How to remove data Overfitting Issue in my training model

1 次查看（过去 30 天）
Naitik
Naitik 2024-4-12
回答： Shreshth 2025-1-17
I am working on project of facial recognisation of endangered species and I am getting mini batch accuracy of 100% , I am using AlexNet for training. The number of images of endangered species I am using is African elephant with count of 470. Even after do argumantation mini batch accuracy is still same. Help me to remove this overfitting.
Dataset Link: Dataset
Data Training Code:
net = alexnet;
layers = [imageInputLayer([227 227 3])
net(2:end-3)
fullyConnectedLayer(1)
softmaxLayer
classificationLayer()
];
opt = trainingOptions('sgdm', 'MaxEpochs', 100, 'InitialLearnRate', 0.0001,'Plots','training-progress');
training = trainNetwork(trainData,layers,opt);

请先登录，再回答此问题。

回答（1 个）

Shreshth
Shreshth 2025-1-17
Hi Naitik,
I found similar question in the community -
https://www.mathworks.com/matlabcentral/answers/1950373-overfitting-deep-neural-network
Also, refer to the following MathWorks documentation for more information on how to avoid overfitting:
https://www.mathworks.com/discovery/overfitting.html
Hope this helps!

类别

在 Help Center 和 File Exchange 中查找有关 Image Data Workflows 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