[PYTORCH] pytorch frontend expand op bug

[Xuxue1]

In pytorch frontend testcase modify input_shape and expand shape

def test_forward_expand():
    torch.set_grad_enabled(False)
    input_shape = [3, 1]

    class Expand1(Module):
        def forward(self, *args):
            return args[0].expand((3, 3, 3, 1))

    input_data = torch.rand(input_shape).float()
    verify_model(Expand1().float().eval(), input_data=input_data)

Traceback (most recent call last):

File "tests/python/frontend/pytorch/test_forward.py", line 2109, in
test_forward_expand()

File "tests/python/frontend/pytorch/test_forward.py", line 912, in test_forward_expand
verify_model(Expand1().float().eval(), input_data=input_data)

File "tests/python/frontend/pytorch/test_forward.py", line 191, in verify_model
assert_shapes_match(baseline_output, compiled_output)

File "tests/python/frontend/pytorch/test_forward.py", line 39, in assert_shapes_match
raise AssertionError(msg.format(tru.shape, est.shape))

AssertionError: Output shapes (3, 3, 3, 1) and (3, 3) don't match

[masahi]

Fixed by #5576