Problem in the TMultiLayerPerceptron graphic outputs

Hi,

When I run the mlpHiggs program, as well as other codes that use this neural network, I get the following curve:


This is for the mlpHiggs Program.

or (for another problem that uses the same NN)

I think this should not look like this knowing that the error evolution is exponential, not linear.
I can’t find the part where this graph is coded to understand where does this come from, does anyone have an idea?

@couet could you give a hand? Thanks!

https://root.cern/doc/master/TMultiLayerPerceptron_8cxx_source.html#l02469

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.