Onnx2torch

Support for new OPS :rocket: :

- [Dropout](https://github.com/onnx/onnx/blob/main/docs/Operators.md#Dropout)
- [LRN](https://github.com/onnx/onnx/blob/main/docs/Operators.md#LRN)

Thanks yan12125 and monicathieu for their contributions.

Include requirements.txt and exclude tests for sdist

You've been waiting, you've asked... And here it is a new release(1.5.0)!

In this release:
- Fixed shape inference of big models (now you can load your lovely big language model, like gptj6B);
- Fixed `BarchNorm` converter according `onnx` specs;
- Added asymmetric padding for both `Conv` and `MaxPool`.

Thank you for your discussions and issues! We always ready to help you!

Minor updates:
* Fix prelu according onnx specs,
* Fix Global average pooling.

- Added operations:
- Einsum,
- Reciprocal,
- Neg,
- Prelu,
- Mean,
- Min,
- Max,
- CumSum,
- Sum.

- Fixes:
- GlobalAveragePool compatible with PyTorch FX QAT,
- Compatibility with torch 1.12,
- Allow to ignore bs+ch dimensions for size input in Resize.
- Check roi only for tf_crop_and_resize mode.

- Code style and workflows:
- [Pre-commit hook](https://pre-commit.com/),
- [Black code-style now is used in project](https://black.readthedocs.io/en/stable/the_black_code_style/index.html),
- Linter and formatter with precommit hook,
- Code style checks for PRs in Github actions.

We thank all contributors for their work! Together we will make onnx2torch great! :star: 10^6

- Add `Pad` operation
- Add `Hardswish`, `Celu`, `Elu`, `Selu`, `Softplus`, `Softsign` functions
- Add `GatherElements` operation
- Add new model tests

Fixes:

- Fix `ConvTranspose` operation