👋 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.

🚀

@DarkLight1337
Hi, I have a question regarding the use of vllm/model_executor/models/minimax_cache.py. We're relying on this module, and its initialization depends on the request_ids_to_seq_ids and finished_requests_ids fields. These fields are part of ModelInput, but they're not available when using DummyRun.
Do you have any suggestions on how to handle this, or how we might construct these two fields?
Previously, I noticed that these fields were included in attn_metadata, which has since been removed.

@ywang96 might have a better idea regarding this.

This pull request has merge conflicts that must be resolved before it can be
merged. Please rebase the PR, @qscqesze.

https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/working-with-forks/syncing-a-fork

Closing as stale