vLLM v0.31.0rc3 Adds Randomized Dummy Input Support for Model Runner V2

Core summary

One sentence to understand this update

vLLM's release candidate v0.31.0rc3 now supports randomized dummy inputs within its Model Runner V2, likely for testing or performance benchmarking.

Impact & opportunity

What this could mean

This update allows for more robust testing and benchmarking of models on vLLM, potentially improving model development and deployment workflows.