fix(transformers): fix v4.50.0 fast ut graph mode and random inputs

[Fzilan]

Fixes

1. Temporarily remove the unnecessary strict graph mode settings.

   The graph mode tests failed during the overall test run, while running them individually caused no errors. The reason is that in the overall test, a previous test may have modified the context configuration (for example, by setting strict graph mode through set_context), but the graph mode context was not properly reset before the next test, which affected subsequent cases.

2. Use specific random seeds in some tests to reproduce results and avoid occasional NaN issues.

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Did you read the contributor guideline?
• Did you make sure to update the documentation with your changes? E.g. record bug fixes or new features in What's New. Here are the
  documentation guidelines
• Did you build and run the code without any errors?
• Did you report the running environment (NPU type/MS version) and performance in the doc? (better record it for data loading, model inference, or training tasks)
• Did you write any new necessary tests?

@wcrzlh @CaitinZhao @vigo999 @zhanghuiyao

[gemini-code-assist]

Summary of Changes

Hello @Fzilan, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request addresses critical fixes for the transformers library, specifically targeting issues within version 4.50.0. The changes primarily involve refining MindSpore context settings in various model tests to enhance compatibility with graph mode and standardizing the generation of random inputs in OwlViT and PoolFormer tests to ensure reproducibility and stability of test outcomes.

Highlights

• MindSpore Context Configuration: The jit_syntax_level=ms.STRICT parameter has been removed from ms.set_context calls across numerous model test files. This adjustment likely resolves compatibility or functional issues when executing tests in MindSpore's graph mode, ensuring smoother operation.
• Reproducible Random Input Generation: A get_rng() helper function, utilizing a fixed seed (random.Random(9)), has been introduced in the OwlViT and PoolFormer model tests. This change guarantees that all random inputs generated for these tests are consistent across different runs, which is essential for reliable and deterministic testing.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[gemini-code-assist]

Code Review

This pull request introduces two main changes to fix unit tests for v4.50.0. First, it removes jit_syntax_level=ms.STRICT from ms.set_context in numerous test files, likely to address compatibility issues in graph mode. Second, it makes random inputs deterministic in the owlvit and poolformer model tests by introducing a seeded random number generator.

While making the tests deterministic is a good practice, the current implementation of get_rng() creates a new random.Random instance with the same seed on every call. This leads to different parts of the tests receiving identical sequences of 'random' numbers, which is likely not the intended behavior and can hide bugs. I've added comments with suggestions to use functools.lru_cache to ensure a single RNG instance is created and reused, thus providing proper random number sequences across the tests.

[zhanghuiyao]

后面可以考虑放到上面的函数装饰方法中进行选择