Bert Slice Support

[ntarafdar]

I accidentally made the issue in the wrong project (pytorch2.0) , so I'm copying it here.

Sprint on making sure all slice configs are supported in Bert.

These are the shapes, I think they should be supported but please can we check and report any issues.
Update the Status column for the shapes that are supported and add these shapes to unit test.

As a side also make sure these shapes are swept on as we already have a sweep test for slice

CC: @ayerofieiev-tt

aten.slice.Tensor

	ATen Input Variations	Status
0	Tensor<[1, 1, 1, 256]> self = ?, int<> dim = 3, Optional[int]<> start = 0, Optional[int]<> end = -1	Unknown
1	Tensor<[1, 256]> self = ?, int<> dim = 0, Optional[int]<> start = 0, Optional[int]<> end = -1	Unknown
2	Tensor<[1, 512]> self = ?, int<> dim = 0, Optional[int]<> start = 0, Optional[int]<> end = -1	Unknown
3	Tensor<[1, 512]> self = ?, int<> dim = 1, Optional[int]<> start = 0, Optional[int]<> end = 256	Unknown