TensorFlow (GPU) libcudart.so.7.5 错误 -- 无法打开共享对象文件

TensorFlow (GPU) libcudart.so.7.5 error -- cannot open shared object file

在 Ubuntu 14.04 上使用 GPU 安装 TensorFlow。对于 Ubuntu/UNIX 环境还很陌生,所以很多东西我都不懂。在搜索此问题时,我遇到了在安装了 CUDA 7.5 时无法导入 CUDA 7.0 的实例(因为 TensorFlow 不支持 7.5)。我安装了 CUDA 7.0,但它似乎在寻找 7.5 - 这是为什么?我的 .bashrc 文件中有以下几行:

export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:/usr/local/cuda-7.0/lib64"
export CUDA_HOME=/usr/local/cuda-7.0

但我仍然看到下面的问题...有人有什么建议吗?

>>> import tensorflow as tf
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/ford/tensorflow/local/lib/python2.7/site-packages/tensorflow/__init__.py", line 23, in <module>
    from tensorflow.python import *
  File "/home/ford/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/__init__.py", line 49, in <module>
    from tensorflow import contrib
  File "/home/ford/tensorflow/local/lib/python2.7/site-packages/tensorflow/contrib/__init__.py", line 23, in <module>
    from tensorflow.contrib import layers
  File "/home/ford/tensorflow/local/lib/python2.7/site-packages/tensorflow/contrib/layers/__init__.py", line 68, in <module>
    from tensorflow.contrib.layers.python.layers import *
  File "/home/ford/tensorflow/local/lib/python2.7/site-packages/tensorflow/contrib/layers/python/layers/__init__.py", line 22, in <module>
    from tensorflow.contrib.layers.python.layers.initializers import *
  File "/home/ford/tensorflow/local/lib/python2.7/site-packages/tensorflow/contrib/layers/python/layers/initializers.py", line 24, in <module>
    from tensorflow.python.ops import random_ops
  File "/home/ford/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/ops/random_ops.py", line 23, in <module>
    from tensorflow.python.framework import ops
  File "/home/ford/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/framework/ops.py", line 39, in <module>
    from tensorflow.python.framework import versions
  File "/home/ford/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/framework/versions.py", line 22, in <module>
    from tensorflow.python import pywrap_tensorflow
  File "/home/ford/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/pywrap_tensorflow.py", line 28, in <module>
    _pywrap_tensorflow = swig_import_helper()
  File "/home/ford/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/pywrap_tensorflow.py", line 24, in swig_import_helper
    _mod = imp.load_module('_pywrap_tensorflow', fp, pathname, description)
ImportError: libcudart.so.7.5: cannot open shared object file: No such file or directory

我遇到了同样的问题。您是否尝试过将 CUDA 7.5 与 7 一起安装?这对我有用,我将 PATH 变量更改为包含 7.5 而不是 7.0

如果您使用的是最新的 r0.7 二进制文件,根据我的经验,它们现在默认构建为支持 CUDA 7.5...而 r0.6 二进制文件适用于 CUDA 7.0

您可以从源代码构建 tensorflow 以使用 CUDA 7.0,请参阅此处的文档 Configure TensorFlow's canonical view of Cuda libraries

$ ./configure
Please specify the location of python. [Default is /usr/bin/python]:
Do you wish to build TensorFlow with GPU support? [y/N] y
GPU support will be enabled for TensorFlow

Please specify the Cuda SDK version you want to use, e.g. 7.0. [Leave
empty to use system default]: 7.0

Please specify the location where CUDA 7.0 toolkit is installed. Refer to
README.md for more details. [default is: /usr/local/cuda]: /usr/local/cuda

Please specify the Cudnn version you want to use. [Leave empty to use system
default]: 4.0.4

Please specify the location where the cuDNN 4.0.4 library is installed. Refer to
README.md for more details. [default is: /usr/local/cuda]: /usr/local/cudnn-r4-rc/

Please specify a list of comma-separated Cuda compute capabilities you want to
build with. You can find the compute capability of your device at: 
https://developer.nvidia.com/cuda-gpus.
Please note that each additional compute capability significantly increases your
build time and binary size. [Default is: \"3.5,5.2\"]: 3.5

Setting up Cuda include
Setting up Cuda lib64
Setting up Cuda bin
Setting up Cuda nvvm
Configuration finished

您需要从 SOURCE 安装 tensorflow 以配置设置。通过 pip 或 easy_install 安装将不起作用。

使用 pip 将 tensorflow 从 0.6.0 升级到 0.7.1 后,我遇到了同样的问题。