libcublas.so.9.0 没有这样的文件或目录答案

【问题标题】：libcublas.so.9.0 No such file or directorylibcublas.so.9.0 没有这样的文件或目录
【发布时间】：2021-07-09 17:24:31
【问题描述】：

导入tensorflow时出错。

我的环境

Ubuntu 18.04 桌面版
Nvidia 显卡驱动 470.42.01
CUDA 11.4
Python 3.6.14 (Anaconda)
TensorFlow-gpu 1.12.0

CUDA 环境变量设置完成 (11.4) 来自 /usr/local/CUDA-11.4 的 /usr/local/CUDA 的链接符号链接

（使用旧版本的原因是项目的SDK是以前写的。）

运行Python后导入tensorflow时，出现如下错误：

>>> import tensorflow
Traceback (most recent call last):
  File "/home/(username)/anaconda3/envs/mdk/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow.py", line 58, in <module>
    from tensorflow.python.pywrap_tensorflow_internal import *
  File "/home/(username)/anaconda3/envs/mdk/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 28, in <module>
    _pywrap_tensorflow_internal = swig_import_helper()
  File "/home/(username)/anaconda3/envs/mdk/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 24, in swig_import_helper
    _mod = imp.load_module('_pywrap_tensorflow_internal', fp, pathname, description)
  File "/home/(username)/anaconda3/envs/mdk/lib/python3.6/imp.py", line 243, in load_module
    return load_dynamic(name, filename, file)
  File "/home/(username)/anaconda3/envs/mdk/lib/python3.6/imp.py", line 343, in load_dynamic
    return _load(spec)
ImportError: libcublas.so.9.0: cannot open shared object file: No such file or directory

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/(username)/anaconda3/envs/mdk/lib/python3.6/site-packages/tensorflow/__init__.py", line 24, in <module>
    from tensorflow.python import pywrap_tensorflow  # pylint: disable=unused-import
  File "/home/(username)/anaconda3/envs/mdk/lib/python3.6/site-packages/tensorflow/python/__init__.py", line 49, in <module>
    from tensorflow.python import pywrap_tensorflow
  File "/home/(username)/anaconda3/envs/mdk/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow.py", line 74, in <module>
    raise ImportError(msg)
ImportError: Traceback (most recent call last):
  File "/home/(username)/anaconda3/envs/mdk/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow.py", line 58, in <module>
    from tensorflow.python.pywrap_tensorflow_internal import *
  File "/home/(username)/anaconda3/envs/mdk/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 28, in <module>
    _pywrap_tensorflow_internal = swig_import_helper()
  File "/home/(username)/anaconda3/envs/mdk/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 24, in swig_import_helper
    _mod = imp.load_module('_pywrap_tensorflow_internal', fp, pathname, description)
  File "/home/(username)/anaconda3/envs/mdk/lib/python3.6/imp.py", line 243, in load_module
    return load_dynamic(name, filename, file)
  File "/home/(username)/anaconda3/envs/mdk/lib/python3.6/imp.py", line 343, in load_dynamic
    return _load(spec)
ImportError: libcublas.so.9.0: cannot open shared object file: No such file or directory


Failed to load the native TensorFlow runtime.

See https://www.tensorflow.org/install/errors

for some common reasons and solutions.  Include the entire stack trace
above this error message when asking for help.

我安装了 CUDA 9.0 来修复这个错误。但是，7.0 版出现了同样的错误。

我需要安装我需要的所有 CUDA 版本吗？或者有没有其他方法可以解决这个问题？

【问题讨论】：

如果您安装了 CUDA 9.0，您会得到一个不同的错误（很可能是关于 cudnn），与 7.0 版的错误不同。

标签： python python-3.x tensorflow ubuntu-18.04

【解决方案1】：

根据 Tensorflow 文档，TF v1.12 需要 CUDA 9.0。查看屏幕截图以获取信息，

安装适当的 CUDA 版本后，将 CUDA®、CUPTI 和 cuDNN 安装目录添加到 %PATH% 环境变量。按照提到的安装步骤here 并根据要求更改 CUDA 版本，即 CUDA 9.0。

【讨论】：