Allow arbitrary number of input channels in ResNet encoder (not only RGB)

With https://github.com/mapbox/robosat/pull/46 we are changing our model architecture from training the encoder and decoder from scratch to using a pre-trained ResNet for the encoder. The pre-trained ResNet uses three channels (RGB) for the input layer through.

We need to be able to add arbitrary channels, say, RGB + water mask + elevation + lidar. To to this we need to construct a wrapper module architecture extending the ResNet architecture, copying weights over, and initializing the new channels with zero. In addition the channel-wise mean and std dev needs to be adapted.

Tasks

• Figure out how to extend the ResNet input channels
• Figure out how to copy over the pre-trained ResNet parts
• Let users construct a model with arbitrary channels
• Adapt mean and std dev

Start by locating the ResNet encoder construction and the code that applies channel-wise mean and standard deviation; the issue does not name specific files or tests. Review how the existing three-channel model is built and determine how arbitrary input channels, pretrained weights, and normalization should be represented. Done means users can construct the encoder with additional channels while preserving pretrained RGB behavior.