The upstream uses gcc 8 by default, and they have no interest to support higher gcc version. However, building paddlepaddle from the source using gcc 8 need a lot of hack and patch, and you might even need to build its deps with gcc 8 too.
Search Criteria
Package Details: python-paddlepaddle 2.2.2-1
Package Actions
Git Clone URL: | https://aur.archlinux.org/python-paddlepaddle.git (read-only, click to copy) |
---|---|
Package Base: | python-paddlepaddle |
Description: | PArallel Distributed Deep LEarning: Machine Learning Framework from Industrial Practice |
Upstream URL: | http://www.paddlepaddle.org |
Licenses: | Apache |
Submitter: | hottea |
Maintainer: | None |
Last Packager: | hottea |
Votes: | 0 |
Popularity: | 0.000000 |
First Submitted: | 2021-12-24 00:59 (UTC) |
Last Updated: | 2022-01-21 04:41 (UTC) |
Dependencies (19)
- blas (blis-cblas-openmpAUR, blis-cblasAUR, aocl-blis-aoccAUR, blas-gitAUR, atlas-lapackAUR, blas-aocl-gccAUR, blas-aocl-aoccAUR, blis-gitAUR, blisAUR, openblas-lapackAUR, blas-mklAUR, aocl-blisAUR, blas-openblas)
- python-astor (python-astor-gitAUR)
- python-decorator
- python-numpy (python-numpy-flameAUR, python-numpy-gitAUR, python-numpy-mkl-binAUR, python-numpy-mkl-tbbAUR, python-numpy-mklAUR, python-numpy1AUR)
- python-pillow
- python-protobuf (python-protobuf-gitAUR)
- python-requests
- python-six
- cmake (cmake-gitAUR) (make)
- cuda (cuda11.1AUR, cuda-12.2AUR, cuda12.0AUR, cuda11.4AUR, cuda11.4-versionedAUR, cuda12.0-versionedAUR) (make)
- cudnn (make)
- git (git-gitAUR, git-glAUR) (make)
- nccl (nccl-gitAUR) (make)
- patchelf (patchelf-gitAUR) (make)
- python-pip (make)
- python-setuptools (make)
- python-wheel (make)
- swig (swig-gitAUR) (make)
- wget (wget-gitAUR, wurlAUR) (make)
Required by (2)
- python-deepxde (optional)
- python-safetensors (optional)
Sources (1)
Latest Comments
hottea commented on 2022-05-15 01:58 (UTC)
firoz commented on 2022-01-19 11:50 (UTC)
You could use the wheel from pypi as a form of binary package.
hottea commented on 2021-12-24 01:17 (UTC)
This PKGBUILD
builds python-paddlepaddle
and python-paddlepaddle-cuda
from the source. It takes about 2.5 hours to build it. To reduce the building time, you could adjust _CUDA_ARCH_LIST
. If you don't need python-paddlepaddle-cuda
, you could comment out the cuda part.
See also here if you'd like to build python-paddlepaddle-git
and python-paddlepaddle-cuda-git
from the source.
Pinned Comments
hottea commented on 2021-12-24 01:17 (UTC)
This
PKGBUILD
buildspython-paddlepaddle
andpython-paddlepaddle-cuda
from the source. It takes about 2.5 hours to build it. To reduce the building time, you could adjust_CUDA_ARCH_LIST
. If you don't needpython-paddlepaddle-cuda
, you could comment out the cuda part. See also here if you'd like to buildpython-paddlepaddle-git
andpython-paddlepaddle-cuda-git
from the source.