3.9CUDA使用过程中的问题

摘要

本文主要是对使用GPU进行数据集的训练中，出现的CUDA水土不服问题进行解决

文中有对具体出现的问题的描述和具体问题的解决方法

• [x] Edit By Porter, 积水成渊,蛟龙生焉。

一、问题 [libcublas.so.9.0]

1.1 问题描述

• ImportError: libcublas.so.9.0: cannot open shared object file: No such file or directory

"/home/porter/文档/github workspace/MyKaggle/project/venv/bin/python" "/home/porter/文档/github workspace/MyKaggle/project/cat and dog classiy/dog_cat_class.py"
Using TensorFlow backend.
Traceback (most recent call last):
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/pywrap_tensorflow.py", line 58, in <module>
    from tensorflow.python.pywrap_tensorflow_internal import *
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 28, in <module>
    _pywrap_tensorflow_internal = swig_import_helper()
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 24, in swig_import_helper
    _mod = imp.load_module('_pywrap_tensorflow_internal', fp, pathname, description)
  File "/usr/lib/python3.5/imp.py", line 242, in load_module
    return load_dynamic(name, filename, file)
  File "/usr/lib/python3.5/imp.py", line 342, in load_dynamic
    return _load(spec)
ImportError: libcublas.so.9.0: cannot open shared object file: No such file or directory

1.2 问题解决

• 步骤一：

在~/.bashrc中添加如下的路径

# add manul configure cuda
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda-9.0/lib64
export PATH=$PATH:/usr/local/cuda-9.0/bin
export CUDA_HOME=$CUDA_HOME:/usr/local/cuda-9.0

export PATH="$PATH:$HOME/bin"

• 步骤二：

检查 /usr/local/cuda-9.0/lib64 下是否有 libcublas.so.9.0，没有的话找这个文件链接进去。

如果有，终端输入:

sudo ldconfig /usr/local/cuda-9.0/lib64