Fix tensor.nonzero() function overload warning

[nelson-liu]

When running the CRF, I currently get this warning (reproduced it in py.test, too):

========================================== warnings summary ===========================================
tests/modules/conditional_random_field_test.py::TestConditionalRandomField::test_viterbi_tags
  /home/nfliu/git/allennlp/tests/modules/conditional_random_field_test.py:1: UserWarning: This overload of nonzero is deprecated:
  	nonzero()
  Consider using one of the following signatures instead:
  	nonzero(*, bool as_tuple) (Triggered internally at  /pytorch/torch/csrc/utils/python_arg_parser.cpp:766.)
    import itertools

-- Docs: https://docs.pytest.org/en/stable/warnings.html
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! _pytest.outcomes.Exit: Quitting debugger !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
============================ 2 passed, 951 deselected, 1 warning in 13.40s ============================

The fix is pretty simple--just provide the as_tuple keyword explicitly (and keep it at its default value, False)

[epwalsh]

LGTM!