ein tests fail on Windows only

[JasonGross]

this seems like an autogenerated name issue?

================================== FAILURES ===================================
____ EinTest.test_ein_unbound_usage_before_usage_fails_with_implicit_sizes ____

>       lambda: ein.array(lambda i: torch.exp(i) + A[i]),
        """dimension dim is unbound""",
    )

gbmi\utils\ein_test.py:77: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
gbmi\utils\ein.py:189: in array
    return _apply(
gbmi\utils\ein.py:137: in _apply
    return _apply_single_dim(
gbmi\utils\ein.py:115: in _apply_single_dim
    xs = f(dim)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

i = d0

>       lambda: ein.array(lambda i: torch.exp(i) + A[i]),
        """dimension dim is unbound""",
    )
E   ValueError: dimension d0 is unbound

gbmi\utils\ein_test.py:77: ValueError

During handling of the above exception, another exception occurred:

self = <gbmi.utils.ein_test.EinTest testMethod=test_ein_unbound_usage_before_usage_fails_with_implicit_sizes>

    def test_ein_unbound_usage_before_usage_fails_with_implicit_sizes(self):
        A = torch.tensor([1, 2, 3])
>       self.assertExpectedRaisesInline(
            ValueError,
            lambda: ein.array(lambda i: torch.exp(i) + A[i]),
            """dimension dim is unbound""",
        )

gbmi\utils\ein_test.py:75: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
.venv\Lib\site-packages\expecttest\__init__.py:376: in assertExpectedRaisesInline
    self.assertExpectedInline(str(e), expect, skip=1)
.venv\Lib\site-packages\expecttest\__init__.py:351: in assertExpectedInline
    assert_expected_inline(
.venv\Lib\site-packages\expecttest\__init__.py:316: in assert_expected_inline
    assert_eq(expect, actual, msg=help_text)
.venv\Lib\site-packages\expecttest\__init__.py:388: in assertMultiLineEqualMaybeCppStack
    self.assertMultiLineEqual(expect, actual, *args, **kwargs)
E   AssertionError: 'dimension dim is unbound' != 'dimension d0 is unbound'
E   - dimension dim is unbound
E   ?            ^^
E   + dimension d0 is unbound
E   ?            ^
E    : To accept the new output, re-run test with envvar EXPECTTEST_ACCEPT=1 (we recommend staging/committing your changes before doing this)

[JasonGross]

Furthermore, this seems to depend on the specific version of torch