[ { "model_name_or_path": "elastic/distilbert-base-uncased-finetuned-conll03-english", "task": "token-classification", "dataset": { "path": "conll2003", "eval_split": "validation", "data_keys": { "primary": "tokens", "secondary": null }, "ref_keys": [ "ner_tags" ], "name": null, "calibration_split": "train" }, "quantization_approach": "static", "operators_to_quantize": [ "Add", "MatMul" ], "node_exclusion": [], "aware_training": false, "per_channel": false, "calibration": { "method": "minmax", "num_calibration_samples": 100, "calibration_histogram_percentile": null, "calibration_moving_average": null, "calibration_moving_average_constant": null }, "framework": "onnxruntime", "framework_args": { "opset": 11, "optimization_level": 1 }, "hardware": "Architecture: x86_64\nCPU op-mode(s): 32-bit, 64-bit\nByte Order: Little Endian\nAddress sizes: 46 bits physical, 48 bits virtual\nCPU(s): 8\nOn-line CPU(s) list: 0-7\nThread(s) per core: 2\nCore(s) per socket: 4\nSocket(s): 1\nNUMA node(s): 1\nVendor ID: GenuineIntel\nCPU family: 6\nModel: 85\nModel name: Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz\nStepping: 7\nCPU MHz: 3100.244\nBogoMIPS: 4999.99\nHypervisor vendor: KVM\nVirtualization type: full\nL1d cache: 128 KiB\nL1i cache: 128 KiB\nL2 cache: 4 MiB\nL3 cache: 35.8 MiB\nNUMA node0 CPU(s): 0-7\nVulnerability Itlb multihit: KVM: Vulnerable\nVulnerability L1tf: Mitigation; PTE Inversion\nVulnerability Mds: Vulnerable: Clear CPU buffers attempted, no microcode; SMT Host state unknown\nVulnerability Meltdown: Mitigation; PTI\nVulnerability Spec store bypass: Vulnerable\nVulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization\nVulnerability Spectre v2: Mitigation; Retpolines, STIBP disabled, RSB filling\nVulnerability Srbds: Not affected\nVulnerability Tsx async abort: Not affected\nFlags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss ht syscall nx pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc cpuid aperfmperf tsc_known_freq pni pclmulqdq ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm abm 3dnowprefetch invpcid_single pti fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid mpx avx512f avx512dq rdseed adx smap clflushopt clwb avx512cd avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves ida arat pku ospke avx512_vnni\n", "versions": { "transformers": "4.20.1", "optimum": "1.2.3.dev0", "optimum_hash": "5ac9c0d9fd7e7cca55b2f9935b961ed5b6c50112" }, "evaluation": { "time": [ { "batch_size": 4, "input_length": 64, "baseline": { "nb_forwards": 25, "throughput": 8.33, "latency_mean": 120.23236996, "latency_std": 0.989423927986037, "latency_50": 120.445322, "latency_90": 121.0641136, "latency_95": 121.63786520000001, "latency_99": 122.31954252, "latency_999": 122.47620775200001 }, "optimized": { "nb_forwards": 53, "throughput": 17.67, "latency_mean": 56.94031537735849, "latency_std": 2.2044830948358625, "latency_50": 56.199388, "latency_90": 60.3284648, "latency_95": 60.6057082, "latency_99": 61.70255691999999, "latency_999": 62.529690292000005 } }, { "batch_size": 8, "input_length": 64, "baseline": { "nb_forwards": 19, "throughput": 6.33, "latency_mean": 166.53055257894738, "latency_std": 1.575841987426849, "latency_50": 166.638572, "latency_90": 168.272883, "latency_95": 168.7129504, "latency_99": 169.52801488, "latency_999": 169.711404388 }, "optimized": { "nb_forwards": 24, "throughput": 8.0, "latency_mean": 129.002869375, "latency_std": 0.6157854643813875, "latency_50": 129.063924, "latency_90": 129.7084936, "latency_95": 129.9355643, "latency_99": 130.24102448, "latency_999": 130.313872748 } }, { "batch_size": 1, "input_length": 64, "baseline": { "nb_forwards": 70, "throughput": 23.33, "latency_mean": 43.048573857142856, "latency_std": 1.1204473128323003, "latency_50": 42.845755, "latency_90": 43.8944438, "latency_95": 44.3052485, "latency_99": 46.73122168000001, "latency_999": 49.909082367999986 }, "optimized": { "nb_forwards": 143, "throughput": 47.67, "latency_mean": 21.113699776223775, "latency_std": 0.1930452254945551, "latency_50": 21.085728, "latency_90": 21.3874956, "latency_95": 21.4500651, "latency_99": 21.640094780000002, "latency_999": 21.648399938 } } ], "others": { "baseline": { "precision": 0.936836221352711, "recall": 0.9533560864618885, "f1": 0.9450239639131661, "accuracy": 0.9880421708059153 }, "optimized": { "precision": 0.07350512058143377, "recall": 0.25312855517633676, "f1": 0.1139272913466462, "accuracy": 0.3629802589683719 } } }, "max_eval_samples": 1000, "time_benchmark_args": { "duration": 3, "warmup_runs": 1 }, "model_type": "distilbert" }, { "model_name_or_path": "elastic/distilbert-base-uncased-finetuned-conll03-english", "task": "token-classification", "dataset": { "path": "conll2003", "eval_split": "validation", "data_keys": { "primary": "tokens", "secondary": null }, "ref_keys": [ "ner_tags" ], "name": null, "calibration_split": "train" }, "quantization_approach": "static", "operators_to_quantize": [ "Add" ], "node_exclusion": [], "aware_training": false, "per_channel": false, "calibration": { "method": "minmax", "num_calibration_samples": 100, "calibration_histogram_percentile": null, "calibration_moving_average": null, "calibration_moving_average_constant": null }, "framework": "onnxruntime", "framework_args": { "opset": 11, "optimization_level": 1 }, "hardware": "Architecture: x86_64\nCPU op-mode(s): 32-bit, 64-bit\nByte Order: Little Endian\nAddress sizes: 46 bits physical, 48 bits virtual\nCPU(s): 8\nOn-line CPU(s) list: 0-7\nThread(s) per core: 2\nCore(s) per socket: 4\nSocket(s): 1\nNUMA node(s): 1\nVendor ID: GenuineIntel\nCPU family: 6\nModel: 85\nModel name: Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz\nStepping: 7\nCPU MHz: 3100.091\nBogoMIPS: 4999.99\nHypervisor vendor: KVM\nVirtualization type: full\nL1d cache: 128 KiB\nL1i cache: 128 KiB\nL2 cache: 4 MiB\nL3 cache: 35.8 MiB\nNUMA node0 CPU(s): 0-7\nVulnerability Itlb multihit: KVM: Vulnerable\nVulnerability L1tf: Mitigation; PTE Inversion\nVulnerability Mds: Vulnerable: Clear CPU buffers attempted, no microcode; SMT Host state unknown\nVulnerability Meltdown: Mitigation; PTI\nVulnerability Spec store bypass: Vulnerable\nVulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization\nVulnerability Spectre v2: Mitigation; Retpolines, STIBP disabled, RSB filling\nVulnerability Srbds: Not affected\nVulnerability Tsx async abort: Not affected\nFlags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss ht syscall nx pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc cpuid aperfmperf tsc_known_freq pni pclmulqdq ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm abm 3dnowprefetch invpcid_single pti fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid mpx avx512f avx512dq rdseed adx smap clflushopt clwb avx512cd avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves ida arat pku ospke avx512_vnni\n", "versions": { "transformers": "4.20.1", "optimum": "1.2.3.dev0", "optimum_hash": "5ac9c0d9fd7e7cca55b2f9935b961ed5b6c50112" }, "evaluation": { "time": [ { "batch_size": 1, "input_length": 64, "baseline": { "nb_forwards": 69, "throughput": 23.0, "latency_mean": 43.50449917391305, "latency_std": 1.1458006326491226, "latency_50": 43.443712, "latency_90": 44.833304, "latency_95": 45.4732784, "latency_99": 46.1717674, "latency_999": 46.293552340000005 }, "optimized": { "nb_forwards": 80, "throughput": 26.67, "latency_mean": 37.9267952125, "latency_std": 0.11734822683861629, "latency_50": 37.9285515, "latency_90": 38.085207600000004, "latency_95": 38.111036399999996, "latency_99": 38.2064807, "latency_999": 38.22722057 } }, { "batch_size": 8, "input_length": 64, "baseline": { "nb_forwards": 14, "throughput": 4.67, "latency_mean": 214.81155885714287, "latency_std": 0.6229026122307055, "latency_50": 214.6879675, "latency_90": 215.571702, "latency_95": 215.72494925, "latency_99": 215.90999385, "latency_999": 215.951628885 }, "optimized": { "nb_forwards": 12, "throughput": 4.0, "latency_mean": 256.95122358333333, "latency_std": 1.2773226309110695, "latency_50": 257.0572985, "latency_90": 258.7638351, "latency_95": 258.84763195, "latency_99": 258.86815838999996, "latency_999": 258.872776839 } }, { "batch_size": 4, "input_length": 64, "baseline": { "nb_forwards": 26, "throughput": 8.67, "latency_mean": 119.1024813076923, "latency_std": 1.5917975126134987, "latency_50": 118.759877, "latency_90": 120.792844, "latency_95": 121.9356475, "latency_99": 123.13953675, "latency_999": 123.40581367499999 }, "optimized": { "nb_forwards": 23, "throughput": 7.67, "latency_mean": 130.78132304347827, "latency_std": 0.5922745467393132, "latency_50": 130.955147, "latency_90": 131.512009, "latency_95": 131.5393553, "latency_99": 131.74930052, "latency_999": 131.801985152 } } ], "others": { "baseline": { "precision": 0.936836221352711, "recall": 0.9533560864618885, "f1": 0.9450239639131661, "accuracy": 0.9880421708059153 }, "optimized": { "precision": 0.06477812995245642, "recall": 0.18600682593856654, "f1": 0.09609168380840435, "accuracy": 0.3400551899808958 } } }, "max_eval_samples": 1000, "time_benchmark_args": { "duration": 3, "warmup_runs": 1 }, "model_type": "distilbert" }, { "model_name_or_path": "elastic/distilbert-base-uncased-finetuned-conll03-english", "task": "token-classification", "dataset": { "path": "conll2003", "eval_split": "validation", "data_keys": { "primary": "tokens", "secondary": null }, "ref_keys": [ "ner_tags" ], "name": null, "calibration_split": "train" }, "quantization_approach": "dynamic", "operators_to_quantize": [ "Add", "MatMul" ], "node_exclusion": [], "aware_training": false, "per_channel": false, "calibration": { "method": "minmax", "num_calibration_samples": 100, "calibration_histogram_percentile": null, "calibration_moving_average": null, "calibration_moving_average_constant": null }, "framework": "onnxruntime", "framework_args": { "opset": 11, "optimization_level": 1 }, "hardware": "Architecture: x86_64\nCPU op-mode(s): 32-bit, 64-bit\nByte Order: Little Endian\nAddress sizes: 46 bits physical, 48 bits virtual\nCPU(s): 8\nOn-line CPU(s) list: 0-7\nThread(s) per core: 2\nCore(s) per socket: 4\nSocket(s): 1\nNUMA node(s): 1\nVendor ID: GenuineIntel\nCPU family: 6\nModel: 85\nModel name: Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz\nStepping: 7\nCPU MHz: 3100.009\nBogoMIPS: 4999.99\nHypervisor vendor: KVM\nVirtualization type: full\nL1d cache: 128 KiB\nL1i cache: 128 KiB\nL2 cache: 4 MiB\nL3 cache: 35.8 MiB\nNUMA node0 CPU(s): 0-7\nVulnerability Itlb multihit: KVM: Vulnerable\nVulnerability L1tf: Mitigation; PTE Inversion\nVulnerability Mds: Vulnerable: Clear CPU buffers attempted, no microcode; SMT Host state unknown\nVulnerability Meltdown: Mitigation; PTI\nVulnerability Spec store bypass: Vulnerable\nVulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization\nVulnerability Spectre v2: Mitigation; Retpolines, STIBP disabled, RSB filling\nVulnerability Srbds: Not affected\nVulnerability Tsx async abort: Not affected\nFlags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss ht syscall nx pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc cpuid aperfmperf tsc_known_freq pni pclmulqdq ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm abm 3dnowprefetch invpcid_single pti fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid mpx avx512f avx512dq rdseed adx smap clflushopt clwb avx512cd avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves ida arat pku ospke avx512_vnni\n", "versions": { "transformers": "4.20.1", "optimum": "1.2.3.dev0", "optimum_hash": "5ac9c0d9fd7e7cca55b2f9935b961ed5b6c50112" }, "evaluation": { "time": [ { "batch_size": 1, "input_length": 64, "baseline": { "nb_forwards": 53, "throughput": 17.67, "latency_mean": 57.63860111320755, "latency_std": 0.5448611043553628, "latency_50": 57.65361, "latency_90": 58.180421, "latency_95": 58.392744, "latency_99": 58.71634352, "latency_999": 58.721444252 }, "optimized": { "nb_forwards": 244, "throughput": 81.33, "latency_mean": 12.298368512295083, "latency_std": 0.4560740565346141, "latency_50": 12.2116125, "latency_90": 13.001667200000002, "latency_95": 13.1330103, "latency_99": 13.2790208, "latency_999": 13.414312331 } }, { "batch_size": 4, "input_length": 64, "baseline": { "nb_forwards": 26, "throughput": 8.67, "latency_mean": 119.50429169230769, "latency_std": 0.4639465722921096, "latency_50": 119.446385, "latency_90": 119.95197, "latency_95": 120.05153425, "latency_99": 120.7893855, "latency_999": 121.00299195000001 }, "optimized": { "nb_forwards": 76, "throughput": 25.33, "latency_mean": 39.91599960526316, "latency_std": 0.883213781232674, "latency_50": 39.8835755, "latency_90": 41.0755615, "latency_95": 41.48617225, "latency_99": 42.00973875, "latency_999": 42.412953375 } }, { "batch_size": 8, "input_length": 64, "baseline": { "nb_forwards": 19, "throughput": 6.33, "latency_mean": 165.83700805263157, "latency_std": 1.7394953701654086, "latency_50": 165.801757, "latency_90": 168.0285054, "latency_95": 168.19460990000002, "latency_99": 168.78632678, "latency_999": 168.919463078 }, "optimized": { "nb_forwards": 40, "throughput": 13.33, "latency_mean": 75.448955425, "latency_std": 1.2544431966810392, "latency_50": 75.414968, "latency_90": 77.1854282, "latency_95": 77.5299735, "latency_99": 77.80073465000001, "latency_999": 77.95147686499999 } } ], "others": { "baseline": { "precision": 0.936836221352711, "recall": 0.9533560864618885, "f1": 0.9450239639131661, "accuracy": 0.9880421708059153 }, "optimized": { "precision": 0.9368008948545862, "recall": 0.9527872582480091, "f1": 0.9447264523406655, "accuracy": 0.9879006580343876 } } }, "max_eval_samples": 1000, "time_benchmark_args": { "duration": 3, "warmup_runs": 1 }, "model_type": "distilbert" }, { "model_name_or_path": "elastic/distilbert-base-uncased-finetuned-conll03-english", "task": "token-classification", "dataset": { "path": "conll2003", "eval_split": "validation", "data_keys": { "primary": "tokens", "secondary": null }, "ref_keys": [ "ner_tags" ], "name": null, "calibration_split": "train" }, "quantization_approach": "dynamic", "operators_to_quantize": [ "Add" ], "node_exclusion": [], "aware_training": false, "per_channel": false, "calibration": { "method": "minmax", "num_calibration_samples": 100, "calibration_histogram_percentile": null, "calibration_moving_average": null, "calibration_moving_average_constant": null }, "framework": "onnxruntime", "framework_args": { "opset": 11, "optimization_level": 1 }, "hardware": "Architecture: x86_64\nCPU op-mode(s): 32-bit, 64-bit\nByte Order: Little Endian\nAddress sizes: 46 bits physical, 48 bits virtual\nCPU(s): 8\nOn-line CPU(s) list: 0-7\nThread(s) per core: 2\nCore(s) per socket: 4\nSocket(s): 1\nNUMA node(s): 1\nVendor ID: GenuineIntel\nCPU family: 6\nModel: 85\nModel name: Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz\nStepping: 7\nCPU MHz: 2638.487\nBogoMIPS: 4999.99\nHypervisor vendor: KVM\nVirtualization type: full\nL1d cache: 128 KiB\nL1i cache: 128 KiB\nL2 cache: 4 MiB\nL3 cache: 35.8 MiB\nNUMA node0 CPU(s): 0-7\nVulnerability Itlb multihit: KVM: Vulnerable\nVulnerability L1tf: Mitigation; PTE Inversion\nVulnerability Mds: Vulnerable: Clear CPU buffers attempted, no microcode; SMT Host state unknown\nVulnerability Meltdown: Mitigation; PTI\nVulnerability Spec store bypass: Vulnerable\nVulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization\nVulnerability Spectre v2: Mitigation; Retpolines, STIBP disabled, RSB filling\nVulnerability Srbds: Not affected\nVulnerability Tsx async abort: Not affected\nFlags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss ht syscall nx pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc cpuid aperfmperf tsc_known_freq pni pclmulqdq ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm abm 3dnowprefetch invpcid_single pti fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid mpx avx512f avx512dq rdseed adx smap clflushopt clwb avx512cd avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves ida arat pku ospke avx512_vnni\n", "versions": { "transformers": "4.20.1", "optimum": "1.2.3.dev0", "optimum_hash": "5ac9c0d9fd7e7cca55b2f9935b961ed5b6c50112" }, "evaluation": { "time": [ { "batch_size": 1, "input_length": 64, "baseline": { "nb_forwards": 69, "throughput": 23.0, "latency_mean": 43.50526027536232, "latency_std": 1.1770353674252074, "latency_50": 43.267983, "latency_90": 45.0357992, "latency_95": 45.6057136, "latency_99": 46.708998679999986, "latency_999": 47.814713768000004 }, "optimized": { "nb_forwards": 102, "throughput": 34.0, "latency_mean": 29.424613480392157, "latency_std": 0.14890697595200564, "latency_50": 29.3912705, "latency_90": 29.646715, "latency_95": 29.68545545, "latency_99": 29.80756655, "latency_999": 29.811399894 } }, { "batch_size": 4, "input_length": 64, "baseline": { "nb_forwards": 26, "throughput": 8.67, "latency_mean": 119.6179461923077, "latency_std": 1.4057848288153165, "latency_50": 119.394914, "latency_90": 121.3817145, "latency_95": 121.8577975, "latency_99": 122.802906, "latency_999": 123.0513933 }, "optimized": { "nb_forwards": 28, "throughput": 9.33, "latency_mean": 107.42320235714286, "latency_std": 0.9405205161982765, "latency_50": 107.1847235, "latency_90": 107.6445599, "latency_95": 108.2160214, "latency_99": 111.05779109000001, "latency_999": 111.916852709 } }, { "batch_size": 8, "input_length": 64, "baseline": { "nb_forwards": 14, "throughput": 4.67, "latency_mean": 214.6487932857143, "latency_std": 0.9053003539723654, "latency_50": 214.552057, "latency_90": 215.54495519999998, "latency_95": 216.14476715, "latency_99": 216.93365343000002, "latency_999": 217.11115284299999 }, "optimized": { "nb_forwards": 15, "throughput": 5.0, "latency_mean": 211.41319233333334, "latency_std": 1.1447515204122778, "latency_50": 211.02957, "latency_90": 213.090243, "latency_95": 213.19109559999998, "latency_99": 213.37423912, "latency_999": 213.415446412 } } ], "others": { "baseline": { "precision": 0.936836221352711, "recall": 0.9533560864618885, "f1": 0.9450239639131661, "accuracy": 0.9880421708059153 }, "optimized": { "precision": 0.936836221352711, "recall": 0.9533560864618885, "f1": 0.9450239639131661, "accuracy": 0.9880421708059153 } } }, "max_eval_samples": 1000, "time_benchmark_args": { "duration": 3, "warmup_runs": 1 }, "model_type": "distilbert" } ]