Description: Serve Llama 3.1 70b on Tensor Processing Units (TPUs) in GKE, using vLLM.
コンソール