Pytorch upsample support

[m-glowacki]

Adding PyTorch Upsample Layer Support and Fixing Output Shape Batch Dimension

Description:
This PR introduces support for PyTorch Upsample layers and addresses a bug related to output shapes in pytorch convolution and pooling handlers:

Commit 1 (bug fix):

Description: Fixing output shapes by removing the unexpected batch dimension.
Details: Resolved an issue where the batch dimension was present in output shapes for each handler, which was not expected input for parse_data_format.

Commit 2 (adding layer support):

Description: Adding support for PyTorch Upsample layers.
Details: Implemented PyTorch Upsample handler and revised the Resize layer to handle channel_first input (default to Pytorch).

[jmitrevs]

Please fix the pre-commit and then we can try running the pytests.

[jmitrevs]

It still fails the pre-commit.

[m-glowacki]

Any advice on how to de-bug this? Seems to be failing the black linter hook, however I already ran this locally to make sure if was compliant with the standard and pushed those changes. There are files being pointed to which have not been modified by this PR.

[vloncar]

pre-commit.ci autofix

[m-glowacki]

I see a couple of the pytests have failed due to time out, can this be reviewed and re-ran if needed?

[vloncar]

I tried this out, but it has some issues and incorrect changes (Conv2D-related commit in particular). Due to the style changes black made, I ended up making a clean reimplementation of this in a new PR. Closing.