Pruning

When I prune a model with TAO toolkit how are the kernel norms computed? Is it something similar to tf.norm  |  TensorFlow v2.10.0 or is it something else going on in the background?

Best regards,
Karl

Refer to DetectNet_v2 — TAO Toolkit 3.22.05 documentation

  • -n, –normalizer : Specify max to normalize by dividing each norm by the maximum norm within a layer; specify L2 to normalize by dividing by the L2 norm of the vector comprising all kernel norms. The default value is max .

Sorry, I should have been more specific. I know about the normalization criteria, but I’m rather interested in how a single kernel norm is defined. Like if it is defined as the L2 norm of the kernel elements or something else.

As mentioned above, specify max to normalize by dividing each norm by the maximum norm within a layer; specify L2 to normalize by dividing by the L2 norm of the vector comprising all kernel norms. The default value is max .

There is no update from you for a period, assuming this is not an issue anymore.
Hence we are closing this topic. If need further support, please open a new one.
Thanks

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.