Are you going to support vllm for inference?

[denadai2]

Hi,
greeeeat paper and results first of all! Are you considering adding the support for vllm in your adapter? This would immensely speed up the inference and deployment in production

thanks