[BugFix] skip loading lm_head for llama if word embeddings are tied

[prashantgupta24]

Check if tie_word_embeddings is true, that means lm_head is tied to embeddings we should be able to skip loading lm_head

Similar to what was done for gemma here - #3553

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.
Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which consists a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of default ones by unblocking the steps in your fast-check build on Buildkite UI.

Once the PR is approved and ready to go, please make sure to run full CI as it is required to merge (or just use auto-merge).

To run full CI, you can do one of these:

• Comment /ready on the PR
• Add ready label to the PR
• Enable auto-merge.

🚀

[DarkLight1337]

Seems that it has already been done in #6758. Would you mind if the other PR is adopted instead?

[prashantgupta24]

Closing in favor of #6758