CNN, resnet50 - plot accuracy

HK
2021 11 19
1 个回答
回答已采纳
更新时间：2021 11 24
6 次查看（30 天）

0 个投票

Can anyone help me with plotting the training progress (accuracy and loss)?
There seems to be a problem with the last line, but I'm not sure what the problem is. Help!
outputFolder = fullfile('Caltech')
rootFolder = fullfile(outputFolder, '101_ObjectCategories')
categories = {'normal_NB_resized', 'slow_OB_resized'}
imds = imageDatastore(fullfile(rootFolder,categories), 'LabelSource','foldernames')
tbl = countEachLabel(imds)
minSetCount = min(tbl{:,2})
imds = splitEachLabel(imds, minSetCount, 'randomize')
countEachLabel(imds)
normal = find(imds.Labels == 'normal', 1)
slow = find(imds.Labels == 'slow', 1)
net = resnet50();
net.Layers(1);
net.Layers(end);
numel(net.Layers(end).ClassNames);
[trainingSet, testSet] = splitEachLabel(imds, 0.7, 'randomize');
imageSize = net.Layers(1).InputSize;
augmentedTrainingSet = augmentedImageDatastore(imageSize,...
trainingSet, 'ColorPreprocessing', 'gray2rgb');
augmentedTestSet = augmentedImageDatastore(imageSize,...
testSet, 'ColorPreprocessing', 'gray2rgb');
w1 = net.Layers(2).Weights;
w1 = mat2gray(w1);
featureLayer = 'fc1000';
trainingFeatures = activations(net,...
augmentedTrainingSet, featureLayer, 'MiniBatchSize', 64, 'OutputAs', 'columns');
trainingLables = trainingSet.Labels;
classifier = fitcecoc(trainingFeatures, trainingLables,...
'Learner', 'Linear', 'Coding', 'onevsall', 'ObservationsIn', 'columns');
testFeatures = activations(net,...
augmentedTestSet, featureLayer, 'MiniBatchSize', 64, 'OutputAs', 'columns');
predictLabels = predict(classifier, testFeatures, 'ObservationsIn', 'columns');
testLables = testSet.Labels
confMat = confusionmat(testLables, predictLabels)
confMat = bsxfun(@rdivide, confMat, sum(confMat,2))
mean(diag(confMat))
newImage = imread(fullfile('test.jpg'));
ds = augmentedImageDatastore(imageSize,...
newImage, 'ColorPreprocessing', 'gray2rgb');
imageFeatures = activations(net,...
ds, featureLayer, 'MiniBatchSize', 64, 'OutputAs', 'columns');
label = predict(classifier, imageFeatures, 'ObservationsIn', 'columns');
sprintf('loaded image belongs to %s class', label)
options = trainingOptions('sgdm', ...
'InitialLearnRate',0.01, ...
'MaxEpochs',4, ...
'Shuffle','every-epoch', ...
'ValidationData',imds, ...
'ValidationFrequency',30, ...
'Verbose',false, ...
'Plots','training-progress');
netT = trainNetwork(imds,layers,options);

请先登录，再回答此问题。

 采纳的回答

Prateek Rai
Prateek Rai 2021-11-22
编辑：Prateek Rai 2021-11-24

0 个投票

Hi,
There are 2 facts about the code which is important to note:
(1) Since, your dataset is big that is why it is taking more time to perform dataset involving tasks and hence you are not able to reach the last line of the code i.e. training the network and plotting the training progress. Thus, it seems that the network training started already without showing the training progress but actually training has not started.
(2) Secondly, if your image datastore is somehow created and your last line started executing then also error will come stating: "Unrecognized function or variable 'layers'. It is because you have not declared variable 'layers' but used it in the last line of the code.

更多回答（0 个）

类别

在 帮助中心 和 File Exchange 中查找有关 Classification Ensembles 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by