IlyasMoutawwakil's picture
Upload cuda_inference_transformers_text-generation_openai-community/gpt2/benchmark.json with huggingface_hub
00d3361 verified
raw
history blame
13.2 kB
{
"config": {
"name": "cuda_inference_transformers_text-generation_openai-community/gpt2",
"backend": {
"name": "pytorch",
"version": "2.2.2+rocm5.7",
"_target_": "optimum_benchmark.backends.pytorch.backend.PyTorchBackend",
"task": "text-generation",
"model": "openai-community/gpt2",
"library": "transformers",
"device": "cuda",
"device_ids": "0",
"seed": 42,
"inter_op_num_threads": null,
"intra_op_num_threads": null,
"hub_kwargs": {
"revision": "main",
"force_download": false,
"local_files_only": false,
"trust_remote_code": false
},
"no_weights": true,
"device_map": null,
"torch_dtype": null,
"eval_mode": true,
"to_bettertransformer": false,
"low_cpu_mem_usage": null,
"attn_implementation": null,
"cache_implementation": null,
"autocast_enabled": false,
"autocast_dtype": null,
"torch_compile": false,
"torch_compile_target": "forward",
"torch_compile_config": {},
"quantization_scheme": null,
"quantization_config": {},
"deepspeed_inference": false,
"deepspeed_inference_config": {},
"peft_type": null,
"peft_config": {}
},
"scenario": {
"name": "inference",
"_target_": "optimum_benchmark.scenarios.inference.scenario.InferenceScenario",
"iterations": 1,
"duration": 1,
"warmup_runs": 1,
"input_shapes": {
"batch_size": 1,
"num_choices": 2,
"sequence_length": 2
},
"new_tokens": null,
"latency": true,
"memory": true,
"energy": false,
"forward_kwargs": {},
"generate_kwargs": {
"max_new_tokens": 2,
"min_new_tokens": 2
},
"call_kwargs": {
"num_inference_steps": 2
}
},
"launcher": {
"name": "process",
"_target_": "optimum_benchmark.launchers.process.launcher.ProcessLauncher",
"device_isolation": true,
"device_isolation_action": "error",
"start_method": "spawn"
},
"environment": {
"cpu": " AMD EPYC 7763 64-Core Processor",
"cpu_count": 128,
"cpu_ram_mb": 1082015.236096,
"system": "Linux",
"machine": "x86_64",
"platform": "Linux-5.15.0-84-generic-x86_64-with-glibc2.35",
"processor": "x86_64",
"python_version": "3.10.12",
"gpu": [
"Advanced Micro Devices, Inc. [AMD/ATI]"
],
"gpu_count": 1,
"gpu_vram_mb": 68702699520,
"optimum_benchmark_version": "0.2.0",
"optimum_benchmark_commit": "47f521462daae132cf29116243bfd11ec14edd99",
"transformers_version": "4.40.2",
"transformers_commit": null,
"accelerate_version": "0.30.1",
"accelerate_commit": null,
"diffusers_version": "0.27.2",
"diffusers_commit": null,
"optimum_version": null,
"optimum_commit": null,
"timm_version": "0.9.16",
"timm_commit": null,
"peft_version": null,
"peft_commit": null
}
},
"report": {
"prefill": {
"memory": {
"unit": "MB",
"max_ram": 1050.365952,
"max_global_vram": 1118.420992,
"max_process_vram": 284234.760192,
"max_reserved": 725.614592,
"max_allocated": 512.832512
},
"latency": {
"unit": "s",
"count": 55,
"total": 0.4992275886535644,
"mean": 0.009076865248246626,
"stdev": 0.0005884350876972205,
"p50": 0.008990035057067872,
"p90": 0.009392849159240723,
"p95": 0.009486656379699707,
"p99": 0.011014386253356937,
"values": [
0.012680255889892578,
0.009546353340148925,
0.009377392768859863,
0.009213554382324218,
0.009207473754882813,
0.009131954193115235,
0.009174513816833495,
0.009219953536987305,
0.009161714553833009,
0.009204593658447266,
0.009302352905273438,
0.009221234321594239,
0.009204434394836427,
0.008516118049621582,
0.009132594108581542,
0.009239314079284668,
0.009461071968078614,
0.0093348331451416,
0.009345394134521484,
0.00940315341949463,
0.009595312118530274,
0.009370833396911622,
0.009420112609863281,
0.009289072990417481,
0.009011474609375,
0.008990035057067872,
0.008992754936218261,
0.008999155044555665,
0.008983155250549317,
0.008968754768371582,
0.008968114852905273,
0.00897259521484375,
0.008930515289306641,
0.008927474975585938,
0.00894267463684082,
0.008930035591125488,
0.008923475265502929,
0.008948115348815918,
0.008924275398254395,
0.008985876083374023,
0.008962354660034179,
0.008942835807800293,
0.008950196266174316,
0.008946036338806152,
0.008974194526672363,
0.00904091453552246,
0.008969555854797364,
0.008897396087646484,
0.008377718925476074,
0.008369558334350586,
0.008616597175598144,
0.008452597618103028,
0.008274039268493653,
0.008252119064331055,
0.008049400329589843
]
},
"throughput": {
"unit": "tokens/s",
"value": 220.34038682973056
},
"energy": null,
"efficiency": null
},
"decode": {
"memory": {
"unit": "MB",
"max_ram": 1050.365952,
"max_global_vram": 1118.416896,
"max_process_vram": 313638.334464,
"max_reserved": 725.614592,
"max_allocated": 513.056256
},
"latency": {
"unit": "s",
"count": 55,
"total": 0.5032158823013306,
"mean": 0.009149379678206011,
"stdev": 0.00032693065600031236,
"p50": 0.009107154846191406,
"p90": 0.009594415664672852,
"p95": 0.009638319492340086,
"p99": 0.009881854724884034,
"values": [
0.009618831634521484,
0.009683791160583497,
0.009298514366149903,
0.009284112930297852,
0.009246514320373536,
0.009219794273376465,
0.009280274391174317,
0.009234993934631348,
0.009274032592773438,
0.009304754257202149,
0.009279793739318847,
0.009263153076171876,
0.008555156707763672,
0.009209874153137208,
0.009210353851318359,
0.009602191925048829,
0.009472752571105957,
0.009378032684326172,
0.009402193069458008,
0.009825551033020019,
0.00959499168395996,
0.009466993331909179,
0.009593551635742187,
0.009225873947143555,
0.009116754531860352,
0.009114354133605958,
0.00994795036315918,
0.009075474739074708,
0.00908987522125244,
0.009106353759765625,
0.009107154846191406,
0.009058514595031738,
0.009053874969482421,
0.009097715377807618,
0.009071795463562012,
0.00904907512664795,
0.009091634750366211,
0.009047314643859864,
0.00909659481048584,
0.00905803394317627,
0.009211793899536132,
0.00905307388305664,
0.009047794342041015,
0.009072593688964844,
0.009047795295715332,
0.009087155342102051,
0.00907435417175293,
0.009073394775390625,
0.008772915840148925,
0.008625237464904786,
0.008668596267700196,
0.008554197311401368,
0.00858075714111328,
0.008272756576538086,
0.008364917755126954
]
},
"throughput": {
"unit": "tokens/s",
"value": 109.29702724896401
},
"energy": null,
"efficiency": null
},
"per_token": {
"memory": null,
"latency": {
"unit": "s",
"count": 54,
"total": 0.9823414936065678,
"mean": 0.01819150914086236,
"stdev": 0.0006239384991006326,
"p50": 0.01811542797088623,
"p90": 0.01895425853729248,
"p95": 0.019113671779632566,
"p99": 0.019265988407135008,
"values": [
0.019278623580932616,
0.018813346862792967,
0.018531267166137694,
0.018503908157348634,
0.01839206886291504,
0.018487747192382813,
0.01850678825378418,
0.018472387313842775,
0.018525508880615233,
0.0186455078125,
0.018514627456665038,
0.017851272583007814,
0.017738311767578126,
0.01835734748840332,
0.018812545776367186,
0.01901494598388672,
0.018774465560913087,
0.018797666549682616,
0.019211744308471678,
0.019254783630371093,
0.018909345626831053,
0.019060863494873046,
0.01862038803100586,
0.018179428100585936,
0.01814390754699707,
0.018973506927490235,
0.018113988876342774,
0.01809494972229004,
0.018116867065429688,
0.018100870132446288,
0.018060550689697266,
0.018036869049072264,
0.018058950424194337,
0.018049030303955077,
0.018009510040283203,
0.018046310424804687,
0.01802566909790039,
0.018057350158691407,
0.01807670783996582,
0.01820566940307617,
0.018031110763549806,
0.018035110473632814,
0.01805030822753906,
0.018053829193115236,
0.018167589187622072,
0.018077667236328126,
0.018002952575683594,
0.01718679428100586,
0.017026796340942382,
0.017358951568603516,
0.017031436920166017,
0.016850635528564452,
0.01659639549255371,
0.016446317672729492
]
},
"throughput": {
"unit": "tokens/s",
"value": 54.97070046562368
},
"energy": null,
"efficiency": null
}
}
}