MATLAB CUDNN

    技术2022-07-10  140

    问题:

    环境:

    Windows 10Matlab2018bgpuDevice: GeForce RTX 2080 Tinvcc --version: Cuda compilation tools, release 10.2, V10.2.89

    matlab 运行Deep Learning Toolbox Model for AlexNet Network 代码 报错

    net = alexnet layer = 'conv1' channels = 1:36 I = deepDreamImage(net,layer,channels,... 'PyramidLevels',1,... 'Verbose',0) figure for i = 1:36 subplot(6,6,i) imshow(I(:,:,:,i)) end

    代码来自博客:如何可视化卷积神经网络特征(基于matlab R2018b)

    警告: The CUDA driver must recompile the GPU libraries because your device is more recent than the libraries. Recompiling can take several minutes. Learn more. 错误使用 nnet.internal.cnngpu.convolveForward2D Unexpected error calling cuDNN: CUDNN_STATUS_EXECUTION_FAILED.

    解决

    方法1:

    在matlab命令行输入:

    >> gpuDevice ans = CUDADevice - 属性: Name: 'GeForce RTX 2080 Ti' Index: 1 ComputeCapability: '7.5' SupportsDouble: 1 DriverVersion: 10.2000 >> nvcc --version 未定义函数或变量 'nvcc'。

    在matlab中配置CUDA环境变量

    操作: 点击:主页 ,设置路径,添加文件夹,选择下面两个路径: C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2\bin C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2\libnvvp 添加完成后点击保存。

    之后再次运行代码,可以成功运行啦。

    方法2:(玄学但是有用…)

    在代码中随意添加一行:将0.0001存入GUP中,之后再重新运行net = trainNetwork( ),不会出现报错问题。

    ile = gpuArray(0.0001); net = trainNetwork(XTrain,YTrain,lgraph_1,options);

    参考网址:https://www.mathworks.com/matlabcentral/answers/391368-cudnn-error-on-surface-book-2-when-running-cnn-example?s_tid=srchtitle

    Processed: 0.008, SQL: 9