Implement lowering of torch.aten.all.dim

[mmakevic]

Lowering of torch.aten.all.dim to linalg.

Per PyTorch documentation:

This function matches the behaviour of NumPy in returning output of dtype bool for all supported dtypes except uint8. For uint8 the dtype of output is uint8 itself.

Since there is no support for ui8 in torch-mlir currently (#1384 (comment)) implementation returns failure for that case.

[newling]

Looks good but I think the tests can be more targeted (smaller tensor, set values by hand). Can you test the empty tensor case returns 'True' ?

[mmakevic]

@newling done :)

[newling]

Thanks for updating the tests, just the isa<.> comment remaining

[newling]

Just the little comment about the extra brackets and formatting, but approving now as it looks good to me otherwise.

[mmakevic]

@newling can you merge this PR? I do not have write access :)

[newling]

My view of the PR:

@mmakevic Let me know if this looks correct ("authored by" email look ok?) and I'll merge it if no one else comments in the next day.

[mmakevic]

@newling Yeah, everything looks okay. Thanks!