MATLAB: Learning Rate Drop Period

deep learningmachine learningtraining options

Hi!

I am using the 'trainNetwork' function for my net training.

Is there any way to set the 'LearnRateDropPeriod' to be less than 1 epoch? I need it to be updated more frequently.

Thanks in advance!

Best Answer

For most tasks, you can control the training algorithm details using the trainingOptions and trainNetwork functions. If the trainingOptions function does not provide the options you need for your task (for example, a custom learn rate schedule), then you can define your own custom training loop using automatic differentiation.

For an example showing how to train a network with a custom learn rate schedule, see Train Network Using Custom Training Loop.

Refer to Define Custom Training Loops, Loss Functions, and Networks, Solver Options.

Related Solutions

MATLAB: How to change learning rate and momentum in ‘trainlm’ neural network

use

options = trainingOptions('sgdm',...
    'MaxEpochs',60, ...
    'LearnRateDropFactor',0.1,...
    'LearnRateDropPeriod',1,...
    'LearnRateSchedule','piecewise',...
    'ValidationData',{valImages,valLabels},...
    'ValidationFrequency',50,...
    'Momentum',0.85,...
    'ValidationPatience',5,...
    'InitialLearnRate',0.1,...
    'MiniBatchSize',128,...
    'Verbose',true,...
    'executionenvironment','cpu',...
    'Shuffle','every-epoch',...
    'Plots','training-progress');
net = trainNetwork(...,...,...,options);

and choose the parameters accordingly.

For default values, see the docs of trainingOptions.

In terms of the learning rate and momentum, I typically start with a large one just to test the general behaviour and then I drop the learning rate to get more accurate results. Its difficult to give you some concrete advice because the optimal learning rate depends on your specific problem.

MATLAB: Drop Down and geoshow

I am not seeing the issue with your code, but it could be improved a bit. Your initial value on the dropdown is the first one - 'Harta1', and if you opened the dropdown menu and pressed it again callback won't be executed, but that's not an issue because you have already loaded Harta1 in the startupFcn.

Better way to do it is by having an intial option that tells you to choose the map, as a result, you don't have to have the startupFcn anymore.

% Value changed function: SelecteazaHartaDropDown
function SelecteazaHartaDropDownValueChanged(app, event)
     value = app.SelecteazaHartaDropDown.Value;
     switch value
         case 'Please select the map'
             cla(app.UIAxes)
         case 'Harta1'
             s=shaperead('C:\Users\Vostro\Desktop\App_Maps\Map1\map1.shp');
             geoshow(app.UIAxes,s);
         case 'Harta2'
             s=shaperead('C:\Users\Vostro\Desktop\App_Maps\Map2\map2.shp');
             geoshow(app.UIAxes,s);
     end
end

Best Answer

Related Solutions

MATLAB: How to change learning rate and momentum in ‘trainlm’ neural network

MATLAB: Drop Down and geoshow

Related Question