Fine-tuned Mixtral 8x22B is ready. Curious what the easiest way to do inference is? Spinning up some GPUs for vLLM right now, is there something better?

→ View original post on X — @mattshumer_