refactored interface to reduce complexity so conv and convt layers forward passes have to call setup explicit now and there is only one ()-operator

histogram is equalized or unmodified.  This way, if you are looking at
particularly dark or badly contrasted images you can toggle this mode and maybe
get a better view of what you are labeling.

min and max object height, it's now min and max object size.  This way, if you
have objects that are short and wide (i.e. objects where the relevant dimension
is width rather than height) you will get sensible behavior out of the random
cropper.

everything in it, it now makes a python file as before but an additional binary
file with all the weights in it.  This way, if you are working with a network
with a very large number of weights you won't end up with a crazy large python
script.

integers.

object so that users of the deep learning MMOD tool don't get spurious errors
about impossibly labeled objects during training.

objects that also provide a .clean() method.

the behavior of dlib's add_prev layers.

when the converter is run.

Noticed compiler warning C4800: 'double': forcing value to bool 'true' or 'false' (performance warning) (#538)

- suggesting to change the test to what perhaps was the original intention

Moved shape_predictor_trainer to its own header in order to support the
use of shape_predictor with DLIB_ISO_CPP_ONLY defined (i.e. without
threading API wrappers).

So I just removed the error check that was preventing the conversion from
proceeding in that case.  I also added more useful output messages about
setting input tensor dimensions.