[
  {
    "50.00 percentile latency (ns)": 157412484,
    "90.00 percentile latency (ns)": 158332005,
    "90th percentile latency (ns)": 158332005,
    "95.00 percentile latency (ns)": 158500433,
    "97.00 percentile latency (ns)": 158638704,
    "99.00 percentile latency (ns)": 158823013,
    "99.90 percentile latency (ns)": 159134571,
    "Max latency (ns)": 159408210,
    "Mean latency (ns)": 157118855,
    "Min duration satisfied": "Yes",
    "Min latency (ns)": 144042226,
    "Min queries satisfied": "Yes",
    "Mode": "PerformanceOnly",
    "QPS w/ loadgen overhead": 6.36,
    "QPS w/o loadgen overhead": 6.36,
    "Result is": "VALID",
    "SUT name": "LWIS_Server",
    "Scenario": "singlestream",
    "accelerator_frequency": "",
    "accelerator_host_interconnect": "",
    "accelerator_interconnect": "",
    "accelerator_interconnect_topology": "",
    "accelerator_memory_capacity": "16 GB",
    "accelerator_memory_configuration": "GDDR6",
    "accelerator_model_name": "NVIDIA T4",
    "accelerator_on-chip_memories": "",
    "accelerators_per_node": 1,
    "accuracy_log_probability": 0,
    "accuracy_log_rng_seed": 0,
    "accuracy_log_sampling_target": 0,
    "characteristics.90th_percentile_latency_ms": 158.332005,
    "characteristics.90th_percentile_latency_ns": 158332005.0,
    "characteristics.90th_percentile_latency_s": 0.158332005,
    "characteristics.90th_percentile_latency_us": 158332.005,
    "characteristics.enhancing tumor": 0.7802,
    "characteristics.mean": 0.8544,
    "characteristics.tumor core": 0.8697,
    "characteristics.whole tumor": 0.9133,
    "ck_system": "T4x1_TRT",
    "ck_used": false,
    "cooling": "",
    "dataset": "BraTS 2019",
    "dataset_link": "https://www.med.upenn.edu/cbica/brats2019/data.html",
    "dim_x_default": "characteristics.90th_percentile_latency_ms",
    "dim_y_default": "characteristics.mean",
    "dim_y_maximize": true,
    "division": "closed",
    "formal_model": "3d-unet",
    "formal_model_accuracy": 99.0,
    "formal_model_link": "",
    "framework": "TensorRT 8.0.2, CUDA 11.3",
    "host_memory_capacity": "128 GB",
    "host_memory_configuration": "",
    "host_networking": "",
    "host_networking_topology": "",
    "host_processor_caches": "",
    "host_processor_core_count": 4,
    "host_processor_frequency": "",
    "host_processor_interconnect": "",
    "host_processor_model_name": "Intel(R) Xeon(R) D-2123IT CPU @ 2.20GHz",
    "host_processors_per_node": 1,
    "host_storage_capacity": "2 TB",
    "host_storage_type": "M.2",
    "hw_notes": "ECC on",
    "informal_model": "3d-unet-99",
    "input_data_types": "int8",
    "key.accuracy": "characteristics.mean",
    "max_async_queries": 1,
    "max_duration (ms)": 0,
    "max_query_count": 0,
    "min_duration (ms)": 600000,
    "min_query_count": 1024,
    "mlperf_version": 1.1,
    "normalize_cores": 1,
    "normalize_processors": 1,
    "note_code": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Lenovo/code",
    "note_details": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Lenovo/results/T4x1_TRT",
    "number_of_nodes": 1,
    "operating_system": "Ubuntu 20.04.2",
    "other_software_stack": "TensorRT 8.0.2, CUDA 11.3, cuDNN 8.2.1, Driver 470.57.02",
    "performance_issue_same": 0,
    "performance_issue_same_index": 0,
    "performance_issue_unique": 0,
    "performance_sample_count": 16,
    "print_timestamps": 0,
    "problem": false,
    "qsl_rng_seed": 1624344308455410291,
    "retraining": "N",
    "sample_index_rng_seed": 517984244576520566,
    "samples_per_query": 1,
    "schedule_rng_seed": 10051496985653635065,
    "starting_weights_filename": "224_224_160_dyanmic_bs.onnx",
    "status": "available",
    "submitter": "Lenovo",
    "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Lenovo",
    "sw_notes": "",
    "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/T4x1_TRT",
    "system_name": "Lenovo ThinkSystem SE350 server",
    "system_type": "edge",
    "target_latency (ns)": 0,
    "target_qps": 6.25,
    "task": "image segmentation",
    "task2": "image segmentation",
    "total_cores": 4,
    "uid": "ddd0ddf6244dc891",
    "use_accelerator": true,
    "weight_data_types": "int8",
    "weight_transformations": "quantization, affine fusion"
  },
  {
    "50.00 percentile latency (ns)": 157412484,
    "90.00 percentile latency (ns)": 158332005,
    "90th percentile latency (ns)": 158332005,
    "95.00 percentile latency (ns)": 158500433,
    "97.00 percentile latency (ns)": 158638704,
    "99.00 percentile latency (ns)": 158823013,
    "99.90 percentile latency (ns)": 159134571,
    "Max latency (ns)": 159408210,
    "Mean latency (ns)": 157118855,
    "Min duration satisfied": "Yes",
    "Min latency (ns)": 144042226,
    "Min queries satisfied": "Yes",
    "Mode": "PerformanceOnly",
    "QPS w/ loadgen overhead": 6.36,
    "QPS w/o loadgen overhead": 6.36,
    "Result is": "VALID",
    "SUT name": "LWIS_Server",
    "Scenario": "singlestream",
    "accelerator_frequency": "",
    "accelerator_host_interconnect": "",
    "accelerator_interconnect": "",
    "accelerator_interconnect_topology": "",
    "accelerator_memory_capacity": "16 GB",
    "accelerator_memory_configuration": "GDDR6",
    "accelerator_model_name": "NVIDIA T4",
    "accelerator_on-chip_memories": "",
    "accelerators_per_node": 1,
    "accuracy_log_probability": 0,
    "accuracy_log_rng_seed": 0,
    "accuracy_log_sampling_target": 0,
    "characteristics.90th_percentile_latency_ms": 158.332005,
    "characteristics.90th_percentile_latency_ns": 158332005.0,
    "characteristics.90th_percentile_latency_s": 0.158332005,
    "characteristics.90th_percentile_latency_us": 158332.005,
    "characteristics.enhancing tumor": 0.7802,
    "characteristics.mean": 0.8544,
    "characteristics.tumor core": 0.8697,
    "characteristics.whole tumor": 0.9133,
    "ck_system": "T4x1_TRT",
    "ck_used": false,
    "cooling": "",
    "dataset": "BraTS 2019",
    "dataset_link": "https://www.med.upenn.edu/cbica/brats2019/data.html",
    "dim_x_default": "characteristics.90th_percentile_latency_ms",
    "dim_y_default": "characteristics.mean",
    "dim_y_maximize": true,
    "division": "closed",
    "formal_model": "3d-unet",
    "formal_model_accuracy": 99.9,
    "formal_model_link": "",
    "framework": "TensorRT 8.0.2, CUDA 11.3",
    "host_memory_capacity": "128 GB",
    "host_memory_configuration": "",
    "host_networking": "",
    "host_networking_topology": "",
    "host_processor_caches": "",
    "host_processor_core_count": 4,
    "host_processor_frequency": "",
    "host_processor_interconnect": "",
    "host_processor_model_name": "Intel(R) Xeon(R) D-2123IT CPU @ 2.20GHz",
    "host_processors_per_node": 1,
    "host_storage_capacity": "2 TB",
    "host_storage_type": "M.2",
    "hw_notes": "ECC on",
    "informal_model": "3d-unet-99.9",
    "input_data_types": "int8",
    "key.accuracy": "characteristics.mean",
    "max_async_queries": 1,
    "max_duration (ms)": 0,
    "max_query_count": 0,
    "min_duration (ms)": 600000,
    "min_query_count": 1024,
    "mlperf_version": 1.1,
    "normalize_cores": 1,
    "normalize_processors": 1,
    "note_code": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Lenovo/code",
    "note_details": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Lenovo/results/T4x1_TRT",
    "number_of_nodes": 1,
    "operating_system": "Ubuntu 20.04.2",
    "other_software_stack": "TensorRT 8.0.2, CUDA 11.3, cuDNN 8.2.1, Driver 470.57.02",
    "performance_issue_same": 0,
    "performance_issue_same_index": 0,
    "performance_issue_unique": 0,
    "performance_sample_count": 16,
    "print_timestamps": 0,
    "problem": false,
    "qsl_rng_seed": 1624344308455410291,
    "retraining": "N",
    "sample_index_rng_seed": 517984244576520566,
    "samples_per_query": 1,
    "schedule_rng_seed": 10051496985653635065,
    "starting_weights_filename": "224_224_160_dyanmic_bs.onnx",
    "status": "available",
    "submitter": "Lenovo",
    "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Lenovo",
    "sw_notes": "",
    "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/T4x1_TRT",
    "system_name": "Lenovo ThinkSystem SE350 server",
    "system_type": "edge",
    "target_latency (ns)": 0,
    "target_qps": 6.25,
    "task": "image segmentation",
    "task2": "image segmentation",
    "total_cores": 4,
    "uid": "d86bcec415387cc9",
    "use_accelerator": true,
    "weight_data_types": "int8",
    "weight_transformations": "quantization, affine fusion"
  },
  {
    "50.00 percentile latency (ns)": 152125113,
    "90.00 percentile latency (ns)": 153003286,
    "90th percentile latency (ns)": 153003286,
    "95.00 percentile latency (ns)": 153221350,
    "97.00 percentile latency (ns)": 153346005,
    "99.00 percentile latency (ns)": 153516770,
    "99.90 percentile latency (ns)": 154048433,
    "Max latency (ns)": 252825245,
    "Mean latency (ns)": 151976141,
    "Min duration satisfied": "Yes",
    "Min latency (ns)": 129781944,
    "Min queries satisfied": "Yes",
    "Mode": "PerformanceOnly",
    "QPS w/ loadgen overhead": 6.58,
    "QPS w/o loadgen overhead": 6.58,
    "Result is": "VALID",
    "SUT name": "LWIS_Server",
    "Scenario": "singlestream",
    "accelerator_frequency": "",
    "accelerator_host_interconnect": "",
    "accelerator_interconnect": "",
    "accelerator_interconnect_topology": "",
    "accelerator_memory_capacity": "16 GB",
    "accelerator_memory_configuration": "GDDR6",
    "accelerator_model_name": "NVIDIA T4",
    "accelerator_on-chip_memories": "",
    "accelerators_per_node": 1,
    "accuracy_log_probability": 0,
    "accuracy_log_rng_seed": 0,
    "accuracy_log_sampling_target": 0,
    "characteristics.90th_percentile_latency_ms": 153.003286,
    "characteristics.90th_percentile_latency_ns": 153003286.0,
    "characteristics.90th_percentile_latency_s": 0.153003286,
    "characteristics.90th_percentile_latency_us": 153003.286,
    "characteristics.enhancing tumor": 0.7799,
    "characteristics.mean": 0.85383,
    "characteristics.tumor core": 0.8684,
    "characteristics.whole tumor": 0.9132,
    "ck_system": "T4x1_TRT",
    "ck_used": false,
    "cooling": "",
    "dataset": "BraTS 2019",
    "dataset_link": "https://www.med.upenn.edu/cbica/brats2019/data.html",
    "dim_x_default": "characteristics.90th_percentile_latency_ms",
    "dim_y_default": "characteristics.mean",
    "dim_y_maximize": true,
    "division": "closed",
    "formal_model": "3d-unet",
    "formal_model_accuracy": 99.0,
    "formal_model_link": "",
    "framework": "TensorRT 7.2.3, CUDA 11.1",
    "host_memory_capacity": "128 GB",
    "host_memory_configuration": "",
    "host_networking": "",
    "host_networking_topology": "",
    "host_processor_caches": "",
    "host_processor_core_count": 4,
    "host_processor_frequency": "",
    "host_processor_interconnect": "",
    "host_processor_model_name": "Intel(R) Xeon(R) D-2123IT CPU @ 2.20GHz",
    "host_processors_per_node": 1,
    "host_storage_capacity": "2 TB",
    "host_storage_type": "M.2",
    "hw_notes": "ECC on",
    "informal_model": "3d-unet-99",
    "input_data_types": "int8",
    "key.accuracy": "characteristics.mean",
    "max_async_queries": 1,
    "max_duration (ms)": 0,
    "max_query_count": 0,
    "min_duration (ms)": 600000,
    "min_query_count": 1024,
    "mlperf_version": 1.0,
    "normalize_cores": 1,
    "normalize_processors": 1,
    "note_code": "https://github.com/mlcommons/inference_results_v1.0/tree/master/closed/Lenovo/code",
    "note_details": "https://github.com/mlcommons/inference_results_v1.0/tree/master/closed/Lenovo/results/T4x1_TRT",
    "number_of_nodes": 1,
    "operating_system": "Ubuntu 18.04.4",
    "other_software_stack": "TensorRT 7.2.3, CUDA 11.1, cuDNN 8.1.1, Driver 460.32.03, DALI 0.30.0",
    "performance_issue_same": 0,
    "performance_issue_same_index": 0,
    "performance_issue_unique": 0,
    "performance_sample_count": 16,
    "print_timestamps": 0,
    "problem": false,
    "qsl_rng_seed": 7322528924094909334,
    "retraining": "N",
    "sample_index_rng_seed": 1570999273408051088,
    "samples_per_query": 1,
    "schedule_rng_seed": 3507442325620259414,
    "starting_weights_filename": "224_224_160_dyanmic_bs.onnx",
    "status": "available",
    "submitter": "Lenovo",
    "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Lenovo",
    "sw_notes": "",
    "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/T4x1_TRT",
    "system_name": "Lenovo ThinkSystem SE350 server",
    "system_type": "edge",
    "target_latency (ns)": 0,
    "target_qps": 6.25,
    "task": "image segmentation",
    "task2": "image segmentation",
    "total_cores": 4,
    "uid": "b74114a5dc2c964d",
    "use_accelerator": true,
    "weight_data_types": "int8",
    "weight_transformations": "quantization, affine fusion"
  },
  {
    "50.00 percentile latency (ns)": 152125113,
    "90.00 percentile latency (ns)": 153003286,
    "90th percentile latency (ns)": 153003286,
    "95.00 percentile latency (ns)": 153221350,
    "97.00 percentile latency (ns)": 153346005,
    "99.00 percentile latency (ns)": 153516770,
    "99.90 percentile latency (ns)": 154048433,
    "Max latency (ns)": 252825245,
    "Mean latency (ns)": 151976141,
    "Min duration satisfied": "Yes",
    "Min latency (ns)": 129781944,
    "Min queries satisfied": "Yes",
    "Mode": "PerformanceOnly",
    "QPS w/ loadgen overhead": 6.58,
    "QPS w/o loadgen overhead": 6.58,
    "Result is": "VALID",
    "SUT name": "LWIS_Server",
    "Scenario": "singlestream",
    "accelerator_frequency": "",
    "accelerator_host_interconnect": "",
    "accelerator_interconnect": "",
    "accelerator_interconnect_topology": "",
    "accelerator_memory_capacity": "16 GB",
    "accelerator_memory_configuration": "GDDR6",
    "accelerator_model_name": "NVIDIA T4",
    "accelerator_on-chip_memories": "",
    "accelerators_per_node": 1,
    "accuracy_log_probability": 0,
    "accuracy_log_rng_seed": 0,
    "accuracy_log_sampling_target": 0,
    "characteristics.90th_percentile_latency_ms": 153.003286,
    "characteristics.90th_percentile_latency_ns": 153003286.0,
    "characteristics.90th_percentile_latency_s": 0.153003286,
    "characteristics.90th_percentile_latency_us": 153003.286,
    "characteristics.enhancing tumor": 0.7799,
    "characteristics.mean": 0.85383,
    "characteristics.tumor core": 0.8684,
    "characteristics.whole tumor": 0.9132,
    "ck_system": "T4x1_TRT",
    "ck_used": false,
    "cooling": "",
    "dataset": "BraTS 2019",
    "dataset_link": "https://www.med.upenn.edu/cbica/brats2019/data.html",
    "dim_x_default": "characteristics.90th_percentile_latency_ms",
    "dim_y_default": "characteristics.mean",
    "dim_y_maximize": true,
    "division": "closed",
    "formal_model": "3d-unet",
    "formal_model_accuracy": 99.9,
    "formal_model_link": "",
    "framework": "TensorRT 7.2.3, CUDA 11.1",
    "host_memory_capacity": "128 GB",
    "host_memory_configuration": "",
    "host_networking": "",
    "host_networking_topology": "",
    "host_processor_caches": "",
    "host_processor_core_count": 4,
    "host_processor_frequency": "",
    "host_processor_interconnect": "",
    "host_processor_model_name": "Intel(R) Xeon(R) D-2123IT CPU @ 2.20GHz",
    "host_processors_per_node": 1,
    "host_storage_capacity": "2 TB",
    "host_storage_type": "M.2",
    "hw_notes": "ECC on",
    "informal_model": "3d-unet-99.9",
    "input_data_types": "int8",
    "key.accuracy": "characteristics.mean",
    "max_async_queries": 1,
    "max_duration (ms)": 0,
    "max_query_count": 0,
    "min_duration (ms)": 600000,
    "min_query_count": 1024,
    "mlperf_version": 1.0,
    "normalize_cores": 1,
    "normalize_processors": 1,
    "note_code": "https://github.com/mlcommons/inference_results_v1.0/tree/master/closed/Lenovo/code",
    "note_details": "https://github.com/mlcommons/inference_results_v1.0/tree/master/closed/Lenovo/results/T4x1_TRT",
    "number_of_nodes": 1,
    "operating_system": "Ubuntu 18.04.4",
    "other_software_stack": "TensorRT 7.2.3, CUDA 11.1, cuDNN 8.1.1, Driver 460.32.03, DALI 0.30.0",
    "performance_issue_same": 0,
    "performance_issue_same_index": 0,
    "performance_issue_unique": 0,
    "performance_sample_count": 16,
    "print_timestamps": 0,
    "problem": false,
    "qsl_rng_seed": 7322528924094909334,
    "retraining": "N",
    "sample_index_rng_seed": 1570999273408051088,
    "samples_per_query": 1,
    "schedule_rng_seed": 3507442325620259414,
    "starting_weights_filename": "224_224_160_dyanmic_bs.onnx",
    "status": "available",
    "submitter": "Lenovo",
    "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Lenovo",
    "sw_notes": "",
    "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/T4x1_TRT",
    "system_name": "Lenovo ThinkSystem SE350 server",
    "system_type": "edge",
    "target_latency (ns)": 0,
    "target_qps": 6.25,
    "task": "image segmentation",
    "task2": "image segmentation",
    "total_cores": 4,
    "uid": "9d448abb86db7e84",
    "use_accelerator": true,
    "weight_data_types": "int8",
    "weight_transformations": "quantization, affine fusion"
  },
  {
    "50.00 percentile latency (ns)": 155052935,
    "90.00 percentile latency (ns)": 156183216,
    "90th percentile latency (ns)": 156183216,
    "95.00 percentile latency (ns)": 156428840,
    "97.00 percentile latency (ns)": 156557025,
    "99.00 percentile latency (ns)": 156792271,
    "99.90 percentile latency (ns)": 157216609,
    "Max latency (ns)": 157309660,
    "Mean latency (ns)": 154932511,
    "Min duration satisfied": "Yes",
    "Min latency (ns)": 142663016,
    "Min queries satisfied": "Yes",
    "Mode": "Performance",
    "QPS w/ loadgen overhead": 6.45,
    "QPS w/o loadgen overhead": 6.45,
    "Result is": "VALID",
    "SUT name": "LWIS_Server",
    "Scenario": "singlestream",
    "accelerator_frequency": "",
    "accelerator_host_interconnect": "",
    "accelerator_interconnect": "",
    "accelerator_interconnect_topology": "",
    "accelerator_memory_capacity": "16 GB",
    "accelerator_memory_configuration": "GDDR6",
    "accelerator_model_name": "NVIDIA T4",
    "accelerator_on-chip_memories": "",
    "accelerators_per_node": 1,
    "accuracy_log_probability": 0,
    "accuracy_log_rng_seed": 0,
    "accuracy_log_sampling_target": 0,
    "characteristics.90th_percentile_latency_ms": 156.183216,
    "characteristics.90th_percentile_latency_ns": 156183216.0,
    "characteristics.90th_percentile_latency_s": 0.156183216,
    "characteristics.90th_percentile_latency_us": 156183.216,
    "characteristics.enhancing tumor": 0.7793,
    "characteristics.mean": 0.85367,
    "characteristics.tumor core": 0.8692,
    "characteristics.whole tumor": 0.9125,
    "ck_system": "T4x1_TRT",
    "ck_used": false,
    "cooling": "",
    "dataset": "BraTS 2019",
    "dataset_link": "https://www.med.upenn.edu/cbica/brats2019/data.html",
    "dim_x_default": "characteristics.90th_percentile_latency_ms",
    "dim_y_default": "characteristics.mean",
    "dim_y_maximize": true,
    "division": "closed",
    "formal_model": "3d-unet",
    "formal_model_accuracy": 99.0,
    "formal_model_link": "",
    "framework": "TensorRT 7.2, CUDA 11.0 Update 1",
    "host_memory_capacity": "128 GB",
    "host_memory_configuration": "",
    "host_networking": "",
    "host_networking_topology": "",
    "host_processor_caches": "",
    "host_processor_core_count": 4,
    "host_processor_frequency": "",
    "host_processor_interconnect": "",
    "host_processor_model_name": "Intel(R) Xeon(R) D-2123IT CPU @ 2.20GHz",
    "host_processors_per_node": 1,
    "host_storage_capacity": "2 TB",
    "host_storage_type": "M.2",
    "hw_notes": "ECC off",
    "informal_model": "3d-unet-99",
    "input_data_types": "int8",
    "key.accuracy": "characteristics.mean",
    "max_async_queries": 1,
    "max_duration (ms)": 0,
    "max_query_count": 0,
    "min_duration (ms)": 60000,
    "min_query_count": 1024,
    "mlperf_version": 0.7,
    "normalize_cores": 1,
    "normalize_processors": 1,
    "note_code": "https://github.com/mlcommons/inference_results_v0.7/tree/master/closed/Lenovo/code",
    "note_details": "https://github.com/mlcommons/inference_results_v0.7/tree/master/closed/Lenovo/results/T4x1_TRT",
    "number_of_nodes": 1,
    "operating_system": "Ubuntu 18.04.4",
    "other_software_stack": "TensorRT 7.2, CUDA 11.0 Update 1, cuDNN 8.0.2, DALI 0.25.0",
    "performance_issue_same": true,
    "performance_issue_same_index": 0,
    "performance_issue_unique": true,
    "performance_sample_count": 16,
    "print_timestamps": true,
    "problem": false,
    "qsl_rng_seed": 12786827339337101903,
    "retraining": "N",
    "sample_index_rng_seed": 12640797754436136668,
    "samples_per_query": 1,
    "schedule_rng_seed": 3135815929913719677,
    "starting_weights_filename": "224_224_160_dyanmic_bs.onnx",
    "status": "available",
    "submitter": "Lenovo",
    "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Lenovo",
    "sw_notes": "",
    "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/T4x1_TRT",
    "system_name": "LenovoSE350 (1x T4)",
    "system_type": "edge",
    "target_latency (ns)": 0,
    "target_qps": 6.25,
    "task": "image segmentation",
    "task2": "image segmentation",
    "total_cores": 4,
    "uid": "38b4720a0d01301d",
    "use_accelerator": true,
    "weight_data_types": "int8,fp16",
    "weight_transformations": "quantization, affine fusion"
  },
  {
    "50.00 percentile latency (ns)": 155052935,
    "90.00 percentile latency (ns)": 156183216,
    "90th percentile latency (ns)": 156183216,
    "95.00 percentile latency (ns)": 156428840,
    "97.00 percentile latency (ns)": 156557025,
    "99.00 percentile latency (ns)": 156792271,
    "99.90 percentile latency (ns)": 157216609,
    "Max latency (ns)": 157309660,
    "Mean latency (ns)": 154932511,
    "Min duration satisfied": "Yes",
    "Min latency (ns)": 142663016,
    "Min queries satisfied": "Yes",
    "Mode": "Performance",
    "QPS w/ loadgen overhead": 6.45,
    "QPS w/o loadgen overhead": 6.45,
    "Result is": "VALID",
    "SUT name": "LWIS_Server",
    "Scenario": "singlestream",
    "accelerator_frequency": "",
    "accelerator_host_interconnect": "",
    "accelerator_interconnect": "",
    "accelerator_interconnect_topology": "",
    "accelerator_memory_capacity": "16 GB",
    "accelerator_memory_configuration": "GDDR6",
    "accelerator_model_name": "NVIDIA T4",
    "accelerator_on-chip_memories": "",
    "accelerators_per_node": 1,
    "accuracy_log_probability": 0,
    "accuracy_log_rng_seed": 0,
    "accuracy_log_sampling_target": 0,
    "characteristics.90th_percentile_latency_ms": 156.183216,
    "characteristics.90th_percentile_latency_ns": 156183216.0,
    "characteristics.90th_percentile_latency_s": 0.156183216,
    "characteristics.90th_percentile_latency_us": 156183.216,
    "characteristics.enhancing tumor": 0.7793,
    "characteristics.mean": 0.85367,
    "characteristics.tumor core": 0.8692,
    "characteristics.whole tumor": 0.9125,
    "ck_system": "T4x1_TRT",
    "ck_used": false,
    "cooling": "",
    "dataset": "BraTS 2019",
    "dataset_link": "https://www.med.upenn.edu/cbica/brats2019/data.html",
    "dim_x_default": "characteristics.90th_percentile_latency_ms",
    "dim_y_default": "characteristics.mean",
    "dim_y_maximize": true,
    "division": "closed",
    "formal_model": "3d-unet",
    "formal_model_accuracy": 99.9,
    "formal_model_link": "",
    "framework": "TensorRT 7.2, CUDA 11.0 Update 1",
    "host_memory_capacity": "128 GB",
    "host_memory_configuration": "",
    "host_networking": "",
    "host_networking_topology": "",
    "host_processor_caches": "",
    "host_processor_core_count": 4,
    "host_processor_frequency": "",
    "host_processor_interconnect": "",
    "host_processor_model_name": "Intel(R) Xeon(R) D-2123IT CPU @ 2.20GHz",
    "host_processors_per_node": 1,
    "host_storage_capacity": "2 TB",
    "host_storage_type": "M.2",
    "hw_notes": "ECC off",
    "informal_model": "3d-unet-99.9",
    "input_data_types": "int8",
    "key.accuracy": "characteristics.mean",
    "max_async_queries": 1,
    "max_duration (ms)": 0,
    "max_query_count": 0,
    "min_duration (ms)": 60000,
    "min_query_count": 1024,
    "mlperf_version": 0.7,
    "normalize_cores": 1,
    "normalize_processors": 1,
    "note_code": "https://github.com/mlcommons/inference_results_v0.7/tree/master/closed/Lenovo/code",
    "note_details": "https://github.com/mlcommons/inference_results_v0.7/tree/master/closed/Lenovo/results/T4x1_TRT",
    "number_of_nodes": 1,
    "operating_system": "Ubuntu 18.04.4",
    "other_software_stack": "TensorRT 7.2, CUDA 11.0 Update 1, cuDNN 8.0.2, DALI 0.25.0",
    "performance_issue_same": true,
    "performance_issue_same_index": 0,
    "performance_issue_unique": true,
    "performance_sample_count": 16,
    "print_timestamps": true,
    "problem": false,
    "qsl_rng_seed": 12786827339337101903,
    "retraining": "N",
    "sample_index_rng_seed": 12640797754436136668,
    "samples_per_query": 1,
    "schedule_rng_seed": 3135815929913719677,
    "starting_weights_filename": "224_224_160_dyanmic_bs.onnx",
    "status": "available",
    "submitter": "Lenovo",
    "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Lenovo",
    "sw_notes": "",
    "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/T4x1_TRT",
    "system_name": "LenovoSE350 (1x T4)",
    "system_type": "edge",
    "target_latency (ns)": 0,
    "target_qps": 6.25,
    "task": "image segmentation",
    "task2": "image segmentation",
    "total_cores": 4,
    "uid": "d352ef3992284cbf",
    "use_accelerator": true,
    "weight_data_types": "int8,fp16",
    "weight_transformations": "quantization, affine fusion"
  }
]