deep learning_alexnet

Question

Srinidhi Gorityala 2020-5-26

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/533708-deep-learning_alexnet

编辑： Walter Roberson 2020-10-2

采纳的回答： Mohammad Sami

在 MATLAB Online 中打开

Helo, iam working on Alexnet network.... below is the error i have got.could anyone please help me in solving this.

Error using trainNetwork (line 170)

The training images are of size 227x227x3 but the input layer expects images of size 227x227x1.

Error in Untitled1 (line 63)

net = trainNetwork(augimdsTrain,lgraph,options);

clear all;
clc;
close all;
myTrainingFolder = 'C:\Users\Admin\Desktop\Major Project\cnn_dataset';
%testingFolder = 'C:\Users\Be Happy\Documents\MATLAB\gtsrbtest';
imds = imageDatastore(myTrainingFolder,'IncludeSubfolders', true,  'LabelSource', 'foldernames');
[imdsTrain,imdsValidation] = splitEachLabel(imds,0.7);
net = alexnet(); % analyzeNetwork(lgraph)
numClasses = numel(categories(imdsTrain.Labels)); % number of classes = number of folders
imageSize = [227 227]; % you can use here the original dataset size
global GinputSize
GinputSize = imageSize;
lgraph = layerGraph(net.Layers);
lgraph = removeLayers(lgraph, 'fc8');
lgraph = removeLayers(lgraph, 'prob');
lgraph = removeLayers(lgraph, 'output');
% create and add layers
inputLayer = imageInputLayer([imageSize 1], 'Name', net.Layers(1).Name,...
    'DataAugmentation', net.Layers(1).DataAugmentation, ...
    'Normalization', net.Layers(1).Normalization);
lgraph = replaceLayer(lgraph,net.Layers(1).Name,inputLayer);
newConv1_Weights = net.Layers(2).Weights;
newConv1_Weights = mean(newConv1_Weights(:,:,1:3,:), 3); % taking the mean of kernal channels
newConv1 = convolution2dLayer(net.Layers(2).FilterSize(1), net.Layers(2).NumFilters,...
    'Name', net.Layers(2).Name,...
    'NumChannels', inputLayer.InputSize(3),...
    'Stride', net.Layers(2).Stride,...
    'DilationFactor', net.Layers(2).DilationFactor,...
    'Padding', net.Layers(2).PaddingSize,...
    'Weights', newConv1_Weights,...BiasLearnRateFactor 
    'Bias', net.Layers(2).Bias,...
    'BiasLearnRateFactor', net.Layers(2).BiasLearnRateFactor);
lgraph = replaceLayer(lgraph,net.Layers(2).Name,newConv1);
lgraph = addLayers(lgraph, fullyConnectedLayer(numClasses,'Name', 'fc2'));
lgraph = addLayers(lgraph, softmaxLayer('Name', 'softmax'));
lgraph = addLayers(lgraph, classificationLayer('Name','output'));
lgraph = connectLayers(lgraph, 'drop7', 'fc2');
lgraph = connectLayers(lgraph, 'fc2', 'softmax');
lgraph = connectLayers(lgraph, 'softmax', 'output');
% -------------------------------------------------------------------------
augmenter = imageDataAugmenter( ...
    'RandRotation',[-20,20], ...
    'RandXReflection',1,...
    'RandYReflection',1,...
    'RandXTranslation',[-3 3], ...
    'RandYTranslation',[-3 3]);
%augimdsTrain = augmentedImageDatastore([224 224],imdsTrain,'DataAugmentation',augmenter);
%augimdsValidation = augmentedImageDatastore([224 224],imdsValidation,'DataAugmentation',augmenter);
augimdsTrain = augmentedImageDatastore(imageSize,imdsTrain);
augimdsValidation = augmentedImageDatastore(imageSize,imdsValidation);
options = trainingOptions('rmsprop', ...
    'MiniBatchSize',10, ...
    'MaxEpochs',20, ...
    'InitialLearnRate',1e-3, ...
    'Shuffle','every-epoch', ...
    'ValidationData',augimdsValidation, ...
    'ValidationFrequency',3, ...
    'Verbose',false, ...
    'Plots','training-progress');
net = trainNetwork(augimdsTrain,lgraph,options);
[YPred, probs] = classify(net,augimdsValidation);
accuracy = mean(YPred ==imdsValidation.Labels);
figure,
cm=confusionchart (imdsValidation.Labels, YPred);

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Walter Roberson 2020-10-2

User tagged with

https://in.mathworks.com/matlabcentral/answers/513441-how-to-fix-this-issue-through-running-the-program

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Mohammad Sami 2020-5-26

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/533708-deep-learning_alexnet#answer_438823

编辑：Mohammad Sami 2020-5-26

在 MATLAB Online 中打开

You have specified the image as single channel in your code. just change it to 3 channels (RGB).

imageInputLayer([imageSize 1],....

to

imageInputLayer([imageSize 3],....

5 个评论
显示 3更早的评论隐藏 3更早的评论

shivan artosh 2020-9-6

在 MATLAB Online 中打开

Hello Mr. Mohammad

I have the same code, 1. coluld you please write some explannation comment only for the following part of code.

2. i think this code is the same work when we modify AlexNet in (deepNetworkDesigner) am i right or not? or is there any different between 2 ways (manually/deepNetworkDesigner and coding)?

lgraph = layerGraph(net.Layers);
lgraph = removeLayers(lgraph, 'fc8');
lgraph = removeLayers(lgraph, 'prob');
lgraph = removeLayers(lgraph, 'output');
% create and add layers
inputLayer = imageInputLayer([imageSize 1], 'Name', net.Layers(1).Name,...
    'DataAugmentation', net.Layers(1).DataAugmentation, ...
    'Normalization', net.Layers(1).Normalization);
lgraph = replaceLayer(lgraph,net.Layers(1).Name,inputLayer);
newConv1_Weights = net.Layers(2).Weights;
newConv1_Weights = mean(newConv1_Weights(:,:,1:3,:), 3); % taking the mean of kernal channels
newConv1 = convolution2dLayer(net.Layers(2).FilterSize(1), net.Layers(2).NumFilters,...
    'Name', net.Layers(2).Name,...
    'NumChannels', inputLayer.InputSize(3),...
    'Stride', net.Layers(2).Stride,...
    'DilationFactor', net.Layers(2).DilationFactor,...
    'Padding', net.Layers(2).PaddingSize,...
    'Weights', newConv1_Weights,...BiasLearnRateFactor 
    'Bias', net.Layers(2).Bias,...
    'BiasLearnRateFactor', net.Layers(2).BiasLearnRateFactor);
lgraph = replaceLayer(lgraph,net.Layers(2).Name,newConv1);
lgraph = addLayers(lgraph, fullyConnectedLayer(numClasses,'Name', 'fc2'));
lgraph = addLayers(lgraph, softmaxLayer('Name', 'softmax'));
lgraph = addLayers(lgraph, classificationLayer('Name','output'));
lgraph = connectLayers(lgraph, 'drop7', 'fc2');
lgraph = connectLayers(lgraph, 'fc2', 'softmax');
lgraph = connectLayers(lgraph, 'softmax', 'output');
% -------------------------------------------------------------------------

Mohammad Sami 2020-9-7

This code is for transfer learning. That is when you already have a pretrained model that you wish to use for another purpose. The process is explained in detail in MATLAB's documentation.

https://www.mathworks.com/discovery/transfer-learning.html?s_tid=srchtitle

Using Deep Network Designer:

https://www.mathworks.com/help/deeplearning/gs/get-started-with-transfer-learning.html

Usign Manual Method:

https://www.mathworks.com/help/deeplearning/ug/transfer-learning-using-alexnet.html?s_tid=srchtitle

Since the network was trained with different classes then your current purpose, you have to remove the final few layers at least from the last fully connected layer onwards. This is because the final fully connected layer needs to have the same output size as the number of classess in your data.

You then have to have to create new layers for the layers that you removed. The new fully connected layer has the output size which matches the number of classes in your data. Finally you can add these newly connected layers to complete your network.

shivan artosh 2020-10-2