Have always loved your articles. Big fan.
George Pligor: Below is the explanation for the piece of code that you are looking for,
capsLayer.py:41 — This line inits a blank Tensor Layer.
capsLayer.py:46 — This line updates the shape for the current dimension
capsLayer.py:47 — This line adds it to the capsule
Gökçen Eraslan: I guess the language of the article is misleading here. True pooling just brings in translatory invariance. Rotational is acquired by either rotating the kernel or the feeding rotated input images to the network.
To answer your second question, what I tried to mean is that pooling (along with variation of…
Well, it’s intended to be a pun regarding the point that a regular system will take a huge amount of time in training these networks. Without good GPU power it’s nearly impossible to train them in practical times.