[Bugfix] Fix boolean conversion for OpenVINO env variable

[helena-intel]

Two minor fixes for OpenVINO integration:

• improve boolean conversion for VLLM_OPENVINO_ENABLE_QUANTIZED_WEIGHTS. Only set it to True if user sets it to ON, True or 1.
• Only convert models that are exported to OpenVINO on the fly to INT8, keep pre-converted IR models as is, regardless of VLLM_OPENVINO_ENABLE_QUANTIZED_WEIGHTS value. This matches what is said in the message shown to users when loading an IR model: "This IR will be used for inference as-is, all possible options that may affect model conversion are ignored."

@ilya-lavrenov

[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.

🚀

[ilya-lavrenov]

ready

[DarkLight1337]

Sorry for missing this!