1
$\begingroup$

The above article only talks about Convolutional Neural Networks:

One of the first methods of pruning is pruning entire convolutional filters. Using an L1 norm of the weight of all the filters in the network, they rank them. This is then followed by pruning the ‘n’ lowest ranking filters globally. The model is then retrained and this process is repeated.

There also exist methods for implementing structured pruning for a more light-touch approach of regulating the output of the method. This method utilizes a set of particle filters that are the same in number as the number of convolutional filters in the network.

Is "Pruning" only applicable to CNNs?

$\endgroup$
1
  • $\begingroup$ No, you can apply the same concepts to any type of network. Also pruning is quite common when dealing with graphs. $\endgroup$
    – razvanc92
    Jun 29 at 6:52
0
$\begingroup$

No, neural network pruning is applicable to any type of neural network, be it a feed-forward, convolutional or recurrent neural network.

$\endgroup$
1
  • $\begingroup$ To support your claim, maybe you could provide a link to a research paper that applies pruning to e.g. a FFNN (that maybe also includes the implementation). $\endgroup$
    – nbro
    Jun 29 at 9:39
0
$\begingroup$

Yes, it is applicable to CNN, and to a wide range of other architectures, even the hype transformers.

For an extensive survey I recommend to have a look at this paper https://arxiv.org/abs/2003.03033.

$\endgroup$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.