{ "experiment": "R-MATH-INK-06-BEHAVIOR-ROLE-001", "generated_at": "2026-07-23T16:40:18.931480+00:00", "seed": 31, "device": "cuda", "cuda_device": "NVIDIA GeForce GTX 1650", "teacher_adapter": "research\\runs\\math_ink_06_online_casecontext_refined_seed31_20260723\\skeleton_adapter.pt", "base_checkpoint": "C:\\Users\\user\\Desktop\\Aiflow\\math-grid-drawer-simulation\\research\\runs\\math_ink_06_federated_virtual_ce025_family010_seed31_20260723\\math_ink_06_candidate.pt", "split_contract": "CROHME2012 trainData writer fit/validation; testDataGT official held-out", "formulas": { "fit": 1063, "validation": 275, "official_test": 488 }, "role_samples": { "fit": { "identifier_lower": 1573, "multiply_operator": 177, "identifier_upper": 99 }, "validation": { "identifier_lower": 401, "multiply_operator": 44, "identifier_upper": 22 }, "official_test": { "identifier_lower": 568, "identifier_upper": 31, "multiply_operator": 36 } }, "teacher_baseline": { "validation": { "samples": 467, "accuracy": 0.40471091866493225, "macro_f1": 0.22979348338888128, "recall": { "identifier_lower": 0.4314214463840399, "identifier_upper": 0.7272727272727273, "multiply_operator": 0.0 }, "confusion": [ [ 173, 228, 0 ], [ 6, 16, 0 ], [ 16, 28, 0 ] ], "ece": 0.4227110123725495 }, "official_test": { "samples": 635, "accuracy": 0.6078740358352661, "macro_f1": 0.31082213130722774, "recall": { "identifier_lower": 0.6373239436619719, "identifier_upper": 0.7741935483870968, "multiply_operator": 0.0 }, "confusion": [ [ 362, 206, 0 ], [ 7, 24, 0 ], [ 7, 29, 0 ] ], "ece": 0.1908320430970747 } }, "selected_epoch": 42, "validation": { "samples": 467, "accuracy": 0.9850106835365295, "macro_f1": 0.9492270079067833, "recall": { "identifier_lower": 0.9900249376558603, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 397, 3, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.01576937972125217 }, "official_test": { "samples": 635, "accuracy": 0.9385826587677002, "macro_f1": 0.7717349248806314, "recall": { "identifier_lower": 0.9630281690140845, "identifier_upper": 0.5161290322580645, "multiply_operator": 0.9166666666666666 }, "confusion": [ [ 547, 6, 15 ], [ 10, 16, 5 ], [ 3, 0, 33 ] ], "ece": 0.050884793443143006 }, "history": [ { "epoch": 1, "training_loss": 0.8767189346112839, "learning_rate": 0.0009993147673772868, "validation": { "samples": 467, "accuracy": 0.9014989137649536, "macro_f1": 0.5273300379683359, "recall": { "identifier_lower": 0.9975062344139651, "identifier_upper": 0.0, "multiply_operator": 0.4772727272727273 }, "confusion": [ [ 400, 0, 1 ], [ 22, 0, 0 ], [ 23, 0, 21 ] ], "ece": 0.20063943331568002 } }, { "epoch": 2, "training_loss": 0.5778835811055371, "learning_rate": 0.0009972609476841365, "validation": { "samples": 467, "accuracy": 0.9357602000236511, "macro_f1": 0.7668176821254237, "recall": { "identifier_lower": 0.9650872817955112, "identifier_upper": 0.36363636363636365, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 387, 4, 10 ], [ 13, 8, 1 ], [ 2, 0, 42 ] ], "ece": 0.19584264685098568 } }, { "epoch": 3, "training_loss": 0.420009775781322, "learning_rate": 0.0009938441702975688, "validation": { "samples": 467, "accuracy": 0.9057815670967102, "macro_f1": 0.7875140750426791, "recall": { "identifier_lower": 0.8952618453865336, "identifier_upper": 0.9090909090909091, "multiply_operator": 1.0 }, "confusion": [ [ 359, 27, 15 ], [ 0, 20, 2 ], [ 0, 0, 44 ] ], "ece": 0.10827849563156344 } }, { "epoch": 4, "training_loss": 0.3109758900042287, "learning_rate": 0.0009890738003669028, "validation": { "samples": 467, "accuracy": 0.9421841502189636, "macro_f1": 0.8499714367323622, "recall": { "identifier_lower": 0.9376558603491272, "identifier_upper": 0.9090909090909091, "multiply_operator": 1.0 }, "confusion": [ [ 376, 18, 7 ], [ 1, 20, 1 ], [ 0, 0, 44 ] ], "ece": 0.08608015172379027 } }, { "epoch": 5, "training_loss": 0.24503726018577604, "learning_rate": 0.0009829629131445341, "validation": { "samples": 467, "accuracy": 0.9486081600189209, "macro_f1": 0.8615622968817506, "recall": { "identifier_lower": 0.9451371571072319, "identifier_upper": 0.9090909090909091, "multiply_operator": 1.0 }, "confusion": [ [ 379, 17, 5 ], [ 1, 20, 1 ], [ 0, 0, 44 ] ], "ece": 0.06178676626927035 } }, { "epoch": 6, "training_loss": 0.2034695878375086, "learning_rate": 0.0009755282581475769, "validation": { "samples": 467, "accuracy": 0.9528908133506775, "macro_f1": 0.871260526795917, "recall": { "identifier_lower": 0.9501246882793017, "identifier_upper": 0.9090909090909091, "multiply_operator": 1.0 }, "confusion": [ [ 381, 14, 6 ], [ 1, 20, 1 ], [ 0, 0, 44 ] ], "ece": 0.04615953424233066 } }, { "epoch": 7, "training_loss": 0.1742880423530622, "learning_rate": 0.0009667902132486009, "validation": { "samples": 467, "accuracy": 0.95931476354599, "macro_f1": 0.8835361743512445, "recall": { "identifier_lower": 0.9576059850374065, "identifier_upper": 0.9090909090909091, "multiply_operator": 1.0 }, "confusion": [ [ 384, 13, 4 ], [ 1, 20, 1 ], [ 0, 0, 44 ] ], "ece": 0.043069085732812096 } }, { "epoch": 8, "training_loss": 0.1442930147035693, "learning_rate": 0.0009567727288213005, "validation": { "samples": 467, "accuracy": 0.95931476354599, "macro_f1": 0.8831616775779212, "recall": { "identifier_lower": 0.9600997506234414, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 385, 13, 3 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.03680740917091671 } }, { "epoch": 9, "training_loss": 0.13445695731819096, "learning_rate": 0.0009455032620941839, "validation": { "samples": 467, "accuracy": 0.9700214266777039, "macro_f1": 0.9081335240265146, "recall": { "identifier_lower": 0.970074812967581, "identifier_upper": 0.9090909090909091, "multiply_operator": 1.0 }, "confusion": [ [ 389, 9, 3 ], [ 1, 20, 1 ], [ 0, 0, 44 ] ], "ece": 0.03388859508217265 } }, { "epoch": 10, "training_loss": 0.1330791132080355, "learning_rate": 0.0009330127018922195, "validation": { "samples": 467, "accuracy": 0.9700214266777039, "macro_f1": 0.9066096145742163, "recall": { "identifier_lower": 0.970074812967581, "identifier_upper": 0.9090909090909091, "multiply_operator": 1.0 }, "confusion": [ [ 389, 10, 2 ], [ 1, 20, 1 ], [ 0, 0, 44 ] ], "ece": 0.029228197112581178 } }, { "epoch": 11, "training_loss": 0.11205834575831536, "learning_rate": 0.0009193352839727121, "validation": { "samples": 467, "accuracy": 0.9700214266777039, "macro_f1": 0.907825279474267, "recall": { "identifier_lower": 0.972568578553616, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 390, 9, 2 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.02395394208736132 } }, { "epoch": 12, "training_loss": 0.094981630116685, "learning_rate": 0.0009045084971874737, "validation": { "samples": 467, "accuracy": 0.9764453768730164, "macro_f1": 0.9233519504577847, "recall": { "identifier_lower": 0.9800498753117207, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 393, 7, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.020297527214571592 } }, { "epoch": 13, "training_loss": 0.10087664918989153, "learning_rate": 0.0008885729807284855, "validation": { "samples": 467, "accuracy": 0.9764453768730164, "macro_f1": 0.9204996544371354, "recall": { "identifier_lower": 0.9775561097256857, "identifier_upper": 0.9090909090909091, "multiply_operator": 1.0 }, "confusion": [ [ 392, 7, 2 ], [ 0, 20, 2 ], [ 0, 0, 44 ] ], "ece": 0.016511393352726972 } }, { "epoch": 14, "training_loss": 0.08016723319646343, "learning_rate": 0.0008715724127386972, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9270383370843582, "recall": { "identifier_lower": 0.9850374064837906, "identifier_upper": 0.8636363636363636, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 395, 5, 1 ], [ 2, 19, 1 ], [ 1, 0, 43 ] ], "ece": 0.01823449031824706 } }, { "epoch": 15, "training_loss": 0.07950997507333626, "learning_rate": 0.0008535533905932739, "validation": { "samples": 467, "accuracy": 0.9764453768730164, "macro_f1": 0.9233519504577847, "recall": { "identifier_lower": 0.9800498753117207, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 393, 7, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.013064050552685835 } }, { "epoch": 16, "training_loss": 0.07501162656896691, "learning_rate": 0.0008345653031794292, "validation": { "samples": 467, "accuracy": 0.9764453768730164, "macro_f1": 0.9223558412237658, "recall": { "identifier_lower": 0.9800498753117207, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 393, 6, 2 ], [ 0, 20, 2 ], [ 1, 0, 43 ] ], "ece": 0.011953361475379592 } }, { "epoch": 17, "training_loss": 0.0662038832008355, "learning_rate": 0.0008146601955249188, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9294443739553914, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 394, 6, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.010072189168083834 } }, { "epoch": 18, "training_loss": 0.0708043240675609, "learning_rate": 0.0007938926261462366, "validation": { "samples": 467, "accuracy": 0.9743040800094604, "macro_f1": 0.9185071611968995, "recall": { "identifier_lower": 0.9775561097256857, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 392, 6, 3 ], [ 0, 20, 2 ], [ 1, 0, 43 ] ], "ece": 0.011191716724897188 } }, { "epoch": 19, "training_loss": 0.059418706099460934, "learning_rate": 0.0007723195175075136, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9294443739553914, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 394, 6, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.007749744198634945 } }, { "epoch": 20, "training_loss": 0.05837932388611005, "learning_rate": 0.00075, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9294443739553914, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 394, 6, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.007486029059862731 } }, { "epoch": 21, "training_loss": 0.0658514654075474, "learning_rate": 0.0007269952498697733, "validation": { "samples": 467, "accuracy": 0.9743040800094604, "macro_f1": 0.9185071611968995, "recall": { "identifier_lower": 0.9775561097256857, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 392, 6, 3 ], [ 0, 20, 2 ], [ 1, 0, 43 ] ], "ece": 0.009387401378373017 } }, { "epoch": 22, "training_loss": 0.060872394502920486, "learning_rate": 0.0007033683215379002, "validation": { "samples": 467, "accuracy": 0.9764453768730164, "macro_f1": 0.9227401953472816, "recall": { "identifier_lower": 0.9775561097256857, "identifier_upper": 0.9090909090909091, "multiply_operator": 1.0 }, "confusion": [ [ 392, 6, 3 ], [ 0, 20, 2 ], [ 0, 0, 44 ] ], "ece": 0.0087111114112145 } }, { "epoch": 23, "training_loss": 0.055712475804556634, "learning_rate": 0.0006791839747726503, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9294443739553914, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 394, 6, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.014150080327061207 } }, { "epoch": 24, "training_loss": 0.05611806782743233, "learning_rate": 0.0006545084971874737, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9294443739553914, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 394, 6, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.009501690482484282 } }, { "epoch": 25, "training_loss": 0.05021703952551694, "learning_rate": 0.0006294095225512603, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.926491442280916, "recall": { "identifier_lower": 0.9850374064837906, "identifier_upper": 0.8636363636363636, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 395, 4, 2 ], [ 1, 19, 2 ], [ 1, 0, 43 ] ], "ece": 0.017516016547181962 } }, { "epoch": 26, "training_loss": 0.047697037862379076, "learning_rate": 0.0006039558454088795, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9294443739553914, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 394, 6, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.01044685828449704 } }, { "epoch": 27, "training_loss": 0.04439225993957824, "learning_rate": 0.0005782172325201154, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9294443739553914, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 394, 6, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.01024836830174436 } }, { "epoch": 28, "training_loss": 0.04366479368387782, "learning_rate": 0.0005522642316338267, "validation": { "samples": 467, "accuracy": 0.980728030204773, "macro_f1": 0.9410410793723215, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.9545454545454546, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 394, 6, 1 ], [ 1, 21, 0 ], [ 1, 0, 43 ] ], "ece": 0.011510787011652646 } }, { "epoch": 29, "training_loss": 0.03727660343059016, "learning_rate": 0.0005261679781214719, "validation": { "samples": 467, "accuracy": 0.980728030204773, "macro_f1": 0.9430843870930619, "recall": { "identifier_lower": 0.9850374064837906, "identifier_upper": 0.9545454545454546, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 395, 5, 1 ], [ 1, 21, 0 ], [ 2, 0, 42 ] ], "ece": 0.010586827790899367 } }, { "epoch": 30, "training_loss": 0.03949941129999395, "learning_rate": 0.0005000000000000001, "validation": { "samples": 467, "accuracy": 0.9764453768730164, "macro_f1": 0.9233519504577847, "recall": { "identifier_lower": 0.9800498753117207, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 393, 7, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.010985663576965538 } }, { "epoch": 31, "training_loss": 0.03434795174765838, "learning_rate": 0.00047383202187852816, "validation": { "samples": 467, "accuracy": 0.9828693866729736, "macro_f1": 0.9403964741043392, "recall": { "identifier_lower": 0.9900249376558603, "identifier_upper": 0.8636363636363636, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 397, 3, 1 ], [ 2, 19, 1 ], [ 1, 0, 43 ] ], "ece": 0.012696523661337333 } }, { "epoch": 32, "training_loss": 0.03366550082202825, "learning_rate": 0.00044773576836617336, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9294443739553914, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 394, 6, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.010714981164157616 } }, { "epoch": 33, "training_loss": 0.02854459550032234, "learning_rate": 0.0004217827674798846, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9294443739553914, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 394, 6, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.01113017341197614 } }, { "epoch": 34, "training_loss": 0.02719002439282339, "learning_rate": 0.00039604415459112036, "validation": { "samples": 467, "accuracy": 0.980728030204773, "macro_f1": 0.9357769673206845, "recall": { "identifier_lower": 0.9850374064837906, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 395, 5, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.00673925313164761 } }, { "epoch": 35, "training_loss": 0.0331632192017437, "learning_rate": 0.00037059047744873974, "validation": { "samples": 467, "accuracy": 0.9764453768730164, "macro_f1": 0.9233519504577847, "recall": { "identifier_lower": 0.9800498753117207, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 393, 7, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.013489080056338226 } }, { "epoch": 36, "training_loss": 0.031542465002102234, "learning_rate": 0.0003454915028125264, "validation": { "samples": 467, "accuracy": 0.980728030204773, "macro_f1": 0.9357769673206845, "recall": { "identifier_lower": 0.9850374064837906, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 395, 5, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.009979711013570536 } }, { "epoch": 37, "training_loss": 0.03016119669183984, "learning_rate": 0.0003208160252273499, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9294443739553914, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 394, 6, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.010851028115519135 } }, { "epoch": 38, "training_loss": 0.02717547928194151, "learning_rate": 0.0002966316784621, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9294443739553914, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 394, 6, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.01079029480006835 } }, { "epoch": 39, "training_loss": 0.0310103608408148, "learning_rate": 0.0002730047501302267, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9294443739553914, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 394, 6, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.010976817397607086 } }, { "epoch": 40, "training_loss": 0.026378964485470637, "learning_rate": 0.00025000000000000017, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9294443739553914, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 394, 6, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.011242367301014355 } }, { "epoch": 41, "training_loss": 0.026919757928185493, "learning_rate": 0.00022768048249248665, "validation": { "samples": 467, "accuracy": 0.980728030204773, "macro_f1": 0.9357769673206845, "recall": { "identifier_lower": 0.9850374064837906, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 395, 5, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.008514272331456352 } }, { "epoch": 42, "training_loss": 0.02101406411269328, "learning_rate": 0.00020610737385376354, "validation": { "samples": 467, "accuracy": 0.9850106835365295, "macro_f1": 0.9492270079067833, "recall": { "identifier_lower": 0.9900249376558603, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 397, 3, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.01576937972125217 } }, { "epoch": 43, "training_loss": 0.025051051183624357, "learning_rate": 0.0001853398044750814, "validation": { "samples": 467, "accuracy": 0.980728030204773, "macro_f1": 0.9357769673206845, "recall": { "identifier_lower": 0.9850374064837906, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 395, 5, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.008957387820282159 } }, { "epoch": 44, "training_loss": 0.02337893257506913, "learning_rate": 0.0001654346968205711, "validation": { "samples": 467, "accuracy": 0.980728030204773, "macro_f1": 0.9357769673206845, "recall": { "identifier_lower": 0.9850374064837906, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 395, 5, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.009912979300719327 } }, { "epoch": 45, "training_loss": 0.02451852051437449, "learning_rate": 0.0001464466094067263, "validation": { "samples": 467, "accuracy": 0.980728030204773, "macro_f1": 0.9357769673206845, "recall": { "identifier_lower": 0.9850374064837906, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 395, 5, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.013544975370280626 } }, { "epoch": 46, "training_loss": 0.026250681678198877, "learning_rate": 0.00012842758726130298, "validation": { "samples": 467, "accuracy": 0.980728030204773, "macro_f1": 0.9357769673206845, "recall": { "identifier_lower": 0.9850374064837906, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 395, 5, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.00912011481585609 } }, { "epoch": 47, "training_loss": 0.02889129058082662, "learning_rate": 0.00011142701927151459, "validation": { "samples": 467, "accuracy": 0.980728030204773, "macro_f1": 0.9357769673206845, "recall": { "identifier_lower": 0.9850374064837906, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 395, 5, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.011757288636972202 } }, { "epoch": 48, "training_loss": 0.022403624845621327, "learning_rate": 9.549150281252635e-05, "validation": { "samples": 467, "accuracy": 0.980728030204773, "macro_f1": 0.9357769673206845, "recall": { "identifier_lower": 0.9850374064837906, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 395, 5, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.011716582876258735 } }, { "epoch": 49, "training_loss": 0.025375535765540896, "learning_rate": 8.066471602728805e-05, "validation": { "samples": 467, "accuracy": 0.9828693866729736, "macro_f1": 0.9423654670112596, "recall": { "identifier_lower": 0.9875311720698254, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 396, 4, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.009382125637459077 } }, { "epoch": 50, "training_loss": 0.022656148082484517, "learning_rate": 6.698729810778066e-05, "validation": { "samples": 467, "accuracy": 0.9828693866729736, "macro_f1": 0.9423654670112596, "recall": { "identifier_lower": 0.9875311720698254, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 396, 4, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.00923494300787589 } }, { "epoch": 51, "training_loss": 0.023824987376462357, "learning_rate": 5.449673790581612e-05, "validation": { "samples": 467, "accuracy": 0.9828693866729736, "macro_f1": 0.9423654670112596, "recall": { "identifier_lower": 0.9875311720698254, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 396, 4, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.009213883294823783 } }, { "epoch": 52, "training_loss": 0.023352916523705403, "learning_rate": 4.322727117869952e-05, "validation": { "samples": 467, "accuracy": 0.9828693866729736, "macro_f1": 0.9423654670112596, "recall": { "identifier_lower": 0.9875311720698254, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 396, 4, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.011872639667516796 } }, { "epoch": 53, "training_loss": 0.021687946286104848, "learning_rate": 3.3209786751399136e-05, "validation": { "samples": 467, "accuracy": 0.9828693866729736, "macro_f1": 0.9423654670112596, "recall": { "identifier_lower": 0.9875311720698254, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 396, 4, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.012312911475995239 } }, { "epoch": 54, "training_loss": 0.02064617546330664, "learning_rate": 2.4471741852423238e-05, "validation": { "samples": 467, "accuracy": 0.9828693866729736, "macro_f1": 0.9423654670112596, "recall": { "identifier_lower": 0.9875311720698254, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 396, 4, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.011963830815148915 } }, { "epoch": 55, "training_loss": 0.02330371331606864, "learning_rate": 1.7037086855465846e-05, "validation": { "samples": 467, "accuracy": 0.9828693866729736, "macro_f1": 0.9423654670112596, "recall": { "identifier_lower": 0.9875311720698254, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 396, 4, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.010443179809695252 } }, { "epoch": 56, "training_loss": 0.023487418434663745, "learning_rate": 1.0926199633097158e-05, "validation": { "samples": 467, "accuracy": 0.9828693866729736, "macro_f1": 0.9423654670112596, "recall": { "identifier_lower": 0.9875311720698254, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 396, 4, 1 ], [ 1, 20, 1 ], [ 1, 0, 43 ] ], "ece": 0.010452173984375396 } } ], "checkpoint": "behavior_role_head.pt", "checkpoint_bytes": 75794, "track": "R_noncommercial_only", "product_validation": false, "interpretation_limit": "CROHME 연구용 연속 수식 행동 head이며 상용 checkpoint에 병합할 수 없다. P-track writer/device-disjoint 재학습이 필요하다." }