reports/vllm_openai_9b.json
227 B · 9 lines · json Raw
1 {
2 "n": 2000,
3 "concurrency": 64,
4 "req_per_s": 150.14946557251415,
5 "p50_ms": 411.9391441345215,
6 "p90_ms": 559.4574689865112,
7 "vs_pytorch_max_abs_dp": 0.0038683414459228516,
8 "vs_pytorch_mean_abs_dp": 0.0008296253508888185
9 }