There is an open PR for VLLM to do this via DefussionGemma<a href="https://github.com/vllm-project/vllm/pull/57250" rel="nofollow">https://github.com/vllm-project/vllm/pull/57250
mohsen1 · · focus · HN ↗
<a href="https://github.com/vllm-project/vllm/pull/57250" rel="nofollow">https://github.com/vllm-project/vllm/pull/57250