Add a pass to convert rank-0 tensor to rank-1 tensor (#8298)

Add a pass to convert rank-0 tensor to rank-1 tensor (#8298)

Summary:


Test with following code, and no long see the error/warning to complain rank-0 tensor
```

class Model(torch.nn.Module):
    def forward(self, x, y):
        return x + y

model = Model()
model.eval()

example_inputs = (torch.tensor(1.0), torch.tensor(2.0))
exported_program_manager_aten = torch.export.export(model, example_inputs)
exported_program_manager_edge = executorch.exir.to_edge(
    exported_program_manager_aten
).transform([Rank0ToRank1Pass()])

delegated_module = to_backend(
    CoreMLBackend.__name__, exported_program_manager_edge.exported_program(), []
)
```

Differential Revision: D69281867