Is the vLLM-compatible model Gemma‑3n‑E4B‑IT supported for audio concurrency?

I am running the Gemma-3n-4B-IT audio model using vLLM. It works fine for a single audio input, but when I process multiple audio inputs concurrently using asyncio, I encounter the following error:
AttributeError: 'list' object has no attribute 'squeeze'.
Could someone please help me understand whether vLLM supports concurrency for audio models, and if so, how to resolve this issue?