Refactor GPUModelRunnerBase load_model method to include device param

[Zzhiter]

FIX #13036

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.

💬 Join our developer Slack at https://slack.vllm.ai to discuss your PR in #pr-reviews, coordinate on features in #feat- channels, or join special interest groups in #sig- channels.

Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which starts running only a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of those by going to your fastcheck build on Buildkite UI (linked in the PR checks section) and unblock them. If you do not have permission to unblock, ping simon-mo or khluu to add you in our Buildkite org.

Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.

To run CI, PR reviewers can either: Add ready label to the PR or enable auto-merge.

🚀

[Zzhiter]

@ywang96 Could review this PR pls? thank u

[jeejeelee]

LGTM,I think this is reasonable, @ywang96

[ywang96]

LGTM too - I think we should also include this change in V1? I'm also curious how no other users reported this before since this code is pretty old...

cc @WoosukKwon @youkaichao

[youkaichao]

LGTM too - I think we should also include this change in V1? I'm also curious how no other users reported this before since this code is pretty old...

cc @WoosukKwon @youkaichao

this matters if users have multiple vllm instances in the same process, which is indeed a rare case.

usually, we set the current device in pytorch, and then most of the code do not need to be aware of it.