[V1] Use input_ids as input for text-only models

[WoosukKwon]

Currently, the model runner in V1 always uses input_embeds rather than input_ids for the compatibility with multi-modal models. However, this excludes the embedding layer from the CUDA graph, and thus causes slight performance regression (and larger performance regression when TP > 1). This PR addresses this by using the input_ids for text-only models while keeping using input_embeds for multi-modal models.

[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 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 do one of these:

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

🚀

[WoosukKwon]

Chatted offline with @ywang96, while we also need to support input embeddings for text-only models, this PR is good to go since the API is still not ready for V1.