great overview, nice to see Xception is mentioned as an emerging backbone network - this is what i'm working with right nowhttps://twitter.com/graphific/status/1041627931842420742 …
-
Keras API puts a low boundary on the input image size - ex for rn50 it requires input size to be no smaller than 197
Hi Helena, all modern CNNs operate downsampling on their inputs, which means they start with large feature maps that get smaller as you do down the network. In general, convolution is a downsampling operating unless you explicit pad your inputs.
-
-
Thats means that all CNNs have a minimum input size, which is the minimum size you can call the network on without resulting in empty feature maps. At that size, the output of your network is a 1x1 feature map.
-
For most nets, this tends to be about ~50% of the input size the network was designed for. Downsampling is absolutely necessary when you have large inputs. To get rid of this limitation, you need to create your own network, where you would do less downsampling.
- Show replies
New conversation -
-
-
That's only partially true: for discriminative models yes, but there's no inherent reason why convolutions should equals downsampling (for generative models for instance).
-
*Of course* you can design architectures that don't do downsampling (just don't include maxpooling or strides, and use "same" padding everywhere). And there are use cases for it. I'm just talking about every pre-trained architecture available out there.
- Show replies
New conversation -
-
-
sure, i understand that - my data is 180X180 and i'm using Xception and it works fine
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.