Revert "[torch.compile] Fix RMSNorm + quant fusion in the non-cutlass…

[tlrmchlsmth]

This affects FP8 models running on RoCM as well as Ada Lovelace GPUs when vLLM is built with CUDA < 12.4

#10902 added a call to get_current_vllm_config() during the forward pass, however _current_vllm_config is not set outside of worker initialization, leading to a lot of warning spam like:

WARNING 03-05 22:03:35 [config.py:3600] Current VLLM config is not set.

This reverts commit bd56c98, although a better approach would be to plumb a variable like allow_input_padding into this function.

[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.

🚀