| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "best_supernet_model_checkpoint": null, | |
| "epoch": 12.0, | |
| "global_step": 3216, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "Minimum SubNet": "OrderedDict([(<ElasticityDim.WIDTH: 'width'>, {0: 512, 1: 256, 2: 320, 3: 448, 4: 640, 5: 640, 6: 768, 7: 576, 8: 448, 9: 256, 10: 384, 11: 320, 12: 949, 13: 959, 14: 1110, 15: 1096, 16: 1158, 17: 1062, 18: 1028, 19: 1014, 20: 670, 21: 436, 22: 348, 23: 370})])", | |
| "epoch": 1.0, | |
| "eval_loss": 0.7149681448936462, | |
| "eval_matthews_correlation": 0.3947290430655349, | |
| "eval_runtime": 1.1039, | |
| "eval_samples_per_second": 944.822, | |
| "eval_steps_per_second": 8.153, | |
| "step": 268 | |
| }, | |
| { | |
| "SuperNet": "OrderedDict([(<ElasticityDim.WIDTH: 'width'>, {0: 768, 1: 768, 2: 768, 3: 768, 4: 768, 5: 768, 6: 768, 7: 768, 8: 768, 9: 768, 10: 768, 11: 768, 12: 3072, 13: 3072, 14: 3072, 15: 3072, 16: 3072, 17: 3072, 18: 3072, 19: 3072, 20: 3072, 21: 3072, 22: 3072, 23: 3072})])", | |
| "epoch": 1.0, | |
| "eval_loss": 0.6399039030075073, | |
| "eval_matthews_correlation": 0.5222017375430389, | |
| "eval_runtime": 1.6156, | |
| "eval_samples_per_second": 645.592, | |
| "eval_steps_per_second": 5.571, | |
| "step": 268 | |
| }, | |
| { | |
| "compression_loss": 0.0, | |
| "epoch": 1.87, | |
| "learning_rate": 1.884896652217917e-05, | |
| "loss": 0.8522, | |
| "step": 500 | |
| }, | |
| { | |
| "Minimum SubNet": "OrderedDict([(<ElasticityDim.WIDTH: 'width'>, {0: 512, 1: 256, 2: 320, 3: 448, 4: 640, 5: 640, 6: 768, 7: 576, 8: 448, 9: 256, 10: 384, 11: 320, 12: 949, 13: 959, 14: 1110, 15: 1096, 16: 1158, 17: 1062, 18: 1028, 19: 1014, 20: 670, 21: 436, 22: 348, 23: 370})])", | |
| "epoch": 2.0, | |
| "eval_loss": 0.7286580801010132, | |
| "eval_matthews_correlation": 0.4630242094821006, | |
| "eval_runtime": 1.1443, | |
| "eval_samples_per_second": 911.446, | |
| "eval_steps_per_second": 7.865, | |
| "step": 536 | |
| }, | |
| { | |
| "SuperNet": "OrderedDict([(<ElasticityDim.WIDTH: 'width'>, {0: 768, 1: 768, 2: 768, 3: 768, 4: 768, 5: 768, 6: 768, 7: 768, 8: 768, 9: 768, 10: 768, 11: 768, 12: 3072, 13: 3072, 14: 3072, 15: 3072, 16: 3072, 17: 3072, 18: 3072, 19: 3072, 20: 3072, 21: 3072, 22: 3072, 23: 3072})])", | |
| "epoch": 2.0, | |
| "eval_loss": 0.6622146964073181, | |
| "eval_matthews_correlation": 0.5624034645452709, | |
| "eval_runtime": 1.4549, | |
| "eval_samples_per_second": 716.899, | |
| "eval_steps_per_second": 6.186, | |
| "step": 536 | |
| }, | |
| { | |
| "Minimum SubNet": "OrderedDict([(<ElasticityDim.WIDTH: 'width'>, {0: 512, 1: 256, 2: 320, 3: 448, 4: 640, 5: 640, 6: 768, 7: 576, 8: 448, 9: 256, 10: 384, 11: 320, 12: 949, 13: 959, 14: 1110, 15: 1096, 16: 1158, 17: 1062, 18: 1028, 19: 1014, 20: 670, 21: 436, 22: 348, 23: 370})])", | |
| "epoch": 3.0, | |
| "eval_loss": 0.7320300340652466, | |
| "eval_matthews_correlation": 0.4774992729117021, | |
| "eval_runtime": 1.179, | |
| "eval_samples_per_second": 884.654, | |
| "eval_steps_per_second": 7.634, | |
| "step": 804 | |
| }, | |
| { | |
| "SuperNet": "OrderedDict([(<ElasticityDim.WIDTH: 'width'>, {0: 768, 1: 768, 2: 768, 3: 768, 4: 768, 5: 768, 6: 768, 7: 768, 8: 768, 9: 768, 10: 768, 11: 768, 12: 3072, 13: 3072, 14: 3072, 15: 3072, 16: 3072, 17: 3072, 18: 3072, 19: 3072, 20: 3072, 21: 3072, 22: 3072, 23: 3072})])", | |
| "epoch": 3.0, | |
| "eval_loss": 0.6782493591308594, | |
| "eval_matthews_correlation": 0.557289393094284, | |
| "eval_runtime": 1.4778, | |
| "eval_samples_per_second": 705.795, | |
| "eval_steps_per_second": 6.09, | |
| "step": 804 | |
| }, | |
| { | |
| "compression_loss": 0.0, | |
| "epoch": 3.73, | |
| "learning_rate": 1.5644725411751336e-05, | |
| "loss": 0.3135, | |
| "step": 1000 | |
| }, | |
| { | |
| "Minimum SubNet": "OrderedDict([(<ElasticityDim.WIDTH: 'width'>, {0: 512, 1: 256, 2: 320, 3: 448, 4: 640, 5: 640, 6: 768, 7: 576, 8: 448, 9: 256, 10: 384, 11: 320, 12: 949, 13: 959, 14: 1110, 15: 1096, 16: 1158, 17: 1062, 18: 1028, 19: 1014, 20: 670, 21: 436, 22: 348, 23: 370})])", | |
| "epoch": 4.0, | |
| "eval_loss": 0.8995152711868286, | |
| "eval_matthews_correlation": 0.48304108757820724, | |
| "eval_runtime": 1.126, | |
| "eval_samples_per_second": 926.26, | |
| "eval_steps_per_second": 7.993, | |
| "step": 1072 | |
| }, | |
| { | |
| "SuperNet": "OrderedDict([(<ElasticityDim.WIDTH: 'width'>, {0: 768, 1: 768, 2: 768, 3: 768, 4: 768, 5: 768, 6: 768, 7: 768, 8: 768, 9: 768, 10: 768, 11: 768, 12: 3072, 13: 3072, 14: 3072, 15: 3072, 16: 3072, 17: 3072, 18: 3072, 19: 3072, 20: 3072, 21: 3072, 22: 3072, 23: 3072})])", | |
| "epoch": 4.0, | |
| "eval_loss": 0.7692049741744995, | |
| "eval_matthews_correlation": 0.554912808282685, | |
| "eval_runtime": 1.5919, | |
| "eval_samples_per_second": 655.204, | |
| "eval_steps_per_second": 5.654, | |
| "step": 1072 | |
| }, | |
| { | |
| "Minimum SubNet": "OrderedDict([(<ElasticityDim.WIDTH: 'width'>, {0: 512, 1: 256, 2: 320, 3: 448, 4: 640, 5: 640, 6: 768, 7: 576, 8: 448, 9: 256, 10: 384, 11: 320, 12: 949, 13: 959, 14: 1110, 15: 1096, 16: 1158, 17: 1062, 18: 1028, 19: 1014, 20: 670, 21: 436, 22: 348, 23: 370})])", | |
| "epoch": 5.0, | |
| "eval_loss": 0.826237142086029, | |
| "eval_matthews_correlation": 0.5107428879748117, | |
| "eval_runtime": 1.1006, | |
| "eval_samples_per_second": 947.693, | |
| "eval_steps_per_second": 8.178, | |
| "step": 1340 | |
| }, | |
| { | |
| "SuperNet": "OrderedDict([(<ElasticityDim.WIDTH: 'width'>, {0: 768, 1: 768, 2: 768, 3: 768, 4: 768, 5: 768, 6: 768, 7: 768, 8: 768, 9: 768, 10: 768, 11: 768, 12: 3072, 13: 3072, 14: 3072, 15: 3072, 16: 3072, 17: 3072, 18: 3072, 19: 3072, 20: 3072, 21: 3072, 22: 3072, 23: 3072})])", | |
| "epoch": 5.0, | |
| "eval_loss": 0.6900777816772461, | |
| "eval_matthews_correlation": 0.5834463254140851, | |
| "eval_runtime": 1.7324, | |
| "eval_samples_per_second": 602.048, | |
| "eval_steps_per_second": 5.195, | |
| "step": 1340 | |
| }, | |
| { | |
| "compression_loss": 0.0, | |
| "epoch": 5.6, | |
| "learning_rate": 1.1130738016122495e-05, | |
| "loss": 0.155, | |
| "step": 1500 | |
| }, | |
| { | |
| "Minimum SubNet": "OrderedDict([(<ElasticityDim.WIDTH: 'width'>, {0: 512, 1: 256, 2: 320, 3: 448, 4: 640, 5: 640, 6: 768, 7: 576, 8: 448, 9: 256, 10: 384, 11: 320, 12: 949, 13: 959, 14: 1110, 15: 1096, 16: 1158, 17: 1062, 18: 1028, 19: 1014, 20: 670, 21: 436, 22: 348, 23: 370})])", | |
| "epoch": 6.0, | |
| "eval_loss": 0.8721849918365479, | |
| "eval_matthews_correlation": 0.5075813582663956, | |
| "eval_runtime": 0.8756, | |
| "eval_samples_per_second": 1191.142, | |
| "eval_steps_per_second": 10.278, | |
| "step": 1608 | |
| }, | |
| { | |
| "SuperNet": "OrderedDict([(<ElasticityDim.WIDTH: 'width'>, {0: 768, 1: 768, 2: 768, 3: 768, 4: 768, 5: 768, 6: 768, 7: 768, 8: 768, 9: 768, 10: 768, 11: 768, 12: 3072, 13: 3072, 14: 3072, 15: 3072, 16: 3072, 17: 3072, 18: 3072, 19: 3072, 20: 3072, 21: 3072, 22: 3072, 23: 3072})])", | |
| "epoch": 6.0, | |
| "eval_loss": 0.7215332388877869, | |
| "eval_matthews_correlation": 0.5924834238001306, | |
| "eval_runtime": 1.2201, | |
| "eval_samples_per_second": 854.87, | |
| "eval_steps_per_second": 7.377, | |
| "step": 1608 | |
| }, | |
| { | |
| "Minimum SubNet": "OrderedDict([(<ElasticityDim.WIDTH: 'width'>, {0: 512, 1: 256, 2: 320, 3: 448, 4: 640, 5: 640, 6: 768, 7: 576, 8: 448, 9: 256, 10: 384, 11: 320, 12: 949, 13: 959, 14: 1110, 15: 1096, 16: 1158, 17: 1062, 18: 1028, 19: 1014, 20: 670, 21: 436, 22: 348, 23: 370})])", | |
| "epoch": 7.0, | |
| "eval_loss": 0.9455906748771667, | |
| "eval_matthews_correlation": 0.5053968650343078, | |
| "eval_runtime": 0.88, | |
| "eval_samples_per_second": 1185.181, | |
| "eval_steps_per_second": 10.227, | |
| "step": 1876 | |
| }, | |
| { | |
| "SuperNet": "OrderedDict([(<ElasticityDim.WIDTH: 'width'>, {0: 768, 1: 768, 2: 768, 3: 768, 4: 768, 5: 768, 6: 768, 7: 768, 8: 768, 9: 768, 10: 768, 11: 768, 12: 3072, 13: 3072, 14: 3072, 15: 3072, 16: 3072, 17: 3072, 18: 3072, 19: 3072, 20: 3072, 21: 3072, 22: 3072, 23: 3072})])", | |
| "epoch": 7.0, | |
| "eval_loss": 0.8112756013870239, | |
| "eval_matthews_correlation": 0.5764508680057442, | |
| "eval_runtime": 1.2263, | |
| "eval_samples_per_second": 850.555, | |
| "eval_steps_per_second": 7.339, | |
| "step": 1876 | |
| }, | |
| { | |
| "compression_loss": 0.0, | |
| "epoch": 7.46, | |
| "learning_rate": 6.354385348824488e-06, | |
| "loss": 0.0957, | |
| "step": 2000 | |
| }, | |
| { | |
| "Minimum SubNet": "OrderedDict([(<ElasticityDim.WIDTH: 'width'>, {0: 512, 1: 256, 2: 320, 3: 448, 4: 640, 5: 640, 6: 768, 7: 576, 8: 448, 9: 256, 10: 384, 11: 320, 12: 949, 13: 959, 14: 1110, 15: 1096, 16: 1158, 17: 1062, 18: 1028, 19: 1014, 20: 670, 21: 436, 22: 348, 23: 370})])", | |
| "epoch": 8.0, | |
| "eval_loss": 0.9190700650215149, | |
| "eval_matthews_correlation": 0.5049093009936784, | |
| "eval_runtime": 0.8638, | |
| "eval_samples_per_second": 1207.492, | |
| "eval_steps_per_second": 10.419, | |
| "step": 2144 | |
| }, | |
| { | |
| "SuperNet": "OrderedDict([(<ElasticityDim.WIDTH: 'width'>, {0: 768, 1: 768, 2: 768, 3: 768, 4: 768, 5: 768, 6: 768, 7: 768, 8: 768, 9: 768, 10: 768, 11: 768, 12: 3072, 13: 3072, 14: 3072, 15: 3072, 16: 3072, 17: 3072, 18: 3072, 19: 3072, 20: 3072, 21: 3072, 22: 3072, 23: 3072})])", | |
| "epoch": 8.0, | |
| "eval_loss": 0.7810819149017334, | |
| "eval_matthews_correlation": 0.5885471185335819, | |
| "eval_runtime": 1.2436, | |
| "eval_samples_per_second": 838.718, | |
| "eval_steps_per_second": 7.237, | |
| "step": 2144 | |
| }, | |
| { | |
| "Minimum SubNet": "OrderedDict([(<ElasticityDim.WIDTH: 'width'>, {0: 512, 1: 256, 2: 320, 3: 448, 4: 640, 5: 640, 6: 768, 7: 576, 8: 448, 9: 256, 10: 384, 11: 320, 12: 949, 13: 959, 14: 1110, 15: 1096, 16: 1158, 17: 1062, 18: 1028, 19: 1014, 20: 670, 21: 436, 22: 348, 23: 370})])", | |
| "epoch": 9.0, | |
| "eval_loss": 0.9647462964057922, | |
| "eval_matthews_correlation": 0.49938409054607086, | |
| "eval_runtime": 0.8845, | |
| "eval_samples_per_second": 1179.158, | |
| "eval_steps_per_second": 10.175, | |
| "step": 2412 | |
| }, | |
| { | |
| "SuperNet": "OrderedDict([(<ElasticityDim.WIDTH: 'width'>, {0: 768, 1: 768, 2: 768, 3: 768, 4: 768, 5: 768, 6: 768, 7: 768, 8: 768, 9: 768, 10: 768, 11: 768, 12: 3072, 13: 3072, 14: 3072, 15: 3072, 16: 3072, 17: 3072, 18: 3072, 19: 3072, 20: 3072, 21: 3072, 22: 3072, 23: 3072})])", | |
| "epoch": 9.0, | |
| "eval_loss": 0.8086517453193665, | |
| "eval_matthews_correlation": 0.5598395777855655, | |
| "eval_runtime": 1.2424, | |
| "eval_samples_per_second": 839.53, | |
| "eval_steps_per_second": 7.244, | |
| "step": 2412 | |
| }, | |
| { | |
| "compression_loss": 0.0, | |
| "epoch": 9.33, | |
| "learning_rate": 2.4303047703271643e-06, | |
| "loss": 0.0729, | |
| "step": 2500 | |
| }, | |
| { | |
| "Minimum SubNet": "OrderedDict([(<ElasticityDim.WIDTH: 'width'>, {0: 512, 1: 256, 2: 320, 3: 448, 4: 640, 5: 640, 6: 768, 7: 576, 8: 448, 9: 256, 10: 384, 11: 320, 12: 949, 13: 959, 14: 1110, 15: 1096, 16: 1158, 17: 1062, 18: 1028, 19: 1014, 20: 670, 21: 436, 22: 348, 23: 370})])", | |
| "epoch": 10.0, | |
| "eval_loss": 0.9290460348129272, | |
| "eval_matthews_correlation": 0.49895377962487897, | |
| "eval_runtime": 0.8983, | |
| "eval_samples_per_second": 1161.144, | |
| "eval_steps_per_second": 10.019, | |
| "step": 2680 | |
| }, | |
| { | |
| "SuperNet": "OrderedDict([(<ElasticityDim.WIDTH: 'width'>, {0: 768, 1: 768, 2: 768, 3: 768, 4: 768, 5: 768, 6: 768, 7: 768, 8: 768, 9: 768, 10: 768, 11: 768, 12: 3072, 13: 3072, 14: 3072, 15: 3072, 16: 3072, 17: 3072, 18: 3072, 19: 3072, 20: 3072, 21: 3072, 22: 3072, 23: 3072})])", | |
| "epoch": 10.0, | |
| "eval_loss": 0.8078813552856445, | |
| "eval_matthews_correlation": 0.575435670477595, | |
| "eval_runtime": 1.2176, | |
| "eval_samples_per_second": 856.621, | |
| "eval_steps_per_second": 7.392, | |
| "step": 2680 | |
| }, | |
| { | |
| "Minimum SubNet": "OrderedDict([(<ElasticityDim.WIDTH: 'width'>, {0: 512, 1: 256, 2: 320, 3: 448, 4: 640, 5: 640, 6: 768, 7: 576, 8: 448, 9: 256, 10: 384, 11: 320, 12: 949, 13: 959, 14: 1110, 15: 1096, 16: 1158, 17: 1062, 18: 1028, 19: 1014, 20: 670, 21: 436, 22: 348, 23: 370})])", | |
| "epoch": 11.0, | |
| "eval_loss": 0.9496363997459412, | |
| "eval_matthews_correlation": 0.49816788996099576, | |
| "eval_runtime": 0.9041, | |
| "eval_samples_per_second": 1153.571, | |
| "eval_steps_per_second": 9.954, | |
| "step": 2948 | |
| }, | |
| { | |
| "SuperNet": "OrderedDict([(<ElasticityDim.WIDTH: 'width'>, {0: 768, 1: 768, 2: 768, 3: 768, 4: 768, 5: 768, 6: 768, 7: 768, 8: 768, 9: 768, 10: 768, 11: 768, 12: 3072, 13: 3072, 14: 3072, 15: 3072, 16: 3072, 17: 3072, 18: 3072, 19: 3072, 20: 3072, 21: 3072, 22: 3072, 23: 3072})])", | |
| "epoch": 11.0, | |
| "eval_loss": 0.8123825788497925, | |
| "eval_matthews_correlation": 0.5727969336224868, | |
| "eval_runtime": 1.2393, | |
| "eval_samples_per_second": 841.575, | |
| "eval_steps_per_second": 7.262, | |
| "step": 2948 | |
| }, | |
| { | |
| "compression_loss": 0.0, | |
| "epoch": 11.19, | |
| "learning_rate": 2.5352335693478615e-07, | |
| "loss": 0.0626, | |
| "step": 3000 | |
| }, | |
| { | |
| "Minimum SubNet": "OrderedDict([(<ElasticityDim.WIDTH: 'width'>, {0: 512, 1: 256, 2: 320, 3: 448, 4: 640, 5: 640, 6: 768, 7: 576, 8: 448, 9: 256, 10: 384, 11: 320, 12: 949, 13: 959, 14: 1110, 15: 1096, 16: 1158, 17: 1062, 18: 1028, 19: 1014, 20: 670, 21: 436, 22: 348, 23: 370})])", | |
| "epoch": 12.0, | |
| "eval_loss": 0.9496448040008545, | |
| "eval_matthews_correlation": 0.49816788996099576, | |
| "eval_runtime": 0.8999, | |
| "eval_samples_per_second": 1159.047, | |
| "eval_steps_per_second": 10.001, | |
| "step": 3216 | |
| }, | |
| { | |
| "SuperNet": "OrderedDict([(<ElasticityDim.WIDTH: 'width'>, {0: 768, 1: 768, 2: 768, 3: 768, 4: 768, 5: 768, 6: 768, 7: 768, 8: 768, 9: 768, 10: 768, 11: 768, 12: 3072, 13: 3072, 14: 3072, 15: 3072, 16: 3072, 17: 3072, 18: 3072, 19: 3072, 20: 3072, 21: 3072, 22: 3072, 23: 3072})])", | |
| "epoch": 12.0, | |
| "eval_loss": 0.8131102323532104, | |
| "eval_matthews_correlation": 0.5727969336224868, | |
| "eval_runtime": 1.2698, | |
| "eval_samples_per_second": 821.371, | |
| "eval_steps_per_second": 7.088, | |
| "step": 3216 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "step": 3216, | |
| "total_flos": 6749587903150080.0, | |
| "train_loss": 0.244898120265695, | |
| "train_runtime": 5953.5704, | |
| "train_samples_per_second": 17.235, | |
| "train_steps_per_second": 0.54 | |
| } | |
| ], | |
| "max_steps": 3216, | |
| "min_subnet_acc": null, | |
| "min_subnet_best_acc": null, | |
| "num_train_epochs": 12, | |
| "supernet_acc": null, | |
| "supernet_best_acc": null, | |
| "total_flos": 6749587903150080.0, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |