IlyasMoutawwakil's picture
Upload cuda_inference_transformers_text-generation_openai-community/gpt2/benchmark.json with huggingface_hub
23cecd2 verified
raw
history blame
13.7 kB
{
"config": {
"name": "cuda_inference_transformers_text-generation_openai-community/gpt2",
"backend": {
"name": "pytorch",
"version": "2.3.0+cu121",
"_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": true,
"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 7R32",
"cpu_count": 16,
"cpu_ram_mb": 66697.29792,
"system": "Linux",
"machine": "x86_64",
"platform": "Linux-5.10.215-203.850.amzn2.x86_64-x86_64-with-glibc2.35",
"processor": "x86_64",
"python_version": "3.10.12",
"gpu": [
"NVIDIA A10G"
],
"gpu_count": 1,
"gpu_vram_mb": 24146608128,
"optimum_benchmark_version": "0.2.0",
"optimum_benchmark_commit": null,
"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": 1101.000704,
"max_global_vram": 1365.77024,
"max_process_vram": 0.0,
"max_reserved": 725.614592,
"max_allocated": 522.007552
},
"latency": {
"unit": "s",
"count": 55,
"total": 0.49212377834320087,
"mean": 0.008947705060785467,
"stdev": 0.0003010174520732969,
"p50": 0.008904416084289551,
"p90": 0.009119859504699708,
"p95": 0.009374284934997557,
"p99": 0.010096753273010253,
"values": [
0.010312735557556152,
0.009870847702026368,
0.009912768363952636,
0.008912768363952637,
0.009135871887207032,
0.009114848136901856,
0.008957535743713378,
0.008905887603759766,
0.00890336036682129,
0.008914591789245606,
0.008856096267700196,
0.009047136306762696,
0.008898591995239258,
0.008850175857543946,
0.008603584289550782,
0.009104127883911134,
0.008802207946777343,
0.008907072067260742,
0.00889628791809082,
0.00869212818145752,
0.008723487854003906,
0.008904416084289551,
0.008949088096618653,
0.008821727752685547,
0.008950976371765137,
0.008922304153442383,
0.008982111930847168,
0.008963232040405273,
0.008954784393310548,
0.009067135810852051,
0.008868224143981933,
0.008896479606628418,
0.009065247535705567,
0.008835328102111817,
0.008740511894226075,
0.008905088424682617,
0.008833151817321777,
0.00872652816772461,
0.008872991561889648,
0.00916147232055664,
0.008933024406433105,
0.009123200416564942,
0.008998335838317871,
0.008917119979858399,
0.008866304397583008,
0.008869376182556152,
0.0083372802734375,
0.00878320026397705,
0.008794303894042968,
0.008749567985534668,
0.00881772804260254,
0.008886688232421875,
0.008929344177246093,
0.008860416412353515,
0.008514975547790527
]
},
"throughput": {
"unit": "tokens/s",
"value": 223.52100191201782
},
"energy": {
"unit": "kWh",
"cpu": 1.0580453168208855e-07,
"ram": 5.763297454116349e-08,
"gpu": 1.8734528261946863e-07,
"total": 3.5078278884272067e-07
},
"efficiency": {
"unit": "tokens/kWh",
"value": 5701534.007977608
}
},
"decode": {
"memory": {
"unit": "MB",
"max_ram": 1101.000704,
"max_global_vram": 1365.77024,
"max_process_vram": 0.0,
"max_reserved": 725.614592,
"max_allocated": 522.4576
},
"latency": {
"unit": "s",
"count": 55,
"total": 0.5081237440109253,
"mean": 0.009238613527471368,
"stdev": 0.00021723855759967287,
"p50": 0.00919484806060791,
"p90": 0.00937535400390625,
"p95": 0.009431017208099364,
"p99": 0.01019793914794922,
"values": [
0.010100640296936036,
0.010312159538269043,
0.00919484806060791,
0.00923862361907959,
0.009328927993774414,
0.009158271789550781,
0.00930303955078125,
0.009207967758178712,
0.009209983825683594,
0.00921174430847168,
0.00940940761566162,
0.00926255989074707,
0.009284704208374024,
0.009019968032836913,
0.009179327964782715,
0.009294207572937013,
0.009481439590454102,
0.009075807571411134,
0.009002976417541504,
0.009053888320922852,
0.009184991836547852,
0.009256159782409668,
0.009280735969543458,
0.009366944313049316,
0.009156736373901367,
0.009217280387878418,
0.009380960464477539,
0.009230496406555176,
0.00919324779510498,
0.009278400421142578,
0.009179967880249024,
0.009165184020996094,
0.009223487854003906,
0.009135711669921874,
0.009090208053588867,
0.00918489646911621,
0.009127424240112305,
0.009075200080871582,
0.009242336273193359,
0.009280991554260255,
0.00932470417022705,
0.009142975807189941,
0.009140992164611816,
0.009177696228027344,
0.009122176170349121,
0.009070752143859863,
0.008986944198608399,
0.009360480308532714,
0.008992959976196289,
0.009105695724487304,
0.009149791717529297,
0.009180768013000488,
0.009130623817443848,
0.00940169620513916,
0.009253631591796875
]
},
"throughput": {
"unit": "tokens/s",
"value": 108.24134996300711
},
"energy": {
"unit": "kWh",
"cpu": 9.582681835026529e-08,
"ram": 5.261225732070272e-08,
"gpu": 1.7644031914324238e-07,
"total": 3.2487939481421053e-07
},
"efficiency": {
"unit": "tokens/kWh",
"value": 3078065.325047383
}
},
"per_token": {
"memory": null,
"latency": {
"unit": "s",
"count": 54,
"total": 0.981582845687866,
"mean": 0.018177460105330858,
"stdev": 0.0003815738996282945,
"p50": 0.018142271995544433,
"p90": 0.01839564781188965,
"p95": 0.01846967830657959,
"p99": 0.019745556926727293,
"values": [
0.020174848556518556,
0.019364864349365234,
0.018185375213623046,
0.018476896286010742,
0.018316287994384766,
0.018289663314819335,
0.01815449523925781,
0.018124832153320312,
0.018143199920654298,
0.018319360733032225,
0.018336767196655272,
0.018209888458251954,
0.017907615661621093,
0.017773567199707033,
0.01844326400756836,
0.018362367630004883,
0.017970176696777345,
0.017951744079589844,
0.017763328552246094,
0.01795686340332031,
0.018164735794067383,
0.018289663314819335,
0.01821183967590332,
0.018137088775634767,
0.018159616470336915,
0.01840332794189453,
0.018183168411254884,
0.018206720352172853,
0.01837772750854492,
0.01803775978088379,
0.018120704650878908,
0.018309152603149415,
0.018017248153686525,
0.01784012794494629,
0.01814134407043457,
0.01795055961608887,
0.01785651206970215,
0.018139135360717772,
0.018465791702270508,
0.018288639068603514,
0.018284543991088868,
0.01818623924255371,
0.018119680404663087,
0.01802137565612793,
0.01794867134094238,
0.017375232696533204,
0.01801523208618164,
0.017967103958129883,
0.017870847702026366,
0.017998912811279296,
0.018096063613891603,
0.01808793640136719,
0.0182794246673584,
0.01780531120300293
]
},
"throughput": {
"unit": "tokens/s",
"value": 55.01318634206395
},
"energy": null,
"efficiency": null
}
}
}