Seems test case may be using the wrong model?

[rossbar]

This was caught by the linter - there is a test that defines a mapping that, from the surrounding context, really seems like it's supposed to be used but isn't:

deepcell-tf/deepcell/model_zoo/panopticnet_test.py

Lines 391 to 392 in 17b54eb

	nsc2 = {'0': 2, '1': 3}
	model2 = PanopticNet(num_semantic_classes=nsc1, **shared_kwargs)

FWIW I ran the tests locally using num_semantic_classes=nsc2 and everything still passes. If there is something obvious here that jumps out to others, please LMK - otherwise I will propose to update this.