Update OV / NNCF/ ModelAPI / Datumaro

[sovrasov]

Summary

This PR updates OV-related dependencies of OTX.
Known issues:

• OMZ plugin may fail in some cases, because version was set to opset1 by default for some of operations. This could be updated later, if we face models that show opset - version mismatch.
• ModelAPI integration requires a deeper refactoring
• MoViNet inference in OV doesn't work
• Tile classifier is not handled by PTQ (as before)

How to test

Checklist

• I have added unit tests to cover my changes.​
• I have added integration tests to cover my changes.​
• I have added e2e tests for validation.
• I have added the description of my changes into CHANGELOG in my target branch (e.g., CHANGELOG in develop).​
• I have updated the documentation in my target branch accordingly (e.g., documentation in develop).
• I have linked related issues.

License

• I submit my code changes under the same Apache License that covers the project.
  Feel free to contact the maintainers if that's a concern.
• I have updated the license header for each file (see an example below).

# Copyright (C) 2023 Intel Corporation
# SPDX-License-Identifier: Apache-2.0

[sungmanc]

Small comment, if the PR is still in progress, please use DRAFT PR. we could use CI tests even in under the DRFAT status.

[yunchu]

@sovrasov note that you can skip the automatic trigger of the actions for the test PR by adding string [skip ci] to your commit message.
https://github.blog/changelog/2021-02-08-github-actions-skip-pull-request-and-push-workflows-with-skip-ci/

[sovrasov]

https://github.com/openvinotoolkit/training_extensions/actions/runs/5442410680 E2E Test from f8f3d67. Could you take a look? I have no idea for anomaly, but for detection, maybe data pipeline between normal detection and tiling would be different. Therefore eeabbca could affect it.

In the latest commits, PTQ was switched to POT via config (NCT team members suggested to do that).
Also, the issue with YOLOX tiling preprocessing was fixed.

[sovrasov]

@jaegukhyun could you launch e2e once again? I was unable to reproduce the anomaly deploy accuracy issue on my side

[jaegukhyun]

https://github.com/openvinotoolkit/training_extensions/actions/runs/5460287541 I rerun e2e from deea851
Firstly @GalyaZalesskaya should solve current E2E and @sovrasov you need to update FQ reference for anomaly and segmentation

[jaegukhyun]

Please rebase your branch, Galina's PR for e2e test failures is merged

[jaegukhyun]

https://github.com/openvinotoolkit/training_extensions/actions/runs/5473307857
All tests are passed, I think we can merge this PR. Thanks for your effort

[goodsong81]

Thank you, it's a huge update!