Apply deprecated evaluation_strategy

[muellerzr]

What does this PR do?

This PR applies the deprecated evaluation_strategy from huggingface/transformers#30190.

TL;DR: evaluation_strategy -> eval_strategy

(Should go live in v4.41.0, but opening now so it's here when ready!)

Fixes # (issue)

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Did you make sure to update the documentation with your changes?
• Did you write any new necessary tests?

Who can review?

• ONNX Runtime Training: @JingyaHuang

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

[Rohan138]

@muellerzr @IlyasMoutawwakil could we merge this?

It currently breaks the examples with transformers>=4.41 because evaluation_strategy is set to None (https://github.com/huggingface/transformers/blob/main/src/transformers/training_args.py#1375, which errors at https://github.com/huggingface/optimum/blob/main/optimum/onnxruntime/training_args.py#L129 because EvaluationStrategy expects one of ['no', 'steps', 'epoch']