When running a protocol with the Learn CNN Image Classifier component, I get the following error: 

Input to reshape is a tensor with 371840 values, but the requested shape has 1115520
[[{{node Reshape}}]] [Op:IteratorGetNext]
Traceback (most recent call last):
File "Python Script", line 347, in onProcess
File "Python Script", line 80, in processed_dataset
File "data.py", line 286, in get_one_example
for example in dataset:
File "iterator_ops.py", line 766, in __next__
return self._next_internal()
File "iterator_ops.py", line 749, in _next_internal
ret = gen_dataset_ops.iterator_get_next(
File "gen_dataset_ops.py", line 3016, in iterator_get_next
_ops.raise_from_not_ok_status(e, name)
File "ops.py", line 7164, in raise_from_not_ok_status
raise core._status_to_exception(e) from None # pylint: disable=protected-access
tensorflow.python.framework.errors_impl.InvalidArgumentError: Input to reshape is a tensor with 371840 values, but the requested shape has 1115520
[[{{node Reshape}}]] [Op:IteratorGetNext]

After some time with Google it appeared that it might be due to my image sizes (I literally just copied the Melanoma Model Learning and Inference protocol and substituted my own images - if I use the supplied images with the example I can train the model successfully). 

I adjusted the Network Input Layer Width/Height to set it to the size of my images (not positive this was the correct approach, but ... ) and I am now getting: 

invalid literal for int() with base 10: '\\\$(Width/Height)'
Traceback (most recent call last):
File "Python Script", line 292, in onProcess
ValueError: invalid literal for int() with base 10: '\\\$(Width/Height)'

Please advise: 

1 - whether the correct solution to the first error is to adjust this parameter or look somewhere else

2 - how to make this adjustment or where to look, depending on the answer to #1.