Analysis of each layer in a Deep Learning Convolutional network

Hi, I am new to the Deep Learning toolbox and I need to guidance on how to proceed:

I have trained my network (A modified GoogleNet) and I need to evaluate the performance of the network, therefore I would like to acquire the output after Layer 1, 2 etc to get more in depth on the performance.

Is there a way to take an image A, feed that to the network and acquire the output after layer X?

Thanks!

Answer 1

[Zenin Easa Panthakkalakath]

Hi Andreas,

You can do this by using the following command:

>> activations(net,X,layer)

where 'net' is the trained network, X is the input data and 'layer' is the layer name.

Have a look at the following documentation for more information:

https://www.mathworks.com/help/deeplearning/ref/activations.html

Regards,

Zenin