YOLO V2 Training

I am following the instructions on MATLAB page to train a YOLO network: https://www.mathworks.com/help/vision/ug/train-an-object-detector-using-you-only-look-once.html

At this time, I am just try to reproduce the same network as in the example. I receive an error that the preprocessedValidationData f is not in the correct format in this part of the code:

options = trainingOptions('sgdm', ...

'MiniBatchSize',16, ....

'InitialLearnRate',1e-3, ...

'MaxEpochs',20,...

'CheckpointPath',tempdir, ...

'ValidationData',preprocessedValidationData);

Answer 1

[Pratyush Roy]

Hi,

"preprocessedValidationData" is generated by the preprocessData function call on data and inputSize

preprocessedTrainingData = transform(augmentedTrainingData,@(data)preprocessData(data,inputSize));
preprocessedValidationData = transform(validationData,@(data)preprocessData(data,inputSize));

preprocessData is a supporting function that is defined above the references section in the doc page. So it has to be made sure that the supporting functions are defined in the same script so that the validation data can be generated properly.

Hope this helps!