{ "status": "complete", "updated": "2026-09-06T00:57:14.243848+08:00", "scope": "Rebase 後同碼：4090 與 Orin；CUDA、Online TRT、Offline TRT；3/6/15 秒；bulk/realtime；RAM/VRAM，不含 disk", "git": { "branch": "feat/orin-tensorrt", "head": "a16280f1ea7e1e663399f13a216606ef50b38318", "origin_main": "4ee182123144c445c30c5d1cd8d545ab45357016" }, "build_deploy": { "4090_base_image_seconds": 36.52, "4090_trt_layer_seconds": 459.27, "orin_online_deploy_sh_seconds": 531.2, "orin_online_rollout_seconds": 165.272, "orin_offline_deploy_sh_seconds": 287.57 }, "audio_sha256": { "3s": "11034b66e2ae87bdec98dfe7401642387b85262b2e9c222b9e201030dc397a1c", "6s": "9566be6ccf6c7307830514e0fe5be86250c493efaef2800f27bfccc1ca1850b0", "15s": "4950a411664da365f4b8cfe13838f881430f01f2fed8b2c3335719f6b791ff9f" }, "platforms": { "4090": { "environment": { "platform": "Linux-6.8.0-136-generic-x86_64-with-glibc2.36", "python": "3.10.19", "packages": { "onnxruntime-gpu": "1.23.2", "numpy": "2.2.6", "sherpa-onnx": "1.12.36", "torch": "2.9.1", "kitt-core": "0.3.7", "pipecat-ai": "0.0.108", "tensorrt-cu12": "10.13.3.9" }, "sha256": { "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx": "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994", "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/tokens.txt": "f449eb28dc567533d7fa59be34e2abca8784f771850c78a47fb731a31429a1dc", "/app/src/asr/batch_asr_manager.py": "fb9a00f6cdbc5bd972d3bb82c85b44a70277a593d1694460071bac124ef25815", "/app/src/kitt_stt_service.py": "5283feaee7a213feb529afe1d98308c7fec447ccd18aa9f737a7632ea08b83bc" } }, "source_hashes": [ "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994 /app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "f449eb28dc567533d7fa59be34e2abca8784f771850c78a47fb731a31429a1dc /app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/tokens.txt", "fb9a00f6cdbc5bd972d3bb82c85b44a70277a593d1694460071bac124ef25815 /app/src/asr/batch_asr_manager.py", "5283feaee7a213feb529afe1d98308c7fec447ccd18aa9f737a7632ea08b83bc /app/src/kitt_stt_service.py" ], "graph": { "cuda": { "arguments": { "self_check": false, "mode": "graph", "provider": "cuda", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": null, "cache": "/bench/4090/online-cache", "audio": null, "length": 3, "frames": "50,100,250", "warmup": 10, "repeats": 50, "threads": 2, "profile": false, "dump_context": null, "out": "/bench/4090/graph-cuda.json" }, "utc": "2026-09-05T15:47:29Z", "machine": "x86_64", "ort": "1.23.2", "numpy": "2.2.6", "model_sha256": "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994", "start": { "VmHWM_MiB": 47.33203125, "VmRSS_MiB": 45.93359375, "Pss_MiB": 45.9296875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 95226.10546875, "memory.current_MiB": 28.24609375, "memory.peak_MiB": 30.21875, "gpu_total_memory_util_clock_power": "9569, 0, 210, 25.91" }, "rows": [ { "frames": 50, "p50_ms": 8.229509927332401, "p95_ms": 8.709857007488608, "samples_ms": [ 8.56557791121304, 8.486267179250717, 8.629079908132553, 8.68336483836174, 8.543949108570814, 8.449758170172572, 8.684423984959722, 8.605704875662923, 8.505851961672306, 8.61129304394126, 8.502527140080929, 8.55311006307602, 8.567909942939878, 8.492808090522885, 8.508480153977871, 8.532097097486258, 8.541222894564271, 8.807723876088858, 8.709857007488608, 8.583734976127744, 8.50635813549161, 8.467409992590547, 8.587050018832088, 8.169414009898901, 8.145015919581056, 8.09298106469214, 8.10725288465619, 8.050389820709825, 8.0100919585675, 8.268500911071897, 8.188168983906507, 8.098904974758625, 9.11645800806582, 8.083817781880498, 8.100470993667841, 8.044140180572867, 8.182193851098418, 8.178977994248271, 8.158163167536259, 8.055913029238582, 8.061457891017199, 8.190518943592906, 8.165681036189198, 8.105196990072727, 8.154971059411764, 8.071891032159328, 8.053672034293413, 8.061061147600412, 8.16251221112907, 8.092211093753576 ], "memory": { "VmHWM_MiB": 1327.23828125, "VmRSS_MiB": 1327.23828125, "Pss_MiB": 1322.7001953125, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93539.3984375, "memory.current_MiB": 1114.08203125, "memory.peak_MiB": 1164.02734375, "gpu_total_memory_util_clock_power": "11008, 27, 2535, 80.82" }, "finite": true, "input_sha256": "472ba979ad46ce462cc2cceeb0d218314362876e03b63f06a136a1c19b3dbab7" }, { "frames": 100, "p50_ms": 8.189372951164842, "p95_ms": 8.483339101076126, "samples_ms": [ 8.118561003357172, 10.513596003875136, 8.216292131692171, 8.152917958796024, 8.348300121724606, 8.155470015481114, 8.183712838217616, 8.144141174852848, 8.267386816442013, 8.177608950063586, 8.175559109076858, 8.175501832738519, 8.258444955572486, 8.170560002326965, 8.125284919515252, 8.09020595625043, 8.207384031265974, 8.166659157723188, 8.19231802597642, 8.144388906657696, 8.365570101886988, 8.241906063631177, 8.29226104542613, 8.165286853909492, 8.37348704226315, 8.252982050180435, 8.231566986069083, 8.28067003749311, 8.483339101076126, 8.338260930031538, 8.178003132343292, 8.101746905595064, 8.07374413125217, 8.100806036964059, 8.10988899320364, 8.152798051014543, 8.267235942184925, 8.240442955866456, 8.186427876353264, 8.103700121864676, 8.120785001665354, 8.21356801316142, 8.169061038643122, 8.327547926455736, 8.195739006623626, 8.215670008212328, 8.171134162694216, 8.591526187956333, 8.351088035851717, 8.211666950955987 ], "memory": { "VmHWM_MiB": 1337.52734375, "VmRSS_MiB": 1337.52734375, "Pss_MiB": 1332.9892578125, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93510.10546875, "memory.current_MiB": 1124.0546875, "memory.peak_MiB": 1164.02734375, "gpu_total_memory_util_clock_power": "11008, 93, 2535, 142.32" }, "finite": true, "input_sha256": "dda0d977bf96004405c6cbeac7cd57cfb466001cc340939617503e06dff71881" }, { "frames": 250, "p50_ms": 10.659188032150269, "p95_ms": 11.639205971732736, "samples_ms": [ 10.64682099968195, 13.655182207003236, 10.77385782264173, 10.610255878418684, 10.802691103890538, 10.52300981245935, 10.494997026398778, 10.573450941592455, 10.63136593438685, 10.508706094697118, 10.459396988153458, 10.465481085702777, 10.554869892075658, 10.550484992563725, 10.501255979761481, 11.030755005776882, 10.821118019521236, 11.212273966521025, 11.004111962392926, 11.347366962581873, 11.465393006801605, 10.612449841573834, 11.561820982024074, 10.774947935715318, 10.709288064390421, 11.639205971732736, 10.79412898980081, 10.64964896067977, 10.801638010889292, 10.757154086604714, 10.680319042876363, 10.924251982942224, 11.891711037606001, 10.668727103620768, 10.564223164692521, 10.531435022130609, 10.853680083528161, 10.583750903606415, 10.768330888822675, 10.689751943573356, 10.618708096444607, 11.169383069500327, 10.809922823682427, 10.63038781285286, 10.52229106426239, 10.625130962580442, 10.509300976991653, 10.440737009048462, 10.632460005581379, 10.541865136474371 ], "memory": { "VmHWM_MiB": 1376.15234375, "VmRSS_MiB": 1376.15234375, "Pss_MiB": 1371.6103515625, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93482.05859375, "memory.current_MiB": 1162.66796875, "memory.peak_MiB": 1164.02734375, "gpu_total_memory_util_clock_power": "11008, 94, 2715, 185.69" }, "finite": true, "input_sha256": "726cc736d46a4b8ca3c2d2831abc8c8e3a59489168194774d072d5fc1aa502b5" } ], "load_s": 15.879271634155884, "ready": { "VmHWM_MiB": 1323.16796875, "VmRSS_MiB": 1252.703125, "Pss_MiB": 1248.2080078125, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93999.734375, "memory.current_MiB": 1084.13671875, "memory.peak_MiB": 1164.02734375, "gpu_total_memory_util_clock_power": "10998, 0, 2535, 64.73" }, "providers": [ "CUDAExecutionProvider", "CPUExecutionProvider" ], "end": { "VmHWM_MiB": 1376.15625, "VmRSS_MiB": 1376.15625, "Pss_MiB": 1371.6142578125, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93481.2578125, "memory.current_MiB": 1162.8046875, "memory.peak_MiB": 1177.94140625, "gpu_total_memory_util_clock_power": "11008, 94, 2715, 186.41" } }, "online": { "arguments": { "self_check": false, "mode": "graph", "provider": "trt", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": null, "cache": "/bench/4090/online-cache", "audio": null, "length": 3, "frames": "50,100,250", "warmup": 10, "repeats": 50, "threads": 2, "profile": false, "dump_context": null, "out": "/bench/4090/graph-trt.json" }, "utc": "2026-09-05T15:50:05Z", "machine": "x86_64", "ort": "1.23.2", "numpy": "2.2.6", "model_sha256": "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994", "start": { "VmHWM_MiB": 47.5078125, "VmRSS_MiB": 46.046875, "Pss_MiB": 46.04296875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 94501.1953125, "memory.current_MiB": 28.24609375, "memory.peak_MiB": 30.24609375, "gpu_total_memory_util_clock_power": "9569, 0, 210, 26.12" }, "rows": [ { "frames": 50, "p50_ms": 5.873516434803605, "p95_ms": 6.136801093816757, "samples_ms": [ 6.086000008508563, 6.1285600531846285, 6.096978904679418, 6.127124885097146, 6.1007661279290915, 6.080055143684149, 6.0981279239058495, 6.1312338802963495, 6.1659549828618765, 6.136801093816757, 6.10750587657094, 6.1054101679474115, 8.34188680164516, 6.0909700114279985, 5.871125962585211, 6.006854120641947, 5.825095111504197, 5.981513997539878, 5.888938903808594, 5.8444528840482235, 5.808338988572359, 5.798847880214453, 5.802572006359696, 5.967573961243033, 5.967969074845314, 5.881469929590821, 5.827230168506503, 5.820384947583079, 6.060648942366242, 5.9138089418411255, 5.855967057868838, 5.838267970830202, 5.834272829815745, 5.875906907021999, 5.901786033064127, 5.809814902022481, 5.7994648814201355, 5.775895901024342, 5.787256872281432, 5.960178095847368, 5.849262000992894, 5.831021815538406, 5.809992086142302, 5.7852768804878, 5.792338866740465, 5.814115982502699, 5.819169105961919, 5.789163988083601, 5.762982880696654, 5.731857847422361 ], "memory": { "VmHWM_MiB": 9405.0859375, "VmRSS_MiB": 7529.65625, "Pss_MiB": 7525.1787109375, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 87792.97265625, "memory.current_MiB": 5441.8046875, "memory.peak_MiB": 7331.46875, "gpu_total_memory_util_clock_power": "11106, 35, 2535, 92.88" }, "finite": true, "input_sha256": "472ba979ad46ce462cc2cceeb0d218314362876e03b63f06a136a1c19b3dbab7" }, { "frames": 100, "p50_ms": 5.829924019053578, "p95_ms": 6.171168060973287, "samples_ms": [ 6.044618086889386, 6.178267067298293, 5.87579607963562, 5.8936860878020525, 6.171168060973287, 6.147991865873337, 5.9155740309506655, 5.846730899065733, 5.8493358083069324, 5.80417481251061, 5.757197039201856, 5.953033920377493, 5.904980003833771, 5.826703039929271, 5.797013873234391, 5.761999171227217, 6.135690025985241, 5.853289971128106, 5.78029197640717, 5.785923916846514, 5.763978930190206, 5.766785936430097, 6.037751911208034, 5.83802186883986, 5.778649123385549, 5.7544170413166285, 5.7861339300870895, 6.144549930468202, 5.856813862919807, 5.788216134533286, 5.769744049757719, 5.752953002229333, 5.833144998177886, 5.784881068393588, 5.770666059106588, 5.762910004705191, 5.742010893300176, 5.736347986385226, 5.992823978886008, 5.807476118206978, 5.782525055110455, 5.779606057330966, 5.746114999055862, 5.743993213400245, 5.936675006523728, 5.884533049538732, 5.916571011766791, 6.120651960372925, 6.047774106264114, 6.286510033532977 ], "memory": { "VmHWM_MiB": 9405.0859375, "VmRSS_MiB": 7529.765625, "Pss_MiB": 7525.2841796875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 87757.6484375, "memory.current_MiB": 5441.8359375, "memory.peak_MiB": 7331.46875, "gpu_total_memory_util_clock_power": "11122, 35, 2535, 143.18" }, "finite": true, "input_sha256": "dda0d977bf96004405c6cbeac7cd57cfb466001cc340939617503e06dff71881" }, { "frames": 250, "p50_ms": 10.197574505582452, "p95_ms": 11.199873872101307, "samples_ms": [ 9.849132038652897, 10.926909046247602, 10.355641134083271, 10.685404995456338, 10.687895119190216, 9.952863911166787, 10.13056212104857, 11.016117874532938, 10.05182322114706, 9.943272918462753, 10.89326897636056, 10.709190974012017, 10.137536097317934, 11.397446971386671, 10.605904972180724, 10.288105113431811, 10.500307893380523, 10.446249041706324, 10.00997913070023, 10.337363928556442, 9.96999884955585, 9.834042051807046, 9.820891078561544, 11.573206167668104, 9.92512097582221, 10.007469914853573, 10.774897877126932, 10.153670096769929, 10.063431924208999, 9.964887984097004, 10.03708504140377, 10.048120049759746, 10.297808097675443, 9.992517996579409, 10.020439978688955, 10.858990950509906, 10.122627019882202, 10.06588013842702, 10.627809911966324, 10.299277026206255, 10.132660157978535, 11.199873872101307, 10.684789856895804, 9.950637118890882, 10.58040396310389, 10.48128493130207, 9.975515073165298, 10.127157205715775, 10.241478914394975, 10.267874924466014 ], "memory": { "VmHWM_MiB": 9405.0859375, "VmRSS_MiB": 7529.77734375, "Pss_MiB": 7525.2919921875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 87830.4296875, "memory.current_MiB": 5442.1171875, "memory.peak_MiB": 7331.46875, "gpu_total_memory_util_clock_power": "11154, 95, 2535, 209.24" }, "finite": true, "input_sha256": "726cc736d46a4b8ca3c2d2831abc8c8e3a59489168194774d072d5fc1aa502b5" } ], "load_s": 5.739633965073153, "ready": { "VmHWM_MiB": 9405.0859375, "VmRSS_MiB": 7527.40234375, "Pss_MiB": 7522.9169921875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 87743.82421875, "memory.current_MiB": 5441.26171875, "memory.peak_MiB": 7331.46875, "gpu_total_memory_util_clock_power": "11086, 60, 2535, 67.08" }, "providers": [ "TensorrtExecutionProvider", "CUDAExecutionProvider", "CPUExecutionProvider" ], "end": { "VmHWM_MiB": 9405.0859375, "VmRSS_MiB": 7529.77734375, "Pss_MiB": 7525.2919921875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 87807.05078125, "memory.current_MiB": 5442.02734375, "memory.peak_MiB": 7331.46875, "gpu_total_memory_util_clock_power": "11154, 95, 2535, 209.24" } }, "offline": { "arguments": { "self_check": false, "mode": "graph", "provider": "trt", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": null, "cache": "offline-bundle", "audio": null, "length": 3, "frames": "50,100,250", "warmup": 10, "repeats": 50, "threads": 2, "profile": false, "dump_context": null, "out": "/bench/4090/graph-offline.json" }, "utc": "2026-09-05T16:07:29Z", "machine": "x86_64", "ort": "1.23.2", "numpy": "2.2.6", "model_sha256": "1ff7f921979915c747be05b820cde5384eaa075fe7f80b1a670d660bc356aac3", "start": { "VmHWM_MiB": 46.171875, "VmRSS_MiB": 46.171875, "Pss_MiB": 46.16796875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 94413.37890625, "memory.current_MiB": 28.21875, "memory.peak_MiB": 28.25, "gpu_total_memory_util_clock_power": "9569, 0, 210, 25.55" }, "rows": [ { "frames": 50, "p50_ms": 5.901396041736007, "p95_ms": 6.223426898941398, "samples_ms": [ 6.4985009375959635, 8.082402171567082, 6.223426898941398, 6.163043901324272, 6.138547090813518, 6.125693907961249, 6.197542883455753, 6.157869007438421, 6.13187113776803, 6.14659208804369, 6.128563079982996, 6.174637004733086, 5.950530990958214, 5.904756020754576, 5.905631929636002, 5.896731046959758, 5.87701122276485, 6.090237991884351, 5.97098795697093, 5.905753001570702, 5.833066999912262, 5.8332018088549376, 5.938209127634764, 5.950902123004198, 5.908404942601919, 5.896531045436859, 5.909203086048365, 5.913108820095658, 5.898036062717438, 5.82310906611383, 5.806558066979051, 5.755770951509476, 5.7900710962712765, 5.830199923366308, 5.849949084222317, 5.850642919540405, 5.799623904749751, 5.8634409215301275, 5.976327927783132, 5.813699914142489, 5.804511019960046, 5.778578110039234, 5.796416895464063, 5.7861399836838245, 5.855737021192908, 5.827226908877492, 5.783218890428543, 5.7536461390554905, 5.770208081230521, 5.915169138461351 ], "memory": { "VmHWM_MiB": 1559.69921875, "VmRSS_MiB": 946.8046875, "Pss_MiB": 942.3271484375, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 94234.0234375, "memory.current_MiB": 569.22265625, "memory.peak_MiB": 1194.1328125, "gpu_total_memory_util_clock_power": "11100, 1, 2535, 92.89" }, "finite": true, "input_sha256": "472ba979ad46ce462cc2cceeb0d218314362876e03b63f06a136a1c19b3dbab7" }, { "frames": 100, "p50_ms": 5.793152027763426, "p95_ms": 6.011463003233075, "samples_ms": [ 6.011463003233075, 8.266900200396776, 5.789669929072261, 5.818888079375029, 5.876495968550444, 6.049096817150712, 5.797808989882469, 5.7691780384629965, 5.735898157581687, 5.720118992030621, 5.752597004175186, 5.930606042966247, 5.8508929796516895, 5.852136993780732, 5.813412135466933, 5.809849128127098, 5.977617111057043, 5.780626088380814, 5.833730800077319, 5.781392101198435, 5.827598040923476, 5.794598953798413, 5.766500951722264, 5.771917989477515, 5.753468023613095, 5.711003206670284, 5.7506191078573465, 5.942365154623985, 5.791705101728439, 5.756594939157367, 5.734442034736276, 5.724290851503611, 5.802405998110771, 5.9143968392163515, 5.805246066302061, 5.79790910705924, 5.802742904052138, 5.815773969516158, 5.986895877867937, 5.788308102637529, 5.760642001405358, 5.743528949096799, 5.722452886402607, 5.764899076893926, 5.772707983851433, 5.755288992077112, 5.756276194006205, 5.761770065873861, 5.814885953441262, 5.984595976769924 ], "memory": { "VmHWM_MiB": 1559.69921875, "VmRSS_MiB": 969.4765625, "Pss_MiB": 964.9990234375, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 94197.984375, "memory.current_MiB": 592.26953125, "memory.peak_MiB": 1194.1328125, "gpu_total_memory_util_clock_power": "11116, 66, 2535, 140.12" }, "finite": true, "input_sha256": "dda0d977bf96004405c6cbeac7cd57cfb466001cc340939617503e06dff71881" }, { "frames": 250, "p50_ms": 10.060626547783613, "p95_ms": 10.72664582170546, "samples_ms": [ 9.775890968739986, 13.440693961456418, 9.827678091824055, 9.78352501988411, 10.289052966982126, 10.72664582170546, 9.816800942644477, 10.855173924937844, 10.422261198982596, 10.560437105596066, 10.642173932865262, 9.978265035897493, 9.896415984258056, 10.26989403180778, 10.02198294736445, 9.905637940391898, 10.021117981523275, 10.19646693021059, 10.085739893838763, 10.268521960824728, 10.195394977927208, 9.952198946848512, 9.920611046254635, 10.569974081590772, 10.06051292642951, 10.267843026667833, 10.304447961971164, 9.864755906164646, 9.799650171771646, 10.29696292243898, 9.842331055551767, 9.815114084631205, 10.02031797543168, 9.800008032470942, 9.943492012098432, 10.00177999958396, 10.42952504940331, 10.022844886407256, 10.409324895590544, 10.563523974269629, 10.033812839537859, 10.04157797433436, 10.535720968618989, 9.978103917092085, 9.929524036124349, 10.285760974511504, 10.13919711112976, 10.060740169137716, 10.193349095061421, 10.168188950046897 ], "memory": { "VmHWM_MiB": 1559.69921875, "VmRSS_MiB": 1011.703125, "Pss_MiB": 1007.2177734375, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93638.55078125, "memory.current_MiB": 634.31640625, "memory.peak_MiB": 1194.1328125, "gpu_total_memory_util_clock_power": "11148, 66, 2535, 209.44" }, "finite": true, "input_sha256": "726cc736d46a4b8ca3c2d2831abc8c8e3a59489168194774d072d5fc1aa502b5" } ], "load_s": 1.5077310029882938, "ready": { "VmHWM_MiB": 1559.69921875, "VmRSS_MiB": 922.22265625, "Pss_MiB": 917.7373046875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 94251.46875, "memory.current_MiB": 547.078125, "memory.peak_MiB": 1194.1328125, "gpu_total_memory_util_clock_power": "11080, 1, 2535, 66.97" }, "providers": [ "TensorrtExecutionProvider", "CUDAExecutionProvider", "CPUExecutionProvider" ], "end": { "VmHWM_MiB": 1559.69921875, "VmRSS_MiB": 1011.703125, "Pss_MiB": 1007.2177734375, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93509.53515625, "memory.current_MiB": 634.37109375, "memory.peak_MiB": 1194.1328125, "gpu_total_memory_util_clock_power": "11148, 66, 2535, 211.06" } } }, "provider_proof": { "cuda": { "arguments": { "self_check": false, "mode": "graph", "provider": "cuda", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": null, "cache": "/bench/4090/online-cache", "audio": null, "length": 3, "frames": "50", "warmup": 0, "repeats": 1, "threads": 2, "profile": true, "dump_context": null, "out": "/bench/4090/proof-cuda.json" }, "utc": "2026-09-05T16:07:49Z", "machine": "x86_64", "ort": "1.23.2", "numpy": "2.2.6", "model_sha256": "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994", "start": { "VmHWM_MiB": 47.32421875, "VmRSS_MiB": 45.81640625, "Pss_MiB": 45.8125, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 94404.7421875, "memory.current_MiB": 28.44140625, "memory.peak_MiB": 30.39453125, "gpu_total_memory_util_clock_power": "9569, 0, 210, 27.34" }, "rows": [ { "frames": 50, "p50_ms": 103.25849894434214, "p95_ms": 103.25849894434214, "samples_ms": [ 103.25849894434214 ], "memory": { "VmHWM_MiB": 1333.08984375, "VmRSS_MiB": 1291.1171875, "Pss_MiB": 1286.5791015625, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93148.26171875, "memory.current_MiB": 1068.02734375, "memory.peak_MiB": 1165.91015625, "gpu_total_memory_util_clock_power": "11006, 0, 2535, 52.40" }, "finite": true, "input_sha256": "472ba979ad46ce462cc2cceeb0d218314362876e03b63f06a136a1c19b3dbab7" } ], "load_s": 1.2709995079785585, "ready": { "VmHWM_MiB": 1333.08984375, "VmRSS_MiB": 1235.96484375, "Pss_MiB": 1231.4697265625, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93182.55859375, "memory.current_MiB": 1061.28125, "memory.peak_MiB": 1165.91015625, "gpu_total_memory_util_clock_power": "10996, 0, 2535, 51.10" }, "providers": [ "CUDAExecutionProvider", "CPUExecutionProvider" ], "node_provider_counts": { "CUDAExecutionProvider": 2908, "CPUExecutionProvider": 785 }, "end": { "VmHWM_MiB": 1333.08984375, "VmRSS_MiB": 1304.578125, "Pss_MiB": 1300.0439453125, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93132.3203125, "memory.current_MiB": 1083.8984375, "memory.peak_MiB": 1165.91015625, "gpu_total_memory_util_clock_power": "11006, 0, 2535, 52.40" } }, "online": { "arguments": { "self_check": false, "mode": "graph", "provider": "trt", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": null, "cache": "/bench/4090/online-cache", "audio": null, "length": 3, "frames": "50", "warmup": 0, "repeats": 1, "threads": 2, "profile": true, "dump_context": null, "out": "/bench/4090/proof-online.json" }, "utc": "2026-09-05T16:07:52Z", "machine": "x86_64", "ort": "1.23.2", "numpy": "2.2.6", "model_sha256": "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994", "start": { "VmHWM_MiB": 46.9921875, "VmRSS_MiB": 45.8359375, "Pss_MiB": 45.83203125, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93907.1328125, "memory.current_MiB": 28.2578125, "memory.peak_MiB": 30.21875, "gpu_total_memory_util_clock_power": "9569, 0, 2535, 67.70" }, "rows": [ { "frames": 50, "p50_ms": 35.45449813827872, "p95_ms": 35.45449813827872, "samples_ms": [ 35.45449813827872 ], "memory": { "VmHWM_MiB": 9405.22265625, "VmRSS_MiB": 7529.5546875, "Pss_MiB": 7525.0732421875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 88793.27734375, "memory.current_MiB": 5441.625, "memory.peak_MiB": 7331.55078125, "gpu_total_memory_util_clock_power": "11106, 3, 2535, 67.96" }, "finite": true, "input_sha256": "472ba979ad46ce462cc2cceeb0d218314362876e03b63f06a136a1c19b3dbab7" } ], "load_s": 5.647201908985153, "ready": { "VmHWM_MiB": 9405.22265625, "VmRSS_MiB": 7527.265625, "Pss_MiB": 7522.7802734375, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 88668.21875, "memory.current_MiB": 5441.27734375, "memory.peak_MiB": 7331.55078125, "gpu_total_memory_util_clock_power": "11086, 60, 2535, 67.19" }, "providers": [ "TensorrtExecutionProvider", "CUDAExecutionProvider", "CPUExecutionProvider" ], "node_provider_counts": { "TensorrtExecutionProvider": 1 }, "end": { "VmHWM_MiB": 9405.22265625, "VmRSS_MiB": 7529.62109375, "Pss_MiB": 7525.1357421875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 88769.921875, "memory.current_MiB": 5441.67578125, "memory.peak_MiB": 7331.55078125, "gpu_total_memory_util_clock_power": "11106, 3, 2535, 67.81" } }, "offline": { "arguments": { "self_check": false, "mode": "graph", "provider": "trt", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": null, "cache": "/app/offline-bundle", "audio": null, "length": 3, "frames": "50", "warmup": 0, "repeats": 1, "threads": 2, "profile": true, "dump_context": null, "out": "/bench/4090/proof-offline.json" }, "utc": "2026-09-05T16:08:00Z", "machine": "x86_64", "ort": "1.23.2", "numpy": "2.2.6", "model_sha256": "1ff7f921979915c747be05b820cde5384eaa075fe7f80b1a670d660bc356aac3", "start": { "VmHWM_MiB": 46.1484375, "VmRSS_MiB": 46.1484375, "Pss_MiB": 46.14453125, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93438.6015625, "memory.current_MiB": 28.28125, "memory.peak_MiB": 28.3125, "gpu_total_memory_util_clock_power": "9569, 0, 2535, 66.11" }, "rows": [ { "frames": 50, "p50_ms": 36.38270800001919, "p95_ms": 36.38270800001919, "samples_ms": [ 36.38270800001919 ], "memory": { "VmHWM_MiB": 1558.4765625, "VmRSS_MiB": 941.3046875, "Pss_MiB": 936.8232421875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93371.03515625, "memory.current_MiB": 564.17578125, "memory.peak_MiB": 1194.21875, "gpu_total_memory_util_clock_power": "11100, 1, 2535, 69.18" }, "finite": true, "input_sha256": "472ba979ad46ce462cc2cceeb0d218314362876e03b63f06a136a1c19b3dbab7" } ], "load_s": 1.5445389528758824, "ready": { "VmHWM_MiB": 1558.4765625, "VmRSS_MiB": 922.14453125, "Pss_MiB": 917.6591796875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93253.8671875, "memory.current_MiB": 546.96875, "memory.peak_MiB": 1194.21875, "gpu_total_memory_util_clock_power": "11080, 1, 2535, 69.18" }, "providers": [ "TensorrtExecutionProvider", "CUDAExecutionProvider", "CPUExecutionProvider" ], "node_provider_counts": { "TensorrtExecutionProvider": 1 }, "end": { "VmHWM_MiB": 1558.4765625, "VmRSS_MiB": 942.67578125, "Pss_MiB": 938.1904296875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93389.84765625, "memory.current_MiB": 565.26953125, "memory.peak_MiB": 1194.21875, "gpu_total_memory_util_clock_power": "11100, 1, 2535, 69.26" } } }, "recognizer": { "cuda": { "3": { "arguments": { "self_check": false, "mode": "recognizer", "provider": "cuda", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/tokens.txt", "cache": "/bench/4090/online-cache", "audio": "/bench/audio", "length": 3, "frames": "50,100,250", "warmup": 10, "repeats": 3, "threads": 2, "profile": false, "dump_context": null, "out": "/bench/4090/recognizer-cuda-3s.json" }, "utc": "2026-09-05T15:46:40Z", "machine": "x86_64", "ort": "1.23.2", "numpy": "2.2.6", "model_sha256": "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994", "start": { "VmHWM_MiB": 46.8046875, "VmRSS_MiB": 45.78125, "Pss_MiB": 45.77734375, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 95004.640625, "memory.current_MiB": 47.0625, "memory.peak_MiB": 49.14453125, "gpu_total_memory_util_clock_power": "9569, 0, 210, 26.26" }, "rows": [ { "seconds": 3, "p50_ms": 12.386747170239687, "p95_ms": 12.596443993970752, "samples_ms": [ 12.596443993970752, 12.386747170239687, 12.301173992455006 ], "texts": [ "帮我导航到内湖科学园区，然后。", "帮我导航到内湖科学园区，然后。", "帮我导航到内湖科学园区，然后。" ], "memory": { "VmHWM_MiB": 1351.671875, "VmRSS_MiB": 1342.87109375, "Pss_MiB": 1338.3291015625, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93876.23046875, "memory.current_MiB": 1298.73828125, "memory.peak_MiB": 1314.2890625, "gpu_total_memory_util_clock_power": "11008, 33, 2535, 67.30" }, "audio_sha256": "11034b66e2ae87bdec98dfe7401642387b85262b2e9c222b9e201030dc397a1c" } ], "load_s": 3.0649817681405693, "ready": { "VmHWM_MiB": 1351.671875, "VmRSS_MiB": 1289.90234375, "Pss_MiB": 1285.4072265625, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 94200.78125, "memory.current_MiB": 1231.68359375, "memory.peak_MiB": 1297.125, "gpu_total_memory_util_clock_power": "10998, 0, 2535, 65.08" }, "after_warmup": { "VmHWM_MiB": 1351.671875, "VmRSS_MiB": 1342.87109375, "Pss_MiB": 1338.3291015625, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93903.91796875, "memory.current_MiB": 1299.0546875, "memory.peak_MiB": 1299.29296875, "gpu_total_memory_util_clock_power": "11008, 0, 2535, 65.54" }, "end": { "VmHWM_MiB": 1351.671875, "VmRSS_MiB": 1342.87109375, "Pss_MiB": 1338.3291015625, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93860.734375, "memory.current_MiB": 1298.55859375, "memory.peak_MiB": 1314.2890625, "gpu_total_memory_util_clock_power": "11008, 33, 2535, 67.30" } }, "6": { "arguments": { "self_check": false, "mode": "recognizer", "provider": "cuda", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/tokens.txt", "cache": "/bench/4090/online-cache", "audio": "/bench/audio", "length": 6, "frames": "50,100,250", "warmup": 10, "repeats": 3, "threads": 2, "profile": false, "dump_context": null, "out": "/bench/4090/recognizer-cuda-6s.json" }, "utc": "2026-09-05T15:46:47Z", "machine": "x86_64", "ort": "1.23.2", "numpy": "2.2.6", "model_sha256": "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994", "start": { "VmHWM_MiB": 47.01953125, "VmRSS_MiB": 45.828125, "Pss_MiB": 45.82421875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 94453.93359375, "memory.current_MiB": 28.3046875, "memory.peak_MiB": 30.26171875, "gpu_total_memory_util_clock_power": "9569, 0, 2535, 64.85" }, "rows": [ { "seconds": 6, "p50_ms": 16.97598583996296, "p95_ms": 24.54869798384607, "samples_ms": [ 24.54869798384607, 16.97598583996296, 16.853041015565395 ], "texts": [ "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥。", "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥。", "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥。" ], "memory": { "VmHWM_MiB": 1354.2265625, "VmRSS_MiB": 1319.7578125, "Pss_MiB": 1315.2158203125, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93211.39453125, "memory.current_MiB": 1086.046875, "memory.peak_MiB": 1175.34765625, "gpu_total_memory_util_clock_power": "11008, 0, 2535, 67.24" }, "audio_sha256": "9566be6ccf6c7307830514e0fe5be86250c493efaef2800f27bfccc1ca1850b0" } ], "load_s": 1.2816854468546808, "ready": { "VmHWM_MiB": 1354.2265625, "VmRSS_MiB": 1266.4765625, "Pss_MiB": 1261.9814453125, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93214.42578125, "memory.current_MiB": 1079.37890625, "memory.peak_MiB": 1175.34765625, "gpu_total_memory_util_clock_power": "10998, 0, 2535, 65.32" }, "after_warmup": { "VmHWM_MiB": 1354.2265625, "VmRSS_MiB": 1319.6796875, "Pss_MiB": 1315.1376953125, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93216.7265625, "memory.current_MiB": 1086.4140625, "memory.peak_MiB": 1175.34765625, "gpu_total_memory_util_clock_power": "11008, 0, 2535, 65.55" }, "end": { "VmHWM_MiB": 1354.2265625, "VmRSS_MiB": 1319.7578125, "Pss_MiB": 1315.2158203125, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93186.29296875, "memory.current_MiB": 1085.97265625, "memory.peak_MiB": 1175.34765625, "gpu_total_memory_util_clock_power": "11008, 0, 2535, 67.24" } }, "15": { "arguments": { "self_check": false, "mode": "recognizer", "provider": "cuda", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/tokens.txt", "cache": "/bench/4090/online-cache", "audio": "/bench/audio", "length": 15, "frames": "50,100,250", "warmup": 10, "repeats": 3, "threads": 2, "profile": false, "dump_context": null, "out": "/bench/4090/recognizer-cuda-15s.json" }, "utc": "2026-09-05T15:46:49Z", "machine": "x86_64", "ort": "1.23.2", "numpy": "2.2.6", "model_sha256": "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994", "start": { "VmHWM_MiB": 47.53125, "VmRSS_MiB": 45.78125, "Pss_MiB": 45.77734375, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 94655.12109375, "memory.current_MiB": 28.3046875, "memory.peak_MiB": 30.89453125, "gpu_total_memory_util_clock_power": "9569, 0, 2535, 65.27" }, "rows": [ { "seconds": 15, "p50_ms": 29.035150073468685, "p95_ms": 36.99248400516808, "samples_ms": [ 36.99248400516808, 29.035150073468685, 29.002567986026406 ], "texts": [ "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥，接着把冷气打开，温度调到24度，风量开大。帮我导航到内湖科学园区，然后选最快的路线。不要走。", "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥，接着把冷气打开，温度调到24度，风量开大。帮我导航到内湖科学园区，然后选最快的路线。不要走。", "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥，接着把冷气打开，温度调到24度，风量开大。帮我导航到内湖科学园区，然后选最快的路线。不要走。" ], "memory": { "VmHWM_MiB": 1352.10546875, "VmRSS_MiB": 1245.0234375, "Pss_MiB": 1240.4814453125, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93686.58984375, "memory.current_MiB": 1010.69140625, "memory.peak_MiB": 1173.48828125, "gpu_total_memory_util_clock_power": "11008, 26, 2535, 68.30" }, "audio_sha256": "4950a411664da365f4b8cfe13838f881430f01f2fed8b2c3335719f6b791ff9f" } ], "load_s": 1.3584592249244452, "ready": { "VmHWM_MiB": 1352.10546875, "VmRSS_MiB": 1191.58203125, "Pss_MiB": 1187.0869140625, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93699.828125, "memory.current_MiB": 1004.3671875, "memory.peak_MiB": 1173.48828125, "gpu_total_memory_util_clock_power": "10998, 0, 2535, 65.05" }, "after_warmup": { "VmHWM_MiB": 1352.10546875, "VmRSS_MiB": 1244.9453125, "Pss_MiB": 1240.4033203125, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93687.5703125, "memory.current_MiB": 1011.06640625, "memory.peak_MiB": 1173.48828125, "gpu_total_memory_util_clock_power": "11008, 26, 2535, 66.04" }, "end": { "VmHWM_MiB": 1352.10546875, "VmRSS_MiB": 1245.0234375, "Pss_MiB": 1240.4814453125, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93656.83984375, "memory.current_MiB": 1010.91796875, "memory.peak_MiB": 1173.48828125, "gpu_total_memory_util_clock_power": "11008, 26, 2535, 68.30" } } }, "online": { "3": { "arguments": { "self_check": false, "mode": "recognizer", "provider": "trt", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/tokens.txt", "cache": "/bench/4090/online-cache", "audio": "/bench/audio", "length": 3, "frames": "50,100,250", "warmup": 10, "repeats": 3, "threads": 2, "profile": false, "dump_context": null, "out": "/bench/4090/recognizer-trt-3s.json" }, "utc": "2026-09-05T15:47:49Z", "machine": "x86_64", "ort": "1.23.2", "numpy": "2.2.6", "model_sha256": "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994", "start": { "VmHWM_MiB": 47.0234375, "VmRSS_MiB": 45.81640625, "Pss_MiB": 45.8125, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 94387.1640625, "memory.current_MiB": 28.2734375, "memory.peak_MiB": 30.49609375, "gpu_total_memory_util_clock_power": "9569, 0, 2715, 82.74" }, "rows": [ { "seconds": 3, "p50_ms": 10.011387057602406, "p95_ms": 10.148542001843452, "samples_ms": [ 10.011387057602406, 9.828076930716634, 10.148542001843452 ], "texts": [ "帮我导航到内湖科学园区，然后。", "帮我导航到内湖科学园区，然后。", "帮我导航到内湖科学园区，然后。" ], "memory": { "VmHWM_MiB": 10902.328125, "VmRSS_MiB": 8965.25390625, "Pss_MiB": 8960.7314453125, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 88376.92578125, "memory.current_MiB": 7876.56640625, "memory.peak_MiB": 9766.52734375, "gpu_total_memory_util_clock_power": "11494, 6, 2535, 67.39" }, "audio_sha256": "11034b66e2ae87bdec98dfe7401642387b85262b2e9c222b9e201030dc397a1c" } ], "load_s": 40.634328824002296, "ready": { "VmHWM_MiB": 10902.328125, "VmRSS_MiB": 8955.25, "Pss_MiB": 8950.7275390625, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 88027.75390625, "memory.current_MiB": 7874.390625, "memory.peak_MiB": 9766.52734375, "gpu_total_memory_util_clock_power": "11474, 1, 2535, 66.75" }, "after_warmup": { "VmHWM_MiB": 10902.328125, "VmRSS_MiB": 8965.25390625, "Pss_MiB": 8960.7314453125, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 88383.3046875, "memory.current_MiB": 7876.51953125, "memory.peak_MiB": 9766.52734375, "gpu_total_memory_util_clock_power": "11494, 1, 2535, 66.21" }, "end": { "VmHWM_MiB": 10902.328125, "VmRSS_MiB": 8965.25390625, "Pss_MiB": 8960.7314453125, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 88274.796875, "memory.current_MiB": 7876.58984375, "memory.peak_MiB": 9766.52734375, "gpu_total_memory_util_clock_power": "11494, 6, 2535, 67.33" } }, "6": { "arguments": { "self_check": false, "mode": "recognizer", "provider": "trt", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/tokens.txt", "cache": "/bench/4090/online-cache", "audio": "/bench/audio", "length": 6, "frames": "50,100,250", "warmup": 10, "repeats": 3, "threads": 2, "profile": false, "dump_context": null, "out": "/bench/4090/recognizer-trt-6s.json" }, "utc": "2026-09-05T15:48:54Z", "machine": "x86_64", "ort": "1.23.2", "numpy": "2.2.6", "model_sha256": "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994", "start": { "VmHWM_MiB": 46.46875, "VmRSS_MiB": 45.96875, "Pss_MiB": 45.96484375, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 94829.375, "memory.current_MiB": 30.86328125, "memory.peak_MiB": 32.796875, "gpu_total_memory_util_clock_power": "9569, 0, 210, 26.40" }, "rows": [ { "seconds": 6, "p50_ms": 13.902554986998439, "p95_ms": 15.958416974171996, "samples_ms": [ 15.958416974171996, 13.902554986998439, 13.855116907507181 ], "texts": [ "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥。", "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥。", "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥。" ], "memory": { "VmHWM_MiB": 9422.64453125, "VmRSS_MiB": 7549.97265625, "Pss_MiB": 7545.4873046875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 88940.41796875, "memory.current_MiB": 5465.13671875, "memory.peak_MiB": 7355.37890625, "gpu_total_memory_util_clock_power": "11112, 2, 2535, 70.99" }, "audio_sha256": "9566be6ccf6c7307830514e0fe5be86250c493efaef2800f27bfccc1ca1850b0" } ], "load_s": 7.0790270131547, "ready": { "VmHWM_MiB": 9422.64453125, "VmRSS_MiB": 7539.70703125, "Pss_MiB": 7535.2216796875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 88922.98828125, "memory.current_MiB": 5462.72265625, "memory.peak_MiB": 7355.37890625, "gpu_total_memory_util_clock_power": "11076, 1, 2535, 68.98" }, "after_warmup": { "VmHWM_MiB": 9422.64453125, "VmRSS_MiB": 7549.97265625, "Pss_MiB": 7545.4873046875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 88919.08984375, "memory.current_MiB": 5465.13671875, "memory.peak_MiB": 7355.37890625, "gpu_total_memory_util_clock_power": "11096, 1, 2535, 69.09" }, "end": { "VmHWM_MiB": 9422.64453125, "VmRSS_MiB": 7549.97265625, "Pss_MiB": 7545.4873046875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 88937.95703125, "memory.current_MiB": 5465.39453125, "memory.peak_MiB": 7355.37890625, "gpu_total_memory_util_clock_power": "11112, 2, 2535, 71.03" } }, "15": { "arguments": { "self_check": false, "mode": "recognizer", "provider": "trt", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/tokens.txt", "cache": "/bench/4090/online-cache", "audio": "/bench/audio", "length": 15, "frames": "50,100,250", "warmup": 10, "repeats": 3, "threads": 2, "profile": false, "dump_context": null, "out": "/bench/4090/recognizer-trt-15s.json" }, "utc": "2026-09-05T15:49:48Z", "machine": "x86_64", "ort": "1.23.2", "numpy": "2.2.6", "model_sha256": "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994", "start": { "VmHWM_MiB": 47.7265625, "VmRSS_MiB": 46.00390625, "Pss_MiB": 46.0, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 94976.0546875, "memory.current_MiB": 28.03125, "memory.peak_MiB": 30.171875, "gpu_total_memory_util_clock_power": "9569, 0, 210, 26.30" }, "rows": [ { "seconds": 15, "p50_ms": 28.323370032012463, "p95_ms": 28.837248915806413, "samples_ms": [ 28.837248915806413, 28.261404018849134, 28.323370032012463 ], "texts": [ "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥，接着把冷气打开，温度调到24度，风量开大。帮我导航到内湖科学园区，然后选最快的路线。不要走。", "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥，接着把冷气打开，温度调到24度，风量开大。帮我导航到内湖科学园区，然后选最快的路线。不要走。", "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥，接着把冷气打开，温度调到24度，风量开大。帮我导航到内湖科学园区，然后选最快的路线。不要走。" ], "memory": { "VmHWM_MiB": 9423.36328125, "VmRSS_MiB": 7549.69140625, "Pss_MiB": 7545.2060546875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 88776.84765625, "memory.current_MiB": 5451.30859375, "memory.peak_MiB": 7341.2734375, "gpu_total_memory_util_clock_power": "11128, 1, 2535, 72.14" }, "audio_sha256": "4950a411664da365f4b8cfe13838f881430f01f2fed8b2c3335719f6b791ff9f" } ], "load_s": 5.682397730182856, "ready": { "VmHWM_MiB": 9423.36328125, "VmRSS_MiB": 7539.609375, "Pss_MiB": 7535.1240234375, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 88764.671875, "memory.current_MiB": 5448.953125, "memory.peak_MiB": 7341.2734375, "gpu_total_memory_util_clock_power": "11076, 1, 2535, 67.75" }, "after_warmup": { "VmHWM_MiB": 9423.36328125, "VmRSS_MiB": 7549.69140625, "Pss_MiB": 7545.2060546875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 88764.7109375, "memory.current_MiB": 5451.1875, "memory.peak_MiB": 7341.2734375, "gpu_total_memory_util_clock_power": "11096, 1, 2535, 69.31" }, "end": { "VmHWM_MiB": 9423.36328125, "VmRSS_MiB": 7549.69140625, "Pss_MiB": 7545.2060546875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 88761.34375, "memory.current_MiB": 5451.03125, "memory.peak_MiB": 7341.2734375, "gpu_total_memory_util_clock_power": "11128, 1, 2535, 72.14" } } }, "offline": { "3": { "arguments": { "self_check": false, "mode": "recognizer", "provider": "trt", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/tokens.txt", "cache": "offline-bundle", "audio": "/bench/audio", "length": 3, "frames": "50,100,250", "warmup": 10, "repeats": 3, "threads": 2, "profile": false, "dump_context": null, "out": "/bench/4090/recognizer-offline-3s.json" }, "utc": "2026-09-05T15:50:55Z", "machine": "x86_64", "ort": "1.23.2", "numpy": "2.2.6", "model_sha256": "1ff7f921979915c747be05b820cde5384eaa075fe7f80b1a670d660bc356aac3", "start": { "VmHWM_MiB": 45.9453125, "VmRSS_MiB": 45.9453125, "Pss_MiB": 45.94140625, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 94504.5, "memory.current_MiB": 28.2265625, "memory.peak_MiB": 28.2578125, "gpu_total_memory_util_clock_power": "9569, 0, 210, 27.76" }, "rows": [ { "seconds": 3, "p50_ms": 9.80040617287159, "p95_ms": 10.120285907760262, "samples_ms": [ 10.120285907760262, 9.80040617287159, 9.699732065200806 ], "texts": [ "帮我导航到内湖科学园区，然后。", "帮我导航到内湖科学园区，然后。", "帮我导航到内湖科学园区，然后。" ], "memory": { "VmHWM_MiB": 1579.0390625, "VmRSS_MiB": 963.53515625, "Pss_MiB": 959.0498046875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 94141.9140625, "memory.current_MiB": 575.5625, "memory.peak_MiB": 1204.09765625, "gpu_total_memory_util_clock_power": "11090, 1, 2535, 70.19" }, "audio_sha256": "11034b66e2ae87bdec98dfe7401642387b85262b2e9c222b9e201030dc397a1c" } ], "load_s": 1.4903482249937952, "ready": { "VmHWM_MiB": 1579.0390625, "VmRSS_MiB": 930.15625, "Pss_MiB": 925.6708984375, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93826.04296875, "memory.current_MiB": 549.76953125, "memory.peak_MiB": 1204.09765625, "gpu_total_memory_util_clock_power": "11070, 1, 2535, 69.47" }, "after_warmup": { "VmHWM_MiB": 1579.0390625, "VmRSS_MiB": 962.98828125, "Pss_MiB": 958.5029296875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93951.40625, "memory.current_MiB": 574.87109375, "memory.peak_MiB": 1204.09765625, "gpu_total_memory_util_clock_power": "11090, 1, 2535, 69.47" }, "end": { "VmHWM_MiB": 1579.0390625, "VmRSS_MiB": 963.54296875, "Pss_MiB": 959.0576171875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 94125.328125, "memory.current_MiB": 575.68359375, "memory.peak_MiB": 1204.09765625, "gpu_total_memory_util_clock_power": "11090, 1, 2535, 70.19" } }, "6": { "arguments": { "self_check": false, "mode": "recognizer", "provider": "trt", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/tokens.txt", "cache": "offline-bundle", "audio": "/bench/audio", "length": 6, "frames": "50,100,250", "warmup": 10, "repeats": 3, "threads": 2, "profile": false, "dump_context": null, "out": "/bench/4090/recognizer-offline-6s.json" }, "utc": "2026-09-05T15:50:58Z", "machine": "x86_64", "ort": "1.23.2", "numpy": "2.2.6", "model_sha256": "1ff7f921979915c747be05b820cde5384eaa075fe7f80b1a670d660bc356aac3", "start": { "VmHWM_MiB": 45.9609375, "VmRSS_MiB": 45.9609375, "Pss_MiB": 45.95703125, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 94758.8828125, "memory.current_MiB": 28.51171875, "memory.peak_MiB": 28.54296875, "gpu_total_memory_util_clock_power": "9569, 0, 2535, 70.73" }, "rows": [ { "seconds": 6, "p50_ms": 14.215043978765607, "p95_ms": 22.356424015015364, "samples_ms": [ 22.356424015015364, 14.200367964804173, 14.215043978765607 ], "texts": [ "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥。", "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥。", "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥。" ], "memory": { "VmHWM_MiB": 1579.6796875, "VmRSS_MiB": 975.51953125, "Pss_MiB": 971.0341796875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93668.9609375, "memory.current_MiB": 587.55859375, "memory.peak_MiB": 1203.8515625, "gpu_total_memory_util_clock_power": "11106, 13, 2535, 70.37" }, "audio_sha256": "9566be6ccf6c7307830514e0fe5be86250c493efaef2800f27bfccc1ca1850b0" } ], "load_s": 1.4986548479646444, "ready": { "VmHWM_MiB": 1579.6796875, "VmRSS_MiB": 930.171875, "Pss_MiB": 925.6865234375, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93711.99609375, "memory.current_MiB": 549.65234375, "memory.peak_MiB": 1203.8515625, "gpu_total_memory_util_clock_power": "11070, 27, 2535, 69.38" }, "after_warmup": { "VmHWM_MiB": 1579.6796875, "VmRSS_MiB": 962.99609375, "Pss_MiB": 958.5107421875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93686.1875, "memory.current_MiB": 574.7578125, "memory.peak_MiB": 1203.8515625, "gpu_total_memory_util_clock_power": "11090, 27, 2535, 69.59" }, "end": { "VmHWM_MiB": 1579.6796875, "VmRSS_MiB": 975.5234375, "Pss_MiB": 971.0380859375, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93684.234375, "memory.current_MiB": 587.4609375, "memory.peak_MiB": 1203.8515625, "gpu_total_memory_util_clock_power": "11106, 13, 2535, 70.37" } }, "15": { "arguments": { "self_check": false, "mode": "recognizer", "provider": "trt", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/tokens.txt", "cache": "offline-bundle", "audio": "/bench/audio", "length": 15, "frames": "50,100,250", "warmup": 10, "repeats": 3, "threads": 2, "profile": false, "dump_context": null, "out": "/bench/4090/recognizer-offline-15s.json" }, "utc": "2026-09-05T15:51:36Z", "machine": "x86_64", "ort": "1.23.2", "numpy": "2.2.6", "model_sha256": "1ff7f921979915c747be05b820cde5384eaa075fe7f80b1a670d660bc356aac3", "start": { "VmHWM_MiB": 46.09375, "VmRSS_MiB": 46.09375, "Pss_MiB": 46.08984375, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 94469.13671875, "memory.current_MiB": 28.203125, "memory.peak_MiB": 28.234375, "gpu_total_memory_util_clock_power": "9569, 0, 210, 24.20" }, "rows": [ { "seconds": 15, "p50_ms": 29.86208302900195, "p95_ms": 36.1810689792037, "samples_ms": [ 36.1810689792037, 29.86208302900195, 28.097413014620543 ], "texts": [ "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥，接着把冷气打开，温度调到24度，风量开大。帮我导航到内湖科学园区，然后选最快的路线。不要走。", "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥，接着把冷气打开，温度调到24度，风量开大。帮我导航到内湖科学园区，然后选最快的路线。不要走。", "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥，接着把冷气打开，温度调到24度，风量开大。帮我导航到内湖科学园区，然后选最快的路线。不要走。" ], "memory": { "VmHWM_MiB": 1578.4609375, "VmRSS_MiB": 993.6875, "Pss_MiB": 989.2021484375, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93979.6796875, "memory.current_MiB": 605.65625, "memory.peak_MiB": 1203.75390625, "gpu_total_memory_util_clock_power": "11122, 6, 2535, 68.54" }, "audio_sha256": "4950a411664da365f4b8cfe13838f881430f01f2fed8b2c3335719f6b791ff9f" } ], "load_s": 1.602380754193291, "ready": { "VmHWM_MiB": 1578.4609375, "VmRSS_MiB": 930.05078125, "Pss_MiB": 925.5654296875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 94208.08984375, "memory.current_MiB": 549.8984375, "memory.peak_MiB": 1203.75390625, "gpu_total_memory_util_clock_power": "11070, 1, 2535, 64.69" }, "after_warmup": { "VmHWM_MiB": 1578.4609375, "VmRSS_MiB": 962.90625, "Pss_MiB": 958.4208984375, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 94014.47265625, "memory.current_MiB": 574.71484375, "memory.peak_MiB": 1203.75390625, "gpu_total_memory_util_clock_power": "11090, 1, 2535, 64.69" }, "end": { "VmHWM_MiB": 1578.4609375, "VmRSS_MiB": 993.6875, "Pss_MiB": 989.2021484375, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93968.8515625, "memory.current_MiB": 605.6953125, "memory.peak_MiB": 1203.75390625, "gpu_total_memory_util_clock_power": "11122, 6, 2535, 68.54" } } } }, "service": { "cuda": { "rounds": [ { "environment": { "target": "ws://127.0.0.1:19242/stt/sensevoice", "uri": "ws://127.0.0.1:19242/stt/sensevoice", "warmup_runs": 3, "single_connection_series": true }, "generated": "2026-09-05T15:56:03", "results": [ { "feed": "bulk", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 25.69, "server_rtf": 0.01, "server_ttft_ms": 22.0, "client_final_latency_ms": 32.02, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.02, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "bulk", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 30.88, "server_rtf": 0.007, "server_ttft_ms": 134.88, "client_final_latency_ms": 150.05, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.05, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "bulk", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 54.1, "server_rtf": 0.004, "server_ttft_ms": 24.9, "client_final_latency_ms": 60.22, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.13, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開。幫我導航到內湖科學園區，然後選最快的路線，不要走。", "error": null }, { "feed": "realtime", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 36.53, "server_rtf": 0.043, "server_ttft_ms": 21.43, "client_final_latency_ms": 39.06, "interim_count": 7, "max_interim_gap_s": 0.518, "feed_secs": 3.15, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "realtime", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 22.24, "server_rtf": 0.051, "server_ttft_ms": 32.27, "client_final_latency_ms": 23.92, "interim_count": 11, "max_interim_gap_s": 1.051, "feed_secs": 6.3, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "realtime", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 34.92, "server_rtf": 0.063, "server_ttft_ms": 39.51, "client_final_latency_ms": 37.67, "interim_count": 27, "max_interim_gap_s": 1.048, "feed_secs": 15.77, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null } ] }, { "environment": { "target": "ws://127.0.0.1:19242/stt/sensevoice", "uri": "ws://127.0.0.1:19242/stt/sensevoice", "warmup_runs": 3, "single_connection_series": true }, "generated": "2026-09-05T15:57:12", "results": [ { "feed": "bulk", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 24.29, "server_rtf": 0.011, "server_ttft_ms": 33.06, "client_final_latency_ms": 30.67, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.03, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "bulk", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 26.86, "server_rtf": 0.005, "server_ttft_ms": 20.34, "client_final_latency_ms": 32.93, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.04, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "bulk", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 40.5, "server_rtf": 0.003, "server_ttft_ms": 19.98, "client_final_latency_ms": 46.69, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.12, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開。幫我導航到內湖科學園區，然後選最快的路線，不要走。", "error": null }, { "feed": "realtime", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 24.39, "server_rtf": 0.031, "server_ttft_ms": 32.63, "client_final_latency_ms": 64.16, "interim_count": 5, "max_interim_gap_s": 0.511, "feed_secs": 3.15, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "realtime", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 32.37, "server_rtf": 0.051, "server_ttft_ms": 24.31, "client_final_latency_ms": 34.4, "interim_count": 11, "max_interim_gap_s": 1.013, "feed_secs": 6.29, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "realtime", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 42.28, "server_rtf": 0.058, "server_ttft_ms": 44.67, "client_final_latency_ms": 45.99, "interim_count": 27, "max_interim_gap_s": 1.035, "feed_secs": 15.76, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null } ] }, { "environment": { "target": "ws://127.0.0.1:19242/stt/sensevoice", "uri": "ws://127.0.0.1:19242/stt/sensevoice", "warmup_runs": 3, "single_connection_series": true }, "generated": "2026-09-05T15:58:20", "results": [ { "feed": "bulk", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 22.7, "server_rtf": 0.01, "server_ttft_ms": 29.2, "client_final_latency_ms": 28.33, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.03, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "bulk", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 26.49, "server_rtf": 0.006, "server_ttft_ms": 30.86, "client_final_latency_ms": 32.54, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.06, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "bulk", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 41.95, "server_rtf": 0.003, "server_ttft_ms": 25.83, "client_final_latency_ms": 48.52, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.12, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開。幫我導航到內湖科學園區，然後選最快的路線，不要走。", "error": null }, { "feed": "realtime", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 39.9, "server_rtf": 0.05, "server_ttft_ms": 24.67, "client_final_latency_ms": 43.16, "interim_count": 6, "max_interim_gap_s": 0.531, "feed_secs": 3.16, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "realtime", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 23.18, "server_rtf": 0.039, "server_ttft_ms": 21.8, "client_final_latency_ms": 25.05, "interim_count": 8, "max_interim_gap_s": 2.442, "feed_secs": 6.26, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "realtime", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 32.39, "server_rtf": 0.027, "server_ttft_ms": 33.03, "client_final_latency_ms": 35.52, "interim_count": 14, "max_interim_gap_s": 9.235, "feed_secs": 15.58, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null } ] } ], "summary": [ { "feed": "bulk", "seconds": 3, "median_ms": 24.29, "min_ms": 22.7, "max_ms": 25.69, "samples_ms": [ 25.69, 24.29, 22.7 ], "vs_cuda_percent": 0.0 }, { "feed": "bulk", "seconds": 6, "median_ms": 26.86, "min_ms": 26.49, "max_ms": 30.88, "samples_ms": [ 30.88, 26.86, 26.49 ], "vs_cuda_percent": 0.0 }, { "feed": "bulk", "seconds": 15, "median_ms": 41.95, "min_ms": 40.5, "max_ms": 54.1, "samples_ms": [ 54.1, 40.5, 41.95 ], "vs_cuda_percent": 0.0 }, { "feed": "realtime", "seconds": 3, "median_ms": 36.53, "min_ms": 24.39, "max_ms": 39.9, "samples_ms": [ 36.53, 24.39, 39.9 ], "vs_cuda_percent": 0.0 }, { "feed": "realtime", "seconds": 6, "median_ms": 23.18, "min_ms": 22.24, "max_ms": 32.37, "samples_ms": [ 22.24, 32.37, 23.18 ], "vs_cuda_percent": 0.0 }, { "feed": "realtime", "seconds": 15, "median_ms": 34.92, "min_ms": 32.39, "max_ms": 42.28, "samples_ms": [ 34.92, 42.28, 32.39 ], "vs_cuda_percent": 0.0 } ], "memory": { "ready": { "utc": 1788623683.768108, "VmHWM_MiB": 2322.05859375, "VmRSS_MiB": 2240.6796875, "Pss_MiB": 2233.419921875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93274.15625, "memory.current_MiB": 1872.7890625, "memory.peak_MiB": 1960.53125 }, "end": { "utc": 1788623901.2642698, "VmHWM_MiB": 2415.5078125, "VmRSS_MiB": 2369.203125, "Pss_MiB": 2362.005859375, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 92609.2109375, "memory.current_MiB": 2002.87890625, "memory.peak_MiB": 2412.32421875 } }, "gpu": { "before": { "total_MiB": 24564, "used_MiB": 9569, "processes": [ "760190, 1444 MiB", "2409067, 1444 MiB", "530235, 3514 MiB", "833763, 3102 MiB" ] }, "ready": { "total_MiB": 24564, "used_MiB": 12042, "processes": [ "760190, 1444 MiB", "2409067, 1444 MiB", "530235, 3514 MiB", "833763, 3102 MiB", "1225523, 2468 MiB" ] }, "end": { "total_MiB": 24564, "used_MiB": 12064, "processes": [ "760190, 1444 MiB", "2409067, 1444 MiB", "530235, 3514 MiB", "833763, 3102 MiB", "1225523, 2490 MiB" ] } } }, "online": { "rounds": [ { "environment": { "target": "ws://127.0.0.1:19242/stt/sensevoice", "uri": "ws://127.0.0.1:19242/stt/sensevoice", "warmup_runs": 3, "single_connection_series": true }, "generated": "2026-09-05T15:59:44", "results": [ { "feed": "bulk", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 13.12, "server_rtf": 0.005, "server_ttft_ms": 13.16, "client_final_latency_ms": 26.42, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.02, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "bulk", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 19.86, "server_rtf": 0.003, "server_ttft_ms": 13.93, "client_final_latency_ms": 33.28, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.05, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "bulk", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 42.53, "server_rtf": 0.002, "server_ttft_ms": 19.69, "client_final_latency_ms": 48.73, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.13, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null }, { "feed": "realtime", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 10.81, "server_rtf": 0.024, "server_ttft_ms": 16.73, "client_final_latency_ms": 14.86, "interim_count": 7, "max_interim_gap_s": 0.52, "feed_secs": 3.15, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "realtime", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 22.08, "server_rtf": 0.029, "server_ttft_ms": 21.72, "client_final_latency_ms": 26.06, "interim_count": 11, "max_interim_gap_s": 1.015, "feed_secs": 6.31, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "realtime", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 20.98, "server_rtf": 0.033, "server_ttft_ms": 10.8, "client_final_latency_ms": 24.58, "interim_count": 28, "max_interim_gap_s": 1.028, "feed_secs": 15.74, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null } ] }, { "environment": { "target": "ws://127.0.0.1:19242/stt/sensevoice", "uri": "ws://127.0.0.1:19242/stt/sensevoice", "warmup_runs": 3, "single_connection_series": true }, "generated": "2026-09-05T16:00:55", "results": [ { "feed": "bulk", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 16.17, "server_rtf": 0.006, "server_ttft_ms": 14.48, "client_final_latency_ms": 24.09, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.02, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "bulk", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 17.13, "server_rtf": 0.003, "server_ttft_ms": 11.41, "client_final_latency_ms": 32.48, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.04, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "bulk", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 33.84, "server_rtf": 0.002, "server_ttft_ms": 26.29, "client_final_latency_ms": 40.42, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.13, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null }, { "feed": "realtime", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 10.8, "server_rtf": 0.021, "server_ttft_ms": 9.75, "client_final_latency_ms": 14.61, "interim_count": 7, "max_interim_gap_s": 0.521, "feed_secs": 3.15, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "realtime", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 18.2, "server_rtf": 0.016, "server_ttft_ms": 25.32, "client_final_latency_ms": 2666.39, "interim_count": 7, "max_interim_gap_s": 1.001, "feed_secs": 6.27, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "realtime", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 21.47, "server_rtf": 0.03, "server_ttft_ms": 29.22, "client_final_latency_ms": 23.81, "interim_count": 26, "max_interim_gap_s": 2.776, "feed_secs": 15.74, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null } ] }, { "environment": { "target": "ws://127.0.0.1:19242/stt/sensevoice", "uri": "ws://127.0.0.1:19242/stt/sensevoice", "warmup_runs": 3, "single_connection_series": true }, "generated": "2026-09-05T16:02:08", "results": [ { "feed": "bulk", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 20.9, "server_rtf": 0.007, "server_ttft_ms": 15.52, "client_final_latency_ms": 30.45, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.04, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "bulk", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 17.24, "server_rtf": 0.003, "server_ttft_ms": 21.04, "client_final_latency_ms": 29.89, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.06, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "bulk", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 33.58, "server_rtf": 0.002, "server_ttft_ms": 13.65, "client_final_latency_ms": 54.82, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.13, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開。幫我導航到內湖科學園區，然後選最快的路線，不要走。", "error": null }, { "feed": "realtime", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 18.64, "server_rtf": 0.023, "server_ttft_ms": 14.4, "client_final_latency_ms": 35.65, "interim_count": 6, "max_interim_gap_s": 0.528, "feed_secs": 3.17, "text": "幫我導航到內湖科學園區，後。", "error": null }, { "feed": "realtime", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 12.29, "server_rtf": 0.028, "server_ttft_ms": 27.53, "client_final_latency_ms": 23.49, "interim_count": 11, "max_interim_gap_s": 1.039, "feed_secs": 6.29, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "realtime", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 34.35, "server_rtf": 0.035, "server_ttft_ms": 21.31, "client_final_latency_ms": 42.32, "interim_count": 28, "max_interim_gap_s": 1.213, "feed_secs": 15.77, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開。幫我導航到內湖科學園區，然後選最快的路線，不要走。", "error": null } ] } ], "summary": [ { "feed": "bulk", "seconds": 3, "median_ms": 16.17, "min_ms": 13.12, "max_ms": 20.9, "samples_ms": [ 13.12, 16.17, 20.9 ], "vs_cuda_percent": -33.4293948126801 }, { "feed": "bulk", "seconds": 6, "median_ms": 17.24, "min_ms": 17.13, "max_ms": 19.86, "samples_ms": [ 19.86, 17.13, 17.24 ], "vs_cuda_percent": -35.815338793745354 }, { "feed": "bulk", "seconds": 15, "median_ms": 33.84, "min_ms": 33.58, "max_ms": 42.53, "samples_ms": [ 42.53, 33.84, 33.58 ], "vs_cuda_percent": -19.332538736591175 }, { "feed": "realtime", "seconds": 3, "median_ms": 10.81, "min_ms": 10.8, "max_ms": 18.64, "samples_ms": [ 10.81, 10.8, 18.64 ], "vs_cuda_percent": -70.4078839310156 }, { "feed": "realtime", "seconds": 6, "median_ms": 18.2, "min_ms": 12.29, "max_ms": 22.08, "samples_ms": [ 22.08, 18.2, 12.29 ], "vs_cuda_percent": -21.48403796376187 }, { "feed": "realtime", "seconds": 15, "median_ms": 21.47, "min_ms": 20.98, "max_ms": 34.35, "samples_ms": [ 20.98, 21.47, 34.35 ], "vs_cuda_percent": -38.51660939289806 } ], "memory": { "ready": { "utc": 1788623916.2496142, "VmHWM_MiB": 9670.0859375, "VmRSS_MiB": 8994.94140625, "Pss_MiB": 8987.685546875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 87492.08984375, "memory.current_MiB": 6751.15625, "memory.peak_MiB": 7438.3515625 }, "end": { "utc": 1788624129.1755688, "VmHWM_MiB": 9670.0859375, "VmRSS_MiB": 9145.74609375, "Pss_MiB": 9138.484375, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 87900.45703125, "memory.current_MiB": 6902.36328125, "memory.peak_MiB": 7438.3515625 } }, "gpu": { "before": { "total_MiB": 24564, "used_MiB": 9569, "processes": [ "760190, 1444 MiB", "2409067, 1444 MiB", "530235, 3514 MiB", "833763, 3102 MiB" ] }, "ready": { "total_MiB": 24564, "used_MiB": 12122, "processes": [ "760190, 1444 MiB", "2409067, 1444 MiB", "530235, 3514 MiB", "833763, 3102 MiB", "1253578, 2548 MiB" ] }, "end": { "total_MiB": 24564, "used_MiB": 12230, "processes": [ "760190, 1444 MiB", "2409067, 1444 MiB", "530235, 3514 MiB", "833763, 3102 MiB", "1253578, 2656 MiB" ] } } }, "offline": { "rounds": [ { "environment": { "target": "ws://127.0.0.1:19242/stt/sensevoice", "uri": "ws://127.0.0.1:19242/stt/sensevoice", "warmup_runs": 3, "single_connection_series": true }, "generated": "2026-09-05T16:03:42", "results": [ { "feed": "bulk", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 11.97, "server_rtf": 0.004, "server_ttft_ms": 22.89, "client_final_latency_ms": 27.52, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.03, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "bulk", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 19.54, "server_rtf": 0.005, "server_ttft_ms": 32.55, "client_final_latency_ms": 32.34, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.07, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "bulk", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 39.89, "server_rtf": 0.001, "server_ttft_ms": 122.49, "client_final_latency_ms": 128.15, "interim_count": 0, "max_interim_gap_s": null, "feed_secs": 0.08, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null }, { "feed": "realtime", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 10.41, "server_rtf": 0.02, "server_ttft_ms": 9.98, "client_final_latency_ms": 14.99, "interim_count": 6, "max_interim_gap_s": 0.521, "feed_secs": 3.15, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "realtime", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 23.21, "server_rtf": 0.032, "server_ttft_ms": 9.93, "client_final_latency_ms": 25.92, "interim_count": 12, "max_interim_gap_s": 1.019, "feed_secs": 6.3, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "realtime", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 23.8, "server_rtf": 0.036, "server_ttft_ms": 28.74, "client_final_latency_ms": 27.37, "interim_count": 28, "max_interim_gap_s": 1.026, "feed_secs": 15.73, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null } ] }, { "environment": { "target": "ws://127.0.0.1:19242/stt/sensevoice", "uri": "ws://127.0.0.1:19242/stt/sensevoice", "warmup_runs": 3, "single_connection_series": true }, "generated": "2026-09-05T16:04:51", "results": [ { "feed": "bulk", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 12.95, "server_rtf": 0.005, "server_ttft_ms": 19.97, "client_final_latency_ms": 26.09, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.03, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "bulk", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 16.34, "server_rtf": 0.003, "server_ttft_ms": 14.34, "client_final_latency_ms": 31.75, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.05, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "bulk", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 33.34, "server_rtf": 0.002, "server_ttft_ms": 20.39, "client_final_latency_ms": 39.67, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.11, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開。幫我導航到內湖科學園區，然後選最快的路線，不要走。", "error": null }, { "feed": "realtime", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 12.86, "server_rtf": 0.024, "server_ttft_ms": 9.75, "client_final_latency_ms": 21.87, "interim_count": 7, "max_interim_gap_s": 0.524, "feed_secs": 3.17, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "realtime", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 20.48, "server_rtf": 0.022, "server_ttft_ms": 9.91, "client_final_latency_ms": 39.8, "interim_count": 11, "max_interim_gap_s": 1.041, "feed_secs": 6.31, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "realtime", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 26.41, "server_rtf": 0.032, "server_ttft_ms": 21.85, "client_final_latency_ms": 33.51, "interim_count": 28, "max_interim_gap_s": 1.023, "feed_secs": 15.78, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null } ] }, { "environment": { "target": "ws://127.0.0.1:19242/stt/sensevoice", "uri": "ws://127.0.0.1:19242/stt/sensevoice", "warmup_runs": 3, "single_connection_series": true }, "generated": "2026-09-05T16:05:59", "results": [ { "feed": "bulk", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 12.46, "server_rtf": 0.005, "server_ttft_ms": 20.56, "client_final_latency_ms": 19.77, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.03, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "bulk", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 16.95, "server_rtf": 0.003, "server_ttft_ms": 13.51, "client_final_latency_ms": 29.65, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.04, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "bulk", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 33.53, "server_rtf": 0.002, "server_ttft_ms": 11.03, "client_final_latency_ms": 39.46, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.12, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開。幫我導航到內湖科學園區，然後選最快的路線，不要走。", "error": null }, { "feed": "realtime", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 11.25, "server_rtf": 0.02, "server_ttft_ms": 9.43, "client_final_latency_ms": 15.26, "interim_count": 6, "max_interim_gap_s": 0.52, "feed_secs": 3.17, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "realtime", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 12.47, "server_rtf": 0.022, "server_ttft_ms": 10.51, "client_final_latency_ms": 22.95, "interim_count": 10, "max_interim_gap_s": 1.479, "feed_secs": 6.3, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "realtime", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 23.51, "server_rtf": 0.03, "server_ttft_ms": 9.98, "client_final_latency_ms": 36.3, "interim_count": 28, "max_interim_gap_s": 1.054, "feed_secs": 15.71, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null } ] } ], "summary": [ { "feed": "bulk", "seconds": 3, "median_ms": 12.46, "min_ms": 11.97, "max_ms": 12.95, "samples_ms": [ 11.97, 12.95, 12.46 ], "vs_cuda_percent": -48.70317002881844 }, { "feed": "bulk", "seconds": 6, "median_ms": 16.95, "min_ms": 16.34, "max_ms": 19.54, "samples_ms": [ 19.54, 16.34, 16.95 ], "vs_cuda_percent": -36.89501116902457 }, { "feed": "bulk", "seconds": 15, "median_ms": 33.53, "min_ms": 33.34, "max_ms": 39.89, "samples_ms": [ 39.89, 33.34, 33.53 ], "vs_cuda_percent": -20.07151370679381 }, { "feed": "realtime", "seconds": 3, "median_ms": 11.25, "min_ms": 10.41, "max_ms": 12.86, "samples_ms": [ 10.41, 12.86, 11.25 ], "vs_cuda_percent": -69.20339447029839 }, { "feed": "realtime", "seconds": 6, "median_ms": 20.48, "min_ms": 12.47, "max_ms": 23.21, "samples_ms": [ 23.21, 20.48, 12.47 ], "vs_cuda_percent": -11.647972389991368 }, { "feed": "realtime", "seconds": 15, "median_ms": 23.8, "min_ms": 23.51, "max_ms": 26.41, "samples_ms": [ 23.8, 26.41, 23.51 ], "vs_cuda_percent": -31.844215349369986 } ], "memory": { "ready": { "utc": 1788624154.5370853, "VmHWM_MiB": 2860.3125, "VmRSS_MiB": 1957.09375, "Pss_MiB": 1949.8828125, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 93081.63671875, "memory.current_MiB": 1412.390625, "memory.peak_MiB": 2326.5625 }, "end": { "utc": 1788624359.696178, "VmHWM_MiB": 2860.3125, "VmRSS_MiB": 2098.26953125, "Pss_MiB": 1953.3271484375, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 92444.0546875, "memory.current_MiB": 1553.64453125, "memory.peak_MiB": 2326.5625 } }, "gpu": { "before": { "total_MiB": 24564, "used_MiB": 9569, "processes": [ "760190, 1444 MiB", "2409067, 1444 MiB", "530235, 3514 MiB", "833763, 3102 MiB" ] }, "ready": { "total_MiB": 24564, "used_MiB": 12114, "processes": [ "760190, 1444 MiB", "2409067, 1444 MiB", "530235, 3514 MiB", "833763, 3102 MiB", "1283122, 2540 MiB" ] }, "end": { "total_MiB": 24564, "used_MiB": 12222, "processes": [ "760190, 1444 MiB", "2409067, 1444 MiB", "530235, 3514 MiB", "833763, 3102 MiB", "1283122, 2648 MiB" ] } } } }, "offline_build": { "arguments": { "self_check": false, "mode": "graph", "provider": "trt", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": null, "cache": "offline-bundle", "audio": null, "length": 3, "frames": "50", "warmup": 0, "repeats": 1, "threads": 2, "profile": false, "dump_context": "/app/offline-bundle/model.onnx", "out": "/bench/4090/offline-build.json" }, "utc": "2026-09-05T15:50:15Z", "machine": "x86_64", "ort": "1.23.2", "numpy": "2.2.6", "model_sha256": "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994", "start": { "VmHWM_MiB": 46.85546875, "VmRSS_MiB": 45.90234375, "Pss_MiB": 45.072265625, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 92963.33984375, "memory.current_MiB": 28.67578125, "memory.peak_MiB": 30.65234375, "gpu_total_memory_util_clock_power": "9569, 0, 2700, 85.82" }, "rows": [ { "frames": 50, "p50_ms": 36.904952954500914, "p95_ms": 36.904952954500914, "samples_ms": [ 36.904952954500914 ], "memory": { "VmHWM_MiB": 10884.12109375, "VmRSS_MiB": 8946.68359375, "Pss_MiB": 8941.2763671875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 87939.9296875, "memory.current_MiB": 7792.63671875, "memory.peak_MiB": 9682.98046875, "gpu_total_memory_util_clock_power": "11504, 1, 2715, 84.52" }, "finite": true, "input_sha256": "472ba979ad46ce462cc2cceeb0d218314362876e03b63f06a136a1c19b3dbab7" } ], "load_s": 22.836240782169625, "ready": { "VmHWM_MiB": 10884.12109375, "VmRSS_MiB": 8944.46875, "Pss_MiB": 8939.0576171875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 87960.12109375, "memory.current_MiB": 7792.36328125, "memory.peak_MiB": 9682.98046875, "gpu_total_memory_util_clock_power": "11484, 1, 2715, 83.94" }, "providers": [ "TensorrtExecutionProvider", "CUDAExecutionProvider", "CPUExecutionProvider" ], "end": { "VmHWM_MiB": 10884.12109375, "VmRSS_MiB": 8946.75, "Pss_MiB": 8941.3388671875, "MemTotal_MiB": 128413.15234375, "MemAvailable_MiB": 87903.57421875, "memory.current_MiB": 7792.703125, "memory.peak_MiB": 9682.98046875, "gpu_total_memory_util_clock_power": "11504, 1, 2715, 84.44" } }, "recheck_15s": { "cuda": { "environment": { "target": "ws://127.0.0.1:19242/stt/sensevoice", "uri": "ws://127.0.0.1:19242/stt/sensevoice", "warmup_runs": 3, "single_connection_series": true }, "generated": "2026-09-05T16:09:39", "results": [ { "feed": "bulk", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 39.17, "server_rtf": 0.003, "server_ttft_ms": 20.63, "client_final_latency_ms": 45.64, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.11, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null }, { "feed": "realtime", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 49.22, "server_rtf": 0.056, "server_ttft_ms": 21.47, "client_final_latency_ms": 51.75, "interim_count": 27, "max_interim_gap_s": 1.024, "feed_secs": 15.79, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開。幫我導航到內湖科學園區，然後選最快的路線，不要走。", "error": null } ] }, "online": { "environment": { "target": "ws://127.0.0.1:19242/stt/sensevoice", "uri": "ws://127.0.0.1:19242/stt/sensevoice", "warmup_runs": 3, "single_connection_series": true }, "generated": "2026-09-05T16:11:26", "results": [ { "feed": "bulk", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 39.4, "server_rtf": 0.002, "server_ttft_ms": 22.06, "client_final_latency_ms": 161.88, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.13, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null }, { "feed": "realtime", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 24.55, "server_rtf": 0.03, "server_ttft_ms": 10.71, "client_final_latency_ms": 27.14, "interim_count": 27, "max_interim_gap_s": 1.046, "feed_secs": 15.8, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null } ] }, "offline": { "environment": { "target": "ws://127.0.0.1:19242/stt/sensevoice", "uri": "ws://127.0.0.1:19242/stt/sensevoice", "warmup_runs": 3, "single_connection_series": true }, "generated": "2026-09-05T16:13:03", "results": [ { "feed": "bulk", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 34.08, "server_rtf": 0.002, "server_ttft_ms": 14.14, "client_final_latency_ms": 39.94, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.11, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null }, { "feed": "realtime", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 38.5, "server_rtf": 0.032, "server_ttft_ms": 22.99, "client_final_latency_ms": 50.01, "interim_count": 27, "max_interim_gap_s": 1.019, "feed_secs": 15.77, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null } ] } } }, "orin": { "environment": { "platform": "Linux-5.15.148-tegra-aarch64-with-glibc2.35", "python": "3.10.12", "packages": { "onnxruntime-gpu": "1.24.0", "numpy": "1.26.4", "sherpa-onnx": "1.13.0+cuda", "torch": "2.8.0", "kitt-core": "0.3.7", "pipecat-ai": "0.0.108" }, "sha256": { "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx": "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994", "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/tokens.txt": "f449eb28dc567533d7fa59be34e2abca8784f771850c78a47fb731a31429a1dc", "/app/src/asr/batch_asr_manager.py": "fb9a00f6cdbc5bd972d3bb82c85b44a70277a593d1694460071bac124ef25815", "/app/src/kitt_stt_service.py": "5283feaee7a213feb529afe1d98308c7fec447ccd18aa9f737a7632ea08b83bc" } }, "source_hashes": [ "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994 /app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "f449eb28dc567533d7fa59be34e2abca8784f771850c78a47fb731a31429a1dc /app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/tokens.txt", "fb9a00f6cdbc5bd972d3bb82c85b44a70277a593d1694460071bac124ef25815 /app/src/asr/batch_asr_manager.py", "5283feaee7a213feb529afe1d98308c7fec447ccd18aa9f737a7632ea08b83bc /app/src/kitt_stt_service.py" ], "graph": { "cuda": { "arguments": { "self_check": false, "mode": "graph", "provider": "cuda", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": null, "cache": "/bench/online-cache", "audio": null, "length": 3, "frames": "50,100,250", "warmup": 10, "repeats": 50, "threads": 2, "profile": false, "dump_context": null, "out": "/bench/result.json" }, "utc": "2026-09-05T16:31:21Z", "machine": "aarch64", "ort": "1.24.0", "numpy": "1.26.4", "model_sha256": "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994", "start": { "VmHWM_MiB": 44.12109375, "VmRSS_MiB": 42.47265625, "Pss_MiB": 32.55859375, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 19807.51953125, "memory.current_MiB": 52.16796875 }, "rows": [ { "frames": 50, "p50_ms": 36.29870247095823, "p95_ms": 36.812099162489176, "samples_ms": [ 36.163431126624346, 36.09613887965679, 36.01940767839551, 36.23773017898202, 36.355451215058565, 36.42274299636483, 36.58820502460003, 36.462516989558935, 36.63687501102686, 35.83397902548313, 36.230241879820824, 36.7313651368022, 36.16077592596412, 35.6446779333055, 35.7745261862874, 36.614124197512865, 36.590893752872944, 36.16323880851269, 36.3168609328568, 36.33362939581275, 36.16176778450608, 36.28054400905967, 35.95339506864548, 36.065516993403435, 36.3426199182868, 35.89576808735728, 36.037649028003216, 35.88633006438613, 36.379996221512556, 35.96098115667701, 36.37295588850975, 36.408633925020695, 35.93016695231199, 36.00904205814004, 35.90207314118743, 35.68566916510463, 35.58132378384471, 36.53195407241583, 36.821473855525255, 36.432857159525156, 36.493685096502304, 36.42517700791359, 36.812099162489176, 36.46744694560766, 36.507636308670044, 37.66203112900257, 36.80902719497681, 35.45429138466716, 36.64823621511459, 36.04734409600496 ], "memory": { "VmHWM_MiB": 2260.2734375, "VmRSS_MiB": 2072.63671875, "Pss_MiB": 885.2431640625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17873.48046875, "memory.current_MiB": 2007.265625 }, "finite": true, "input_sha256": "472ba979ad46ce462cc2cceeb0d218314362876e03b63f06a136a1c19b3dbab7" }, { "frames": 100, "p50_ms": 36.99308377690613, "p95_ms": 37.513721734285355, "samples_ms": [ 37.19777287915349, 36.98393004015088, 37.42671897634864, 37.16609487310052, 37.56648674607277, 37.002393044531345, 37.513721734285355, 37.066133227199316, 37.38848119974136, 36.68657084926963, 36.579121835529804, 36.71348234638572, 36.87996789813042, 37.21959609538317, 36.698410753160715, 37.07509208470583, 36.681003868579865, 37.16721525415778, 36.83209978044033, 37.11429191753268, 36.45919496193528, 36.51932021602988, 36.357889883220196, 37.362501956522465, 36.59237176179886, 36.45046008750796, 37.256140261888504, 37.30727266520262, 37.707759998738766, 37.05896716564894, 36.72125982120633, 37.340679205954075, 37.0544558390975, 36.68913384899497, 36.91478539258242, 36.498712841421366, 36.619666032493114, 36.99385188519955, 36.99231566861272, 36.99820302426815, 36.84227680787444, 37.034167908132076, 36.605075001716614, 37.300010211765766, 37.21771016716957, 36.65889659896493, 36.412063986063, 36.91481798887253, 37.3829179443419, 36.93475294858217 ], "memory": { "VmHWM_MiB": 2260.2734375, "VmRSS_MiB": 2072.8046875, "Pss_MiB": 885.478515625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17873.49609375, "memory.current_MiB": 2007.265625 }, "finite": true, "input_sha256": "dda0d977bf96004405c6cbeac7cd57cfb466001cc340939617503e06dff71881" }, { "frames": 250, "p50_ms": 44.97032589279115, "p95_ms": 45.04207707941532, "samples_ms": [ 44.989566784352064, 44.99231884256005, 44.97577715665102, 45.03903677687049, 44.946529902517796, 45.04207707941532, 44.962977059185505, 44.99910632148385, 44.874120969325304, 44.977346900850534, 44.93955709040165, 44.99216191470623, 44.998690020293, 45.01827200874686, 44.8872079141438, 45.0334707275033, 44.969923328608274, 44.959107879549265, 44.88269705325365, 44.97318668290973, 44.98643474653363, 44.962947722524405, 44.878185261040926, 44.990690890699625, 44.986019376665354, 44.83908275142312, 44.96090020984411, 45.01868784427643, 45.03503907471895, 45.022433158010244, 44.970597606152296, 44.845420867204666, 45.02054676413536, 45.01043586060405, 44.9452237226069, 44.97005417943001, 45.050336979329586, 45.04400072619319, 44.96883787214756, 44.95565500110388, 44.99866021797061, 44.944263994693756, 44.913000892847776, 44.949895702302456, 44.94090424850583, 44.94320787489414, 44.92192808538675, 44.936231803148985, 45.029090251773596, 44.96963694691658 ], "memory": { "VmHWM_MiB": 2260.2734375, "VmRSS_MiB": 2072.703125, "Pss_MiB": 885.533203125, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17872.83984375, "memory.current_MiB": 2007.28515625 }, "finite": true, "input_sha256": "726cc736d46a4b8ca3c2d2831abc8c8e3a59489168194774d072d5fc1aa502b5" } ], "load_s": 3.0156567422673106, "ready": { "VmHWM_MiB": 2260.2734375, "VmRSS_MiB": 2034.76171875, "Pss_MiB": 866.3447265625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17885.40234375, "memory.current_MiB": 1991.3125 }, "providers": [ "CUDAExecutionProvider", "CPUExecutionProvider" ], "end": { "VmHWM_MiB": 2260.2734375, "VmRSS_MiB": 2072.703125, "Pss_MiB": 885.537109375, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17872.83984375, "memory.current_MiB": 2007.28515625 } }, "online": { "arguments": { "self_check": false, "mode": "graph", "provider": "trt", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": null, "cache": "/bench/online-cache", "audio": null, "length": 3, "frames": "50,100,250", "warmup": 10, "repeats": 50, "threads": 2, "profile": false, "dump_context": null, "out": "/bench/result.json" }, "utc": "2026-09-05T16:31:38Z", "machine": "aarch64", "ort": "1.24.0", "numpy": "1.26.4", "model_sha256": "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994", "start": { "VmHWM_MiB": 44.17578125, "VmRSS_MiB": 42.52734375, "Pss_MiB": 32.609375, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18663.703125, "memory.current_MiB": 1219.03125 }, "rows": [ { "frames": 50, "p50_ms": 19.974726485088468, "p95_ms": 20.069952122867107, "samples_ms": [ 20.072864834219217, 20.069952122867107, 20.082656294107437, 19.99702863395214, 20.063937176018953, 20.008068066090345, 20.055392757058144, 19.980870187282562, 19.998149015009403, 19.955398980528116, 19.99696483835578, 19.951208028942347, 20.008324645459652, 19.911882001906633, 19.994597882032394, 19.95655195787549, 19.99638881534338, 19.954022951424122, 19.985573831945658, 19.909386057406664, 20.035203080624342, 19.87268403172493, 19.95424786582589, 19.895531237125397, 19.969062879681587, 19.891787320375443, 20.00704500824213, 19.890635274350643, 19.98176658526063, 19.898314960300922, 20.000644959509373, 19.884460140019655, 19.98419826850295, 19.90903401747346, 19.994982052594423, 19.9172580614686, 20.014404319226742, 19.92375310510397, 19.972485955804586, 19.87102022394538, 20.015875808894634, 19.928873050957918, 20.06051316857338, 19.953384064137936, 20.01248486340046, 19.929833244532347, 19.97696701437235, 19.895882345736027, 19.970727153122425, 19.861420150846243 ], "memory": { "VmHWM_MiB": 8629.703125, "VmRSS_MiB": 7015.62109375, "Pss_MiB": 5547.685546875, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 13100.640625, "memory.current_MiB": 7979.57421875 }, "finite": true, "input_sha256": "472ba979ad46ce462cc2cceeb0d218314362876e03b63f06a136a1c19b3dbab7" }, { "frames": 100, "p50_ms": 20.391424652189016, "p95_ms": 20.496650133281946, "samples_ms": [ 20.446285139769316, 20.546311978250742, 20.492522977292538, 20.390896126627922, 20.407790783792734, 20.374320913106203, 20.45086119323969, 20.365457981824875, 20.428877789527178, 20.33211477100849, 20.496650133281946, 20.419342909008265, 20.425166934728622, 20.361040718853474, 20.406320225447416, 20.361170172691345, 20.402448251843452, 20.366610027849674, 20.406126976013184, 20.34520311281085, 20.482762716710567, 20.374609157443047, 20.417422987520695, 20.415439270436764, 20.383120980113745, 20.357906818389893, 20.407472271472216, 20.349138882011175, 20.390448160469532, 20.355953834950924, 20.50199406221509, 20.405744202435017, 20.44558059424162, 20.3378121368587, 20.400336012244225, 20.36408195272088, 20.39195317775011, 20.345202647149563, 20.388911943882704, 20.3412352129817, 20.47607582062483, 20.388689823448658, 20.406671799719334, 20.406640134751797, 20.389809273183346, 20.34936286509037, 20.408431999385357, 20.347507670521736, 20.38084901869297, 20.349939819425344 ], "memory": { "VmHWM_MiB": 8629.703125, "VmRSS_MiB": 7031.578125, "Pss_MiB": 5547.6650390625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 13100.65234375, "memory.current_MiB": 7979.48046875 }, "finite": true, "input_sha256": "dda0d977bf96004405c6cbeac7cd57cfb466001cc340939617503e06dff71881" }, { "frames": 250, "p50_ms": 39.474494056776166, "p95_ms": 39.57104589790106, "samples_ms": [ 39.564421866089106, 39.57642102614045, 39.57104589790106, 39.52186414971948, 39.518472738564014, 39.55987887457013, 39.53664796426892, 39.526824839413166, 39.539114106446505, 39.55850517377257, 39.573736023157835, 39.45521404966712, 39.53396202996373, 39.490860886871815, 39.50260393321514, 39.50234688818455, 39.49085995554924, 39.45534164085984, 39.51575513929129, 39.506155997514725, 39.46852497756481, 39.44327915087342, 39.505323860794306, 39.47028471156955, 39.47313316166401, 39.472909178584814, 39.54692091792822, 39.45947112515569, 39.47262233123183, 39.40104227513075, 39.46455800905824, 39.47575716301799, 39.46228604763746, 39.41777907311916, 39.47243094444275, 39.451856166124344, 39.44484889507294, 39.43627281114459, 39.43400178104639, 39.4521770067513, 39.47224002331495, 39.38558790832758, 39.4611363299191, 39.44859188050032, 39.49908493086696, 39.473230950534344, 39.52913172543049, 39.53668288886547, 39.495854172855616, 39.42545782774687 ], "memory": { "VmHWM_MiB": 8629.703125, "VmRSS_MiB": 7063.578125, "Pss_MiB": 5547.6728515625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 13100.21484375, "memory.current_MiB": 7979.515625 }, "finite": true, "input_sha256": "726cc736d46a4b8ca3c2d2831abc8c8e3a59489168194774d072d5fc1aa502b5" } ], "load_s": 66.89491306478158, "ready": { "VmHWM_MiB": 8629.703125, "VmRSS_MiB": 6994.671875, "Pss_MiB": 5547.076171875, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 13100.9921875, "memory.current_MiB": 7979.32421875 }, "providers": [ "TensorrtExecutionProvider", "CUDAExecutionProvider", "CPUExecutionProvider" ], "end": { "VmHWM_MiB": 8629.703125, "VmRSS_MiB": 7063.578125, "Pss_MiB": 5547.6767578125, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 13100.21484375, "memory.current_MiB": 7979.515625 } }, "offline": { "arguments": { "self_check": false, "mode": "graph", "provider": "trt", "model": "/app/offline-bundle/model.onnx", "tokens": null, "cache": "/app/offline-bundle", "audio": null, "length": 3, "frames": "50,100,250", "warmup": 10, "repeats": 50, "threads": 2, "profile": false, "dump_context": null, "out": "/bench/result.json" }, "utc": "2026-09-05T16:33:11Z", "machine": "aarch64", "ort": "1.24.0", "numpy": "1.26.4", "model_sha256": "4cc9c486d3156800a4a5eb0257678aae4e1803c57eca66fd9623e55e6fd2a6db", "start": { "VmHWM_MiB": 44.48046875, "VmRSS_MiB": 44.46484375, "Pss_MiB": 34.4599609375, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18581.1640625, "memory.current_MiB": 2550.25390625 }, "rows": [ { "frames": 50, "p50_ms": 19.246540730819106, "p95_ms": 19.309161230921745, "samples_ms": [ 19.266059156507254, 20.264179911464453, 19.273641984909773, 19.299177918583155, 19.271147903054953, 19.26522795110941, 19.27076280117035, 19.29261814802885, 19.3229210563004, 19.272299949079752, 19.240941386669874, 19.255117047578096, 19.23905313014984, 19.221101887524128, 19.25857225432992, 19.238764885812998, 19.283914007246494, 19.231341779232025, 19.258219748735428, 19.23386985436082, 19.26477998495102, 19.234573002904654, 19.246572628617287, 19.227310083806515, 19.250317011028528, 19.255404826253653, 19.237901084125042, 19.256876315921545, 19.22721415758133, 19.226829055696726, 19.278634805232286, 19.24919616430998, 19.27198003977537, 19.24218889325857, 19.236749038100243, 19.238220993429422, 19.309161230921745, 19.236109219491482, 19.246508833020926, 19.223277922719717, 19.24788486212492, 19.273482728749514, 19.236620981246233, 19.199055153876543, 19.175279885530472, 19.153521861881018, 19.230092875659466, 19.175536930561066, 19.213455263525248, 19.154288806021214 ], "memory": { "VmHWM_MiB": 2261.8828125, "VmRSS_MiB": 2129.94921875, "Pss_MiB": 838.1552734375, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17857.046875, "memory.current_MiB": 3497.8671875 }, "finite": true, "input_sha256": "472ba979ad46ce462cc2cceeb0d218314362876e03b63f06a136a1c19b3dbab7" }, { "frames": 100, "p50_ms": 19.881692249327898, "p95_ms": 19.95645510032773, "samples_ms": [ 19.900203216820955, 21.752229891717434, 19.95645510032773, 19.90948198363185, 19.91732232272625, 19.929257221519947, 19.95818316936493, 19.901355262845755, 19.915114156901836, 19.864683970808983, 19.899881910532713, 19.86202783882618, 19.882890861481428, 19.841293338686228, 19.862444140017033, 19.87185189500451, 19.916937220841646, 19.865483976900578, 19.881644286215305, 19.8638211004436, 19.87223606556654, 19.8672441765666, 19.86221969127655, 19.88868322223425, 19.916169811040163, 19.85969301313162, 19.874188117682934, 19.87437903881073, 19.873612094670534, 19.899371080100536, 19.874668214470148, 19.89549770951271, 19.85482918098569, 19.866188056766987, 19.91216978058219, 19.882699940353632, 19.879436120390892, 19.893771037459373, 19.867179915308952, 19.88046010956168, 19.91216978058219, 19.91367293521762, 19.927081186324358, 19.899082835763693, 19.91489017382264, 19.862636923789978, 19.864364992827177, 19.848589785397053, 19.88174021244049, 19.841677974909544 ], "memory": { "VmHWM_MiB": 2261.8828125, "VmRSS_MiB": 2169.9296875, "Pss_MiB": 862.232421875, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17833.13671875, "memory.current_MiB": 3521.9453125 }, "finite": true, "input_sha256": "dda0d977bf96004405c6cbeac7cd57cfb466001cc340939617503e06dff71881" }, { "frames": 250, "p50_ms": 38.584522204473615, "p95_ms": 38.774514105170965, "samples_ms": [ 38.51273190230131, 39.812854025512934, 38.71204890310764, 38.50399609655142, 38.58623234555125, 38.51669980213046, 38.57637569308281, 38.576279766857624, 38.52981887757778, 38.5615611448884, 38.599384017288685, 38.52764517068863, 38.57778571546078, 38.525693118572235, 38.587352726608515, 38.58601022511721, 38.59727410599589, 38.564282935112715, 38.56940194964409, 38.55177201330662, 38.57234679162502, 38.59122563153505, 38.5778178460896, 38.540379609912634, 38.58873015269637, 38.53196417912841, 38.62677561119199, 38.557339925318956, 38.61743211746216, 38.56297116726637, 38.60498499125242, 38.58303418383002, 38.594488985836506, 38.48332678899169, 38.63323898985982, 38.590296637266874, 38.71793858706951, 38.70267700403929, 38.81239891052246, 38.5962831787765, 38.47804991528392, 38.57062058523297, 38.6999580077827, 38.774514105170965, 38.652120158076286, 38.438755087554455, 38.57045993208885, 38.67429494857788, 38.66114281117916, 38.61135430634022 ], "memory": { "VmHWM_MiB": 2261.8828125, "VmRSS_MiB": 2244.69921875, "Pss_MiB": 904.9560546875, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17787.48046875, "memory.current_MiB": 3564.828125 }, "finite": true, "input_sha256": "726cc736d46a4b8ca3c2d2831abc8c8e3a59489168194774d072d5fc1aa502b5" } ], "load_s": 8.179328948259354, "ready": { "VmHWM_MiB": 2261.8828125, "VmRSS_MiB": 2096.21875, "Pss_MiB": 823.318359375, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17870.4609375, "memory.current_MiB": 3484.6171875 }, "providers": [ "TensorrtExecutionProvider", "CUDAExecutionProvider", "CPUExecutionProvider" ], "end": { "VmHWM_MiB": 2261.8828125, "VmRSS_MiB": 2244.87109375, "Pss_MiB": 904.9638671875, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17781.51171875, "memory.current_MiB": 3564.7578125 } } }, "provider_proof": { "cuda": { "arguments": { "self_check": false, "mode": "graph", "provider": "cuda", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": null, "cache": "/bench/online-cache", "audio": null, "length": 3, "frames": "50", "warmup": 0, "repeats": 1, "threads": 2, "profile": true, "dump_context": null, "out": "/bench/result.json" }, "utc": "2026-09-05T16:31:33Z", "machine": "aarch64", "ort": "1.24.0", "numpy": "1.26.4", "model_sha256": "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994", "start": { "VmHWM_MiB": 44.15625, "VmRSS_MiB": 42.5078125, "Pss_MiB": 32.517578125, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18676.08203125, "memory.current_MiB": 1213.890625 }, "rows": [ { "frames": 50, "p50_ms": 269.0139818005264, "p95_ms": 269.0139818005264, "samples_ms": [ 269.0139818005264 ], "memory": { "VmHWM_MiB": 2309.63671875, "VmRSS_MiB": 2294.1953125, "Pss_MiB": 1106.734375, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17648.1953125, "memory.current_MiB": 2237.953125 }, "finite": true, "input_sha256": "472ba979ad46ce462cc2cceeb0d218314362876e03b63f06a136a1c19b3dbab7" } ], "load_s": 2.7495639421977103, "ready": { "VmHWM_MiB": 2309.63671875, "VmRSS_MiB": 2253.0, "Pss_MiB": 1084.728515625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17651.875, "memory.current_MiB": 2231.06640625 }, "providers": [ "CUDAExecutionProvider", "CPUExecutionProvider" ], "node_provider_counts": { "CUDAExecutionProvider": 2908, "CPUExecutionProvider": 785 }, "end": { "VmHWM_MiB": 2309.63671875, "VmRSS_MiB": 2307.921875, "Pss_MiB": 1120.2890625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17634.17578125, "memory.current_MiB": 2254.078125 } }, "online": { "arguments": { "self_check": false, "mode": "graph", "provider": "trt", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": null, "cache": "/bench/online-cache", "audio": null, "length": 3, "frames": "50", "warmup": 0, "repeats": 1, "threads": 2, "profile": true, "dump_context": null, "out": "/bench/result.json" }, "utc": "2026-09-05T16:32:53Z", "machine": "aarch64", "ort": "1.24.0", "numpy": "1.26.4", "model_sha256": "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994", "start": { "VmHWM_MiB": 44.46484375, "VmRSS_MiB": 42.5859375, "Pss_MiB": 32.677734375, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18365.78125, "memory.current_MiB": 2716.63671875 }, "rows": [ { "frames": 50, "p50_ms": 42.35797002911568, "p95_ms": 42.35797002911568, "samples_ms": [ 42.35797002911568 ], "memory": { "VmHWM_MiB": 8590.87109375, "VmRSS_MiB": 6979.1796875, "Pss_MiB": 5510.435546875, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 13339.171875, "memory.current_MiB": 7805.3359375 }, "finite": true, "input_sha256": "472ba979ad46ce462cc2cceeb0d218314362876e03b63f06a136a1c19b3dbab7" } ], "load_s": 14.997931960970163, "ready": { "VmHWM_MiB": 8590.87109375, "VmRSS_MiB": 6958.015625, "Pss_MiB": 5508.765625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 13359.421875, "memory.current_MiB": 7785.21484375 }, "providers": [ "TensorrtExecutionProvider", "CUDAExecutionProvider", "CPUExecutionProvider" ], "node_provider_counts": { "TensorrtExecutionProvider": 1 }, "end": { "VmHWM_MiB": 8590.87109375, "VmRSS_MiB": 6979.1796875, "Pss_MiB": 5510.47265625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 13339.171875, "memory.current_MiB": 7805.3359375 } }, "offline": { "arguments": { "self_check": false, "mode": "graph", "provider": "trt", "model": "/app/offline-bundle/model.onnx", "tokens": null, "cache": "/app/offline-bundle", "audio": null, "length": 3, "frames": "50", "warmup": 0, "repeats": 1, "threads": 2, "profile": true, "dump_context": null, "out": "/bench/result.json" }, "utc": "2026-09-05T16:33:25Z", "machine": "aarch64", "ort": "1.24.0", "numpy": "1.26.4", "model_sha256": "4cc9c486d3156800a4a5eb0257678aae4e1803c57eca66fd9623e55e6fd2a6db", "start": { "VmHWM_MiB": 44.48828125, "VmRSS_MiB": 44.47265625, "Pss_MiB": 34.55078125, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18562.2578125, "memory.current_MiB": 2778.234375 }, "rows": [ { "frames": 50, "p50_ms": 42.28810127824545, "p95_ms": 42.28810127824545, "samples_ms": [ 42.28810127824545 ], "memory": { "VmHWM_MiB": 2263.95703125, "VmRSS_MiB": 2128.1640625, "Pss_MiB": 836.40234375, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17841.01171875, "memory.current_MiB": 3496.23046875 }, "finite": true, "input_sha256": "472ba979ad46ce462cc2cceeb0d218314362876e03b63f06a136a1c19b3dbab7" } ], "load_s": 7.9305646400898695, "ready": { "VmHWM_MiB": 2263.95703125, "VmRSS_MiB": 2098.2109375, "Pss_MiB": 825.22265625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17848.94921875, "memory.current_MiB": 3486.60546875 }, "providers": [ "TensorrtExecutionProvider", "CUDAExecutionProvider", "CPUExecutionProvider" ], "node_provider_counts": { "TensorrtExecutionProvider": 1 }, "end": { "VmHWM_MiB": 2263.95703125, "VmRSS_MiB": 2128.39453125, "Pss_MiB": 836.548828125, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17840.02734375, "memory.current_MiB": 3496.3125 } } }, "recognizer": { "cuda": { "3": { "arguments": { "self_check": false, "mode": "recognizer", "provider": "cuda", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/tokens.txt", "cache": "/bench/online-cache", "audio": "/bench/audio", "length": 3, "frames": "50,100,250", "warmup": 10, "repeats": 3, "threads": 2, "profile": false, "dump_context": null, "out": "/bench/result.json" }, "utc": "2026-09-05T15:56:11Z", "machine": "aarch64", "ort": "1.24.0", "numpy": "1.26.4", "model_sha256": "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994", "start": { "VmHWM_MiB": 44.53125, "VmRSS_MiB": 42.65234375, "Pss_MiB": 32.791015625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18743.91015625, "memory.current_MiB": 73.77734375 }, "rows": [ { "seconds": 3, "p50_ms": 45.21427908912301, "p95_ms": 45.33975198864937, "samples_ms": [ 45.33975198864937, 45.21427908912301, 44.48544094339013 ], "texts": [ "帮我导航到内湖科学园区，然后。", "帮我导航到内湖科学园区，然后。", "帮我导航到内湖科学园区，然后。" ], "memory": { "VmHWM_MiB": 2305.90234375, "VmRSS_MiB": 2286.6796875, "Pss_MiB": 1096.01171875, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17741.27734375, "memory.current_MiB": 1091.46484375 }, "audio_sha256": "11034b66e2ae87bdec98dfe7401642387b85262b2e9c222b9e201030dc397a1c" } ], "load_s": 2.871950440108776, "ready": { "VmHWM_MiB": 2305.90234375, "VmRSS_MiB": 2249.09765625, "Pss_MiB": 1077.267578125, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17738.60546875, "memory.current_MiB": 1082.87109375 }, "after_warmup": { "VmHWM_MiB": 2305.90234375, "VmRSS_MiB": 2286.68359375, "Pss_MiB": 1096.009765625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17740.9609375, "memory.current_MiB": 1091.57421875 }, "end": { "VmHWM_MiB": 2305.90234375, "VmRSS_MiB": 2286.6796875, "Pss_MiB": 1096.01171875, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17741.27734375, "memory.current_MiB": 1091.46484375 } }, "6": { "arguments": { "self_check": false, "mode": "recognizer", "provider": "cuda", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/tokens.txt", "cache": "/bench/online-cache", "audio": "/bench/audio", "length": 6, "frames": "50,100,250", "warmup": 10, "repeats": 3, "threads": 2, "profile": false, "dump_context": null, "out": "/bench/result.json" }, "utc": "2026-09-05T15:56:16Z", "machine": "aarch64", "ort": "1.24.0", "numpy": "1.26.4", "model_sha256": "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994", "start": { "VmHWM_MiB": 46.296875, "VmRSS_MiB": 44.6484375, "Pss_MiB": 34.6953125, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18753.5078125, "memory.current_MiB": 91.71484375 }, "rows": [ { "seconds": 6, "p50_ms": 54.43830508738756, "p95_ms": 81.13740617409348, "samples_ms": [ 81.13740617409348, 54.43830508738756, 54.2326713912189 ], "texts": [ "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥。", "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥。", "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥。" ], "memory": { "VmHWM_MiB": 2287.26953125, "VmRSS_MiB": 2128.37890625, "Pss_MiB": 937.5078125, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17911.6328125, "memory.current_MiB": 937.86328125 }, "audio_sha256": "9566be6ccf6c7307830514e0fe5be86250c493efaef2800f27bfccc1ca1850b0" } ], "load_s": 2.756065231282264, "ready": { "VmHWM_MiB": 2287.26953125, "VmRSS_MiB": 2086.69140625, "Pss_MiB": 914.931640625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17912.31640625, "memory.current_MiB": 933.36328125 }, "after_warmup": { "VmHWM_MiB": 2287.26953125, "VmRSS_MiB": 2128.171875, "Pss_MiB": 937.423828125, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17911.83203125, "memory.current_MiB": 937.86328125 }, "end": { "VmHWM_MiB": 2287.26953125, "VmRSS_MiB": 2128.37890625, "Pss_MiB": 937.5078125, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17911.6328125, "memory.current_MiB": 937.86328125 } }, "15": { "arguments": { "self_check": false, "mode": "recognizer", "provider": "cuda", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/tokens.txt", "cache": "/bench/online-cache", "audio": "/bench/audio", "length": 15, "frames": "50,100,250", "warmup": 10, "repeats": 3, "threads": 2, "profile": false, "dump_context": null, "out": "/bench/result.json" }, "utc": "2026-09-05T15:56:20Z", "machine": "aarch64", "ort": "1.24.0", "numpy": "1.26.4", "model_sha256": "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994", "start": { "VmHWM_MiB": 46.3046875, "VmRSS_MiB": 44.65625, "Pss_MiB": 34.787109375, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18758.93359375, "memory.current_MiB": 91.87109375 }, "rows": [ { "seconds": 15, "p50_ms": 87.73670298978686, "p95_ms": 107.93649218976498, "samples_ms": [ 107.93649218976498, 87.49062102288008, 87.73670298978686 ], "texts": [ "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥，接着把冷气打开，温度调到24度，风量开大。帮我导航到内湖科学园区，然后选最快的路线。不要走。", "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥，接着把冷气打开，温度调到24度，风量开大。帮我导航到内湖科学园区，然后选最快的路线。不要走。", "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥，接着把冷气打开，温度调到24度，风量开大。帮我导航到内湖科学园区，然后选最快的路线。不要走。" ], "memory": { "VmHWM_MiB": 2337.26953125, "VmRSS_MiB": 2337.26953125, "Pss_MiB": 1146.6875, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17705.5859375, "memory.current_MiB": 1147.44921875 }, "audio_sha256": "4950a411664da365f4b8cfe13838f881430f01f2fed8b2c3335719f6b791ff9f" } ], "load_s": 2.820148492231965, "ready": { "VmHWM_MiB": 2295.58984375, "VmRSS_MiB": 2295.58984375, "Pss_MiB": 1123.912109375, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17710.33203125, "memory.current_MiB": 1142.6875 }, "after_warmup": { "VmHWM_MiB": 2337.1015625, "VmRSS_MiB": 2337.1015625, "Pss_MiB": 1146.435546875, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17705.65625, "memory.current_MiB": 1147.07421875 }, "end": { "VmHWM_MiB": 2337.26953125, "VmRSS_MiB": 2337.26953125, "Pss_MiB": 1146.6875, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17705.5859375, "memory.current_MiB": 1147.44921875 } } }, "online": { "3": { "arguments": { "self_check": false, "mode": "recognizer", "provider": "trt", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/tokens.txt", "cache": "/bench/online-cache", "audio": "/bench/audio", "length": 3, "frames": "50,100,250", "warmup": 10, "repeats": 3, "threads": 2, "profile": false, "dump_context": null, "out": "/bench/result.json" }, "utc": "2026-09-05T15:57:12Z", "machine": "aarch64", "ort": "1.24.0", "numpy": "1.26.4", "model_sha256": "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994", "start": { "VmHWM_MiB": 46.28515625, "VmRSS_MiB": 44.63671875, "Pss_MiB": 34.77734375, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18742.40625, "memory.current_MiB": 552.8671875 }, "rows": [ { "seconds": 3, "p50_ms": 28.106088284403086, "p95_ms": 28.504072688519955, "samples_ms": [ 28.504072688519955, 28.106088284403086, 28.027432039380074 ], "texts": [ "帮我导航到内湖科学园区，然后。", "帮我导航到内湖科学园区，然后。", "帮我导航到内湖科学园区，然后。" ], "memory": { "VmHWM_MiB": 8649.32421875, "VmRSS_MiB": 7029.03515625, "Pss_MiB": 5564.865234375, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 13499.1640625, "memory.current_MiB": 6721.2265625 }, "audio_sha256": "11034b66e2ae87bdec98dfe7401642387b85262b2e9c222b9e201030dc397a1c" } ], "load_s": 66.58305451786146, "ready": { "VmHWM_MiB": 8649.32421875, "VmRSS_MiB": 7001.546875, "Pss_MiB": 5561.640625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 13521.28515625, "memory.current_MiB": 6699.1015625 }, "after_warmup": { "VmHWM_MiB": 8649.32421875, "VmRSS_MiB": 7029.03515625, "Pss_MiB": 5564.865234375, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 13499.1640625, "memory.current_MiB": 6721.2265625 }, "end": { "VmHWM_MiB": 8649.32421875, "VmRSS_MiB": 7029.03515625, "Pss_MiB": 5564.865234375, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 13499.1640625, "memory.current_MiB": 6721.2265625 } }, "6": { "arguments": { "self_check": false, "mode": "recognizer", "provider": "trt", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/tokens.txt", "cache": "/bench/online-cache", "audio": "/bench/audio", "length": 6, "frames": "50,100,250", "warmup": 10, "repeats": 3, "threads": 2, "profile": false, "dump_context": null, "out": "/bench/result.json" }, "utc": "2026-09-05T15:58:22Z", "machine": "aarch64", "ort": "1.24.0", "numpy": "1.26.4", "model_sha256": "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994", "start": { "VmHWM_MiB": 44.30078125, "VmRSS_MiB": 42.65234375, "Pss_MiB": 32.7041015625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18765.609375, "memory.current_MiB": 1436.4140625 }, "rows": [ { "seconds": 6, "p50_ms": 39.52630516141653, "p95_ms": 54.66651590541005, "samples_ms": [ 54.66651590541005, 39.52630516141653, 39.51395396143198 ], "texts": [ "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥。", "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥。", "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥。" ], "memory": { "VmHWM_MiB": 8601.890625, "VmRSS_MiB": 7000.1484375, "Pss_MiB": 5519.109375, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 13491.02734375, "memory.current_MiB": 6778.9921875 }, "audio_sha256": "9566be6ccf6c7307830514e0fe5be86250c493efaef2800f27bfccc1ca1850b0" } ], "load_s": 14.710078530944884, "ready": { "VmHWM_MiB": 8601.890625, "VmRSS_MiB": 6955.86328125, "Pss_MiB": 5514.31640625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 13529.6953125, "memory.current_MiB": 6740.87109375 }, "after_warmup": { "VmHWM_MiB": 8601.890625, "VmRSS_MiB": 6984.1484375, "Pss_MiB": 5519.109375, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 13507.02734375, "memory.current_MiB": 6762.9921875 }, "end": { "VmHWM_MiB": 8601.890625, "VmRSS_MiB": 7000.1484375, "Pss_MiB": 5519.025390625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 13490.53515625, "memory.current_MiB": 6778.9921875 } }, "15": { "arguments": { "self_check": false, "mode": "recognizer", "provider": "trt", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/tokens.txt", "cache": "/bench/online-cache", "audio": "/bench/audio", "length": 15, "frames": "50,100,250", "warmup": 10, "repeats": 3, "threads": 2, "profile": false, "dump_context": null, "out": "/bench/result.json" }, "utc": "2026-09-05T15:58:39Z", "machine": "aarch64", "ort": "1.24.0", "numpy": "1.26.4", "model_sha256": "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994", "start": { "VmHWM_MiB": 44.5390625, "VmRSS_MiB": 42.66015625, "Pss_MiB": 32.794921875, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18745.3828125, "memory.current_MiB": 1509.46484375 }, "rows": [ { "seconds": 15, "p50_ms": 81.88698580488563, "p95_ms": 89.79894407093525, "samples_ms": [ 89.79894407093525, 81.88698580488563, 81.79856790229678 ], "texts": [ "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥，接着把冷气打开，温度调到24度，风量开大。帮我导航到内湖科学园区，然后选最快的路线。不要走。", "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥，接着把冷气打开，温度调到24度，风量开大。帮我导航到内湖科学园区，然后选最快的路线。不要走。", "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥，接着把冷气打开，温度调到24度，风量开大。帮我导航到内湖科学园区，然后选最快的路线。不要走。" ], "memory": { "VmHWM_MiB": 8601.8515625, "VmRSS_MiB": 7015.89453125, "Pss_MiB": 5518.923828125, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 13410.921875, "memory.current_MiB": 6862.86328125 }, "audio_sha256": "4950a411664da365f4b8cfe13838f881430f01f2fed8b2c3335719f6b791ff9f" } ], "load_s": 14.863332874607295, "ready": { "VmHWM_MiB": 8601.8515625, "VmRSS_MiB": 6955.61328125, "Pss_MiB": 5514.12890625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 13465.3359375, "memory.current_MiB": 6808.7421875 }, "after_warmup": { "VmHWM_MiB": 8601.8515625, "VmRSS_MiB": 6983.8984375, "Pss_MiB": 5518.921875, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 13442.921875, "memory.current_MiB": 6830.86328125 }, "end": { "VmHWM_MiB": 8601.8515625, "VmRSS_MiB": 7015.89453125, "Pss_MiB": 5518.923828125, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 13410.921875, "memory.current_MiB": 6862.86328125 } } }, "offline": { "3": { "arguments": { "self_check": false, "mode": "recognizer", "provider": "trt", "model": "/app/offline-bundle/model.onnx", "tokens": "/app/offline-bundle/tokens.txt", "cache": "offline-bundle", "audio": "/bench/audio", "length": 3, "frames": "50,100,250", "warmup": 10, "repeats": 3, "threads": 2, "profile": false, "dump_context": null, "out": "/bench/result.json" }, "utc": "2026-09-05T15:59:43Z", "machine": "aarch64", "ort": "1.24.0", "numpy": "1.26.4", "model_sha256": "4cc9c486d3156800a4a5eb0257678aae4e1803c57eca66fd9623e55e6fd2a6db", "start": { "VmHWM_MiB": 42.546875, "VmRSS_MiB": 42.53125, "Pss_MiB": 32.666015625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18744.16015625, "memory.current_MiB": 2588.48046875 }, "rows": [ { "seconds": 3, "p50_ms": 27.810572180896997, "p95_ms": 28.070413041859865, "samples_ms": [ 28.070413041859865, 27.810572180896997, 27.76234783232212 ], "texts": [ "帮我导航到内湖科学园区，然后。", "帮我导航到内湖科学园区，然后。", "帮我导航到内湖科学园区，然后。" ], "memory": { "VmHWM_MiB": 2273.27734375, "VmRSS_MiB": 2130.9453125, "Pss_MiB": 845.373046875, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18029.03515625, "memory.current_MiB": 3315.328125 }, "audio_sha256": "11034b66e2ae87bdec98dfe7401642387b85262b2e9c222b9e201030dc397a1c" } ], "load_s": 7.9453939837403595, "ready": { "VmHWM_MiB": 2273.27734375, "VmRSS_MiB": 2087.0390625, "Pss_MiB": 825.400390625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18044.3515625, "memory.current_MiB": 3299.953125 }, "after_warmup": { "VmHWM_MiB": 2273.27734375, "VmRSS_MiB": 2130.41796875, "Pss_MiB": 844.8203125, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18029.5234375, "memory.current_MiB": 3314.703125 }, "end": { "VmHWM_MiB": 2273.27734375, "VmRSS_MiB": 2130.9453125, "Pss_MiB": 845.380859375, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18029.03515625, "memory.current_MiB": 3315.328125 } }, "6": { "arguments": { "self_check": false, "mode": "recognizer", "provider": "trt", "model": "/app/offline-bundle/model.onnx", "tokens": "/app/offline-bundle/tokens.txt", "cache": "offline-bundle", "audio": "/bench/audio", "length": 6, "frames": "50,100,250", "warmup": 10, "repeats": 3, "threads": 2, "profile": false, "dump_context": null, "out": "/bench/result.json" }, "utc": "2026-09-05T15:59:53Z", "machine": "aarch64", "ort": "1.24.0", "numpy": "1.26.4", "model_sha256": "4cc9c486d3156800a4a5eb0257678aae4e1803c57eca66fd9623e55e6fd2a6db", "start": { "VmHWM_MiB": 44.55078125, "VmRSS_MiB": 44.53515625, "Pss_MiB": 34.58203125, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18739.3984375, "memory.current_MiB": 2590.8515625 }, "rows": [ { "seconds": 6, "p50_ms": 39.595384150743484, "p95_ms": 52.709611132740974, "samples_ms": [ 52.709611132740974, 39.595384150743484, 38.95317483693361 ], "texts": [ "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥。", "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥。", "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥。" ], "memory": { "VmHWM_MiB": 2275.328125, "VmRSS_MiB": 2163.16796875, "Pss_MiB": 861.40625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18002.3671875, "memory.current_MiB": 3331.37109375 }, "audio_sha256": "9566be6ccf6c7307830514e0fe5be86250c493efaef2800f27bfccc1ca1850b0" } ], "load_s": 7.931633648928255, "ready": { "VmHWM_MiB": 2275.328125, "VmRSS_MiB": 2088.890625, "Pss_MiB": 827.248046875, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18034.1171875, "memory.current_MiB": 3301.74609375 }, "after_warmup": { "VmHWM_MiB": 2275.328125, "VmRSS_MiB": 2132.53515625, "Pss_MiB": 846.908203125, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18018.796875, "memory.current_MiB": 3316.87109375 }, "end": { "VmHWM_MiB": 2275.328125, "VmRSS_MiB": 2163.16796875, "Pss_MiB": 861.41015625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18002.3671875, "memory.current_MiB": 3331.37109375 } }, "15": { "arguments": { "self_check": false, "mode": "recognizer", "provider": "trt", "model": "/app/offline-bundle/model.onnx", "tokens": "/app/offline-bundle/tokens.txt", "cache": "offline-bundle", "audio": "/bench/audio", "length": 15, "frames": "50,100,250", "warmup": 10, "repeats": 3, "threads": 2, "profile": false, "dump_context": null, "out": "/bench/result.json" }, "utc": "2026-09-05T16:00:02Z", "machine": "aarch64", "ort": "1.24.0", "numpy": "1.26.4", "model_sha256": "4cc9c486d3156800a4a5eb0257678aae4e1803c57eca66fd9623e55e6fd2a6db", "start": { "VmHWM_MiB": 42.5546875, "VmRSS_MiB": 42.5390625, "Pss_MiB": 32.58984375, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18748.109375, "memory.current_MiB": 2589.02734375 }, "rows": [ { "seconds": 15, "p50_ms": 81.2513642013073, "p95_ms": 86.93114714697003, "samples_ms": [ 86.93114714697003, 81.2513642013073, 80.75357088819146 ], "texts": [ "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥，接着把冷气打开，温度调到24度，风量开大。帮我导航到内湖科学园区，然后选最快的路线。不要走。", "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥，接着把冷气打开，温度调到24度，风量开大。帮我导航到内湖科学园区，然后选最快的路线。不要走。", "帮我导航到内湖科学园区，然后选最快的路线，不要走高架桥，接着把冷气打开，温度调到24度，风量开大。帮我导航到内湖科学园区，然后选最快的路线。不要走。" ], "memory": { "VmHWM_MiB": 2275.328125, "VmRSS_MiB": 2195.58984375, "Pss_MiB": 877.9765625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17984.2265625, "memory.current_MiB": 3347.98046875 }, "audio_sha256": "4950a411664da365f4b8cfe13838f881430f01f2fed8b2c3335719f6b791ff9f" } ], "load_s": 7.93012819159776, "ready": { "VmHWM_MiB": 2275.328125, "VmRSS_MiB": 2089.0078125, "Pss_MiB": 827.38671875, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18026.4375, "memory.current_MiB": 3301.98046875 }, "after_warmup": { "VmHWM_MiB": 2275.328125, "VmRSS_MiB": 2131.89453125, "Pss_MiB": 846.16015625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18016.1015625, "memory.current_MiB": 3316.10546875 }, "end": { "VmHWM_MiB": 2275.328125, "VmRSS_MiB": 2195.58984375, "Pss_MiB": 877.98046875, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 17984.2265625, "memory.current_MiB": 3347.98046875 } } } }, "service": { "cuda": { "rounds": [ { "environment": { "target": "ws://127.0.0.1:19342/stt/sensevoice", "uri": "ws://127.0.0.1:19342/stt/sensevoice", "warmup_runs": 3, "single_connection_series": true }, "generated": "2026-09-05T16:20:59", "results": [ { "feed": "bulk", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 96.26, "server_rtf": 0.021, "server_ttft_ms": 185.38, "client_final_latency_ms": 175.86, "interim_count": 0, "max_interim_gap_s": null, "feed_secs": 0.02, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "bulk", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 111.8, "server_rtf": 0.013, "server_ttft_ms": 214.37, "client_final_latency_ms": 180.91, "interim_count": 0, "max_interim_gap_s": null, "feed_secs": 0.04, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "bulk", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 138.05, "server_rtf": 0.011, "server_ttft_ms": 201.47, "client_final_latency_ms": 298.39, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.1, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null }, { "feed": "realtime", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 83.96, "server_rtf": 0.139, "server_ttft_ms": 71.3, "client_final_latency_ms": 90.33, "interim_count": 7, "max_interim_gap_s": 0.527, "feed_secs": 3.16, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "realtime", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 84.01, "server_rtf": 0.14, "server_ttft_ms": 72.08, "client_final_latency_ms": 98.03, "interim_count": 11, "max_interim_gap_s": 1.032, "feed_secs": 6.35, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "realtime", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 88.75, "server_rtf": 0.149, "server_ttft_ms": 71.21, "client_final_latency_ms": 105.45, "interim_count": 28, "max_interim_gap_s": 1.026, "feed_secs": 15.86, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null } ] }, { "environment": { "target": "ws://127.0.0.1:19342/stt/sensevoice", "uri": "ws://127.0.0.1:19342/stt/sensevoice", "warmup_runs": 3, "single_connection_series": true }, "generated": "2026-09-05T16:22:13", "results": [ { "feed": "bulk", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 133.31, "server_rtf": 0.02, "server_ttft_ms": 523.75, "client_final_latency_ms": 499.18, "interim_count": 0, "max_interim_gap_s": null, "feed_secs": 0.03, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "bulk", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 97.35, "server_rtf": 0.022, "server_ttft_ms": 100.16, "client_final_latency_ms": 184.51, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.04, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "bulk", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 131.55, "server_rtf": 0.01, "server_ttft_ms": 104.97, "client_final_latency_ms": 269.62, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.11, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null }, { "feed": "realtime", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 84.63, "server_rtf": 0.135, "server_ttft_ms": 71.46, "client_final_latency_ms": 89.57, "interim_count": 6, "max_interim_gap_s": 0.509, "feed_secs": 3.17, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "realtime", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 86.05, "server_rtf": 0.133, "server_ttft_ms": 70.3, "client_final_latency_ms": 101.03, "interim_count": 11, "max_interim_gap_s": 1.02, "feed_secs": 6.37, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "realtime", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 86.38, "server_rtf": 0.145, "server_ttft_ms": 71.64, "client_final_latency_ms": 104.71, "interim_count": 27, "max_interim_gap_s": 1.033, "feed_secs": 15.87, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null } ] }, { "environment": { "target": "ws://127.0.0.1:19342/stt/sensevoice", "uri": "ws://127.0.0.1:19342/stt/sensevoice", "warmup_runs": 3, "single_connection_series": true }, "generated": "2026-09-05T16:23:27", "results": [ { "feed": "bulk", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 95.83, "server_rtf": 0.025, "server_ttft_ms": 135.17, "client_final_latency_ms": 717.47, "interim_count": 0, "max_interim_gap_s": null, "feed_secs": 0.03, "text": "幫我導航到內湖科學園區。然後。", "error": null }, { "feed": "bulk", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 98.95, "server_rtf": 0.022, "server_ttft_ms": 97.85, "client_final_latency_ms": 165.25, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.05, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "bulk", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 124.84, "server_rtf": 0.01, "server_ttft_ms": 202.49, "client_final_latency_ms": 277.5, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.09, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null }, { "feed": "realtime", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 84.65, "server_rtf": 0.135, "server_ttft_ms": 70.94, "client_final_latency_ms": 89.33, "interim_count": 6, "max_interim_gap_s": 0.515, "feed_secs": 3.18, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "realtime", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 92.39, "server_rtf": 0.14, "server_ttft_ms": 76.69, "client_final_latency_ms": 109.5, "interim_count": 10, "max_interim_gap_s": 1.032, "feed_secs": 6.35, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "realtime", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 125.89, "server_rtf": 0.132, "server_ttft_ms": 71.53, "client_final_latency_ms": 147.38, "interim_count": 25, "max_interim_gap_s": 1.314, "feed_secs": 15.87, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null } ] } ], "summary": [ { "feed": "bulk", "seconds": 3, "median_ms": 96.26, "min_ms": 95.83, "max_ms": 133.31, "samples_ms": [ 96.26, 133.31, 95.83 ], "vs_cuda_percent": 0.0 }, { "feed": "bulk", "seconds": 6, "median_ms": 98.95, "min_ms": 97.35, "max_ms": 111.8, "samples_ms": [ 111.8, 97.35, 98.95 ], "vs_cuda_percent": 0.0 }, { "feed": "bulk", "seconds": 15, "median_ms": 131.55, "min_ms": 124.84, "max_ms": 138.05, "samples_ms": [ 138.05, 131.55, 124.84 ], "vs_cuda_percent": 0.0 }, { "feed": "realtime", "seconds": 3, "median_ms": 84.63, "min_ms": 83.96, "max_ms": 84.65, "samples_ms": [ 83.96, 84.63, 84.65 ], "vs_cuda_percent": 0.0 }, { "feed": "realtime", "seconds": 6, "median_ms": 86.05, "min_ms": 84.01, "max_ms": 92.39, "samples_ms": [ 84.01, 86.05, 92.39 ], "vs_cuda_percent": 0.0 }, { "feed": "realtime", "seconds": 15, "median_ms": 88.75, "min_ms": 86.38, "max_ms": 125.89, "samples_ms": [ 88.75, 86.38, 125.89 ], "vs_cuda_percent": 0.0 } ], "memory": { "ready": { "utc": 1788625183.4753273, "VmHWM_MiB": 3898.6171875, "VmRSS_MiB": 3894.796875, "Pss_MiB": 1655.85546875, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 16081.5390625, "memory.current_MiB": 3807.69921875 }, "end": { "utc": 1788625408.753695, "VmHWM_MiB": 4069.36328125, "VmRSS_MiB": 4021.7578125, "Pss_MiB": 1760.474609375, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 15959.765625, "memory.current_MiB": 3936.51953125 } } }, "online": { "rounds": [ { "environment": { "target": "ws://127.0.0.1:19342/stt/sensevoice", "uri": "ws://127.0.0.1:19342/stt/sensevoice", "warmup_runs": 3, "single_connection_series": true }, "generated": "2026-09-05T16:04:18", "results": [ { "feed": "bulk", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 55.55, "server_rtf": 0.01, "server_ttft_ms": 131.45, "client_final_latency_ms": 106.54, "interim_count": 0, "max_interim_gap_s": null, "feed_secs": 0.03, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "bulk", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 61.1, "server_rtf": 0.011, "server_ttft_ms": 84.16, "client_final_latency_ms": 171.15, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.03, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "bulk", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 119.26, "server_rtf": 0.007, "server_ttft_ms": 42.23, "client_final_latency_ms": 242.82, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.13, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null }, { "feed": "realtime", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 46.62, "server_rtf": 0.069, "server_ttft_ms": 38.27, "client_final_latency_ms": 61.62, "interim_count": 6, "max_interim_gap_s": 0.509, "feed_secs": 3.16, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "realtime", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 45.45, "server_rtf": 0.067, "server_ttft_ms": 38.28, "client_final_latency_ms": 85.34, "interim_count": 12, "max_interim_gap_s": 1.02, "feed_secs": 6.34, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "realtime", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 112.91, "server_rtf": 0.077, "server_ttft_ms": 38.31, "client_final_latency_ms": 157.1, "interim_count": 28, "max_interim_gap_s": 1.024, "feed_secs": 15.8, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null } ] }, { "environment": { "target": "ws://127.0.0.1:19342/stt/sensevoice", "uri": "ws://127.0.0.1:19342/stt/sensevoice", "warmup_runs": 3, "single_connection_series": true }, "generated": "2026-09-05T16:05:33", "results": [ { "feed": "bulk", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 52.87, "server_rtf": 0.02, "server_ttft_ms": 43.91, "client_final_latency_ms": 90.76, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.03, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "bulk", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 57.59, "server_rtf": 0.011, "server_ttft_ms": 72.74, "client_final_latency_ms": 160.11, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.03, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "bulk", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 107.56, "server_rtf": 0.006, "server_ttft_ms": 89.98, "client_final_latency_ms": 257.05, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.09, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null }, { "feed": "realtime", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 44.91, "server_rtf": 0.06, "server_ttft_ms": 38.17, "client_final_latency_ms": 61.47, "interim_count": 5, "max_interim_gap_s": 0.808, "feed_secs": 3.15, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "realtime", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 44.38, "server_rtf": 0.068, "server_ttft_ms": 38.29, "client_final_latency_ms": 86.37, "interim_count": 11, "max_interim_gap_s": 1.027, "feed_secs": 6.32, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "realtime", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 68.59, "server_rtf": 0.076, "server_ttft_ms": 38.21, "client_final_latency_ms": 92.25, "interim_count": 26, "max_interim_gap_s": 1.311, "feed_secs": 15.88, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null } ] }, { "environment": { "target": "ws://127.0.0.1:19342/stt/sensevoice", "uri": "ws://127.0.0.1:19342/stt/sensevoice", "warmup_runs": 3, "single_connection_series": true }, "generated": "2026-09-05T16:06:58", "results": [ { "feed": "bulk", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 54.17, "server_rtf": 0.01, "server_ttft_ms": 138.43, "client_final_latency_ms": 126.34, "interim_count": 0, "max_interim_gap_s": null, "feed_secs": 0.02, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "bulk", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 58.72, "server_rtf": 0.011, "server_ttft_ms": 53.53, "client_final_latency_ms": 149.48, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.05, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "bulk", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 109.34, "server_rtf": 0.006, "server_ttft_ms": 58.12, "client_final_latency_ms": 245.69, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.1, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null }, { "feed": "realtime", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 58.38, "server_rtf": 0.06, "server_ttft_ms": 38.12, "client_final_latency_ms": 579.38, "interim_count": 6, "max_interim_gap_s": 0.521, "feed_secs": 3.15, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "realtime", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 44.99, "server_rtf": 0.067, "server_ttft_ms": 38.19, "client_final_latency_ms": 86.91, "interim_count": 12, "max_interim_gap_s": 1.027, "feed_secs": 6.34, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "realtime", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 69.03, "server_rtf": 0.075, "server_ttft_ms": 37.85, "client_final_latency_ms": 94.2, "interim_count": 26, "max_interim_gap_s": 1.102, "feed_secs": 15.83, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null } ] } ], "summary": [ { "feed": "bulk", "seconds": 3, "median_ms": 54.17, "min_ms": 52.87, "max_ms": 55.55, "samples_ms": [ 55.55, 52.87, 54.17 ], "vs_cuda_percent": -43.72532723872844 }, { "feed": "bulk", "seconds": 6, "median_ms": 58.72, "min_ms": 57.59, "max_ms": 61.1, "samples_ms": [ 61.1, 57.59, 58.72 ], "vs_cuda_percent": -40.65689742294089 }, { "feed": "bulk", "seconds": 15, "median_ms": 109.34, "min_ms": 107.56, "max_ms": 119.26, "samples_ms": [ 119.26, 107.56, 109.34 ], "vs_cuda_percent": -16.88331432915242 }, { "feed": "realtime", "seconds": 3, "median_ms": 46.62, "min_ms": 44.91, "max_ms": 58.38, "samples_ms": [ 46.62, 44.91, 58.38 ], "vs_cuda_percent": -44.91315136476427 }, { "feed": "realtime", "seconds": 6, "median_ms": 44.99, "min_ms": 44.38, "max_ms": 45.45, "samples_ms": [ 45.45, 44.38, 44.99 ], "vs_cuda_percent": -47.71644392794886 }, { "feed": "realtime", "seconds": 15, "median_ms": 69.03, "min_ms": 68.59, "max_ms": 112.91, "samples_ms": [ 112.91, 68.59, 69.03 ], "vs_cuda_percent": -22.21971830985915 } ], "memory": { "ready": { "utc": 1788624171.9871628, "VmHWM_MiB": 9368.82421875, "VmRSS_MiB": 8582.1796875, "Pss_MiB": 6162.4365234375, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18754.93359375, "memory.current_MiB": 8487.48828125 }, "end": { "utc": 1788624418.7563136, "VmHWM_MiB": 9368.82421875, "VmRSS_MiB": 8760.10546875, "Pss_MiB": 6265.8359375, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18674.76953125, "memory.current_MiB": 8592.06640625 } } }, "offline": { "rounds": [ { "environment": { "target": "ws://127.0.0.1:19342/stt/sensevoice", "uri": "ws://127.0.0.1:19342/stt/sensevoice", "warmup_runs": 3, "single_connection_series": true }, "generated": "2026-09-05T16:25:35", "results": [ { "feed": "bulk", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 59.33, "server_rtf": 0.011, "server_ttft_ms": 127.7, "client_final_latency_ms": 116.13, "interim_count": 0, "max_interim_gap_s": null, "feed_secs": 0.01, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "bulk", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 65.61, "server_rtf": 0.012, "server_ttft_ms": 81.4, "client_final_latency_ms": 164.26, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.03, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "bulk", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 125.06, "server_rtf": 0.007, "server_ttft_ms": 44.29, "client_final_latency_ms": 254.66, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.12, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null }, { "feed": "realtime", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 47.78, "server_rtf": 0.069, "server_ttft_ms": 38.25, "client_final_latency_ms": 61.39, "interim_count": 7, "max_interim_gap_s": 0.519, "feed_secs": 3.17, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "realtime", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 45.75, "server_rtf": 0.068, "server_ttft_ms": 38.49, "client_final_latency_ms": 92.34, "interim_count": 12, "max_interim_gap_s": 1.017, "feed_secs": 6.33, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "realtime", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 68.03, "server_rtf": 0.078, "server_ttft_ms": 37.23, "client_final_latency_ms": 87.78, "interim_count": 28, "max_interim_gap_s": 1.019, "feed_secs": 15.84, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開。幫我導航到內湖科學園區，然後選最快的路線，不要走。", "error": null } ] }, { "environment": { "target": "ws://127.0.0.1:19342/stt/sensevoice", "uri": "ws://127.0.0.1:19342/stt/sensevoice", "warmup_runs": 3, "single_connection_series": true }, "generated": "2026-09-05T16:26:57", "results": [ { "feed": "bulk", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 54.1, "server_rtf": 0.02, "server_ttft_ms": 52.36, "client_final_latency_ms": 89.6, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.04, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "bulk", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 59.19, "server_rtf": 0.011, "server_ttft_ms": 72.04, "client_final_latency_ms": 150.85, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.03, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "bulk", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 100.35, "server_rtf": 0.006, "server_ttft_ms": 220.49, "client_final_latency_ms": 258.63, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.1, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開。幫我導航到內湖科學園區，然後選最快的路線，不要走。", "error": null }, { "feed": "realtime", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 42.3, "server_rtf": 0.069, "server_ttft_ms": 38.24, "client_final_latency_ms": 75.82, "interim_count": 6, "max_interim_gap_s": 0.512, "feed_secs": 3.17, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "realtime", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 45.65, "server_rtf": 0.068, "server_ttft_ms": 38.69, "client_final_latency_ms": 81.54, "interim_count": 12, "max_interim_gap_s": 1.025, "feed_secs": 6.31, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "realtime", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 66.83, "server_rtf": 0.075, "server_ttft_ms": 42.89, "client_final_latency_ms": 87.42, "interim_count": 27, "max_interim_gap_s": 1.024, "feed_secs": 15.79, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null } ] }, { "environment": { "target": "ws://127.0.0.1:19342/stt/sensevoice", "uri": "ws://127.0.0.1:19342/stt/sensevoice", "warmup_runs": 3, "single_connection_series": true }, "generated": "2026-09-05T16:28:10", "results": [ { "feed": "bulk", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 51.91, "server_rtf": 0.01, "server_ttft_ms": 157.94, "client_final_latency_ms": 230.52, "interim_count": 0, "max_interim_gap_s": null, "feed_secs": 0.02, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "bulk", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 59.23, "server_rtf": 0.011, "server_ttft_ms": 70.03, "client_final_latency_ms": 143.9, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.05, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "bulk", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 107.73, "server_rtf": 0.006, "server_ttft_ms": 51.09, "client_final_latency_ms": 239.97, "interim_count": 1, "max_interim_gap_s": null, "feed_secs": 0.1, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開大。幫我導航到內湖科學園區，然後選最快的路線。不要走。", "error": null }, { "feed": "realtime", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 45.98, "server_rtf": 0.068, "server_ttft_ms": 37.97, "client_final_latency_ms": 59.57, "interim_count": 7, "max_interim_gap_s": 0.516, "feed_secs": 3.16, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "realtime", "target_s": 6.0, "duration_s": 6.0, "status": "ok", "server_latency_ms": 45.53, "server_rtf": 0.064, "server_ttft_ms": 47.34, "client_final_latency_ms": 88.96, "interim_count": 10, "max_interim_gap_s": 1.022, "feed_secs": 6.31, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋。", "error": null }, { "feed": "realtime", "target_s": 15.0, "duration_s": 15.0, "status": "ok", "server_latency_ms": 67.63, "server_rtf": 0.071, "server_ttft_ms": 37.01, "client_final_latency_ms": 88.04, "interim_count": 26, "max_interim_gap_s": 1.271, "feed_secs": 15.79, "text": "幫我導航到內湖科學園區，然後選最快的路線，不要走高架橋，接着把冷氣打開，溫度調到24度，風量開。幫我導航到內湖科學園區，然後選最快的路線，不要走。", "error": null } ] } ], "summary": [ { "feed": "bulk", "seconds": 3, "median_ms": 54.1, "min_ms": 51.91, "max_ms": 59.33, "samples_ms": [ 59.33, 54.1, 51.91 ], "vs_cuda_percent": -43.7980469561604 }, { "feed": "bulk", "seconds": 6, "median_ms": 59.23, "min_ms": 59.19, "max_ms": 65.61, "samples_ms": [ 65.61, 59.19, 59.23 ], "vs_cuda_percent": -40.14148559878728 }, { "feed": "bulk", "seconds": 15, "median_ms": 107.73, "min_ms": 100.35, "max_ms": 125.06, "samples_ms": [ 125.06, 100.35, 107.73 ], "vs_cuda_percent": -18.107183580387687 }, { "feed": "realtime", "seconds": 3, "median_ms": 45.98, "min_ms": 42.3, "max_ms": 47.78, "samples_ms": [ 47.78, 42.3, 45.98 ], "vs_cuda_percent": -45.669384379061796 }, { "feed": "realtime", "seconds": 6, "median_ms": 45.65, "min_ms": 45.53, "max_ms": 45.75, "samples_ms": [ 45.75, 45.65, 45.53 ], "vs_cuda_percent": -46.949447995351534 }, { "feed": "realtime", "seconds": 15, "median_ms": 67.63, "min_ms": 66.83, "max_ms": 68.03, "samples_ms": [ 68.03, 66.83, 67.63 ], "vs_cuda_percent": -23.797183098591557 } ], "memory": { "ready": { "utc": 1788625462.1301136, "VmHWM_MiB": 4188.4921875, "VmRSS_MiB": 3776.890625, "Pss_MiB": 1443.2216796875, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 16205.67578125, "memory.current_MiB": 3632.59375 }, "end": { "utc": 1788625690.8206105, "VmHWM_MiB": 4188.4921875, "VmRSS_MiB": 3976.3515625, "Pss_MiB": 1567.931640625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 16022.421875, "memory.current_MiB": 3831.87890625 } } } }, "offline_build": { "arguments": { "self_check": false, "mode": "graph", "provider": "trt", "model": "/app/models/sherpa-onnx-sense-voice-kitt-wake-lora-v1/model.onnx", "tokens": null, "cache": "offline-bundle", "audio": null, "length": 3, "frames": "50", "warmup": 0, "repeats": 1, "threads": 2, "profile": false, "dump_context": "/app/offline-bundle/model.onnx", "out": "/bench/result.json" }, "utc": "2026-09-05T15:53:37Z", "machine": "aarch64", "ort": "1.24.0", "numpy": "1.26.4", "model_sha256": "719559b0512db928e18eb3c130a122a55dc10166fb6bd2c8c90f2b6f51064994", "start": { "VmHWM_MiB": 44.44140625, "VmRSS_MiB": 42.79296875, "Pss_MiB": 32.8505859375, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 18649.12890625, "memory.current_MiB": 2643.90234375 }, "rows": [ { "frames": 50, "p50_ms": 42.72760916501284, "p95_ms": 42.72760916501284, "samples_ms": [ 42.72760916501284 ], "memory": { "VmHWM_MiB": 8632.85546875, "VmRSS_MiB": 7016.390625, "Pss_MiB": 5547.4189453125, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 13518.74609375, "memory.current_MiB": 8613.78515625 }, "finite": true, "input_sha256": "472ba979ad46ce462cc2cceeb0d218314362876e03b63f06a136a1c19b3dbab7" } ], "load_s": 67.48600840009749, "ready": { "VmHWM_MiB": 8632.85546875, "VmRSS_MiB": 6995.4140625, "Pss_MiB": 5546.8251953125, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 13538.97265625, "memory.current_MiB": 8593.66015625 }, "providers": [ "TensorrtExecutionProvider", "CUDAExecutionProvider", "CPUExecutionProvider" ], "end": { "VmHWM_MiB": 8632.85546875, "VmRSS_MiB": 7016.390625, "Pss_MiB": 5547.4248046875, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 13518.74609375, "memory.current_MiB": 8613.78125 } }, "online_cold_service": { "result": { "environment": { "target": "ws://127.0.0.1:19342/stt/sensevoice", "uri": "ws://127.0.0.1:19342/stt/sensevoice", "warmup_runs": 0, "single_connection_series": true }, "generated": "2026-09-05T16:30:32", "results": [ { "feed": "bulk", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 60.1, "server_rtf": 0.012, "server_ttft_ms": 229.67, "client_final_latency_ms": 217.74, "interim_count": 0, "max_interim_gap_s": null, "feed_secs": 0.02, "text": "幫我導航到內湖科學園區，然後。", "error": null }, { "feed": "realtime", "target_s": 3.0, "duration_s": 3.0, "status": "ok", "server_latency_ms": 49.07, "server_rtf": 0.072, "server_ttft_ms": 39.82, "client_final_latency_ms": 61.98, "interim_count": 7, "max_interim_gap_s": 0.518, "feed_secs": 3.16, "text": "幫我導航到內湖科學園區，然後。", "error": null } ] }, "memory": { "ready": { "utc": 1788625803.9486284, "VmHWM_MiB": 10275.578125, "VmRSS_MiB": 8649.59765625, "Pss_MiB": 6139.5009765625, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 11418.671875, "memory.current_MiB": 9053.33203125 }, "end": { "utc": 1788625833.3184311, "VmHWM_MiB": 10275.578125, "VmRSS_MiB": 8701.29296875, "Pss_MiB": 6163.4169921875, "MemTotal_MiB": 62840.6796875, "MemAvailable_MiB": 11343.11328125, "memory.current_MiB": 9104.48828125 } } } } }, "history_4090_previous_cycle_run": { "source": "2026-09-04-orin-tensorrt-deploy-recheck-20260905.json", "service": { "cuda": [ { "feed": "bulk", "seconds": 3, "median_ms": 22.45, "min_ms": 21.72, "max_ms": 23.1, "samples_ms": [ 22.45, 23.1, 21.72 ] }, { "feed": "bulk", "seconds": 6, "median_ms": 26.83, "min_ms": 26.65, "max_ms": 32.92, "samples_ms": [ 32.92, 26.65, 26.83 ] }, { "feed": "bulk", "seconds": 15, "median_ms": 72.64, "min_ms": 69.4, "max_ms": 99.66, "samples_ms": [ 99.66, 72.64, 69.4 ] }, { "feed": "realtime", "seconds": 3, "median_ms": 28.21, "min_ms": 21.07, "max_ms": 42.77, "samples_ms": [ 42.77, 28.21, 21.07 ] }, { "feed": "realtime", "seconds": 6, "median_ms": 30.55, "min_ms": 23.32, "max_ms": 33.77, "samples_ms": [ 23.32, 30.55, 33.77 ] }, { "feed": "realtime", "seconds": 15, "median_ms": 74.55, "min_ms": 67.57, "max_ms": 85.37, "samples_ms": [ 74.55, 67.57, 85.37 ] } ], "online": [ { "feed": "bulk", "seconds": 3, "median_ms": 13.34, "min_ms": 13.18, "max_ms": 16.32, "samples_ms": [ 13.18, 16.32, 13.34 ] }, { "feed": "bulk", "seconds": 6, "median_ms": 17.97, "min_ms": 16.9, "max_ms": 19.47, "samples_ms": [ 19.47, 16.9, 17.97 ] }, { "feed": "bulk", "seconds": 15, "median_ms": 75.1, "min_ms": 73.87, "max_ms": 108.68, "samples_ms": [ 75.1, 73.87, 108.68 ] }, { "feed": "realtime", "seconds": 3, "median_ms": 17.98, "min_ms": 11.29, "max_ms": 21.24, "samples_ms": [ 17.98, 11.29, 21.24 ] }, { "feed": "realtime", "seconds": 6, "median_ms": 13.37, "min_ms": 12.29, "max_ms": 21.1, "samples_ms": [ 13.37, 12.29, 21.1 ] }, { "feed": "realtime", "seconds": 15, "median_ms": 54.44, "min_ms": 53.66, "max_ms": 56.77, "samples_ms": [ 53.66, 56.77, 54.44 ] } ] } }, "excluded_attempts": [ "4090 初次服務啟動缺少 audio bind mount，送音前即失敗", "Orin 臨時 CUDA 首次 port-forward 探測後退出，client 未送出有效音訊", "Orin registry HEAD 暫時回 502，臨時 Pod 改用節點既有相同 digest image 後重跑" ], "scripts": [ "deploy.sh", "tests/batch_decode/provider_recheck.py", "tests/batch_decode/hardware_profile.py", "gitignored service-memory.py" ], "final_orin_deployment": { "verified_at": "2026-09-06T00:55:37+08:00", "image": "registry.devops.foxconn.com/srdc/agent/stt/kitt-stt:v5.8.0-main-a16280f-offline-20260906-orin@sha256:1dd50b2f99d40ab5fbe8128e9a0f02405caf888cdc2af568f9dbb8d0b55fb73e", "local_image_sha256": "9bbe96c2c9742644788eb3f8efb921fb2618df14048ea7a0f321cdc2db98d1ac", "pod": "kitt-stt-6dfdfc9cd6-qj56s", "pod_status": "Ready", "restarts": 0, "init_container": "verify-offline-trt-bundle", "init_exit_code": 0, "bundle": "offline-a16280f", "context_model_bytes": 12273, "context_model_sha256": "4cc9c486d3156800a4a5eb0257678aae4e1803c57eca66fd9623e55e6fd2a6db", "engine_bytes": 984975172, "engine_sha256": "5cc07fdc131f87e25c41adabc01583b2abe1b7dcb5658c4950ea01632d5a215d", "runtime_maps": [ "libnvinfer.so.10.3.0", "libnvinfer_plugin.so.10.3.0", "libonnxruntime_providers_tensorrt.so" ], "deploy_sh_seconds": 287.57, "rollout_seconds": 112.308, "service_smoke": { "source": "orin/final-service-smoke/results.json", "ok": 6, "expected": 6, "bulk_ms": { "3": 59.87, "6": 66.18, "15": 124.27 }, "realtime_ms": { "3": 48.0, "6": 45.98, "15": 68.02 } } } }