ef3ae58a4d3174b99ec59a2abfdad",
"md5": "51c907fda6f3cd5e30f8417f22fe03ff",
"sha256": "18ce3c70b49517efb2da1af28bc93474e1cbe06e74f0099a8a7241ebfe0d06a4"
},
"downloads": -1,
"filename": "faiss_gpu_cu12-1.12.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl",
"has_sig": false,
"md5_digest": "51c907fda6f3cd5e30f8417f22fe03ff",
"packagetype": "bdist_wheel",
"python_version": "cp310",
"requires_python": "<3.13,>=3.9",
"size": 48143038,
"upload_time": "2025-08-24T14:25:19",
"upload_time_iso_8601": "2025-08-24T14:25:19.955226Z",
"url": "https://files.pythonhosted.org/packages/4f/6a/2b4e10ed0ef29870add4f4861708cb2ef3ae58a4d3174b99ec59a2abfdad/faiss_gpu_cu12-1.12.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": null,
"digests": {
"blake2b_256": "cd4891177aaabbc147fe88af3646b9448df5c32ea9de3e845d2068513fec5090",
"md5": "bfd01623e3fd31dd39e476513e103adc",
"sha256": "ecc9c515fdf6a215d8a781bc9e931eca4e2d40217f928594088e42d0d7ff8d52"
},
"downloads": -1,
"filename": "faiss_gpu_cu12-1.12.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl",
"has_sig": false,
"md5_digest": "bfd01623e3fd31dd39e476513e103adc",
"packagetype": "bdist_wheel",
"python_version": "cp311",
"requires_python": "<3.13,>=3.9",
"size": 48142980,
"upload_time": "2025-08-24T14:25:24",
"upload_time_iso_8601": "2025-08-24T14:25:24.016367Z",
"url": "https://files.pythonhosted.org/packages/cd/48/91177aaabbc147fe88af3646b9448df5c32ea9de3e845d2068513fec5090/faiss_gpu_cu12-1.12.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": null,
"digests": {
"blake2b_256": "f283183805c6c057e7fdfbfa6746140662c14c2bb122c23b7d74958d1e30981d",
"md5": "eb3ef5dc3e0359c66b763723ff35f0e6",
"sha256": "5d3302e745e546578186ab101ea522b3555c94ef6e5738613f7f1bd55489b0e2"
},
"downloads": -1,
"filename": "faiss_gpu_cu12-1.12.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl",
"has_sig": false,
"md5_digest": "eb3ef5dc3e0359c66b763723ff35f0e6",
"packagetype": "bdist_wheel",
"python_version": "cp312",
"requires_python": "<3.13,>=3.9",
"size": 48142334,
"upload_time": "2025-08-24T14:25:27",
"upload_time_iso_8601": "2025-08-24T14:25:27.704196Z",
"url": "https://files.pythonhosted.org/packages/f2/83/183805c6c057e7fdfbfa6746140662c14c2bb122c23b7d74958d1e30981d/faiss_gpu_cu12-1.12.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": null,
"digests": {
"blake2b_256": "e606bd001ca3b46e5807b5f2696144542256a00f39756e8c61bc3dda261cb4a2",
"md5": "e97d9057961c6e6acd8c64075514c397",
"sha256": "7a66b6062e27fcaf50e75e2de493cafa630d8f6c6d0dbdd1919dcd39079e5777"
},
"downloads": -1,
"filename": "faiss_gpu_cu12-1.12.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl",
"has_sig": false,
"md5_digest": "e97d9057961c6e6acd8c64075514c397",
"packagetype": "bdist_wheel",
"python_version": "cp39",
"requires_python": "<3.13,>=3.9",
"size": 48143321,
"upload_time": "2025-08-24T14:25:32",
"upload_time_iso_8601": "2025-08-24T14:25:32.031137Z",
"url": "https://files.pythonhosted.org/packages/e6/06/bd001ca3b46e5807b5f2696144542256a00f39756e8c61bc3dda261cb4a2/faiss_gpu_cu12-1.12.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2025-08-24 14:25:19",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "Di-Is",
"github_project": "faiss-gpu-wheels",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"lcname": "faiss-gpu-cu12"
}
