Upload cpu_training_transformers_text-generation_hf-internal-testing/tiny-random-LlamaForCausalLM/benchmark.json with huggingface_hub
Browse files
cpu_training_transformers_text-generation_hf-internal-testing/tiny-random-LlamaForCausalLM/benchmark.json
ADDED
|
@@ -0,0 +1,227 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"config": {
|
| 3 |
+
"name": "cpu_training_transformers_text-generation_hf-internal-testing/tiny-random-LlamaForCausalLM",
|
| 4 |
+
"backend": {
|
| 5 |
+
"name": "pytorch",
|
| 6 |
+
"version": "2.5.1",
|
| 7 |
+
"_target_": "optimum_benchmark.backends.pytorch.backend.PyTorchBackend",
|
| 8 |
+
"model": "hf-internal-testing/tiny-random-LlamaForCausalLM",
|
| 9 |
+
"processor": "hf-internal-testing/tiny-random-LlamaForCausalLM",
|
| 10 |
+
"task": "text-generation",
|
| 11 |
+
"library": "transformers",
|
| 12 |
+
"model_type": "llama",
|
| 13 |
+
"device": "cpu",
|
| 14 |
+
"device_ids": null,
|
| 15 |
+
"seed": 42,
|
| 16 |
+
"inter_op_num_threads": null,
|
| 17 |
+
"intra_op_num_threads": null,
|
| 18 |
+
"model_kwargs": {},
|
| 19 |
+
"processor_kwargs": {},
|
| 20 |
+
"no_weights": true,
|
| 21 |
+
"tp_plan": null,
|
| 22 |
+
"device_map": null,
|
| 23 |
+
"torch_dtype": null,
|
| 24 |
+
"eval_mode": true,
|
| 25 |
+
"to_bettertransformer": false,
|
| 26 |
+
"low_cpu_mem_usage": null,
|
| 27 |
+
"attn_implementation": null,
|
| 28 |
+
"cache_implementation": null,
|
| 29 |
+
"allow_tf32": false,
|
| 30 |
+
"autocast_enabled": false,
|
| 31 |
+
"autocast_dtype": null,
|
| 32 |
+
"torch_compile": false,
|
| 33 |
+
"torch_compile_target": "forward",
|
| 34 |
+
"torch_compile_config": {},
|
| 35 |
+
"quantization_scheme": null,
|
| 36 |
+
"quantization_config": {},
|
| 37 |
+
"deepspeed_inference": false,
|
| 38 |
+
"deepspeed_inference_config": {},
|
| 39 |
+
"peft_type": null,
|
| 40 |
+
"peft_config": {}
|
| 41 |
+
},
|
| 42 |
+
"scenario": {
|
| 43 |
+
"name": "training",
|
| 44 |
+
"_target_": "optimum_benchmark.scenarios.training.scenario.TrainingScenario",
|
| 45 |
+
"max_steps": 5,
|
| 46 |
+
"warmup_steps": 2,
|
| 47 |
+
"dataset_shapes": {
|
| 48 |
+
"dataset_size": 500,
|
| 49 |
+
"sequence_length": 16,
|
| 50 |
+
"num_choices": 1
|
| 51 |
+
},
|
| 52 |
+
"training_arguments": {
|
| 53 |
+
"per_device_train_batch_size": 2,
|
| 54 |
+
"gradient_accumulation_steps": 1,
|
| 55 |
+
"output_dir": "./trainer_output",
|
| 56 |
+
"eval_strategy": "no",
|
| 57 |
+
"save_strategy": "no",
|
| 58 |
+
"do_train": true,
|
| 59 |
+
"use_cpu": false,
|
| 60 |
+
"max_steps": 5,
|
| 61 |
+
"do_eval": false,
|
| 62 |
+
"do_predict": false,
|
| 63 |
+
"report_to": "none",
|
| 64 |
+
"skip_memory_metrics": true,
|
| 65 |
+
"ddp_find_unused_parameters": false
|
| 66 |
+
},
|
| 67 |
+
"latency": true,
|
| 68 |
+
"memory": true,
|
| 69 |
+
"energy": true
|
| 70 |
+
},
|
| 71 |
+
"launcher": {
|
| 72 |
+
"name": "process",
|
| 73 |
+
"_target_": "optimum_benchmark.launchers.process.launcher.ProcessLauncher",
|
| 74 |
+
"device_isolation": false,
|
| 75 |
+
"device_isolation_action": null,
|
| 76 |
+
"numactl": false,
|
| 77 |
+
"numactl_kwargs": {},
|
| 78 |
+
"start_method": "spawn"
|
| 79 |
+
},
|
| 80 |
+
"environment": {
|
| 81 |
+
"cpu": " AMD EPYC 7742 64-Core Processor",
|
| 82 |
+
"cpu_count": 128,
|
| 83 |
+
"cpu_ram_mb": 540671.643648,
|
| 84 |
+
"system": "Linux",
|
| 85 |
+
"machine": "x86_64",
|
| 86 |
+
"platform": "Linux-5.4.0-166-generic-x86_64-with-glibc2.31",
|
| 87 |
+
"processor": "x86_64",
|
| 88 |
+
"python_version": "3.10.18",
|
| 89 |
+
"gpu": [
|
| 90 |
+
"NVIDIA A100-SXM4-80GB",
|
| 91 |
+
"NVIDIA A100-SXM4-80GB",
|
| 92 |
+
"NVIDIA A100-SXM4-80GB",
|
| 93 |
+
"NVIDIA DGX Display",
|
| 94 |
+
"NVIDIA A100-SXM4-80GB"
|
| 95 |
+
],
|
| 96 |
+
"gpu_count": 5,
|
| 97 |
+
"gpu_vram_mb": 347892350976,
|
| 98 |
+
"optimum_benchmark_version": "0.6.0.dev0",
|
| 99 |
+
"optimum_benchmark_commit": "61a08086def388b3e78bbf6b42ed20ab4af3f8db",
|
| 100 |
+
"transformers_version": "4.45.1",
|
| 101 |
+
"transformers_commit": "61a08086def388b3e78bbf6b42ed20ab4af3f8db",
|
| 102 |
+
"accelerate_version": "1.10.0",
|
| 103 |
+
"accelerate_commit": "61a08086def388b3e78bbf6b42ed20ab4af3f8db",
|
| 104 |
+
"diffusers_version": "0.34.0",
|
| 105 |
+
"diffusers_commit": "61a08086def388b3e78bbf6b42ed20ab4af3f8db",
|
| 106 |
+
"optimum_version": null,
|
| 107 |
+
"optimum_commit": null,
|
| 108 |
+
"timm_version": "1.0.19",
|
| 109 |
+
"timm_commit": "61a08086def388b3e78bbf6b42ed20ab4af3f8db",
|
| 110 |
+
"peft_version": "0.17.0",
|
| 111 |
+
"peft_commit": "61a08086def388b3e78bbf6b42ed20ab4af3f8db"
|
| 112 |
+
},
|
| 113 |
+
"print_report": true,
|
| 114 |
+
"log_report": true
|
| 115 |
+
},
|
| 116 |
+
"report": {
|
| 117 |
+
"overall": {
|
| 118 |
+
"memory": {
|
| 119 |
+
"unit": "MB",
|
| 120 |
+
"max_ram": 764.45696,
|
| 121 |
+
"max_global_vram": null,
|
| 122 |
+
"max_process_vram": null,
|
| 123 |
+
"max_reserved": null,
|
| 124 |
+
"max_allocated": null
|
| 125 |
+
},
|
| 126 |
+
"latency": {
|
| 127 |
+
"unit": "s",
|
| 128 |
+
"values": [
|
| 129 |
+
0.1563250799663365,
|
| 130 |
+
0.012235323898494244,
|
| 131 |
+
0.011144319083541632,
|
| 132 |
+
0.014274328015744686,
|
| 133 |
+
0.014069293159991503
|
| 134 |
+
],
|
| 135 |
+
"count": 5,
|
| 136 |
+
"total": 0.20804834412410855,
|
| 137 |
+
"mean": 0.04160966882482171,
|
| 138 |
+
"p50": 0.014069293159991503,
|
| 139 |
+
"p90": 0.09950477918609979,
|
| 140 |
+
"p95": 0.1279149295762181,
|
| 141 |
+
"p99": 0.1506430498883128,
|
| 142 |
+
"stdev": 0.057369518695879806,
|
| 143 |
+
"stdev_": 137.8754513461947
|
| 144 |
+
},
|
| 145 |
+
"throughput": {
|
| 146 |
+
"unit": "samples/s",
|
| 147 |
+
"value": 240.32875729197414
|
| 148 |
+
},
|
| 149 |
+
"energy": {
|
| 150 |
+
"unit": "kWh",
|
| 151 |
+
"cpu": 1.2375592967146077e-05,
|
| 152 |
+
"ram": 2.0760073413276463e-05,
|
| 153 |
+
"gpu": 3.035057984135392e-05,
|
| 154 |
+
"total": 6.348624622177646e-05
|
| 155 |
+
},
|
| 156 |
+
"efficiency": {
|
| 157 |
+
"unit": "samples/kWh",
|
| 158 |
+
"value": 157514.4317883752
|
| 159 |
+
}
|
| 160 |
+
},
|
| 161 |
+
"warmup": {
|
| 162 |
+
"memory": {
|
| 163 |
+
"unit": "MB",
|
| 164 |
+
"max_ram": 764.45696,
|
| 165 |
+
"max_global_vram": null,
|
| 166 |
+
"max_process_vram": null,
|
| 167 |
+
"max_reserved": null,
|
| 168 |
+
"max_allocated": null
|
| 169 |
+
},
|
| 170 |
+
"latency": {
|
| 171 |
+
"unit": "s",
|
| 172 |
+
"values": [
|
| 173 |
+
0.1563250799663365,
|
| 174 |
+
0.012235323898494244
|
| 175 |
+
],
|
| 176 |
+
"count": 2,
|
| 177 |
+
"total": 0.16856040386483073,
|
| 178 |
+
"mean": 0.08428020193241537,
|
| 179 |
+
"p50": 0.08428020193241537,
|
| 180 |
+
"p90": 0.14191610435955226,
|
| 181 |
+
"p95": 0.14912059216294438,
|
| 182 |
+
"p99": 0.15488418240565807,
|
| 183 |
+
"stdev": 0.07204487803392112,
|
| 184 |
+
"stdev_": 85.48256456681749
|
| 185 |
+
},
|
| 186 |
+
"throughput": {
|
| 187 |
+
"unit": "samples/s",
|
| 188 |
+
"value": 47.46073108851372
|
| 189 |
+
},
|
| 190 |
+
"energy": null,
|
| 191 |
+
"efficiency": null
|
| 192 |
+
},
|
| 193 |
+
"train": {
|
| 194 |
+
"memory": {
|
| 195 |
+
"unit": "MB",
|
| 196 |
+
"max_ram": 764.45696,
|
| 197 |
+
"max_global_vram": null,
|
| 198 |
+
"max_process_vram": null,
|
| 199 |
+
"max_reserved": null,
|
| 200 |
+
"max_allocated": null
|
| 201 |
+
},
|
| 202 |
+
"latency": {
|
| 203 |
+
"unit": "s",
|
| 204 |
+
"values": [
|
| 205 |
+
0.011144319083541632,
|
| 206 |
+
0.014274328015744686,
|
| 207 |
+
0.014069293159991503
|
| 208 |
+
],
|
| 209 |
+
"count": 3,
|
| 210 |
+
"total": 0.03948794025927782,
|
| 211 |
+
"mean": 0.013162646753092607,
|
| 212 |
+
"p50": 0.014069293159991503,
|
| 213 |
+
"p90": 0.014233321044594049,
|
| 214 |
+
"p95": 0.014253824530169368,
|
| 215 |
+
"p99": 0.014270227318629623,
|
| 216 |
+
"stdev": 0.001429625769044644,
|
| 217 |
+
"stdev_": 10.86123327520545
|
| 218 |
+
},
|
| 219 |
+
"throughput": {
|
| 220 |
+
"unit": "samples/s",
|
| 221 |
+
"value": 455.8353735801867
|
| 222 |
+
},
|
| 223 |
+
"energy": null,
|
| 224 |
+
"efficiency": null
|
| 225 |
+
}
|
| 226 |
+
}
|
| 227 |
+
}
|