What if we use l1 loss? I attempt to use NTK to get the convergence rate of a NN. Here is the original l2 loss version: https://rajatvd.github.io/NTK/. When I relpace it to l1 loss, I find the convergence rate is a constant. Is there anything wrong?

  • $\begingroup$ Can you briefly describe in 1-2 sentences what NTK is? $\endgroup$
    – nbro
    Commented May 18 at 13:42


You must log in to answer this question.

Browse other questions tagged .