Guang Yang
Guang Yang
I agree, there should be no backpropagation through the EM algorithm, other than learning the beta_a and beta_v. Will modify. Thanks.
Hi, @1kaiser , Thank you for your contribution. I look through your patch, seems like most of the changes are straightforward. And, there is no change impacting in the core...
@Germanunkol Thank you for the comments and suggestions. I indeed found that the NaN is due to the activations from the primary capsules, e.g., capsule activations initialized from regular convolution...
Thank you!
HI, @MichaelChirico , Could you advice me on defining .N and `:=` in bpp_viewer.r? ``` .N = NULL `:=` = function(...) NULL ``` I think we want .N and `:=`...