vLLM adds support for Transformers pooling models

The vLLM inference runtime now supports Transformers pooling models via a recent pull request merge. This addition expands the library's ability to handle diverse architecture types for high-throughput serving.