I use NetTrain to train a model,Then I got net and LossEvolutionPlot

{net, LossEvolutionPlot} = NetTrain[net,File["training_normalization_ForTF.h5"], MeanSquaredLossLayer[], {"TrainedNet", "LossEvolutionPlot"}, BatchSize -> 4096, TargetDevice -> "GPU", ValidationSet -> Scaled[0.22222], MaxTrainingRounds -> 200,

The expression of LossEvolutionPlot is:

p = Uncompress@Import@"https://wolfr.am/lNslN7ev"

enter image description here

It does not have picks,so how to make the plot more detailed?


You can format the plot using Show like this:

Show[p, FrameTicks -> Automatic, FrameLabel -> {"batch", "loss"}, 
 ImagePadding -> Automatic]

enter image description here

  • $\begingroup$ but MSE loss always bigger than zero. $\endgroup$
    – partida
    May 20 '17 at 23:59
  • $\begingroup$ @partida I think it is in log scale. $\endgroup$ May 21 '17 at 0:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.