Infrorror:libcublas.so.10.0:无法打开共享对象文件:没有此类文件或目录

我从以下链接下载了cuda 10.0CUDA 10.0

然后我使用以下命令安装了它：

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

选择了我通过链接cudnn电源链接下载的正确版本后之后，我添加了cudnn的include和lib文件如下：

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*

修改了lib的.bashrc和cuda 10.0的路径，如果没有它，则需要将它们添加到.bashrc

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}}

在所有这些步骤之后，我成功地在Python3中导入了TensorFlow。。

从此答案中复制：https：//github.com/tensorflow/tensorflow/issues/26182#issuecomment-684993950

... libcublas.so.10坐落在/usr/local/cuda-10.2/lib64（来自Nvidia的惊喜 - 安装10.1安装10.2件东西），但只有/usr/local/cuda在路径中包括哪个点to/usr/local/cuda-10.1。

fix 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：该修复程序已知可以在CUDA 10.1，v10.1.243（用nvcc -V打印您的版本）。

cuda 10.1每个TensorFlow文档）抛出can't find libcublas.so.10.0错误。lib存在于/usr/local/cuda-10.1/targets/x86_64-linux/lib/中，但被错误命名。

还有另一个（丢失的）Stackoverflow帖子说这是包裹的依赖性问题，可以用额外的CLI标志来固定。这似乎并没有解决我的问题。

测试的解决方法是修改说明以降级为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似乎也遭受了相同的问题

# 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;

测试Python中的GPU可见性

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

TensorFlow Import in tensorflow上的未来胜地

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

两个解决方案：

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

更新：

您还需要正确的TensorFlow版本才能与CUDA版本匹配

TensorFlow/CUDA版本组合：

• TensorFlow v2.x不支持CUDA 9（Ubuntu 18.4默认值）
• TensorFlow v2.1.0与CUDA 10.1
一起使用
• TensorFlow V2.0.0与CUDA 10.0
一起使用

请参阅完整列表：https：//www.tensorflow.org/install/source#tested_build_configurations

您可能需要用匹配您的CUDA

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

然后将cuda添加到$路径和$ ld_library_path in〜/.bashrc

〜/.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

我在我的conda环境上安装了正确的CUDA和tensorflow-gpu==1.14.0版本，但是我仍然以某种方式收到此错误消息。这篇文章帮助我最终解决了它。

我以前已经通过pip安装了tensorflow-gpu-创建新环境并通过conda安装CC_11解决了我的问题。

conda install -c anaconda tensorflow-gpu=1.14.0

当安装CUDA和TensorFlow的版本不兼容时，就会发生此错误。使用CUDA 9运行TensorFlow版本1.13.0时，我遇到了类似的Improterror。由于我在使用PIP的虚拟环境上安装了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

现在一切都起作用。

正如Calderbot提到的那样，您也可以做到这一点

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

我也有相同的问题。我通过将以下命令添加到' .bashrc '文件。

export ld_library_path = $ ld_library_path：/usr/local/cuda-10.0.0/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配置我的系统。

问题是由您当前的CUDA版本（10.1）引起的（正如我们可以从顶部看到图像的右角）。

您可以从官方TF网站上看到，TF和CUDA之间的信件是：图表的TF网站

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TF升级到2.1，也可以使用：

conda install cudatoolkit=10.0.130

然后它也会自动降级您的cudnn。

如果某人仍在遇到这个问题，则可以存在libcublas.so.10，但名称为libcublas.so.10.0

因此，您可以通过运行：

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

在 /usr/local/cuda-10.0/lib64

更改我的tensorflow版本解决了我的问题。

检查此问题1Https：//github.com/tensorflow/tensorflow/issues/26182）

官方Tensorflow-GPU二进制文件（由PIP或CONDA下载的二进制文件）是使用CUDA 9.0，自tf 1.5以来的Cudnn 7和Cuda 10.0，Cuda 10.0，Cudnn 7构建的，自TF 1.13以来。这些写在发行说明中。如果使用官方二进制文件，则必须使用CUDA的匹配版本。

您的计算机CUDA是否有能力？

在Linux中，您可以验证系统是否具有CUDA能力的GPU：

$ lspci | grep -i nvidia

如果您看不到任何设置，请通过在命令行中输入update-pciids（通常在/sbin中）来更新Linux维护的PCI硬件数据库，然后重新运行以前的LSPCI命令。

在此页面中，您有指令安装CUDA：https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html

如果您的计算机没有CUDA能力，则可以安装另一个TensorFlow或编译TensorFlow代码的分布：https：//www.tensorflow.org/install/source/source

cuda-10.2安装似乎已将libcublas.so.10移至一个目录更深。在我的cuda-9.1的非根安装中，上述库位于我在安装过程中指定的--toolkitpath的lib64文件夹中。现在它在lib64/lib64/文件夹中。因此，我要做的就是向LD_LIBRARY_PATH添加更多路径。

1. 我是通过从下面的网站上理解的。它说CUDA工具套件安装不正确。https://saturncloud.io/blog/how-to-fix-the-tensorflow-importerror-libcublasso80-error/
2. 然后我重新安装CUDA工具包：https://developer.nvidia.com/cuda-downloads?target_os=linux& target_ark = x86_64＆amp; distribution = ubuntu＆amp; target_version = 22.04＆amp；
3. 然后我去网站：解决安装问题：https://forums.developer.nvidia.com/t/driver-in--inknown-state-after-awter-awter-to-install-cuda-ubuntu-22-04/217106/2

amin，

尝试从TensorFlow模型包运行Imagenet教程时，我会遇到相同的错误 - https://github.com/tensorflow/models/tree/master/master/master/tutorials/image/image/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

我认为某个地方的版本不兼容，可能的张力可能仍然依赖于CUDA库提供的旧版本的二进制文件。去存储二进制文件并创建命名为10.0但Targets 10.1或库的默认版本的链接似乎为我解决了问题。

 # 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

现在我能够成功运行教程

 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.

我在尝试安装SPCONV时面临类似的问题。

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10.1解决了问题：

$ conda install -c anaconda cudatoolkit=10.1