ImportError：libcublas.so.10.0：无法打开共享对象文件：没有这样的文件或目录答案

【问题标题】：ImportError: libcublas.so.10.0: cannot open shared object file: No such file or directoryImportError：libcublas.so.10.0：无法打开共享对象文件：没有这样的文件或目录
【发布时间】：2019-08-08 23:28:09
【问题描述】：

我已经在 Ubuntu 18.04 上安装了 Cuda 10.1 和 cudnn，它似乎正确安装为 nvcc 和 nvidia-smi 类型，我得到了正确的响应：

user:~$ nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2019 NVIDIA Corporation
Built on Fri_Feb__8_19:08:17_PST_2019
Cuda compilation tools, release 10.1, V10.1.105
user:~$ nvidia-smi 
Mon Mar 18 14:36:47 2019       
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 418.43       Driver Version: 418.43       CUDA Version: 10.1     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  Quadro K5200        Off  | 00000000:03:00.0  On |                  Off |
| 26%   39C    P8    14W / 150W |    225MiB /  8118MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID   Type   Process name                             Usage      |
|=============================================================================|
|    0      1538      G   /usr/lib/xorg/Xorg                            32MiB |
|    0      1583      G   /usr/bin/gnome-shell                           5MiB |
|    0      3008      G   /usr/lib/xorg/Xorg                           100MiB |
|    0      3120      G   /usr/bin/gnome-shell                          82MiB |
+-----------------------------------------------------------------------------+

我已经使用以下方法安装了 tensorflow： user:~$ sudo pip3 install --upgrade tensorflow-gpu

The directory '/home/amin/.cache/pip/http' or its parent directory is not owned by the current user and the cache has been disabled. Please check the permissions and owner of that directory. If executing pip with sudo, you may want sudo's -H flag.
The directory '/home/amin/.cache/pip' or its parent directory is not owned by the current user and caching wheels has been disabled. check the permissions and owner of that directory. If executing pip with sudo, you may want sudo's -H flag.
Requirement already up-to-date: tensorflow-gpu in /usr/local/lib/python3.6/dist-packages (1.13.1)
Requirement already satisfied, skipping upgrade: keras-applications>=1.0.6 in /usr/local/lib/python3.6/dist-packages (from tensorflow-gpu) (1.0.7)
Requirement already satisfied, skipping upgrade: protobuf>=3.6.1 in /usr/local/lib/python3.6/dist-packages (from tensorflow-gpu) (3.6.1)
Requirement already satisfied, skipping upgrade: wheel>=0.26 in /usr/local/lib/python3.6/dist-packages (from tensorflow-gpu) (0.32.3)
Requirement already satisfied, skipping upgrade: absl-py>=0.1.6 in /usr/local/lib/python3.6/dist-packages (from tensorflow-gpu) (0.7.0)
Requirement already satisfied, skipping upgrade: keras-preprocessing>=1.0.5 in /usr/local/lib/python3.6/dist-packages (from tensorflow-gpu) (1.0.9)
Requirement already satisfied, skipping upgrade: gast>=0.2.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow-gpu) (0.2.2)
Requirement already satisfied, skipping upgrade: termcolor>=1.1.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow-gpu) (1.1.0)
Requirement already satisfied, skipping upgrade: grpcio>=1.8.6 in /usr/local/lib/python3.6/dist-packages (from tensorflow-gpu) (1.18.0)
Requirement already satisfied, skipping upgrade: tensorflow-estimator<1.14.0rc0,>=1.13.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow-gpu) (1.13.0)
Requirement already satisfied, skipping upgrade: six>=1.10.0 in /usr/lib/python3/dist-packages (from tensorflow-gpu) (1.11.0)
Requirement already satisfied, skipping upgrade: numpy>=1.13.3 in /usr/lib/python3/dist-packages (from tensorflow-gpu) (1.13.3)
Requirement already satisfied, skipping upgrade: astor>=0.6.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow-gpu) (0.7.1)
Requirement already satisfied, skipping upgrade: tensorboard<1.14.0,>=1.13.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow-gpu) (1.13.1)
Requirement already satisfied, skipping upgrade: h5py in /usr/local/lib/python3.6/dist-packages (from keras-applications>=1.0.6->tensorflow-gpu) (2.9.0)
Requirement already satisfied, skipping upgrade: setuptools in /usr/local/lib/python3.6/dist-packages (from protobuf>=3.6.1->tensorflow-gpu) (40.6.3)
Requirement already satisfied, skipping upgrade: mock>=2.0.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow-estimator<1.14.0rc0,>=1.13.0->tensorflow-gpu) (2.0.0)
Requirement already satisfied, skipping upgrade: werkzeug>=0.11.15 in /usr/local/lib/python3.6/dist-packages (from tensorboard<1.14.0,>=1.13.0->tensorflow-gpu) (0.14.1)
Requirement already satisfied, skipping upgrade: markdown>=2.6.8 in /usr/local/lib/python3.6/dist-packages (from tensorboard<1.14.0,>=1.13.0->tensorflow-gpu) (3.0.1)
Requirement already satisfied, skipping upgrade: pbr>=0.11 in /usr/local/lib/python3.6/dist-packages (from mock>=2.0.0->tensorflow-estimator<1.14.0rc0,>=1.13.0->tensorflow-gpu) (5.1.1)

但是，当我尝试导入 tensorflow 时，我收到有关 libcublas.so.10.0 的错误：

user:~$ python3
Python 3.6.7 (default, Oct 22 2018, 11:32:17) 
[GCC 8.2.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import tensorflow as tf
Traceback (most recent call last):
  File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/pywrap_tensorflow.py", line 58, in <module>
    from tensorflow.python.pywrap_tensorflow_internal import *
  File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 28, in <module>
    _pywrap_tensorflow_internal = swig_import_helper()
  File "/usr/local/lib/python3.6/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/python3.6/imp.py", line 243, in load_module
    return load_dynamic(name, filename, file)
  File "/usr/lib/python3.6/imp.py", line 343, in load_dynamic
    return _load(spec)
ImportError: libcublas.so.10.0: 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 "/usr/local/lib/python3.6/dist-packages/tensorflow/__init__.py", line 24, in <module>
    from tensorflow.python import pywrap_tensorflow  # pylint: disable=unused-import
  File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/__init__.py", line 49, in <module>
    from tensorflow.python import pywrap_tensorflow
  File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/pywrap_tensorflow.py", line 74, in <module>
    raise ImportError(msg)
ImportError: Traceback (most recent call last):
  File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/pywrap_tensorflow.py", line 58, in <module>
    from tensorflow.python.pywrap_tensorflow_internal import *
  File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 28, in <module>
    _pywrap_tensorflow_internal = swig_import_helper()
  File "/usr/local/lib/python3.6/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/python3.6/imp.py", line 243, in load_module
    return load_dynamic(name, filename, file)
  File "/usr/lib/python3.6/imp.py", line 343, in load_dynamic
    return _load(spec)
ImportError: libcublas.so.10.0: 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.

我错过了什么？我该如何解决这个问题？

谢谢

【问题讨论】：

您的 TF 期待 CUDA 10.0。您不能使用 CUDA10.1 作为替代品。您必须以某种方式安装 CUDA 10.0。由于您已经安装了 GPU 驱动程序，因此您只需要 CUDA 10.0 工具包。按照 CUDA 10.0 的 CUDA linux 安装指南中的说明进行操作
感谢您的帮助。我需要先卸载 cuda 10.1 还是添加 10.0 就足够了？我可以有多个版本的 cuda 吗？谢谢
您可以安装多个版本。您无需删除 10.1 即可安装/使用 10.0。您可能需要正确设置环境变量
感谢我安装了 CUDA 10.0，现在它可以工作了。
没有nvidia gpu，所以我以为我安装了仅CPU版本的tensorflow，不知何故我收到了这条消息。有什么想法吗？

标签： tensorflow ubuntu-18.04

【解决方案1】：

我从以下链接下载了 cuda 10.0 CUDA 10.0

然后我使用以下命令安装它：

sudo dpkg -i cuda-repo-ubuntu1804_10.0.130-1_amd64.deb
sudo apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/7fa2af80.pub
sudo apt-get update
sudo apt-get install cuda-10-0

然后我通过链接为 CUDA 10.0 安装了 cudnn v7.5.0 CUDNN download，您需要使用帐号登录。

在选择正确的版本后，我通过链接CUDNN power link 下载了之后，我为 cudnn 添加了 include 和 lib 文件，如下所示：

sudo cp -P cuda/targets/ppc64le-linux/include/cudnn.h /usr/local/cuda-10.0/include/
sudo cp -P cuda/targets/ppc64le-linux/lib/libcudnn* /usr/local/cuda-10.0/lib64/
sudo chmod a+r /usr/local/cuda-10.0/lib64/libcudnn*

修改cuda 10.0的lib和路径的.bashrc后，如果没有，需要添加到.bashrc中

export PATH=/usr/local/cuda-10.0/bin${PATH:+:${PATH}}
export LD_LIBRARY_PATH=/usr/local/cuda-10.0/lib64:${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}

经过所有这些步骤，我成功地在 python3 中导入了 tensorflow。

【讨论】：

只是一个小修正，您可能为 CUDA 10.0 安装了 CUDNN 7.5.0
对我来说，这会安装驱动程序版本 410.14 或 410.10，而 cuda 10 应该只适用于驱动程序
这是最好的答案，但是您输入的过程和链接现在正在下载与某些 pytorch 版本不兼容的 10.1 版本
经过数小时的 tensorflow 和 nvidia 文档苦苦挣扎，这是“最佳解决方案”。荣誉
我安装了 cuda-11-0 并且遇到了同样的问题。除了安装了支持 CUDA 10.0 的最新版本 CUDNN-7.6.5 之外，我按照此处的说明进行操作。

【解决方案2】：

如果使用 Cuda 10.1（如 https://www.tensorflow.org/install/gpu 中的指示），问题是 libcublas.so.10 已移出 cuda-10.1 目录并进入 cuda-10.2(!)

从这个答案复制：https://github.com/tensorflow/tensorflow/issues/26182#issuecomment-684993950

... libcublas.so.10 位于 /usr/local/cuda-10.2/lib64 中（来自 nvidia 的惊喜 - 安装 10.1 会安装一些 10.2 的东西）但只有 /usr/local/cuda 在包含路径中到 /usr/local/cuda-10.1.

修复是将其添加到您的包含路径：

export LD_LIBRARY_PATH=/usr/local/cuda-10.2/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}

注意：已知此修复适用于 Cuda 10.1、V10.1.243（使用 nvcc -V 打印您的版本）。

【讨论】：

有史以来最好的答案！你救了我的命！非常感谢！
他们为什么会做出这么奇怪的事情？
我们必须安装一个 cuda-10.2 来解决一个由 cuda-10.1 引起的问题？

【解决方案3】：

CUDA 10.1（根据 tensorflow 文档安装）抛出 can't find libcublas.so.10.0 错误。这些库存在于/usr/local/cuda-10.1/targets/x86_64-linux/lib/ 中，但名称错误。

还有另一个（丢失的）stackoverflow 帖子说这是包的固定依赖问题，可以通过一个额外的 cli 标志来修复。这似乎没有解决我的问题。

经过测试的解决方法是修改指令以降级到 CUDA 10.0

# Uninstall packages from tensorflow installation instructions 
sudo apt-get remove cuda-10-1 \
    libcudnn7 \
    libcudnn7-dev \
    libnvinfer6 \
    libnvinfer-dev \
    libnvinfer-plugin6

# WORKS: Downgrade to CUDA-10.0
sudo apt-get install -y --no-install-recommends \
    cuda-10-0 \
    libcudnn7=7.6.4.38-1+cuda10.0  \
    libcudnn7-dev=7.6.4.38-1+cuda10.0;
sudo apt-get install -y --no-install-recommends \
    libnvinfer6=6.0.1-1+cuda10.0 \
    libnvinfer-dev=6.0.1-1+cuda10.0 \
    libnvinfer-plugin6=6.0.1-1+cuda10.0;

升级到 CUDA-10.2 似乎也遇到了同样的问题

# BROKEN: Upgrade to CUDA-10.2 
# use `apt show -a libcudnn7 libnvinfer7` to find 10.2 compatable version numbers
sudo apt-get install -y --no-install-recommends \
    cuda-10-2 \
    libcudnn7=7.6.5.32-1+cuda10.2  \
    libcudnn7-dev=7.6.5.32-1+cuda10.2;
sudo apt-get install -y --no-install-recommends \
    libnvinfer7=7.0.0-1+cuda10.2 \
    libnvinfer-dev=7.0.0-1+cuda10.2 \
    libnvinfer-plugin7=7.0.0-1+cuda10.2;

在 Python 中测试 GPU 可见性

python3
>>> import tensorflow as tf
>>> tf.test.is_gpu_available()

关于 TensorFlow 导入的未来警告

https://github.com/tensorflow/tensorflow/issues/30427

两种解决方案：

pip3 install tf-nightly-gpu
pip3 install "numpy<1.17"

更新：

您还需要正确的 tensorflow 版本以匹配您的 CUDA 版本

Tensorflow / CUDA 版本组合：

Tensorflow v2.x 不支持 CUDA 9（Ubuntu 18.4 默认）
Tensorflow v2.1.0 适用于 CUDA 10.1
Tensorflow v2.0.0 可与 CUDA 10.0 配合使用

查看完整列表：https://www.tensorflow.org/install/source#tested_build_configurations

您可能需要使用与您的 CUDA 匹配的命名版本重新安装 tensorflow

pip uninstall tensorflow tensorflow-gpu
pip install tensorflow==2.1.0 tensorflow-gpu==2.1.0

然后在 ~/.bashrc 中添加 CUDA 到 $PATH 和 $LD_LIBRARY_PATH

~/.bashrc

# CUDA Environment Setup: https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#environment-setup
for CUDA_BIN_DIR in `find /usr/local/cuda-*/bin   -maxdepth 0`; do export PATH="$PATH:$CUDA_BIN_DIR"; done;
for CUDA_LIB_DIR in `find /usr/local/cuda-*/lib64 -maxdepth 0`; do export LD_LIBRARY_PATH="${LD_LIBRARY_PATH:+${LD_LIBRARY_PATH}:}$CUDA_LIB_DIR"; done;

export            PATH=`echo $PATH            | tr ':' '\n' | awk '!x[$0]++' | tr '\n' ':' | sed 's/:$//g'` # Deduplicate $PATH
export LD_LIBRARY_PATH=`echo $LD_LIBRARY_PATH | tr ':' '\n' | awk '!x[$0]++' | tr '\n' ':' | sed 's/:$//g'` # Deduplicate $LD_LIBRARY_PATH

【讨论】：

这篇文章是正确的，如果你在 Ubuntu 18.04 上针对 Tensorflow 1.15.0 去选择 Cuda 10.0，尽管这里写的是：tensorflow.org/install/gpu
这对我使用 tensorflow 1.15.2 有所帮助。谢谢

【解决方案4】：

当安装的cuda和tensorflow版本不兼容时会出现此错误。我在使用 cuda 9 运行 tensorflow 版本 1.13.0 时遇到了类似的 ImportError。由于我已经使用 pip 在虚拟环境中安装了 tensorflow，因此我只是卸载了 tensorflow 1.13.0 并安装了 tensorflow 1.12.0，如下所示；

    pip uninstall tensorflow-gpu tensorflow-estimator tensorboard
    pip install tensorflow-gpu==1.12.0

现在一切正常。

【讨论】：

这对我有用，但只适用于 python3.6，而不是 3.5 或 3.7。
我收到另一个错误ImportError: libcublas.so.9.0: cannot open shared object file: No such file or directory

【解决方案5】：

正如 CalderBot 所说，您也可以这样做

sudo cp -r /usr/local/cuda-10.2/lib64/libcu* /usr/local/cuda-10.1/lib64/

【讨论】：

【解决方案6】：

我在我的 conda 环境中安装了正确版本的 CUDA 和 tensorflow-gpu==1.14.0，但不知何故我仍然收到此错误消息。 This post帮我终于解决了。

我之前通过pip 安装了tensorflow-gpu - 在创建新环境并通过conda 安装tensorflow-gpu 后解决了我的问题。

conda install -c anaconda tensorflow-gpu=1.14.0

【讨论】：

【解决方案7】：

我遇到了同样的问题。我通过将以下命令添加到“.bashrc”文件来修复它。

导出 LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda-10.0/lib64/

系统配置：

Ubuntu 16.04 LTS
Tensorflow GPU 2.0beta1
Cuda 10.0
cuDNN 7.6.0 for Cuda 10.0

我使用 conda 来配置我的系统。

【讨论】：

【解决方案8】：

问题是由您当前的 cuda 版本 10.1 引起的（我们可以从图片的右上角看到）。

从TF官网可以看到，tf和cuda的对应关系是：TF website for the chart

Version                 cuDNN    CUDA
tensorflow-2.1.0         7.6       10.1
tensorflow-2.0.0         7.4       10.0
tensorflow_gpu-1.14.0    7.4       10.0
tensorflow_gpu-1.13.1    7.4       10.0

因此，您可以将您的 tf 升级到 2.1 或降级您的 cuda：

conda install cudatoolkit=10.0.130

然后它也会自动降级你的cudnn。

【讨论】：

我每次在新设备上重新使用虚拟环境时都会遇到这个问题。因为我想同时使用 PyTorch 和 TF。但是，1.4.0 以后的 PyTorch 需要 cuda 10.1，而 2.1 以下的 TF 不兼容。所以我同时选择1.3.0和1.13.1！

【解决方案9】：

如果有人仍然遇到此问题，libcublas.so.10 可以存在，但名称为 libcublas.so.10.0

所以，你可以通过运行来修复它：

sudo ln libcublas.so.10.0.130 libcublas.so.10

在/usr/local/cuda-10.0/lib64

【讨论】：

【解决方案10】：

你的电脑支持 CUDA 吗？

在 linux 中，您可以验证您的系统是否具有支持 CUDA 的 GPU：

$ lspci | grep -i nvidia

如果您没有看到任何设置，请通过在命令行输入 update-pciids（通常在 /sbin 中找到）更新 Linux 维护的 PCI 硬件数据库，然后重新运行之前的 lspci 命令。

在此页面中，您有安装 CUDA 的说明： https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html

如果您的计算机不支持 CUDA，您可以安装另一个 tensorflow 发行版或编译 tensorflow 代码：https://www.tensorflow.org/install/source

【讨论】：

【解决方案11】：

更改我的 tensorflow 版本解决了我的问题。

检查这个问题1 https://github.com/tensorflow/tensorflow/issues/26182)

官方的 tensorflow-gpu 二进制文件（由 pip 或 conda 下载的那个）是用 cuda 9.0、cudnn 7 自 TF 1.5 和 cuda 10.0、cudnn 7 自 TF 1.13 构建的。这些都写在发行说明中。如果使用官方二进制文件，则必须使用匹配版本的 cuda。

【讨论】：

【解决方案12】：

阿明，

当我尝试从 tensorflow 模型包运行 imagenet 教程时遇到同样的错误 -- https://github.com/tensorflow/models/tree/master/tutorials/image/imagenet

 python3 classify_image.py
 ...
 2019-07-21 22:29:58.367858: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Could not dlopen library 'libcudart.so.10.0'; dlerror: libcudart.so.10.0: cannot open shared object file: No such file or directory
 2019-07-21 22:29:58.367982: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Could not dlopen library 'libcublas.so.10.0'; dlerror: libcublas.so.10.0: cannot open shared object file: No such file or directory
 2019-07-21 22:29:58.368112: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Could not dlopen library 'libcufft.so.10.0'; dlerror: libcufft.so.10.0: cannot open shared object file: No such file or directory
 2019-07-21 22:29:58.368234: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Could not dlopen library 'libcurand.so.10.0'; dlerror: libcurand.so.10.0: cannot open shared object file: No such file or directory
 2019-07-21 22:29:58.368369: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Could not dlopen library 'libcusolver.so.10.0'; dlerror: libcusolver.so.10.0: cannot open shared object file: No such file or directory
 2019-07-21 22:29:58.368498: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Could not dlopen library 'libcusparse.so.10.0'; dlerror: libcusparse.so.10.0: cannot open shared object file: No such file or directory
 2019-07-21 22:29:58.374333: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcudnn.so.7

我认为某处存在版本不兼容，并且可能是 tensorflow，仍然依赖于 cuda 库提供的旧版本的二进制文件。转到存储二进制文件的地方并创建一个名为 10.0 但目标为 10.1 或库的默认版本的链接，似乎可以解决我的问题。

 # cd /usr/lib/x86_64-linux-gnu
 # ln -s libcudart.so.10.1 libcudart.so.10.0
 # ln -s libcublas.so libcublas.so.10.0
 # ln -s libcufft.so libcufft.so.10.0
 # ln -s libcurand.so libcurand.so.10.0
 # ln -s libcusolver.so libcusolver.so.10.0
 # ln -s libcusparse.so libcusparse.so.10.0

现在我可以成功运行教程了

 2019-07-24 21:43:21.172908: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcudart.so.10.0
 2019-07-24 21:43:21.174653: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcublas.so.10.0
 2019-07-24 21:43:21.175826: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcufft.so.10.0
 2019-07-24 21:43:21.182305: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcurand.so.10.0
 2019-07-24 21:43:21.183970: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcusolver.so.10.0
 2019-07-24 21:43:21.206796: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcusparse.so.10.0
 2019-07-24 21:43:21.210685: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcudnn.so.7
 2019-07-24 21:43:21.212694: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1763] Adding visible gpu devices: 0
 2019-07-24 21:43:21.213060: I tensorflow/core/platform/cpu_feature_guard.cc:142]      
 Your CPU supports instructions that this TensorFlow binary was not compiled to use: FMA
 2019-07-24 21:43:21.238541: I tensorflow/core/platform/profile_utils/cpu_utils.cc:94] CPU Frequency: 3214745000 Hz
 2019-07-24 21:43:21.240096: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x557e2b682ce0 executing computations on platform Host. Devices:
 2019-07-24 21:43:21.240162: I tensorflow/compiler/xla/service/service.cc:175]   StreamExecutor device (0): <undefined>, <undefined>
 2019-07-24 21:43:21.355158: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x557e2b652000 executing computations on platform CUDA. Devices:
 2019-07-24 21:43:21.355234: I tensorflow/compiler/xla/service/service.cc:175]   StreamExecutor device (0): GeForce GTX 1060 6GB, Compute Capability 6.1
 2019-07-24 21:43:21.357074: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1640] Found device 0 with properties: 
 name: GeForce GTX 1060 6GB major: 6 minor: 1 memoryClockRate(GHz): 1.7715
 pciBusID: 0000:01:00.0
 2019-07-24 21:43:21.357151: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcudart.so.10.0
 2019-07-24 21:43:21.357207: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcublas.so.10.0
 2019-07-24 21:43:21.357245: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcufft.so.10.0
 2019-07-24 21:43:21.357283: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcurand.so.10.0
 2019-07-24 21:43:21.357321: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcusolver.so.10.0
 2019-07-24 21:43:21.357358: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcusparse.so.10.0
 2019-07-24 21:43:21.357395: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcudnn.so.7
 2019-07-24 21:43:21.360449: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1763] Adding visible gpu devices: 0
 2019-07-24 21:43:21.380616: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcudart.so.10.0
 2019-07-24 21:43:21.385223: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1181] Device interconnect StreamExecutor with strength 1 edge matrix:
 2019-07-24 21:43:21.385272: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1187]      0 
 2019-07-24 21:43:21.385299: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1200] 0:   N 
 2019-07-24 21:43:21.388647: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1326] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 5250 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1060 6GB, pci bus id: 0000:01:00.0, compute capability: 6.1)
 2019-07-24 21:43:32.001598: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcublas.so.10.0
 2019-07-24 21:43:32.532105: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcudnn.so.7
 W0724 21:43:34.981204 140284114071872 deprecation_wrapper.py:119] From classify_image.py:85: The name tf.gfile.GFile is deprecated. Please use tf.io.gfile.GFile instead.

【讨论】：

【解决方案13】：

我在尝试安装 spconv 时遇到了类似的问题。

File "/home/kmario23/anaconda3/envs/py38/lib/python3.8/site-packages/torch/_ops.py", line 105, in load_library
    ctypes.CDLL(path)
  File "/home/kmario23/anaconda3/envs/py38/lib/python3.8/ctypes/__init__.py", line 373, in __init__
    self._handle = _dlopen(self._name, mode)
OSError: libcublas.so.10: cannot open shared object file: No such file or directory

在特定环境中安装cuda工具包版本10.1解决了这个问题：

$ conda install -c anaconda cudatoolkit=10.1

【讨论】：