Variable input resolutions for ResNets

[macaodha]

The current implementation of the ResNet models has a fixed 7x7 average pooling layer before the final FC layer.

If this was changed to nn.AdaptiveAvgPool2d(1) it would allow for variable sized input batches (e.g. like the 320x320 batches used at test time in the 'Identity Mappings in Deep Residual Network' paper)? Although I'm not sure if this will incur a performance hit?

[tonyduan]

+1. Would greatly appreciate the flexibility in variable sized inputs [i.e. not fixed to 224x224].

[fmassa]

This has been implemented in #643