cpu_inference_transformers_text-generation_openai-community/gpt2/benchmark.json

{
    "config": {
        "name": "cpu_inference_transformers_text-generation_openai-community/gpt2",
        "backend": {
            "name": "pytorch",
            "version": "2.3.0+cpu",
            "_target_": "optimum_benchmark.backends.pytorch.backend.PyTorchBackend",
            "task": "text-generation",
            "library": "transformers",
            "model": "openai-community/gpt2",
            "processor": "openai-community/gpt2",
            "device": "cpu",
            "device_ids": null,
            "seed": 42,
            "inter_op_num_threads": null,
            "intra_op_num_threads": null,
            "model_kwargs": {},
            "processor_kwargs": {},
            "hub_kwargs": {},
            "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": false,
            "device_isolation_action": "error",
            "start_method": "spawn"
        },
        "environment": {
            "cpu": " AMD EPYC 7763 64-Core Processor",
            "cpu_count": 4,
            "cpu_ram_mb": 16757.354496,
            "system": "Linux",
            "machine": "x86_64",
            "platform": "Linux-6.5.0-1021-azure-x86_64-with-glibc2.35",
            "processor": "x86_64",
            "python_version": "3.9.19",
            "optimum_benchmark_version": "0.2.0",
            "optimum_benchmark_commit": "90aae7d9a74aeabd69ab3338806fa9a29ef64627",
            "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": "1.0.3",
            "timm_commit": null,
            "peft_version": null,
            "peft_commit": null
        }
    },
    "report": {
        "prefill": {
            "memory": {
                "unit": "MB",
                "max_ram": 952.795136,
                "max_global_vram": null,
                "max_process_vram": null,
                "max_reserved": null,
                "max_allocated": null
            },
            "latency": {
                "unit": "s",
                "count": 11,
                "total": 0.6383326590000138,
                "mean": 0.05803024172727398,
                "stdev": 0.002777207257126519,
                "p50": 0.057320351000015535,
                "p90": 0.060217708999999786,
                "p95": 0.06226782999999614,
                "p99": 0.06390792679999323,
                "values": [
                    0.05933693899999071,
                    0.059465590000002067,
                    0.0597518859999866,
                    0.06431795099999249,
                    0.05554175800000394,
                    0.056214379000010695,
                    0.05601956399999608,
                    0.060217708999999786,
                    0.05613758500001609,
                    0.05400894699999981,
                    0.057320351000015535
                ]
            },
            "throughput": {
                "unit": "tokens/s",
                "value": 34.464788366718246
            },
            "energy": {
                "unit": "kWh",
                "cpu": 2.1882351533866227e-06,
                "ram": 9.144106055632367e-08,
                "gpu": 0.0,
                "total": 2.279676213942946e-06
            },
            "efficiency": {
                "unit": "tokens/kWh",
                "value": 877317.5715777568
            }
        },
        "decode": {
            "memory": {
                "unit": "MB",
                "max_ram": 952.795136,
                "max_global_vram": null,
                "max_process_vram": null,
                "max_reserved": null,
                "max_allocated": null
            },
            "latency": {
                "unit": "s",
                "count": 11,
                "total": 0.39245878199992035,
                "mean": 0.03567807109090185,
                "stdev": 0.0026861941941847167,
                "p50": 0.03431401800000344,
                "p90": 0.038771668999999065,
                "p95": 0.03983036749998803,
                "p99": 0.0406773262999792,
                "values": [
                    0.038771668999999065,
                    0.03431401800000344,
                    0.038533021999995754,
                    0.03319226599998615,
                    0.03310243799998602,
                    0.03294136599998865,
                    0.03390299400001595,
                    0.03794973899999832,
                    0.03510635199998546,
                    0.03375585199998454,
                    0.04088906599997699
                ]
            },
            "throughput": {
                "unit": "tokens/s",
                "value": 28.028421083980824
            },
            "energy": {
                "unit": "kWh",
                "cpu": 1.2419237398806917e-06,
                "ram": 5.190814080657607e-08,
                "gpu": 0.0,
                "total": 1.2938318806872675e-06
            },
            "efficiency": {
                "unit": "tokens/kWh",
                "value": 772897.9436407243
            }
        },
        "per_token": {
            "memory": null,
            "latency": {
                "unit": "s",
                "count": 10,
                "total": 0.9330569309999817,
                "mean": 0.09330569309999817,
                "stdev": 0.004183401697467714,
                "p50": 0.09261663849999024,
                "p90": 0.09815250730000571,
                "p95": 0.09817963464999764,
                "p99": 0.09820133652999118,
                "values": [
                    0.0939672379999763,
                    0.0981464790000075,
                    0.09771918900000287,
                    0.08865479499999651,
                    0.08916938000001551,
                    0.08998324099999877,
                    0.09814281199999186,
                    0.09126603900000418,
                    0.08780099599999858,
                    0.09820676199998957
                ]
            },
            "throughput": {
                "unit": "tokens/s",
                "value": 10.717459640198735
            },
            "energy": null,
            "efficiency": null
        }
    }
}