mudassiruddin21 commited on
Commit
d059788
·
verified ·
1 Parent(s): 4a804fc

Run r2: logs and results

Browse files
runs/r2/engine_logs/20261118-180129-eval-r1.log CHANGED
@@ -458,3 +458,171 @@ INFO: 127.0.0.1:38097 - "POST /v1/completions HTTP/1.1" 200 OK
458
  (APIServer pid=30189) INFO 11-18 18:14:20 [loggers.py:320] Engine 000: Avg prompt throughput: 27.4 tokens/s, Avg generation throughput: 67.2 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
459
  (APIServer pid=30189) INFO 11-18 18:14:20 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 3.76, Accepted throughput: 49.69 tokens/s, Drafted throughput: 53.99 tokens/s, Accepted: 497 tokens, Drafted: 540 tokens, Per-position acceptance rate: 0.978, 0.917, 0.867, Avg Draft acceptance rate: 92.0%
460
  INFO: 127.0.0.1:41457 - "POST /v1/completions HTTP/1.1" 200 OK
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
458
  (APIServer pid=30189) INFO 11-18 18:14:20 [loggers.py:320] Engine 000: Avg prompt throughput: 27.4 tokens/s, Avg generation throughput: 67.2 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
459
  (APIServer pid=30189) INFO 11-18 18:14:20 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 3.76, Accepted throughput: 49.69 tokens/s, Drafted throughput: 53.99 tokens/s, Accepted: 497 tokens, Drafted: 540 tokens, Per-position acceptance rate: 0.978, 0.917, 0.867, Avg Draft acceptance rate: 92.0%
460
  INFO: 127.0.0.1:41457 - "POST /v1/completions HTTP/1.1" 200 OK
461
+ INFO: 127.0.0.1:31996 - "POST /v1/completions HTTP/1.1" 200 OK
462
+ (APIServer pid=30189) INFO 11-18 18:14:30 [loggers.py:320] Engine 000: Avg prompt throughput: 14.6 tokens/s, Avg generation throughput: 68.8 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
463
+ (APIServer pid=30189) INFO 11-18 18:14:30 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 3.81, Accepted throughput: 50.79 tokens/s, Drafted throughput: 54.29 tokens/s, Accepted: 508 tokens, Drafted: 543 tokens, Per-position acceptance rate: 0.978, 0.934, 0.895, Avg Draft acceptance rate: 93.6%
464
+ INFO: 127.0.0.1:18798 - "POST /v1/completions HTTP/1.1" 200 OK
465
+ INFO: 127.0.0.1:27983 - "POST /v1/completions HTTP/1.1" 200 OK
466
+ INFO: 127.0.0.1:31906 - "POST /v1/completions HTTP/1.1" 200 OK
467
+ (APIServer pid=30189) INFO 11-18 18:14:40 [loggers.py:320] Engine 000: Avg prompt throughput: 26.5 tokens/s, Avg generation throughput: 64.5 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
468
+ (APIServer pid=30189) INFO 11-18 18:14:40 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 3.61, Accepted throughput: 46.69 tokens/s, Drafted throughput: 53.69 tokens/s, Accepted: 467 tokens, Drafted: 537 tokens, Per-position acceptance rate: 0.939, 0.872, 0.799, Avg Draft acceptance rate: 87.0%
469
+ INFO: 127.0.0.1:51142 - "POST /v1/completions HTTP/1.1" 200 OK
470
+ INFO: 127.0.0.1:32766 - "POST /v1/completions HTTP/1.1" 200 OK
471
+ INFO: 127.0.0.1:45832 - "POST /v1/completions HTTP/1.1" 200 OK
472
+ (APIServer pid=30189) INFO 11-18 18:14:50 [loggers.py:320] Engine 000: Avg prompt throughput: 23.0 tokens/s, Avg generation throughput: 67.9 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
473
+ (APIServer pid=30189) INFO 11-18 18:14:50 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 3.79, Accepted throughput: 49.99 tokens/s, Drafted throughput: 53.69 tokens/s, Accepted: 500 tokens, Drafted: 537 tokens, Per-position acceptance rate: 0.983, 0.933, 0.877, Avg Draft acceptance rate: 93.1%
474
+ INFO: 127.0.0.1:44358 - "POST /v1/completions HTTP/1.1" 200 OK
475
+ INFO: 127.0.0.1:23484 - "POST /v1/completions HTTP/1.1" 200 OK
476
+ (APIServer pid=30189) INFO 11-18 18:15:00 [loggers.py:320] Engine 000: Avg prompt throughput: 18.1 tokens/s, Avg generation throughput: 68.5 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
477
+ (APIServer pid=30189) INFO 11-18 18:15:00 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 3.80, Accepted throughput: 50.59 tokens/s, Drafted throughput: 54.29 tokens/s, Accepted: 506 tokens, Drafted: 543 tokens, Per-position acceptance rate: 0.989, 0.961, 0.845, Avg Draft acceptance rate: 93.2%
478
+ INFO: 127.0.0.1:25940 - "POST /v1/completions HTTP/1.1" 200 OK
479
+ INFO: 127.0.0.1:53066 - "POST /v1/completions HTTP/1.1" 200 OK
480
+ (APIServer pid=30189) INFO 11-18 18:15:10 [loggers.py:320] Engine 000: Avg prompt throughput: 17.4 tokens/s, Avg generation throughput: 66.0 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
481
+ (APIServer pid=30189) INFO 11-18 18:15:10 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 3.64, Accepted throughput: 47.79 tokens/s, Drafted throughput: 54.29 tokens/s, Accepted: 478 tokens, Drafted: 543 tokens, Per-position acceptance rate: 0.934, 0.895, 0.812, Avg Draft acceptance rate: 88.0%
482
+ INFO: 127.0.0.1:22931 - "GET /metrics HTTP/1.1" 200 OK
483
+ INFO: 127.0.0.1:20155 - "GET /metrics HTTP/1.1" 200 OK
484
+ INFO: 127.0.0.1:31323 - "POST /v1/completions HTTP/1.1" 200 OK
485
+ (APIServer pid=30189) INFO 11-18 18:15:20 [loggers.py:320] Engine 000: Avg prompt throughput: 4.2 tokens/s, Avg generation throughput: 48.7 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
486
+ (APIServer pid=30189) INFO 11-18 18:15:20 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.67, Accepted throughput: 30.50 tokens/s, Drafted throughput: 54.89 tokens/s, Accepted: 305 tokens, Drafted: 549 tokens, Per-position acceptance rate: 0.765, 0.525, 0.377, Avg Draft acceptance rate: 55.6%
487
+ INFO: 127.0.0.1:21275 - "POST /v1/completions HTTP/1.1" 200 OK
488
+ (APIServer pid=30189) INFO 11-18 18:15:30 [loggers.py:320] Engine 000: Avg prompt throughput: 3.8 tokens/s, Avg generation throughput: 47.9 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
489
+ (APIServer pid=30189) INFO 11-18 18:15:30 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.62, Accepted throughput: 29.70 tokens/s, Drafted throughput: 54.89 tokens/s, Accepted: 297 tokens, Drafted: 549 tokens, Per-position acceptance rate: 0.792, 0.525, 0.306, Avg Draft acceptance rate: 54.1%
490
+ (APIServer pid=30189) INFO 11-18 18:15:40 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 47.3 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
491
+ (APIServer pid=30189) INFO 11-18 18:15:40 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.56, Accepted throughput: 28.80 tokens/s, Drafted throughput: 55.49 tokens/s, Accepted: 288 tokens, Drafted: 555 tokens, Per-position acceptance rate: 0.778, 0.519, 0.259, Avg Draft acceptance rate: 51.9%
492
+ INFO: 127.0.0.1:43982 - "POST /v1/completions HTTP/1.1" 200 OK
493
+ (APIServer pid=30189) INFO 11-18 18:15:50 [loggers.py:320] Engine 000: Avg prompt throughput: 4.2 tokens/s, Avg generation throughput: 45.9 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
494
+ (APIServer pid=30189) INFO 11-18 18:15:50 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.53, Accepted throughput: 27.90 tokens/s, Drafted throughput: 54.60 tokens/s, Accepted: 279 tokens, Drafted: 546 tokens, Per-position acceptance rate: 0.758, 0.467, 0.308, Avg Draft acceptance rate: 51.1%
495
+ (APIServer pid=30189) INFO 11-18 18:16:00 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 46.1 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
496
+ (APIServer pid=30189) INFO 11-18 18:16:00 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.49, Accepted throughput: 27.59 tokens/s, Drafted throughput: 55.49 tokens/s, Accepted: 276 tokens, Drafted: 555 tokens, Per-position acceptance rate: 0.719, 0.492, 0.281, Avg Draft acceptance rate: 49.7%
497
+ INFO: 127.0.0.1:51669 - "POST /v1/completions HTTP/1.1" 200 OK
498
+ (APIServer pid=30189) INFO 11-18 18:16:10 [loggers.py:320] Engine 000: Avg prompt throughput: 7.0 tokens/s, Avg generation throughput: 47.6 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
499
+ (APIServer pid=30189) INFO 11-18 18:16:10 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.62, Accepted throughput: 29.50 tokens/s, Drafted throughput: 54.59 tokens/s, Accepted: 295 tokens, Drafted: 546 tokens, Per-position acceptance rate: 0.775, 0.522, 0.324, Avg Draft acceptance rate: 54.0%
500
+ (APIServer pid=30189) INFO 11-18 18:16:20 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 46.9 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
501
+ (APIServer pid=30189) INFO 11-18 18:16:20 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.54, Accepted throughput: 28.40 tokens/s, Drafted throughput: 55.50 tokens/s, Accepted: 284 tokens, Drafted: 555 tokens, Per-position acceptance rate: 0.778, 0.454, 0.303, Avg Draft acceptance rate: 51.2%
502
+ INFO: 127.0.0.1:27803 - "POST /v1/completions HTTP/1.1" 200 OK
503
+ (APIServer pid=30189) INFO 11-18 18:16:30 [loggers.py:320] Engine 000: Avg prompt throughput: 6.7 tokens/s, Avg generation throughput: 49.4 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
504
+ (APIServer pid=30189) INFO 11-18 18:16:30 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.71, Accepted throughput: 31.30 tokens/s, Drafted throughput: 54.90 tokens/s, Accepted: 313 tokens, Drafted: 549 tokens, Per-position acceptance rate: 0.781, 0.552, 0.377, Avg Draft acceptance rate: 57.0%
505
+ (APIServer pid=30189) INFO 11-18 18:16:40 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 42.6 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
506
+ (APIServer pid=30189) INFO 11-18 18:16:40 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.30, Accepted throughput: 24.10 tokens/s, Drafted throughput: 55.49 tokens/s, Accepted: 241 tokens, Drafted: 555 tokens, Per-position acceptance rate: 0.708, 0.405, 0.189, Avg Draft acceptance rate: 43.4%
507
+ INFO: 127.0.0.1:17069 - "POST /v1/completions HTTP/1.1" 200 OK
508
+ (APIServer pid=30189) INFO 11-18 18:16:50 [loggers.py:320] Engine 000: Avg prompt throughput: 4.6 tokens/s, Avg generation throughput: 48.3 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
509
+ (APIServer pid=30189) INFO 11-18 18:16:50 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.65, Accepted throughput: 30.00 tokens/s, Drafted throughput: 54.59 tokens/s, Accepted: 300 tokens, Drafted: 546 tokens, Per-position acceptance rate: 0.808, 0.527, 0.313, Avg Draft acceptance rate: 54.9%
510
+ (APIServer pid=30189) INFO 11-18 18:17:00 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 46.2 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
511
+ (APIServer pid=30189) INFO 11-18 18:17:00 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.48, Accepted throughput: 27.60 tokens/s, Drafted throughput: 55.79 tokens/s, Accepted: 276 tokens, Drafted: 558 tokens, Per-position acceptance rate: 0.715, 0.500, 0.269, Avg Draft acceptance rate: 49.5%
512
+ (APIServer pid=30189) INFO 11-18 18:17:10 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 45.6 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.4%
513
+ (APIServer pid=30189) INFO 11-18 18:17:10 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.48, Accepted throughput: 27.20 tokens/s, Drafted throughput: 55.20 tokens/s, Accepted: 272 tokens, Drafted: 552 tokens, Per-position acceptance rate: 0.696, 0.462, 0.321, Avg Draft acceptance rate: 49.3%
514
+ INFO: 127.0.0.1:34600 - "POST /v1/completions HTTP/1.1" 200 OK
515
+ (APIServer pid=30189) INFO 11-18 18:17:20 [loggers.py:320] Engine 000: Avg prompt throughput: 4.3 tokens/s, Avg generation throughput: 39.7 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
516
+ (APIServer pid=30189) INFO 11-18 18:17:20 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.17, Accepted throughput: 21.50 tokens/s, Drafted throughput: 54.89 tokens/s, Accepted: 215 tokens, Drafted: 549 tokens, Per-position acceptance rate: 0.705, 0.311, 0.158, Avg Draft acceptance rate: 39.2%
517
+ (APIServer pid=30189) INFO 11-18 18:17:30 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 47.0 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.4%
518
+ (APIServer pid=30189) INFO 11-18 18:17:30 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.54, Accepted throughput: 28.50 tokens/s, Drafted throughput: 55.49 tokens/s, Accepted: 285 tokens, Drafted: 555 tokens, Per-position acceptance rate: 0.816, 0.459, 0.265, Avg Draft acceptance rate: 51.4%
519
+ INFO: 127.0.0.1:62451 - "POST /v1/completions HTTP/1.1" 200 OK
520
+ (APIServer pid=30189) INFO 11-18 18:17:40 [loggers.py:320] Engine 000: Avg prompt throughput: 7.0 tokens/s, Avg generation throughput: 47.8 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
521
+ (APIServer pid=30189) INFO 11-18 18:17:40 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.64, Accepted throughput: 29.80 tokens/s, Drafted throughput: 54.59 tokens/s, Accepted: 298 tokens, Drafted: 546 tokens, Per-position acceptance rate: 0.802, 0.500, 0.335, Avg Draft acceptance rate: 54.6%
522
+ (APIServer pid=30189) INFO 11-18 18:17:50 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 43.0 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.4%
523
+ (APIServer pid=30189) INFO 11-18 18:17:50 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.34, Accepted throughput: 24.60 tokens/s, Drafted throughput: 55.19 tokens/s, Accepted: 246 tokens, Drafted: 552 tokens, Per-position acceptance rate: 0.668, 0.418, 0.250, Avg Draft acceptance rate: 44.6%
524
+ INFO: 127.0.0.1:40866 - "POST /v1/completions HTTP/1.1" 200 OK
525
+ (APIServer pid=30189) INFO 11-18 18:18:00 [loggers.py:320] Engine 000: Avg prompt throughput: 3.7 tokens/s, Avg generation throughput: 45.6 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
526
+ (APIServer pid=30189) INFO 11-18 18:18:00 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.50, Accepted throughput: 27.50 tokens/s, Drafted throughput: 54.90 tokens/s, Accepted: 275 tokens, Drafted: 549 tokens, Per-position acceptance rate: 0.699, 0.492, 0.311, Avg Draft acceptance rate: 50.1%
527
+ (APIServer pid=30189) INFO 11-18 18:18:10 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 44.7 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.4%
528
+ (APIServer pid=30189) INFO 11-18 18:18:10 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.43, Accepted throughput: 26.30 tokens/s, Drafted throughput: 55.19 tokens/s, Accepted: 263 tokens, Drafted: 552 tokens, Per-position acceptance rate: 0.723, 0.457, 0.250, Avg Draft acceptance rate: 47.6%
529
+ INFO: 127.0.0.1:32354 - "POST /v1/completions HTTP/1.1" 200 OK
530
+ (APIServer pid=30189) INFO 11-18 18:18:20 [loggers.py:320] Engine 000: Avg prompt throughput: 7.2 tokens/s, Avg generation throughput: 47.1 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
531
+ (APIServer pid=30189) INFO 11-18 18:18:20 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.57, Accepted throughput: 28.70 tokens/s, Drafted throughput: 54.90 tokens/s, Accepted: 287 tokens, Drafted: 549 tokens, Per-position acceptance rate: 0.776, 0.497, 0.295, Avg Draft acceptance rate: 52.3%
532
+ (APIServer pid=30189) INFO 11-18 18:18:30 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 45.3 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.4%
533
+ (APIServer pid=30189) INFO 11-18 18:18:30 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.46, Accepted throughput: 26.90 tokens/s, Drafted throughput: 55.19 tokens/s, Accepted: 269 tokens, Drafted: 552 tokens, Per-position acceptance rate: 0.712, 0.473, 0.277, Avg Draft acceptance rate: 48.7%
534
+ INFO: 127.0.0.1:23920 - "POST /v1/completions HTTP/1.1" 200 OK
535
+ (APIServer pid=30189) INFO 11-18 18:18:40 [loggers.py:320] Engine 000: Avg prompt throughput: 4.2 tokens/s, Avg generation throughput: 47.8 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
536
+ (APIServer pid=30189) INFO 11-18 18:18:40 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.63, Accepted throughput: 29.70 tokens/s, Drafted throughput: 54.60 tokens/s, Accepted: 297 tokens, Drafted: 546 tokens, Per-position acceptance rate: 0.786, 0.511, 0.335, Avg Draft acceptance rate: 54.4%
537
+ (APIServer pid=30189) INFO 11-18 18:18:50 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 43.5 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
538
+ (APIServer pid=30189) INFO 11-18 18:18:50 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.35, Accepted throughput: 25.00 tokens/s, Drafted throughput: 55.49 tokens/s, Accepted: 250 tokens, Drafted: 555 tokens, Per-position acceptance rate: 0.681, 0.427, 0.243, Avg Draft acceptance rate: 45.0%
539
+ INFO: 127.0.0.1:32180 - "POST /v1/completions HTTP/1.1" 200 OK
540
+ (APIServer pid=30189) INFO 11-18 18:19:00 [loggers.py:320] Engine 000: Avg prompt throughput: 7.3 tokens/s, Avg generation throughput: 45.3 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
541
+ (APIServer pid=30189) INFO 11-18 18:19:00 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.48, Accepted throughput: 27.00 tokens/s, Drafted throughput: 54.59 tokens/s, Accepted: 270 tokens, Drafted: 546 tokens, Per-position acceptance rate: 0.742, 0.467, 0.275, Avg Draft acceptance rate: 49.5%
542
+ (APIServer pid=30189) INFO 11-18 18:19:10 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 48.3 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
543
+ (APIServer pid=30189) INFO 11-18 18:19:10 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.62, Accepted throughput: 29.89 tokens/s, Drafted throughput: 55.19 tokens/s, Accepted: 299 tokens, Drafted: 552 tokens, Per-position acceptance rate: 0.766, 0.554, 0.304, Avg Draft acceptance rate: 54.2%
544
+ INFO: 127.0.0.1:28654 - "POST /v1/completions HTTP/1.1" 200 OK
545
+ (APIServer pid=30189) INFO 11-18 18:19:20 [loggers.py:320] Engine 000: Avg prompt throughput: 7.0 tokens/s, Avg generation throughput: 46.9 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
546
+ (APIServer pid=30189) INFO 11-18 18:19:20 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.57, Accepted throughput: 28.69 tokens/s, Drafted throughput: 54.89 tokens/s, Accepted: 287 tokens, Drafted: 549 tokens, Per-position acceptance rate: 0.781, 0.497, 0.290, Avg Draft acceptance rate: 52.3%
547
+ (APIServer pid=30189) INFO 11-18 18:19:30 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 45.4 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
548
+ (APIServer pid=30189) INFO 11-18 18:19:30 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.45, Accepted throughput: 26.89 tokens/s, Drafted throughput: 55.49 tokens/s, Accepted: 269 tokens, Drafted: 555 tokens, Per-position acceptance rate: 0.708, 0.459, 0.286, Avg Draft acceptance rate: 48.5%
549
+ INFO: 127.0.0.1:48221 - "POST /v1/completions HTTP/1.1" 200 OK
550
+ (APIServer pid=30189) INFO 11-18 18:19:40 [loggers.py:320] Engine 000: Avg prompt throughput: 9.0 tokens/s, Avg generation throughput: 42.6 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
551
+ (APIServer pid=30189) INFO 11-18 18:19:40 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.34, Accepted throughput: 24.40 tokens/s, Drafted throughput: 54.59 tokens/s, Accepted: 244 tokens, Drafted: 546 tokens, Per-position acceptance rate: 0.659, 0.429, 0.253, Avg Draft acceptance rate: 44.7%
552
+ (APIServer pid=30189) INFO 11-18 18:19:50 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 47.8 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
553
+ (APIServer pid=30189) INFO 11-18 18:19:50 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.60, Accepted throughput: 29.39 tokens/s, Drafted throughput: 55.19 tokens/s, Accepted: 294 tokens, Drafted: 552 tokens, Per-position acceptance rate: 0.804, 0.511, 0.283, Avg Draft acceptance rate: 53.3%
554
+ INFO: 127.0.0.1:43605 - "POST /v1/completions HTTP/1.1" 200 OK
555
+ (APIServer pid=30189) INFO 11-18 18:20:00 [loggers.py:320] Engine 000: Avg prompt throughput: 7.1 tokens/s, Avg generation throughput: 43.9 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
556
+ (APIServer pid=30189) INFO 11-18 18:20:00 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.42, Accepted throughput: 25.80 tokens/s, Drafted throughput: 54.59 tokens/s, Accepted: 258 tokens, Drafted: 546 tokens, Per-position acceptance rate: 0.714, 0.423, 0.280, Avg Draft acceptance rate: 47.3%
557
+ (APIServer pid=30189) INFO 11-18 18:20:10 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 47.2 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
558
+ (APIServer pid=30189) INFO 11-18 18:20:10 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.55, Accepted throughput: 28.70 tokens/s, Drafted throughput: 55.50 tokens/s, Accepted: 287 tokens, Drafted: 555 tokens, Per-position acceptance rate: 0.800, 0.459, 0.292, Avg Draft acceptance rate: 51.7%
559
+ INFO: 127.0.0.1:35329 - "POST /v1/completions HTTP/1.1" 200 OK
560
+ (APIServer pid=30189) INFO 11-18 18:20:20 [loggers.py:320] Engine 000: Avg prompt throughput: 4.4 tokens/s, Avg generation throughput: 50.5 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
561
+ (APIServer pid=30189) INFO 11-18 18:20:20 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.76, Accepted throughput: 32.20 tokens/s, Drafted throughput: 54.90 tokens/s, Accepted: 322 tokens, Drafted: 549 tokens, Per-position acceptance rate: 0.798, 0.574, 0.388, Avg Draft acceptance rate: 58.7%
562
+ (APIServer pid=30189) INFO 11-18 18:20:30 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 43.8 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
563
+ (APIServer pid=30189) INFO 11-18 18:20:30 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.37, Accepted throughput: 25.30 tokens/s, Drafted throughput: 55.49 tokens/s, Accepted: 253 tokens, Drafted: 555 tokens, Per-position acceptance rate: 0.724, 0.422, 0.222, Avg Draft acceptance rate: 45.6%
564
+ INFO: 127.0.0.1:24814 - "POST /v1/completions HTTP/1.1" 200 OK
565
+ (APIServer pid=30189) INFO 11-18 18:20:40 [loggers.py:320] Engine 000: Avg prompt throughput: 5.4 tokens/s, Avg generation throughput: 46.3 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
566
+ (APIServer pid=30189) INFO 11-18 18:20:40 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.56, Accepted throughput: 28.30 tokens/s, Drafted throughput: 54.29 tokens/s, Accepted: 283 tokens, Drafted: 543 tokens, Per-position acceptance rate: 0.757, 0.519, 0.287, Avg Draft acceptance rate: 52.1%
567
+ (APIServer pid=30189) INFO 11-18 18:20:50 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 44.6 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
568
+ (APIServer pid=30189) INFO 11-18 18:20:50 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.41, Accepted throughput: 26.10 tokens/s, Drafted throughput: 55.49 tokens/s, Accepted: 261 tokens, Drafted: 555 tokens, Per-position acceptance rate: 0.692, 0.454, 0.265, Avg Draft acceptance rate: 47.0%
569
+ (APIServer pid=30189) INFO 11-18 18:21:00 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 46.1 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.4%
570
+ (APIServer pid=30189) INFO 11-18 18:21:00 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.51, Accepted throughput: 27.70 tokens/s, Drafted throughput: 55.19 tokens/s, Accepted: 277 tokens, Drafted: 552 tokens, Per-position acceptance rate: 0.734, 0.467, 0.304, Avg Draft acceptance rate: 50.2%
571
+ INFO: 127.0.0.1:64108 - "POST /v1/completions HTTP/1.1" 200 OK
572
+ (APIServer pid=30189) INFO 11-18 18:21:10 [loggers.py:320] Engine 000: Avg prompt throughput: 4.9 tokens/s, Avg generation throughput: 45.6 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
573
+ (APIServer pid=30189) INFO 11-18 18:21:10 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.50, Accepted throughput: 27.50 tokens/s, Drafted throughput: 54.90 tokens/s, Accepted: 275 tokens, Drafted: 549 tokens, Per-position acceptance rate: 0.760, 0.486, 0.257, Avg Draft acceptance rate: 50.1%
574
+ (APIServer pid=30189) INFO 11-18 18:21:20 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 47.9 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.4%
575
+ (APIServer pid=30189) INFO 11-18 18:21:20 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.60, Accepted throughput: 29.50 tokens/s, Drafted throughput: 55.20 tokens/s, Accepted: 295 tokens, Drafted: 552 tokens, Per-position acceptance rate: 0.755, 0.500, 0.348, Avg Draft acceptance rate: 53.4%
576
+ INFO: 127.0.0.1:46741 - "POST /v1/completions HTTP/1.1" 200 OK
577
+ (APIServer pid=30189) INFO 11-18 18:21:30 [loggers.py:320] Engine 000: Avg prompt throughput: 7.5 tokens/s, Avg generation throughput: 45.8 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
578
+ (APIServer pid=30189) INFO 11-18 18:21:30 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.50, Accepted throughput: 27.40 tokens/s, Drafted throughput: 54.89 tokens/s, Accepted: 274 tokens, Drafted: 549 tokens, Per-position acceptance rate: 0.754, 0.464, 0.279, Avg Draft acceptance rate: 49.9%
579
+ (APIServer pid=30189) INFO 11-18 18:21:40 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 46.8 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.4%
580
+ (APIServer pid=30189) INFO 11-18 18:21:40 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.54, Accepted throughput: 28.40 tokens/s, Drafted throughput: 55.20 tokens/s, Accepted: 284 tokens, Drafted: 552 tokens, Per-position acceptance rate: 0.799, 0.484, 0.261, Avg Draft acceptance rate: 51.4%
581
+ INFO: 127.0.0.1:43348 - "POST /v1/completions HTTP/1.1" 200 OK
582
+ (APIServer pid=30189) INFO 11-18 18:21:50 [loggers.py:320] Engine 000: Avg prompt throughput: 4.2 tokens/s, Avg generation throughput: 50.4 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
583
+ (APIServer pid=30189) INFO 11-18 18:21:50 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.77, Accepted throughput: 32.20 tokens/s, Drafted throughput: 54.60 tokens/s, Accepted: 322 tokens, Drafted: 546 tokens, Per-position acceptance rate: 0.819, 0.582, 0.368, Avg Draft acceptance rate: 59.0%
584
+ (APIServer pid=30189) INFO 11-18 18:22:00 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 46.6 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.4%
585
+ (APIServer pid=30189) INFO 11-18 18:22:00 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.52, Accepted throughput: 28.10 tokens/s, Drafted throughput: 55.49 tokens/s, Accepted: 281 tokens, Drafted: 555 tokens, Per-position acceptance rate: 0.751, 0.476, 0.292, Avg Draft acceptance rate: 50.6%
586
+ INFO: 127.0.0.1:41184 - "POST /v1/completions HTTP/1.1" 200 OK
587
+ (APIServer pid=30189) INFO 11-18 18:22:10 [loggers.py:320] Engine 000: Avg prompt throughput: 5.4 tokens/s, Avg generation throughput: 43.7 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
588
+ (APIServer pid=30189) INFO 11-18 18:22:10 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.41, Accepted throughput: 25.60 tokens/s, Drafted throughput: 54.60 tokens/s, Accepted: 256 tokens, Drafted: 546 tokens, Per-position acceptance rate: 0.725, 0.434, 0.247, Avg Draft acceptance rate: 46.9%
589
+ (APIServer pid=30189) INFO 11-18 18:22:20 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 46.2 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.4%
590
+ (APIServer pid=30189) INFO 11-18 18:22:20 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.51, Accepted throughput: 27.80 tokens/s, Drafted throughput: 55.19 tokens/s, Accepted: 278 tokens, Drafted: 552 tokens, Per-position acceptance rate: 0.717, 0.495, 0.299, Avg Draft acceptance rate: 50.4%
591
+ INFO: 127.0.0.1:57921 - "POST /v1/completions HTTP/1.1" 200 OK
592
+ (APIServer pid=30189) INFO 11-18 18:22:30 [loggers.py:320] Engine 000: Avg prompt throughput: 7.3 tokens/s, Avg generation throughput: 43.2 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
593
+ (APIServer pid=30189) INFO 11-18 18:22:30 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.36, Accepted throughput: 24.80 tokens/s, Drafted throughput: 54.89 tokens/s, Accepted: 248 tokens, Drafted: 549 tokens, Per-position acceptance rate: 0.760, 0.393, 0.202, Avg Draft acceptance rate: 45.2%
594
+ (APIServer pid=30189) INFO 11-18 18:22:40 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 42.3 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.4%
595
+ (APIServer pid=30189) INFO 11-18 18:22:40 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.30, Accepted throughput: 23.90 tokens/s, Drafted throughput: 55.19 tokens/s, Accepted: 239 tokens, Drafted: 552 tokens, Per-position acceptance rate: 0.696, 0.391, 0.212, Avg Draft acceptance rate: 43.3%
596
+ INFO: 127.0.0.1:21104 - "POST /v1/completions HTTP/1.1" 200 OK
597
+ (APIServer pid=30189) INFO 11-18 18:22:50 [loggers.py:320] Engine 000: Avg prompt throughput: 6.1 tokens/s, Avg generation throughput: 51.2 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
598
+ (APIServer pid=30189) INFO 11-18 18:22:50 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.80, Accepted throughput: 32.89 tokens/s, Drafted throughput: 54.89 tokens/s, Accepted: 329 tokens, Drafted: 549 tokens, Per-position acceptance rate: 0.798, 0.607, 0.393, Avg Draft acceptance rate: 59.9%
599
+ (APIServer pid=30189) INFO 11-18 18:23:00 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 49.0 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.4%
600
+ (APIServer pid=30189) INFO 11-18 18:23:00 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.66, Accepted throughput: 30.59 tokens/s, Drafted throughput: 55.19 tokens/s, Accepted: 306 tokens, Drafted: 552 tokens, Per-position acceptance rate: 0.783, 0.522, 0.359, Avg Draft acceptance rate: 55.4%
601
+ INFO: 127.0.0.1:18446 - "POST /v1/completions HTTP/1.1" 200 OK
602
+ (APIServer pid=30189) INFO 11-18 18:23:10 [loggers.py:320] Engine 000: Avg prompt throughput: 5.3 tokens/s, Avg generation throughput: 43.1 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
603
+ (APIServer pid=30189) INFO 11-18 18:23:10 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.37, Accepted throughput: 25.00 tokens/s, Drafted throughput: 54.59 tokens/s, Accepted: 250 tokens, Drafted: 546 tokens, Per-position acceptance rate: 0.709, 0.429, 0.236, Avg Draft acceptance rate: 45.8%
604
+ (APIServer pid=30189) INFO 11-18 18:23:20 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 47.0 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.4%
605
+ (APIServer pid=30189) INFO 11-18 18:23:20 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.54, Accepted throughput: 28.50 tokens/s, Drafted throughput: 55.50 tokens/s, Accepted: 285 tokens, Drafted: 555 tokens, Per-position acceptance rate: 0.762, 0.503, 0.276, Avg Draft acceptance rate: 51.4%
606
+ INFO: 127.0.0.1:28962 - "POST /v1/completions HTTP/1.1" 200 OK
607
+ (APIServer pid=30189) INFO 11-18 18:23:30 [loggers.py:320] Engine 000: Avg prompt throughput: 4.1 tokens/s, Avg generation throughput: 45.0 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
608
+ (APIServer pid=30189) INFO 11-18 18:23:30 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.47, Accepted throughput: 26.80 tokens/s, Drafted throughput: 54.59 tokens/s, Accepted: 268 tokens, Drafted: 546 tokens, Per-position acceptance rate: 0.725, 0.462, 0.286, Avg Draft acceptance rate: 49.1%
609
+ (APIServer pid=30189) INFO 11-18 18:23:40 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 46.1 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.4%
610
+ (APIServer pid=30189) INFO 11-18 18:23:40 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.51, Accepted throughput: 27.70 tokens/s, Drafted throughput: 55.19 tokens/s, Accepted: 277 tokens, Drafted: 552 tokens, Per-position acceptance rate: 0.761, 0.478, 0.266, Avg Draft acceptance rate: 50.2%
611
+ INFO: 127.0.0.1:20457 - "POST /v1/completions HTTP/1.1" 200 OK
612
+ (APIServer pid=30189) INFO 11-18 18:23:50 [loggers.py:320] Engine 000: Avg prompt throughput: 5.7 tokens/s, Avg generation throughput: 43.5 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
613
+ (APIServer pid=30189) INFO 11-18 18:23:50 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.37, Accepted throughput: 25.10 tokens/s, Drafted throughput: 54.90 tokens/s, Accepted: 251 tokens, Drafted: 549 tokens, Per-position acceptance rate: 0.699, 0.421, 0.251, Avg Draft acceptance rate: 45.7%
614
+ (APIServer pid=30189) INFO 11-18 18:24:00 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 44.5 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
615
+ (APIServer pid=30189) INFO 11-18 18:24:00 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.42, Accepted throughput: 26.10 tokens/s, Drafted throughput: 55.19 tokens/s, Accepted: 261 tokens, Drafted: 552 tokens, Per-position acceptance rate: 0.712, 0.451, 0.255, Avg Draft acceptance rate: 47.3%
616
+ INFO: 127.0.0.1:36830 - "POST /v1/completions HTTP/1.1" 200 OK
617
+ (APIServer pid=30189) INFO 11-18 18:24:10 [loggers.py:320] Engine 000: Avg prompt throughput: 6.5 tokens/s, Avg generation throughput: 43.1 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
618
+ (APIServer pid=30189) INFO 11-18 18:24:10 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.37, Accepted throughput: 24.89 tokens/s, Drafted throughput: 54.59 tokens/s, Accepted: 249 tokens, Drafted: 546 tokens, Per-position acceptance rate: 0.736, 0.418, 0.214, Avg Draft acceptance rate: 45.6%
619
+ (APIServer pid=30189) INFO 11-18 18:24:20 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 43.9 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
620
+ (APIServer pid=30189) INFO 11-18 18:24:20 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.37, Accepted throughput: 25.40 tokens/s, Drafted throughput: 55.49 tokens/s, Accepted: 254 tokens, Drafted: 555 tokens, Per-position acceptance rate: 0.708, 0.422, 0.243, Avg Draft acceptance rate: 45.8%
621
+ INFO: 127.0.0.1:49748 - "POST /v1/completions HTTP/1.1" 200 OK
622
+ (APIServer pid=30189) INFO 11-18 18:24:30 [loggers.py:320] Engine 000: Avg prompt throughput: 5.4 tokens/s, Avg generation throughput: 46.9 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
623
+ (APIServer pid=30189) INFO 11-18 18:24:30 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.58, Accepted throughput: 28.70 tokens/s, Drafted throughput: 54.60 tokens/s, Accepted: 287 tokens, Drafted: 546 tokens, Per-position acceptance rate: 0.747, 0.516, 0.313, Avg Draft acceptance rate: 52.6%
624
+ (APIServer pid=30189) INFO 11-18 18:24:40 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 45.9 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
625
+ (APIServer pid=30189) INFO 11-18 18:24:40 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.48, Accepted throughput: 27.40 tokens/s, Drafted throughput: 55.50 tokens/s, Accepted: 274 tokens, Drafted: 555 tokens, Per-position acceptance rate: 0.741, 0.459, 0.281, Avg Draft acceptance rate: 49.4%
626
+ INFO: 127.0.0.1:18833 - "POST /v1/completions HTTP/1.1" 200 OK
627
+ (APIServer pid=30189) INFO 11-18 18:24:50 [loggers.py:320] Engine 000: Avg prompt throughput: 5.7 tokens/s, Avg generation throughput: 46.8 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
628
+ (APIServer pid=30189) INFO 11-18 18:24:50 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.57, Accepted throughput: 28.50 tokens/s, Drafted throughput: 54.59 tokens/s, Accepted: 285 tokens, Drafted: 546 tokens, Per-position acceptance rate: 0.747, 0.495, 0.324, Avg Draft acceptance rate: 52.2%
runs/r2/logs/eval_main.log CHANGED
@@ -9,3 +9,4 @@ PASS
9
  [18:03:51] vLLM ready after 142s
10
  [18:09:27] chat: 40 prompts done
11
  [18:12:07] code: 40 prompts done
 
 
9
  [18:03:51] vLLM ready after 142s
10
  [18:09:27] chat: 40 prompts done
11
  [18:12:07] code: 40 prompts done
12
+ [18:15:11] math: 40 prompts done
runs/r2/logs/stats.log CHANGED
@@ -32,3 +32,4 @@
32
  [2026-11-18 18:01:20] ram 0.2/172 GB used (cache 0.0, shmem 0.0) | cgroup 0.2/172 GB | disk 117 GB used | gpu 0.0/95.6 GiB, 0%, 60 W | load 0.0 | largest python 0.1 GB, python 0.1 GB, python 0.0 GB
33
  [2026-11-18 18:04:21] ram 4.6/172 GB used (cache 0.3, shmem 0.0) | cgroup 4.6/172 GB | disk 117 GB used | gpu 83.1/95.6 GiB, 100%, 338 W | load 0.0 | largest VLLM::EngineCor 4.3 GB, vllm 2.5 GB, python 1.2 GB
34
  [2026-11-18 18:14:23] ram 4.6/172 GB used (cache 0.3, shmem 0.0) | cgroup 4.6/172 GB | disk 117 GB used | gpu 83.1/95.6 GiB, 100%, 290 W | load 0.0 | largest VLLM::EngineCor 4.3 GB, vllm 2.5 GB, python 1.2 GB
 
 
32
  [2026-11-18 18:01:20] ram 0.2/172 GB used (cache 0.0, shmem 0.0) | cgroup 0.2/172 GB | disk 117 GB used | gpu 0.0/95.6 GiB, 0%, 60 W | load 0.0 | largest python 0.1 GB, python 0.1 GB, python 0.0 GB
33
  [2026-11-18 18:04:21] ram 4.6/172 GB used (cache 0.3, shmem 0.0) | cgroup 4.6/172 GB | disk 117 GB used | gpu 83.1/95.6 GiB, 100%, 338 W | load 0.0 | largest VLLM::EngineCor 4.3 GB, vllm 2.5 GB, python 1.2 GB
34
  [2026-11-18 18:14:23] ram 4.6/172 GB used (cache 0.3, shmem 0.0) | cgroup 4.6/172 GB | disk 117 GB used | gpu 83.1/95.6 GiB, 100%, 290 W | load 0.0 | largest VLLM::EngineCor 4.3 GB, vllm 2.5 GB, python 1.2 GB
35
+ [2026-11-18 18:24:55] ram 4.6/172 GB used (cache 0.3, shmem 0.0) | cgroup 4.6/172 GB | disk 117 GB used | gpu 83.1/95.6 GiB, 100%, 292 W | load 0.0 | largest VLLM::EngineCor 4.3 GB, vllm 2.5 GB, python 1.2 GB
runs/r2/logs/sync.log CHANGED
@@ -27,3 +27,4 @@ uploaded sub_bf16 to mudassiruddin21/EMC:submissions/r2/main-bf16 (https://huggi
27
  No files have been modified since last commit. Skipping to prevent empty commit.
28
  uploaded sub_fp8 to mudassiruddin21/EMC:submissions/r2/main-fp8 (https://huggingface.co/datasets/mudassiruddin21/EMC/commit/aecb03559f68008a57d31a963e1ddbbef171dd3f)
29
  uploaded logs/results to mudassiruddin21/EMC:runs/r2
 
 
27
  No files have been modified since last commit. Skipping to prevent empty commit.
28
  uploaded sub_fp8 to mudassiruddin21/EMC:submissions/r2/main-fp8 (https://huggingface.co/datasets/mudassiruddin21/EMC/commit/aecb03559f68008a57d31a963e1ddbbef171dd3f)
29
  uploaded logs/results to mudassiruddin21/EMC:runs/r2
30
+ uploaded logs/results to mudassiruddin21/EMC:runs/r2