vllm-project/tpu-inference

vLLM‑TPU is an open‑source extension that enables high‑throughput inference of large language models on Google TPU hardware. It provides a unified interface tha