Estimate inference throughput in tokens per second from the generated token count and end-to-end latency, for capacity planning.
實際受批大小與排隊影響。