I have a small confusion. When someone says that a convolutional neural network is 48 layers deep,does it mean that the CNN has 48 convolution layers(as in layers that perform convolution only) or 48 layers(which includes layers that perform pooling,ReLU,convolution,fully connected layer etc.)? Kindly help.

  • $\begingroup$ there is no convention, as far as I know. Logically, as you phrased it, it should refer to the number of convolution-pooling layers, but the only way to know for sure is to look at the details of the paper or at the code... $\endgroup$ – Antoine Jun 20 '17 at 12:07
  • 1
    $\begingroup$ As @Antoine said there is no convention, but usually it refers to convolution + non-linearity or even convolution + non-linearity + batch norm $\endgroup$ – Łukasz Grad Jun 20 '17 at 12:58

Your Answer

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

Browse other questions tagged or ask your own question.