CoderHHX comments

Results 6 comments of


                                            CoderHHX

Softmax in routing algorithm incorrect?

> Hi, > I think the softmax in the routing algorithm is being calculated over the wrong dimension. > > Currently the code has: > > ``` > b_ij =...

Why dim=2 here? I guess the softmax is taken over all capsules in the layer above, so shouldn't that be dim=0?

I have the same question and I read some other implementation like Tensorflow and PyTorch for CapsuleNet and I think that softmax logits [10, 100, 1152, 1, 16] should apply...

Why dim=2 here? I guess the softmax is taken over all capsules in the layer above, so shouldn't that be dim=0?

@InnovArul Thanks for your reply! In my opinion, if we want to follow the original paper, we should set the dim equal 0. And as you say that with the...

Is there any error about the softmax operation?

@zzzz94 Thanks for your solution firstly. @h982639009 notice that before setting dim = 0, the dimension of c_ij is 1132 where setting dim = 1, the dimension is 10. This...

Pytorch version

@wuwusky youxiu, Have you compared your re-implementation with the original paper on CUB datasets?

is there any reproducibility problem?

Yeah, I wish the author could release clean and reproducible codes, thanks!