i.e. this is sep conv as found in the API of popular frameworks such as TF. I will release an updated version of the paper to clear that up.
Some people are getting confused with the term "separable convolution" in this context. This is in fact *depthwise* separable convolution.https://twitter.com/fchollet/status/785511714334703616 …
-
-
-
Please read the paper, not just the abstract. The usage of the term is unambiguous in context.
End of conversation
New conversation -
-
-
depth ie across channels?
- End of conversation
New conversation -
-
-
another question I looked through the paper, how is this different from 3d conv networks?
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.