Error in deep learning code

Question

shweta gaikwad 2019-7-28

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/473774-error-in-deep-learning-code

评论： Walter Roberson 2019-7-28

在 MATLAB Online 中打开

I am trying code. And gettiing an error like as follows:

----Error using load

Cannot read file C:\Users\HP\AppData\Local\Temp\imagenet-caffe-alex.mat.

----Error in helperImportMatConvNet (line 19)

matconvnet = load(matfile);

----Error in tranfdeeplearning (line 18)

net = helperImportMatConvNet(cnnMatFile);

%% Load Image Data
location = 'C:\Users\HP\Desktop\data';
imds = imageDatastore(location,'IncludeSubfolders',1,...
    'LabelSource','foldernames');
tbl = countEachLabel(imds);
%% SeriesNetwork object can be used to inspect the network architecture,
% classify new data, and extract network activations from specific layers.
% Location of pre-trained "AlexNet"
cnnURL = 'http://www.vlfeat.org/matconvnet/models/beta16/imagenet-caffe-alex.mat';
cnnMatFile = fullfile(tempdir, 'imagenet-caffe-alex.mat');
if ~exist(cnnMatFile, 'file') 
    disp('Downloading pre-trained CNN model...');
    websave(cnnMatFile, cnnURL);
end
net = helperImportMatConvNet(cnnMatFile);
%% Look at structure of pre-trained network
% Notice the last layer performs 1000 object classification
net.Layers
%% Perform net surgery
layers = net.Layers(1:end-3);
% %optional, Add new fully connected layer for 2 categories.
layers(end+1) = fullyConnectedLayer(64, 'Name', 'special_2');
layers(end+1) = reluLayer;
% introduce another layer : adding non-linearity 
% improving the network's ability to handle data
% Not enough training 1.2 million (total)
layers(end+1) = fullyConnectedLayer(height(tbl), 'Name', 'fc8_2')
% Add the softmax layer and the classification layer which make up the
% remaining portion of the networks classification layers.
layers(end+1) = softmaxLayer
layers(end+1) = classificationLayer()
% Modify image layer to add randcrop data augmentation. This increases the
% diversity of training images. The size of the input images is set to the
% original networks input size.
layers(1) = imageInputLayer([227 227 3]);
%% Setup learning rates for fine-tuning
% fc 8 - bump up learning rate for last layers
layers(end-2).WeightLearnRateFactor = 10;
layers(end-2).WeightL2Factor = 1;
layers(end-2).BiasLearnRateFactor = 20;
layers(end-2).BiasL2Factor = 0;
%% Equalize number of images of each class in training set
minSetCount = min(tbl{:,2});
%Use splitEachLabel method to trim the set.
imds = splitEachLabel(imds, minSetCount);
% Notice that each set now has exactly the same number of images.
countEachLabel(imds)
[trainingDS, testDS] = splitEachLabel(imds, 0.7,'randomize');
% Convert labels to categoricals
trainingDS.Labels = categorical(trainingDS.Labels);
trainingDS.ReadFcn = @readFunctionTrain;
% % Setup test data for validation
% testDS.Labels = categorical(testDS.Labels);
% testDS.ReadFcn = @readFunctionValidation;
%% Fine-tune the Network
miniBatchSize = 50; % lower this if your GPU runs out of memory.
numImages = numel(trainingDS.Files);
% Run training for 5000 iterations. Convert 20000 iterations into the
% number of epochs this will be.
maxEpochs = 20; 
lr = 0.0001;
opts = trainingOptions('sgdm', ...
    'LearnRateSchedule', 'none',...
    'InitialLearnRate', lr,... 
    'MaxEpochs', maxEpochs, ...
    'MiniBatchSize', miniBatchSize);
net = trainNetwork(trainingDS, layers, opts);
save('trainedNet.mat','net')
save('testDS.mat','testDS')
return