[
  {
    "50.00 percentile latency (ns)": 341926540846,
    "90.00 percentile latency (ns)": 614575339639,
    "95.00 percentile latency (ns)": 648605798912,
    "97.00 percentile latency (ns)": 662080953027,
    "99.00 percentile latency (ns)": 675683030080,
    "99.90 percentile latency (ns)": 681853381597,
    "Max latency (ns)": 682335659421,
    "Mean latency (ns)": 341902879695,
    "Min duration satisfied": "Yes",
    "Min latency (ns)": 1225270205,
    "Min queries satisfied": "Yes",
    "Mode": "PerformanceOnly",
    "Result is": "VALID",
    "SUT name": "PySUT",
    "Samples per second": 885.048,
    "Scenario": "offline",
    "accelerator_frequency": "",
    "accelerator_host_interconnect": "",
    "accelerator_interconnect": "",
    "accelerator_interconnect_topology": "",
    "accelerator_memory_capacity": "N/A",
    "accelerator_memory_configuration": "",
    "accelerator_model_name": "N/A",
    "accelerator_on-chip_memories": "",
    "accelerators_per_node": 0,
    "accuracy_log_probability": 0,
    "accuracy_log_rng_seed": 0,
    "accuracy_log_sampling_target": 0,
    "characteristics.accuracy": 75.844,
    "characteristics.good": 37922,
    "characteristics.samples_per_second": 885.048,
    "characteristics.samples_per_second.normalized_per_core": 73.754,
    "characteristics.samples_per_second.normalized_per_processor": 885.048,
    "characteristics.total": 50000,
    "ck_system": "aws-m5zn.6xlarge-tvm",
    "ck_used": true,
    "cooling": "",
    "dataset": "ImageNet 2012",
    "dataset_link": "https://github.com/ctuning/ck/blob/master/docs/mlperf-automation/datasets/imagenet2012.md",
    "dim_x_default": "characteristics.samples_per_second",
    "dim_x_maximize": true,
    "dim_y_default": "characteristics.accuracy",
    "dim_y_maximize": true,
    "division": "open",
    "formal_model": "resnet50",
    "formal_model_accuracy": 99.0,
    "formal_model_link": "https://github.com/mlcommons/ck-mlops/tree/main/package",
    "framework": "TVM v0.8-dev (DNNL)",
    "host_memory_capacity": "384GB",
    "host_memory_configuration": "6 slots / 32GB each / 2934 MT/s per socket",
    "host_networking": "",
    "host_networking_topology": "",
    "host_processor_caches": "L1d cache: 384 KiB; L1i cache: 384 KiB; L2 cache: 12 MiB; L3 cache: 24.8 MiB",
    "host_processor_core_count": 12,
    "host_processor_frequency": "3800 MHz",
    "host_processor_interconnect": "",
    "host_processor_model_name": "Intel(R) Xeon(R) Platinum 8252C CPU @ 3.80GHz",
    "host_processors_per_node": 1,
    "host_storage_capacity": "50 GiB",
    "host_storage_type": "SSD",
    "hw_notes": "https://aws.amazon.com/blogs/aws/new-ec2-m5zn-instances-fastest-intel-xeon-scalable-cpu-in-the-cloud ; vCPU: 24",
    "informal_model": "resnet50",
    "input_data_types": "int8",
    "key.accuracy": "characteristics.accuracy",
    "max_async_queries": 1,
    "max_duration (ms)": 610000,
    "max_query_count": 0,
    "min_duration (ms)": 610000,
    "min_query_count": 1,
    "mlperf_version": 1.1,
    "normalize_cores": 12,
    "normalize_processors": 1,
    "note_code": "https://github.com/mlcommons/inference_results_v1.1/tree/master/open/OctoML/code",
    "note_details": "https://github.com/mlcommons/inference_results_v1.1/tree/master/open/OctoML/results/aws-m5zn.6xlarge-tvm",
    "number_of_nodes": 1,
    "operating_system": "Ubuntu 20.04.2 LTS (Linux-5.8.0-1041-aws-x86_64-with-glibc2.29)",
    "other_software_stack": "5.4.0-45-generic; Python 3.8.10; GCC 9.3.0",
    "performance_issue_same": 0,
    "performance_issue_same_index": 0,
    "performance_issue_unique": 0,
    "performance_sample_count": 1024,
    "print_timestamps": 0,
    "problem": false,
    "qsl_rng_seed": 1624344308455410291,
    "retraining": "no",
    "sample_index_rng_seed": 517984244576520566,
    "samples_per_query": 603900,
    "schedule_rng_seed": 10051496985653635065,
    "starting_weights_filename": "https://zenodo.org/record/4589637/files/resnet50_INT8bit_quantized.pt",
    "status": "available",
    "submitter": "OctoML",
    "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/OctoML",
    "sw_notes": "Automated by MLCommons Collective Knowledge v2.5.8 (https://github.com/mlcommons/ck) and the CK-powered MLPerf submission workflow (https://github.com/mlcommons/ck-mlops/tree/main/module/bench.mlperf.inference)",
    "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/aws-m5zn.6xlarge-tvm",
    "system_name": "Amazon EC2 (m5zn.6xlarge)",
    "system_type": "datacenter",
    "target_latency (ns)": 0,
    "target_qps": 900,
    "task": "image classification",
    "task2": "image classification",
    "total_cores": 12,
    "uid": "5b7d46fca689ab54",
    "use_accelerator": false,
    "weight_data_types": "int8",
    "weight_transformations": "no"
  },
  {
    "50.00 percentile latency (ns)": 339060365460,
    "90.00 percentile latency (ns)": 609590646273,
    "95.00 percentile latency (ns)": 643481901225,
    "97.00 percentile latency (ns)": 656910592424,
    "99.00 percentile latency (ns)": 670442921740,
    "99.90 percentile latency (ns)": 676701658598,
    "Max latency (ns)": 676815643660,
    "Mean latency (ns)": 339704999355,
    "Min duration satisfied": "Yes",
    "Min latency (ns)": 2447993079,
    "Min queries satisfied": "Yes",
    "Mode": "PerformanceOnly",
    "Result is": "VALID",
    "SUT name": "PySUT",
    "Samples per second": 2081.96,
    "Scenario": "offline",
    "accelerator_frequency": "",
    "accelerator_host_interconnect": "",
    "accelerator_interconnect": "",
    "accelerator_interconnect_topology": "",
    "accelerator_memory_capacity": "N/A",
    "accelerator_memory_configuration": "",
    "accelerator_model_name": "N/A",
    "accelerator_on-chip_memories": "",
    "accelerators_per_node": 0,
    "accuracy_log_probability": 0,
    "accuracy_log_rng_seed": 0,
    "accuracy_log_sampling_target": 0,
    "characteristics.accuracy": 75.844,
    "characteristics.good": 37922,
    "characteristics.samples_per_second": 2081.96,
    "characteristics.samples_per_second.normalized_per_core": 52.049,
    "characteristics.samples_per_second.normalized_per_processor": 1040.98,
    "characteristics.total": 50000,
    "ck_system": "gcp-n2-standard-80-tvm",
    "ck_used": true,
    "cooling": "",
    "dataset": "ImageNet 2012",
    "dataset_link": "https://github.com/ctuning/ck/blob/master/docs/mlperf-automation/datasets/imagenet2012.md",
    "dim_x_default": "characteristics.samples_per_second",
    "dim_x_maximize": true,
    "dim_y_default": "characteristics.accuracy",
    "dim_y_maximize": true,
    "division": "open",
    "formal_model": "resnet50",
    "formal_model_accuracy": 99.0,
    "formal_model_link": "https://github.com/mlcommons/ck-mlops/tree/main/package",
    "framework": "TVM v0.8-dev (DNNL)",
    "host_memory_capacity": "320GB",
    "host_memory_configuration": "",
    "host_networking": "",
    "host_networking_topology": "",
    "host_processor_caches": "L1d cache: 1.3 MiB; L1i cache: 1.3 MiB; L2 cache: 40 MiB; L3 cache: 66 MiB",
    "host_processor_core_count": 20,
    "host_processor_frequency": "2800MHz",
    "host_processor_interconnect": "",
    "host_processor_model_name": "Intel(R) Xeon(R) CPU (Intel Cascade Lake CPU platform)",
    "host_processors_per_node": 2,
    "host_storage_capacity": "150 GiB",
    "host_storage_type": "Balanced",
    "hw_notes": "vCPU: 80",
    "informal_model": "resnet50",
    "input_data_types": "int8",
    "key.accuracy": "characteristics.accuracy",
    "max_async_queries": 1,
    "max_duration (ms)": 610000,
    "max_query_count": 0,
    "min_duration (ms)": 610000,
    "min_query_count": 1,
    "mlperf_version": 1.1,
    "normalize_cores": 40,
    "normalize_processors": 2,
    "note_code": "https://github.com/mlcommons/inference_results_v1.1/tree/master/open/OctoML/code",
    "note_details": "https://github.com/mlcommons/inference_results_v1.1/tree/master/open/OctoML/results/gcp-n2-standard-80-tvm",
    "number_of_nodes": 1,
    "operating_system": "Ubuntu 20.04.2 LTS (Linux-5.8.0-1038-gcp-x86_64-with-glibc2.29)",
    "other_software_stack": "5.4.0-45-generic; Python 3.8.10; GCC 9.3.0",
    "performance_issue_same": 0,
    "performance_issue_same_index": 0,
    "performance_issue_unique": 0,
    "performance_sample_count": 1024,
    "print_timestamps": 0,
    "problem": false,
    "qsl_rng_seed": 1624344308455410291,
    "retraining": "no",
    "sample_index_rng_seed": 517984244576520566,
    "samples_per_query": 1409100,
    "schedule_rng_seed": 10051496985653635065,
    "starting_weights_filename": "https://zenodo.org/record/4589637/files/resnet50_INT8bit_quantized.pt",
    "status": "available",
    "submitter": "OctoML",
    "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/OctoML",
    "sw_notes": "Automated by MLCommons Collective Knowledge v2.5.8 (https://github.com/mlcommons/ck) and the CK-powered MLPerf submission workflow (https://github.com/mlcommons/ck-mlops/tree/main/module/bench.mlperf.inference)",
    "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/gcp-n2-standard-80-tvm",
    "system_name": "Google (Google Compute Engine) n2-standard-80",
    "system_type": "datacenter",
    "target_latency (ns)": 0,
    "target_qps": 2100,
    "task": "image classification",
    "task2": "image classification",
    "total_cores": 40,
    "uid": "b322561d64783937",
    "use_accelerator": false,
    "weight_data_types": "int8",
    "weight_transformations": "no"
  }
]