GPU TensorFlow安装问题

Question

我是一个Ubuntu的新用户，花费了一周以上的时间尝试安装gpu tensorflow。我尝试按照网站和博客上的教程进行操作，但每当我在python控制台中导入tensorflow时，总是会显示：

import tensorflow
I tensorflow/stream_executor/dso_loader.cc:128] successfully opened CUDA library libcublas.so.8.0 locally
I tensorflow/stream_executor/dso_loader.cc:128] successfully opened CUDA library libcudnn.so.5.1.5 locally
I tensorflow/stream_executor/dso_loader.cc:128] successfully opened CUDA library libcufft.so.8.0 locally
I tensorflow/stream_executor/dso_loader.cc:119] Couldn't open CUDA   library libcuda.so.1. LD_LIBRARY_PATH:   :/usr/local/cuda/lib64:/usr/local/cuda/extras/CUPTI/lib64
I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:165] hostname: HanGuo
I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:189] libcuda reported version is: Not found: was unable to find libcuda.so DSO loaded into this program
I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:193] kernel reported version is: Permission denied: could not open driver version path for reading: /proc/driver/nvidia/version
I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1091]  LD_LIBRARY_PATH: :/usr/local/cuda/lib64:/usr/local/cuda/extras/CUPTI/lib64
I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1092] failed to find libcuda.so on this system: Failed precondition: could not dlopen DSO: libcuda.so.1; dlerror: libnvidia-fatbinaryloader.so.367.57: cannot open shared object file: No such file or directory
I tensorflow/stream_executor/dso_loader.cc:128] successfully opened CUDA library libcurand.so.8.0 locally

我尝试在谷歌上搜索答案，但还没有真正的解决方案。有人能帮帮我吗？

- Han Guo

---

请提供有关您如何安装Tensorflow和Cuda的更多信息。 - BernardoGO

我从Nvidia网站下载了CUDA .run文件，并使用$ sudo tar -xzvf cudnn-8.0-linux-x64-v5.1.tgz将cuda安装到/usr/local/cuda中。 然后，我使用以下命令将cudnn.h复制到/usr/local/cuda/include中：$ sudo cp cuda/include/cudnn.h /usr/local/cuda/include，并将libcudnn*复制到/usr/local/cuda/lib64中：$ sudo cp cuda/lib64/libcudnn* /usr/local/cuda/lib64。最后，我使用以下命令更改权限：$ sudo chmod a+r /usr/local/cuda/include/cudnn.h /usr/local/cuda/lib64/libcudnn*。TensorFlow是通过git clone https://github.com/tensorflow/tensorflow下载的，并使用./configure进行配置。 - Han Guo

基本上我完全按照(https://alliseesolutions.wordpress.com/2016/09/08/install-gpu-tensorflow-from-sources-w-ubuntu-16-04-and-cuda-8-0-rc/)所说的方式操作。 - Han Guo

你执行了 $ sudo sh cuda_8.0.44_linux.run --override 吗？$ ls /usr/local/cuda 的输出是什么？ - BernardoGO

bin extras jre libnsight nvml pkgconfig share tools doc include lib64 libnvvp nvvm samples src version.txt - Han Guo

Answer 1

根据 $ ls /usr/local/cuda 的输出，您已经成功安装了它。

现在缺少的是 PATH 定义，以便 TensorFlow 可以找到它。

1）打开您的 .bashrc 文件：gedit ~/.bashrc 2）将以下两行附加到文件末尾：

export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:/usr/local/cuda/lib64:/usr/local/cuda/extras/CUPTI/lib64:/usr/local/cuda/include"  
export CUDA_HOME=/usr/local/cuda

3) 保存并关闭文本文件后，您可以返回原始终端并键入此命令以重新加载您的 .bashrc 文件：

source ~/.bashrc

虽然第三部分应该会照顾到这个问题，但如果你想100%确保它能正常工作，请重新启动计算机。

现在，对于TensorFlow，请记得将CUDA路径设置为/usr/local/cuda，版本为8.0。

- BernardoGO

Answer 2

安装libcuda5-dev可能会解决您的问题

- Cro