[XPU] Fix xpu model runner call torch.cuda APIs

[jikunshang]

Purpose

#23693 introduce torch.cuda.Stream in GPUModelRunner, this breaks xpu behavior. this PR fix this issue by forward torch.cuda API to torch.xpu .

Test Plan

CI

Test Result

---

Essential Elements of an Effective PR Description Checklist

• The purpose of the PR, such as "Fix some issue (link existing issues this PR will resolve)".
• The test plan, such as providing test command.
• The test results, such as pasting the results comparison before and after, or e2e results
• (Optional) The necessary documentation update, such as updating supported_models.md and examples for a new model.
• (Optional) Release notes update. If your change is user facing, please update the release notes draft in the Google Doc.

[gemini-code-assist]

Code Review

This pull request aims to fix an issue for XPU devices by patching torch.cuda APIs to forward to torch.xpu. While the intention is correct, the implementation has a critical flaw: it performs monkey-patching on the global torch.cuda module without properly restoring the original attributes. This can lead to global state corruption and unpredictable behavior in other parts of the code. My review includes a suggestion to implement the patching in a safe, contained manner using a try...finally block that saves and restores the original attributes.