[ { "50.00 percentile latency (ns)": 32713227455, "90.00 percentile latency (ns)": 59438380828, "95.00 percentile latency (ns)": 62851244348, "97.00 percentile latency (ns)": 64227084720, "99.00 percentile latency (ns)": 65597801446, "99.90 percentile latency (ns)": 66226802543, "Max latency (ns)": 66282412362, "Mean latency (ns)": 32821442420, "Min duration satisfied": "Yes", "Min latency (ns)": 22196180, "Min queries satisfied": "Yes", "Mode": "Performance", "Result is": "VALID", "SUT name": "LWIS_Server", "Samples per second": 7431.2, "Scenario": "offline", "accelerator_memory_capacity": "16GB", "accelerator_memory_configuration": "GDDR6", "accelerator_model_name": "Nvidia Tesla T4", "accelerators_per_node": 1, "accuracy_log_probability": 0, "accuracy_log_rng_seed": 0, "characteristics.mAP": 22.911, "characteristics.samples_per_second": 7431.2, "characteristics.samples_per_second.normalized_per_core": 7431.2, "characteristics.samples_per_second.normalized_per_processor": 7431.2, "ck_system": "alibaba_cloud_t4", "ck_used": true, "dataset": "COCO 2017 (300x300)", "dataset_link": "https://github.com/ctuning/ck/blob/master/docs/mlperf-automation/datasets/coco2017.md", "dim_x_default": "characteristics.samples_per_second", "dim_x_maximize": true, "dim_y_default": "characteristics.mAP", "dim_y_maximize": true, "division": "closed", "formal_model": "ssd-mobilenet", "formal_model_accuracy": 99.0, "formal_model_link": "https://github.com/mlcommons/ck-mlops/tree/main/package", "framework": "TensorRT 6.0, CUDA 10.1, cuDNN 7.6.3, libjemalloc2, cub 1.8.0, tensorrt-laboratory mlperf branch", "host_memory_capacity": "93GB", "host_processor_core_count": 120, "host_processor_model_name": "AMD EPYC 7V13 64-Core Processor", "host_processor_vcpu_count": 24, "host_processors_per_node": 1, "host_storage_capacity": "500GB", "host_storage_type": "SSD", "informal_model": "ssd-small", "input_data_types": "int8", "key.accuracy": "characteristics.mAP", "max_async_queries": 1, "max_duration (ms)": 0, "max_query_count": 0, "min_duration (ms)": 60000, "min_query_count": 1, "mlperf_version": 0.5, "normalize_cores": 1, "normalize_processors": 1, "note_code": "https://github.com/mlcommons/inference_results_v0.5/tree/master/closed/Alibaba/code", "note_details": "https://github.com/mlcommons/inference_results_v0.5/tree/master/closed/Alibaba/results/alibaba_cloud_t4", "number_of_nodes": 1, "operating_system": "Ubuntu 18.04.5 LTS (Linux-5.4.0-1055-azure-x86_64-with-Ubuntu-18.04-bionic)", "other_software_stack": "docker 18.09.2, python 3.6.8,gcc 5.5.0,onnx 1.3.0, tensorflow 1.13.1, pytorch 1.1.0, torchvision 0.3.0, pycuda 2019.1, sacrebleu 1.3.3, SimpleJSON, OpenCV 4.1.1; GCC 7.5.0; Python 3.7.10", "performance_issue_same": true, "performance_issue_same_index": 0, "performance_issue_unique": true, "performance_sample_count": 256, "print_timestamps": true, "problem": false, "qsl_rng_seed": 3133965575612453542, "retraining": "N", "sample_index_rng_seed": 665484352860916858, "samples_per_query": 492558, "schedule_rng_seed": 3622009729038561421, "starting_weights_filename": "frozen_inference_graph.pb", "status": "available", "submitter": "Alibaba", "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Alibaba", "sw_notes": "Powered by CK v2.5.8 (https://github.com/ctuning/ck)", "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/alibaba_cloud_t4", "system_name": "Microsoft Corporation 7.0 (Virtual Machine)", "target_latency (ns)": 0, "target_qps": 7463, "task": "object detection", "task2": "object detection", "total_cores": 120, "uid": "ab11d858412d4a5e", "use_accelerator": true, "weight_data_types": "int8", "weight_transformations": "quantization, affine fusion" }, { "50.00 percentile latency (ns)": 94475352759, "90.00 percentile latency (ns)": 172261972774, "95.00 percentile latency (ns)": 182064190871, "97.00 percentile latency (ns)": 186006899525, "99.00 percentile latency (ns)": 189952904252, "99.90 percentile latency (ns)": 191686079245, "Max latency (ns)": 191877089037, "Mean latency (ns)": 94884545024, "Min duration satisfied": "Yes", "Min latency (ns)": 78825883, "Min queries satisfied": "Yes", "Mode": "Performance", "Result is": "VALID", "SUT name": "LWIS_Server", "Samples per second": 128.082, "Scenario": "offline", "accelerator_memory_capacity": "16GB", "accelerator_memory_configuration": "GDDR6", "accelerator_model_name": "Nvidia Tesla T4", "accelerators_per_node": 1, "accuracy_log_probability": 0, "accuracy_log_rng_seed": 0, "characteristics.mAP": 20.067, "characteristics.samples_per_second": 128.082, "characteristics.samples_per_second.normalized_per_core": 128.082, "characteristics.samples_per_second.normalized_per_processor": 128.082, "ck_system": "alibaba_cloud_t4", "ck_used": true, "dataset": "COCO 2017 (300x300)", "dataset_link": "https://github.com/ctuning/ck/blob/master/docs/mlperf-automation/datasets/coco2017.md", "dim_x_default": "characteristics.samples_per_second", "dim_x_maximize": true, "dim_y_default": "characteristics.mAP", "dim_y_maximize": true, "division": "closed", "formal_model": "ssd-mobilenet", "formal_model_accuracy": 99.0, "formal_model_link": "https://github.com/mlcommons/ck-mlops/tree/main/package", "framework": "TensorRT 6.0, CUDA 10.1, cuDNN 7.6.3, libjemalloc2, cub 1.8.0, tensorrt-laboratory mlperf branch", "host_memory_capacity": "93GB", "host_processor_core_count": 120, "host_processor_model_name": "AMD EPYC 7V13 64-Core Processor", "host_processor_vcpu_count": 24, "host_processors_per_node": 1, "host_storage_capacity": "500GB", "host_storage_type": "SSD", "informal_model": "ssd-large", "input_data_types": "int8", "key.accuracy": "characteristics.mAP", "max_async_queries": 1, "max_duration (ms)": 0, "max_query_count": 0, "min_duration (ms)": 60000, "min_query_count": 1, "mlperf_version": 0.5, "normalize_cores": 1, "normalize_processors": 1, "note_code": "https://github.com/mlcommons/inference_results_v0.5/tree/master/closed/Alibaba/code", "note_details": "https://github.com/mlcommons/inference_results_v0.5/tree/master/closed/Alibaba/results/alibaba_cloud_t4", "number_of_nodes": 1, "operating_system": "Ubuntu 18.04.5 LTS (Linux-5.4.0-1055-azure-x86_64-with-Ubuntu-18.04-bionic)", "other_software_stack": "docker 18.09.2, python 3.6.8,gcc 5.5.0,onnx 1.3.0, tensorflow 1.13.1, pytorch 1.1.0, torchvision 0.3.0, pycuda 2019.1, sacrebleu 1.3.3, SimpleJSON, OpenCV 4.1.1; GCC 7.5.0; Python 3.7.10", "performance_issue_same": true, "performance_issue_same_index": 0, "performance_issue_unique": true, "performance_sample_count": 64, "print_timestamps": true, "problem": false, "qsl_rng_seed": 3133965575612453542, "retraining": "N", "sample_index_rng_seed": 665484352860916858, "samples_per_query": 24576, "schedule_rng_seed": 3622009729038561421, "starting_weights_filename": "resnet34-ssd1200.pytorch", "status": "available", "submitter": "Alibaba", "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Alibaba", "sw_notes": "Powered by CK v2.5.8 (https://github.com/ctuning/ck)", "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/alibaba_cloud_t4", "system_name": "Microsoft Corporation 7.0 (Virtual Machine)", "target_latency (ns)": 0, "target_qps": 134, "task": "object detection", "task2": "object detection", "total_cores": 120, "uid": "c5432bf4cdb8f2b2", "use_accelerator": true, "weight_data_types": "int8", "weight_transformations": "quantization, affine fusion" } ]