Neural Network Validation Index

When I initialize my network I include the standard 70% training, 15% validation, and 15% testing indices with their respective variable (network.divideParam.trainInd / valInd / testInd). However after training there is no information for the validation. It appears that the validation indices were appended to the training indices.

My network variables appear as follows: network.divideParam.valInd = [1 x 238140] network.userdata.trainingRecord.valInd = []

Why would there be no validation information in the training record?

[Greg Heath]

I cannot help without seeing your code.

Answer 1

[Greg Heath]

You may be using trainbr.