python - Output of neurons in Multiple Layer Perceprton Classifier in scikit-learn -

i working on mlpclassifier of neural_network package of sklearn.

i trained classifier , predicting/running fine. need output values of neurons(nodes) in each layers when predicts class particular input after training, visualisation purposes.

i read api , there attribute - coefs_ returns weight matrix of network couldn't find method or attribute return output of neurons.

so being not mentioned in documentation, suppose not possible directly. there way/tweaking available these output of neurons @ each layer or alternatively direct method of visualisation of mlpclassifier.

note - mlpclassifier not available in stable version of scikit-learn , 0.18 dev version only.

i using python 2.7 , scikit-learn 0.18 dev version.

output of neurons makes sense if have inputs. it's not intrinsic part of model. take inner product of inputs weights "outputs".