tensorflow 1.13.1 不检测 gpu [英] tensorflow 1.13.1 doesn't detect gpu
问题描述
我已经为此苦苦挣扎了一段时间并阅读了我找到的所有帖子.我在 Ubuntu 16.04 上有 gtx 1080 gpu.我的要求是 tensorflow 1.13.1.在我的 conda 环境中，我有
I have been struggling with this for a while and read all the posts I found. I have gtx 1080 gpu on Ubuntu 16.04. My requirement is tensorflow 1.13.1. In my conda env I l have
_libgcc_mutex 0.1 main
_tflow_select 2.1.0 gpu
absl-py 0.7.1 py36_0 conda-forge
astor 0.7.1 py_0 conda-forge
c-ares 1.15.0 h516909a_1001 conda-forge
ca-certificates 2019.6.16 hecc5488_0 conda-forge
certifi 2019.6.16 py36_1 conda-forge
cudatoolkit 10.0.130 0
cudnn 7.6.0 cuda10.0_0
gast 0.2.2 py_0 conda-forge
grpcio 1.16.1 py36hf8bcb03_1
h5py 2.9.0 nompi_py36hf008753_1102 conda-forge
hdf5 1.10.4 nompi_h3c11f04_1106 conda-forge
keras-applications 1.0.7 py_1 conda-forge
keras-preprocessing 1.0.9 py_1 conda-forge
libblas 3.8.0 10_openblas conda-forge
libcblas 3.8.0 10_openblas conda-forge
libffi 3.2.1 he1b5a44_1006 conda-forge
libgcc-ng 9.1.0 hdf63c60_0
libgfortran-ng 7.3.0 hdf63c60_0
liblapack 3.8.0 10_openblas conda-forge
libopenblas 0.3.6 h6e990d7_5 conda-forge
libprotobuf 3.9.0 h8b12597_0 conda-forge
libstdcxx-ng 9.1.0 hdf63c60_0
markdown 3.1.1 py_0 conda-forge
mock 3.0.5 py36_0 conda-forge
ncurses 6.1 hf484d3e_1002 conda-forge
numpy 1.16.4 py36h95a1406_0 conda-forge
openblas 0.3.6 h6e990d7_5 conda-forge
openssl 1.1.1c h516909a_0 conda-forge
pip 19.2.1 py36_0 conda-forge
protobuf 3.9.0 py36he1b5a44_0 conda-forge
python 3.6.7 h357f687_1005 conda-forge
readline 8.0 hf8c457e_0 conda-forge
scipy 1.3.0 py36h921218d_0 conda-forge
setuptools 41.0.1 py36_0 conda-forge
six 1.12.0 py36_1000 conda-forge
sqlite 3.29.0 hcee41ef_0 conda-forge
tensorboard 1.13.1 py36_0 conda-forge
tensorflow 1.13.1 py36_0 conda-forge
tensorflow-estimator 1.13.0 py_0
tensorflow-gpu 1.13.1 h0d30ee6_0
termcolor 1.1.0 py_2 conda-forge
tk 8.6.9 hed695b0_1002 conda-forge
werkzeug 0.15.5 py_0 conda-forge
wheel 0.33.4 py36_0 conda-forge
xz 5.2.4 h14c3975_1001 conda-forge
zlib 1.2.11 h516909a_1005 conda-forge
AFAIK，所有必需的软件都在那里并且是正确的版本，cudnn、cuda-toolkit、tensorflow-gpu.但仍然 tensorflow 没有检测到 gpu.device_lib.list_local_devices() 只报告 CPU.
AFAIK, all the required software is there and are the right versions, cudnn, cuda-toolkit, tensorflow-gpu. But still tensorflow does not detect gpu. device_lib.list_local_devices() only reports CPU.
任何人都可以提出其他错误的建议?
Anyone can suggest what else can be wrong?
推荐答案
问题是 TensorFlow 1.13 的 Anaconda 元数据中的一个错误.它也可以通过安装 tensorflow=*=*gpu* 来修复.
The issue is a bug in the Anaconda metadata for tensorflow 1.13. It can be fixed by also installing tensorflow=*=*gpu*.
这仅对 tensorflow 1.13 是必需的.
This is only necessary for tensorflow 1.13.
这篇关于tensorflow 1.13.1 不检测 gpu的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！
