Chop output_shape to (y, x) format to pass to target_shape.
Force model compiling before training, testing and predicting.
Fix pool2d when dim_ordering is 'tf'.
TODO:
Current mxnet deconvolution operator only accepts target_shape which is no greater than s*(i - 1) + k, where s is strides, i is input size and k is kernel size. Maybe in the future we need to add support for oversized target_shape.
TODO: Current mxnet deconvolution operator only accepts target_shape which is no greater than s*(i - 1) + k, where s is strides, i is input size and k is kernel size. Maybe in the future we need to add support for oversized target_shape.