cuda_training_transformers_fill-mask_hf-internal-testing/tiny-random-BertModel/benchmark.json

{
    "config": {
        "name": "cuda_training_transformers_fill-mask_hf-internal-testing/tiny-random-BertModel",
        "backend": {
            "name": "pytorch",
            "version": "2.8.0",
            "_target_": "optimum_benchmark.backends.pytorch.backend.PyTorchBackend",
            "model": "hf-internal-testing/tiny-random-BertModel",
            "processor": "hf-internal-testing/tiny-random-BertModel",
            "task": "fill-mask",
            "library": "transformers",
            "model_type": "bert",
            "device": "cuda",
            "device_ids": "0",
            "seed": 42,
            "inter_op_num_threads": null,
            "intra_op_num_threads": null,
            "model_kwargs": {},
            "processor_kwargs": {},
            "no_weights": true,
            "tp_plan": null,
            "device_map": null,
            "torch_dtype": null,
            "eval_mode": true,
            "to_bettertransformer": false,
            "low_cpu_mem_usage": null,
            "attn_implementation": null,
            "cache_implementation": null,
            "allow_tf32": false,
            "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": "training",
            "_target_": "optimum_benchmark.scenarios.training.scenario.TrainingScenario",
            "max_steps": 5,
            "warmup_steps": 2,
            "dataset_shapes": {
                "dataset_size": 500,
                "sequence_length": 16,
                "num_choices": 1
            },
            "training_arguments": {
                "per_device_train_batch_size": 2,
                "gradient_accumulation_steps": 1,
                "output_dir": "./trainer_output",
                "eval_strategy": "no",
                "save_strategy": "no",
                "do_train": true,
                "use_cpu": false,
                "max_steps": 5,
                "do_eval": false,
                "do_predict": false,
                "report_to": "none",
                "skip_memory_metrics": true,
                "ddp_find_unused_parameters": false
            },
            "latency": true,
            "memory": true,
            "energy": true
        },
        "launcher": {
            "name": "process",
            "_target_": "optimum_benchmark.launchers.process.launcher.ProcessLauncher",
            "device_isolation": true,
            "device_isolation_action": "error",
            "numactl": false,
            "numactl_kwargs": {},
            "start_method": "spawn"
        },
        "environment": {
            "cpu": " AMD EPYC 7R32",
            "cpu_count": 16,
            "cpu_ram_mb": 66697.248768,
            "system": "Linux",
            "machine": "x86_64",
            "platform": "Linux-5.10.240-238.966.amzn2.x86_64-x86_64-with-glibc2.35",
            "processor": "x86_64",
            "python_version": "3.10.18",
            "gpu": [
                "NVIDIA A10G"
            ],
            "gpu_count": 1,
            "gpu_vram_mb": 24146608128,
            "optimum_benchmark_version": "0.7.0.dev0",
            "optimum_benchmark_commit": "f22b1bee8ed3f82f26b7e707031dd0063f1eaf25",
            "transformers_version": "4.55.2",
            "transformers_commit": "f22b1bee8ed3f82f26b7e707031dd0063f1eaf25",
            "accelerate_version": "1.10.0",
            "accelerate_commit": "f22b1bee8ed3f82f26b7e707031dd0063f1eaf25",
            "diffusers_version": "0.35.0",
            "diffusers_commit": "f22b1bee8ed3f82f26b7e707031dd0063f1eaf25",
            "optimum_version": null,
            "optimum_commit": null,
            "timm_version": "1.0.19",
            "timm_commit": "f22b1bee8ed3f82f26b7e707031dd0063f1eaf25",
            "peft_version": "0.17.0",
            "peft_commit": "f22b1bee8ed3f82f26b7e707031dd0063f1eaf25"
        },
        "print_report": true,
        "log_report": true
    },
    "report": {
        "overall": {
            "memory": {
                "unit": "MB",
                "max_ram": 1420.701696,
                "max_global_vram": 555.024384,
                "max_process_vram": 0.0,
                "max_reserved": 25.165824,
                "max_allocated": 19.698688
            },
            "latency": {
                "unit": "s",
                "values": [
                    0.2423162841796875,
                    0.013017087936401368,
                    0.011868160247802734,
                    0.01183232021331787,
                    0.01198902416229248
                ],
                "count": 5,
                "total": 0.29102287673950195,
                "mean": 0.05820457534790039,
                "p50": 0.01198902416229248,
                "p90": 0.15059660568237307,
                "p95": 0.19645644493103023,
                "p99": 0.23314431632995605,
                "stdev": 0.09205689211515894,
                "stdev_": 158.1609204515564
            },
            "throughput": {
                "unit": "samples/s",
                "value": 171.80779930491718
            },
            "energy": {
                "unit": "kWh",
                "cpu": 5.931765009027501e-06,
                "ram": 2.79002441111105e-06,
                "gpu": 8.178895431999993e-06,
                "total": 1.6900684852138545e-05
            },
            "efficiency": {
                "unit": "samples/kWh",
                "value": 591691.9987259948
            }
        },
        "warmup": {
            "memory": {
                "unit": "MB",
                "max_ram": 1420.701696,
                "max_global_vram": 555.024384,
                "max_process_vram": 0.0,
                "max_reserved": 25.165824,
                "max_allocated": 19.698688
            },
            "latency": {
                "unit": "s",
                "values": [
                    0.2423162841796875,
                    0.013017087936401368
                ],
                "count": 2,
                "total": 0.25533337211608886,
                "mean": 0.12766668605804443,
                "p50": 0.12766668605804443,
                "p90": 0.2193863645553589,
                "p95": 0.2308513243675232,
                "p99": 0.24002329221725466,
                "stdev": 0.11464959812164308,
                "stdev_": 89.80384911809878
            },
            "throughput": {
                "unit": "samples/s",
                "value": 31.33158793031861
            },
            "energy": null,
            "efficiency": null
        },
        "train": {
            "memory": {
                "unit": "MB",
                "max_ram": 1420.701696,
                "max_global_vram": 555.024384,
                "max_process_vram": 0.0,
                "max_reserved": 25.165824,
                "max_allocated": 19.698688
            },
            "latency": {
                "unit": "s",
                "values": [
                    0.011868160247802734,
                    0.01183232021331787,
                    0.01198902416229248
                ],
                "count": 3,
                "total": 0.03568950462341308,
                "mean": 0.011896501541137694,
                "p50": 0.011868160247802734,
                "p90": 0.011964851379394532,
                "p95": 0.011976937770843505,
                "p99": 0.011986606884002686,
                "stdev": 6.70395584063769e-05,
                "stdev_": 0.5635233028345048
            },
            "throughput": {
                "unit": "samples/s",
                "value": 504.3499535768735
            },
            "energy": null,
            "efficiency": null
        }
    }
}