Fix tensor.test_torch_instance_norm

[fabriziocacicia]

[fabriziocacicia]

While investigating the causes of the failure of this tests I came across a failure caused by data_type.default_float_dtpye

E           During the handling of the above exception, another exception occurred:
E             File "/ivy/ivy/utils/exceptions.py", line 210, in _handle_exceptions
E               return fn(*args, **kwargs)
E             File "/ivy/ivy/func_wrapper.py", line 971, in _handle_nestable
E               return fn(*args, **kwargs)
E             File "/ivy/ivy/func_wrapper.py", line 482, in _inputs_to_ivy_arrays
E               return fn(*ivy_args, **ivy_kwargs)
E             File "/ivy/ivy/functional/ivy/data_type.py", line 1122, in default_float_dtype
E               return ivy.FloatDtype(ivy.as_ivy_dtype(ret))
E           During the handling of the above exception, another exception occurred:
E            local variable 'ret' referenced before assignment
E           
E           You can reproduce this example by temporarily adding @reproduce_failure('6.79.1', b'AXicY2JgZAAhEIBSYBoAAIEABw==') as a decorator on your test case

/ivy/ivy/utils/exceptions.py:246: IvyBackendException

[ivy-leaves]

If you are working on an open task, please edit the PR description to link to the issue you've created.

For more information, please check ToDo List Issues Guide.

Thank you :hugs:

[rishabgit]

Hey @fabriziocacicia, haven't checked too deeply, but I'm guessing this draft PR isn't relevant anymore? Closing it for now but lmk if I'm wrong, and we can always open this back up 😄