[VLM] Remove BaseProcessingInfo.get_mm_max_tokens_per_item

[DarkLight1337]

It is somewhat redundant to use BaseProcessingInfo.get_mm_max_tokens_per_item when we can directly get the number of multi-modal tokens from the dummy data now (via PlaceholderRange.get_num_embeds added by #15712).

Although this means we need to generate the dummy data to get the multi-modal token count, this is only called once before profiling starts, so it shouldn't affect the performance much. On the other hand, this simplifies the code somewhat by removing one abstract method. Note however that we still need to compute the number of multimodal tokens for the prompt updates.

This change won't conflict with the current WIP models since it just makes the abstract method unused rather than banning it.

[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.

🚀