Files
qwen38-flash-next-rtx6000d/audit/runtime/benchmark-summary.json
T

12 lines
278 B
JSON

{
"eager": {
"median_decode_tps": 22.001705837292434,
"median_ttft_s": 0.11141436400066596
},
"graph": {
"median_decode_tps": 77.11544284916513,
"median_ttft_s": 0.06954390600003535
},
"speedup": 3.5049756332282227,
"identical_generated_text": true
}