Is there a way to use weights without using gpuArray

7 次查看（过去 30 天）

Marina Ghobrial 2021-7-25

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/885244-is-there-a-way-to-use-weights-without-using-gpuarray

评论： Marina Ghobrial 2021-7-27

I found this code online and was wondering if there is another way to use it without the gpuArray function. I keep recieving this error when I use it and I am not familiar with gpu:

Error using gpuArray

Failed to load graphics driver. Unable to load library 'nvcuda.dll'. The error was:

The specified module could not be found.

Update or reinstall your graphics driver. For more information on GPU support, see GPU Support by Release.

varSize = 21;
conv1 = convolution2dLayer(5,varSize,'Padding',2,'BiasLearnRateFactor',2);
conv1.Weights = gpuArray(single(randn([5 5 3 varSize])*0.0001));
fc1 = fullyConnectedLayer(64,'BiasLearnRateFactor',2);
fc1.Weights = gpuArray(single(randn([64 576])*0.1));
fc2 = fullyConnectedLayer(4,'BiasLearnRateFactor',2);
 fc2.Weights = gpuArray(single(randn([4 64])*0.1));

i want to be able to use convulation neural network and be able to use the weights because I know they help the progrma run faster. so my question is: Is there a way to use weights without using gpuArray?

Thank you

4 个评论
显示 2更早的评论隐藏 2更早的评论

Walter Roberson 2021-7-25

在 MATLAB Online 中打开

varSize = 21;

gpuArray = @(x) x;

conv1 = convolution2dLayer(5,varSize,'Padding',2,'BiasLearnRateFactor',2);

conv1.Weights = gpuArray(single(randn([5 5 3 varSize])*0.0001));

fc1 = fullyConnectedLayer(576,'BiasLearnRateFactor',2);

fc1.Weights = gpuArray(single(randn([576, 64])*0.1));

fc2 = fullyConnectedLayer(4,'BiasLearnRateFactor',2);

fc2.Weights = gpuArray(single(randn([4 64])*0.1))

fc2 =

FullyConnectedLayer with properties: Name: '' Hyperparameters InputSize: 64 OutputSize: 4 Learnable Parameters Weights: [4×64 single] Bias: [] Show all properties

fc1

fc1 =

FullyConnectedLayer with properties: Name: '' Hyperparameters InputSize: 64 OutputSize: 576 Learnable Parameters Weights: [576×64 single] Bias: [] Show all properties

fc2

fc2 =

FullyConnectedLayer with properties: Name: '' Hyperparameters InputSize: 64 OutputSize: 4 Learnable Parameters Weights: [4×64 single] Bias: [] Show all properties

Marina Ghobrial 2021-7-27