julienr / ipynb_playground Goto Github PK

Thanks sir for this awesome work!. Actually I've been looking for something like this for quite long time, so thank you . I have a question if you allow me, If I have a test image and I want to get the model output after activation of the dense layer, so in this case what will be the input of this layer ( K.function(inputs, [convout1.output])) ? the original test image ? or the image after computing the output of all the previous layers? shall I compute them by myself or it will be computed implicitly if the test image fed directly into the dense layer?

This example won't work anymore as keras interface has changed.

Have changes and am willing to send if you'd like.

Hello @julienr , thank you so much for your contribution on visualisation of the neural network.
I've been following this link which uses the weights from mnist_cnn_weights_v1.hdf, however, I was unable to find the file in your repository. Could you point me to the link or where I can get it?
Thank you :)