{ "experiment": "R-MATH-INK-06-BEHAVIOR-ROLE-001", "generated_at": "2026-07-23T21:32:46.157929+00:00", "seed": 17, "device": "cuda", "cuda_device": "NVIDIA GeForce GTX 1650", "teacher_adapter": "research\\runs\\math_ink_06_online_casecontext_refined_seed17_20260723\\skeleton_adapter.pt", "base_checkpoint": "C:\\Users\\user\\Desktop\\Aiflow\\math-grid-drawer-simulation\\research\\runs\\math_ink_06_federated_virtual_ce025_family010_seed17_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 } }, "product_proxy": { "samples": 1590, "role_counts": { "identifier_upper": 781, "identifier_lower": 809 }, "ambiguous_size_samples": 389, "sources": { "hwrt": 438, "uci-uji-pen-v1": 256, "uci-uji-pen-v2": 896 }, "track": "P_approved_synthetic_layout_proxy", "actual_continuous_formula": false, "target_bases": [ "c", "o", "s", "u", "v", "w", "x", "z" ], "lowercase_ratio": 1.0 }, "product_proxy_loss_weight": 0.35, "teacher_baseline": { "validation": { "samples": 467, "accuracy": 0.5117772817611694, "macro_f1": 0.2694663504489392, "recall": { "identifier_lower": 0.5610972568578554, "identifier_upper": 0.6363636363636364, "multiply_operator": 0.0 }, "confusion": [ [ 225, 176, 0 ], [ 8, 14, 0 ], [ 17, 27, 0 ] ], "ece": 0.2943883374417391 }, "official_test": { "samples": 635, "accuracy": 0.7433071136474609, "macro_f1": 0.3644206619715246, "recall": { "identifier_lower": 0.7922535211267606, "identifier_upper": 0.7096774193548387, "multiply_operator": 0.0 }, "confusion": [ [ 450, 118, 0 ], [ 9, 22, 0 ], [ 16, 20, 0 ] ], "ece": 0.07124687355102433 } }, "selected_epoch": 30, "validation": { "samples": 467, "accuracy": 0.980728030204773, "macro_f1": 0.9269215563878485, "recall": { "identifier_lower": 0.9900249376558603, "identifier_upper": 0.8636363636363636, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 397, 3, 1 ], [ 1, 19, 2 ], [ 1, 1, 42 ] ], "ece": 0.009913665229302898 }, "official_test": { "samples": 635, "accuracy": 0.930708646774292, "macro_f1": 0.7361821054646563, "recall": { "identifier_lower": 0.9612676056338029, "identifier_upper": 0.45161290322580644, "multiply_operator": 0.8611111111111112 }, "confusion": [ [ 546, 5, 17 ], [ 10, 14, 7 ], [ 5, 0, 31 ] ], "ece": 0.05347054235840293 }, "history": [ { "epoch": 1, "training_loss": 0.4695802242751037, "learning_rate": 0.0009993147673772868, "validation": { "samples": 467, "accuracy": 0.9314774870872498, "macro_f1": 0.6810897891494907, "recall": { "identifier_lower": 0.9675810473815462, "identifier_upper": 0.13636363636363635, "multiply_operator": 1.0 }, "confusion": [ [ 388, 0, 13 ], [ 15, 3, 4 ], [ 0, 0, 44 ] ], "ece": 0.18756604964739892 } }, { "epoch": 2, "training_loss": 0.26416257603565735, "learning_rate": 0.0009972609476841365, "validation": { "samples": 467, "accuracy": 0.9486081600189209, "macro_f1": 0.805022859517872, "recall": { "identifier_lower": 0.972568578553616, "identifier_upper": 0.4090909090909091, "multiply_operator": 1.0 }, "confusion": [ [ 390, 1, 10 ], [ 11, 9, 2 ], [ 0, 0, 44 ] ], "ece": 0.0989909679372516 } }, { "epoch": 3, "training_loss": 0.1998122558279348, "learning_rate": 0.0009938441702975688, "validation": { "samples": 467, "accuracy": 0.9571734666824341, "macro_f1": 0.852514619883041, "recall": { "identifier_lower": 0.9750623441396509, "identifier_upper": 0.5909090909090909, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 391, 3, 7 ], [ 7, 13, 2 ], [ 1, 0, 43 ] ], "ece": 0.05527584794546986 } }, { "epoch": 4, "training_loss": 0.15762516093153564, "learning_rate": 0.0009890738003669028, "validation": { "samples": 467, "accuracy": 0.9571734666824341, "macro_f1": 0.8743495077355837, "recall": { "identifier_lower": 0.9625935162094763, "identifier_upper": 0.8181818181818182, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 386, 8, 7 ], [ 2, 18, 2 ], [ 1, 0, 43 ] ], "ece": 0.04046737894665883 } }, { "epoch": 5, "training_loss": 0.13147952250744166, "learning_rate": 0.0009829629131445341, "validation": { "samples": 467, "accuracy": 0.9700214266777039, "macro_f1": 0.900130421076447, "recall": { "identifier_lower": 0.9800498753117207, "identifier_upper": 0.7727272727272727, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 393, 4, 4 ], [ 3, 17, 2 ], [ 1, 0, 43 ] ], "ece": 0.026655886021542147 } }, { "epoch": 6, "training_loss": 0.115951776274868, "learning_rate": 0.0009755282581475769, "validation": { "samples": 467, "accuracy": 0.9614561200141907, "macro_f1": 0.8815638656064189, "recall": { "identifier_lower": 0.9675810473815462, "identifier_upper": 0.8181818181818182, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 388, 8, 5 ], [ 2, 18, 2 ], [ 1, 0, 43 ] ], "ece": 0.03733365243572495 } }, { "epoch": 7, "training_loss": 0.10126321473139668, "learning_rate": 0.0009667902132486009, "validation": { "samples": 467, "accuracy": 0.9678800702095032, "macro_f1": 0.8969958923966703, "recall": { "identifier_lower": 0.9750623441396509, "identifier_upper": 0.8181818181818182, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 391, 6, 4 ], [ 2, 18, 2 ], [ 1, 0, 43 ] ], "ece": 0.020665142067610387 } }, { "epoch": 8, "training_loss": 0.09452478858006143, "learning_rate": 0.0009567727288213005, "validation": { "samples": 467, "accuracy": 0.9743040800094604, "macro_f1": 0.9134776995188895, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.8181818181818182, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 394, 4, 3 ], [ 2, 18, 2 ], [ 1, 0, 43 ] ], "ece": 0.017813534384579915 } }, { "epoch": 9, "training_loss": 0.0816046844070545, "learning_rate": 0.0009455032620941839, "validation": { "samples": 467, "accuracy": 0.9721627235412598, "macro_f1": 0.9046206280399721, "recall": { "identifier_lower": 0.9800498753117207, "identifier_upper": 0.8181818181818182, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 393, 6, 2 ], [ 2, 18, 2 ], [ 1, 0, 43 ] ], "ece": 0.016128934220334984 } }, { "epoch": 10, "training_loss": 0.07826958534094648, "learning_rate": 0.0009330127018922195, "validation": { "samples": 467, "accuracy": 0.9721627235412598, "macro_f1": 0.9046206280399721, "recall": { "identifier_lower": 0.9800498753117207, "identifier_upper": 0.8181818181818182, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 393, 6, 2 ], [ 2, 18, 2 ], [ 1, 0, 43 ] ], "ece": 0.017286768024444565 } }, { "epoch": 11, "training_loss": 0.07614069413279995, "learning_rate": 0.0009193352839727121, "validation": { "samples": 467, "accuracy": 0.9743040800094604, "macro_f1": 0.9108412055780478, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.8181818181818182, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 394, 5, 2 ], [ 2, 18, 2 ], [ 1, 0, 43 ] ], "ece": 0.011480572244482401 } }, { "epoch": 12, "training_loss": 0.07079621948849915, "learning_rate": 0.0009045084971874737, "validation": { "samples": 467, "accuracy": 0.9764453768730164, "macro_f1": 0.9147638251518102, "recall": { "identifier_lower": 0.9850374064837906, "identifier_upper": 0.8181818181818182, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 395, 5, 1 ], [ 2, 18, 2 ], [ 1, 0, 43 ] ], "ece": 0.012716759466477046 } }, { "epoch": 13, "training_loss": 0.06981331932402586, "learning_rate": 0.0008885729807284855, "validation": { "samples": 467, "accuracy": 0.9743040800094604, "macro_f1": 0.9108412055780478, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.8181818181818182, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 394, 5, 2 ], [ 2, 18, 2 ], [ 1, 0, 43 ] ], "ece": 0.01353120519745804 } }, { "epoch": 14, "training_loss": 0.06253736879753491, "learning_rate": 0.0008715724127386972, "validation": { "samples": 467, "accuracy": 0.9764453768730164, "macro_f1": 0.9147638251518102, "recall": { "identifier_lower": 0.9850374064837906, "identifier_upper": 0.8181818181818182, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 395, 5, 1 ], [ 2, 18, 2 ], [ 1, 0, 43 ] ], "ece": 0.012102323031857154 } }, { "epoch": 15, "training_loss": 0.060433164074276034, "learning_rate": 0.0008535533905932739, "validation": { "samples": 467, "accuracy": 0.9764453768730164, "macro_f1": 0.9147638251518102, "recall": { "identifier_lower": 0.9850374064837906, "identifier_upper": 0.8181818181818182, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 395, 5, 1 ], [ 2, 18, 2 ], [ 1, 0, 43 ] ], "ece": 0.009877669651074283 } }, { "epoch": 16, "training_loss": 0.06271934157871652, "learning_rate": 0.0008345653031794292, "validation": { "samples": 467, "accuracy": 0.9743040800094604, "macro_f1": 0.908544307628976, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.8181818181818182, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 394, 6, 1 ], [ 2, 18, 2 ], [ 1, 0, 43 ] ], "ece": 0.010271194092643848 } }, { "epoch": 17, "training_loss": 0.06164936625999086, "learning_rate": 0.0008146601955249188, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9262795458775358, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 394, 6, 1 ], [ 0, 20, 2 ], [ 1, 0, 43 ] ], "ece": 0.014615969672941431 } }, { "epoch": 18, "training_loss": 0.05562269583769093, "learning_rate": 0.0007938926261462366, "validation": { "samples": 467, "accuracy": 0.9743040800094604, "macro_f1": 0.9164296279440208, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.8636363636363636, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 394, 6, 1 ], [ 2, 19, 1 ], [ 2, 0, 42 ] ], "ece": 0.00953849047297134 } }, { "epoch": 19, "training_loss": 0.055309959967041404, "learning_rate": 0.0007723195175075136, "validation": { "samples": 467, "accuracy": 0.9743040800094604, "macro_f1": 0.9104400749063671, "recall": { "identifier_lower": 0.9850374064837906, "identifier_upper": 0.8181818181818182, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 395, 5, 1 ], [ 2, 18, 2 ], [ 2, 0, 42 ] ], "ece": 0.01310491915415099 } }, { "epoch": 20, "training_loss": 0.052588789419149996, "learning_rate": 0.00075, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9262795458775358, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9772727272727273 }, "confusion": [ [ 394, 6, 1 ], [ 0, 20, 2 ], [ 1, 0, 43 ] ], "ece": 0.010324389672118767 } }, { "epoch": 21, "training_loss": 0.05353964164129486, "learning_rate": 0.0007269952498697733, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9260882230545152, "recall": { "identifier_lower": 0.9875311720698254, "identifier_upper": 0.8636363636363636, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 396, 4, 1 ], [ 1, 19, 2 ], [ 2, 0, 42 ] ], "ece": 0.010287888492073655 } }, { "epoch": 22, "training_loss": 0.051659728876767584, "learning_rate": 0.0007033683215379002, "validation": { "samples": 467, "accuracy": 0.9764453768730164, "macro_f1": 0.9141019334534439, "recall": { "identifier_lower": 0.9850374064837906, "identifier_upper": 0.8636363636363636, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 395, 5, 1 ], [ 1, 19, 2 ], [ 1, 1, 42 ] ], "ece": 0.01312589043445081 } }, { "epoch": 23, "training_loss": 0.04992970377076813, "learning_rate": 0.0006791839747726503, "validation": { "samples": 467, "accuracy": 0.9764453768730164, "macro_f1": 0.9195477003802385, "recall": { "identifier_lower": 0.9850374064837906, "identifier_upper": 0.8636363636363636, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 395, 5, 1 ], [ 1, 19, 2 ], [ 2, 0, 42 ] ], "ece": 0.009554800505844507 } }, { "epoch": 24, "training_loss": 0.053961525392518486, "learning_rate": 0.0006545084971874737, "validation": { "samples": 467, "accuracy": 0.9764453768730164, "macro_f1": 0.9219537372512715, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 394, 6, 1 ], [ 0, 20, 2 ], [ 2, 0, 42 ] ], "ece": 0.012457899277120515 } }, { "epoch": 25, "training_loss": 0.04929926841757817, "learning_rate": 0.0006294095225512603, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9227902073474655, "recall": { "identifier_lower": 0.9850374064837906, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 395, 5, 1 ], [ 0, 20, 2 ], [ 1, 1, 42 ] ], "ece": 0.008438708652121402 } }, { "epoch": 26, "training_loss": 0.04608062488780045, "learning_rate": 0.0006039558454088795, "validation": { "samples": 467, "accuracy": 0.9764453768730164, "macro_f1": 0.9166988346916883, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 394, 6, 1 ], [ 0, 20, 2 ], [ 1, 1, 42 ] ], "ece": 0.012565350727601343 } }, { "epoch": 27, "training_loss": 0.0441696440124536, "learning_rate": 0.0005782172325201154, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9203820766839513, "recall": { "identifier_lower": 0.9875311720698254, "identifier_upper": 0.8636363636363636, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 396, 4, 1 ], [ 1, 19, 2 ], [ 1, 1, 42 ] ], "ece": 0.011315091112917064 } }, { "epoch": 28, "training_loss": 0.0467573282674202, "learning_rate": 0.0005522642316338267, "validation": { "samples": 467, "accuracy": 0.9764453768730164, "macro_f1": 0.9166988346916883, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 394, 6, 1 ], [ 0, 20, 2 ], [ 1, 1, 42 ] ], "ece": 0.013571739072491401 } }, { "epoch": 29, "training_loss": 0.04851798766069753, "learning_rate": 0.0005261679781214719, "validation": { "samples": 467, "accuracy": 0.9764453768730164, "macro_f1": 0.9141019334534439, "recall": { "identifier_lower": 0.9850374064837906, "identifier_upper": 0.8636363636363636, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 395, 5, 1 ], [ 1, 19, 2 ], [ 1, 1, 42 ] ], "ece": 0.011841069713424537 } }, { "epoch": 30, "training_loss": 0.04493519555594242, "learning_rate": 0.0005000000000000001, "validation": { "samples": 467, "accuracy": 0.980728030204773, "macro_f1": 0.9269215563878485, "recall": { "identifier_lower": 0.9900249376558603, "identifier_upper": 0.8636363636363636, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 397, 3, 1 ], [ 1, 19, 2 ], [ 1, 1, 42 ] ], "ece": 0.009913665229302898 } }, { "epoch": 31, "training_loss": 0.04379846392008204, "learning_rate": 0.00047383202187852816, "validation": { "samples": 467, "accuracy": 0.9764453768730164, "macro_f1": 0.9166988346916883, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 394, 6, 1 ], [ 0, 20, 2 ], [ 1, 1, 42 ] ], "ece": 0.01218321381668197 } }, { "epoch": 32, "training_loss": 0.042996837788129585, "learning_rate": 0.00044773576836617336, "validation": { "samples": 467, "accuracy": 0.9764453768730164, "macro_f1": 0.9141019334534439, "recall": { "identifier_lower": 0.9850374064837906, "identifier_upper": 0.8636363636363636, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 395, 5, 1 ], [ 1, 19, 2 ], [ 1, 1, 42 ] ], "ece": 0.013857337607995873 } }, { "epoch": 33, "training_loss": 0.04039572703934541, "learning_rate": 0.0004217827674798846, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9203820766839513, "recall": { "identifier_lower": 0.9875311720698254, "identifier_upper": 0.8636363636363636, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 396, 4, 1 ], [ 1, 19, 2 ], [ 1, 1, 42 ] ], "ece": 0.012063260662852376 } }, { "epoch": 34, "training_loss": 0.042098714903703295, "learning_rate": 0.00039604415459112036, "validation": { "samples": 467, "accuracy": 0.9743040800094604, "macro_f1": 0.9080648483623827, "recall": { "identifier_lower": 0.9825436408977556, "identifier_upper": 0.8636363636363636, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 394, 6, 1 ], [ 1, 19, 2 ], [ 1, 1, 42 ] ], "ece": 0.015506909390579637 } }, { "epoch": 35, "training_loss": 0.04239705811041009, "learning_rate": 0.00037059047744873974, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9203820766839513, "recall": { "identifier_lower": 0.9875311720698254, "identifier_upper": 0.8636363636363636, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 396, 4, 1 ], [ 1, 19, 2 ], [ 1, 1, 42 ] ], "ece": 0.010795353270132424 } }, { "epoch": 36, "training_loss": 0.042145438765276655, "learning_rate": 0.0003454915028125264, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9203820766839513, "recall": { "identifier_lower": 0.9875311720698254, "identifier_upper": 0.8636363636363636, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 396, 4, 1 ], [ 1, 19, 2 ], [ 1, 1, 42 ] ], "ece": 0.013025458414588437 } }, { "epoch": 37, "training_loss": 0.039371089365817534, "learning_rate": 0.0003208160252273499, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9203820766839513, "recall": { "identifier_lower": 0.9875311720698254, "identifier_upper": 0.8636363636363636, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 396, 4, 1 ], [ 1, 19, 2 ], [ 1, 1, 42 ] ], "ece": 0.011519250068731807 } }, { "epoch": 38, "training_loss": 0.04198872989981145, "learning_rate": 0.0002966316784621, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9203820766839513, "recall": { "identifier_lower": 0.9875311720698254, "identifier_upper": 0.8636363636363636, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 396, 4, 1 ], [ 1, 19, 2 ], [ 1, 1, 42 ] ], "ece": 0.013143556901195869 } }, { "epoch": 39, "training_loss": 0.04180180780561416, "learning_rate": 0.0002730047501302267, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9203820766839513, "recall": { "identifier_lower": 0.9875311720698254, "identifier_upper": 0.8636363636363636, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 396, 4, 1 ], [ 1, 19, 2 ], [ 1, 1, 42 ] ], "ece": 0.013186807601712869 } }, { "epoch": 40, "training_loss": 0.04202993525780887, "learning_rate": 0.00025000000000000017, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9203820766839513, "recall": { "identifier_lower": 0.9875311720698254, "identifier_upper": 0.8636363636363636, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 396, 4, 1 ], [ 1, 19, 2 ], [ 1, 1, 42 ] ], "ece": 0.011174082918664313 } }, { "epoch": 41, "training_loss": 0.04020663223593049, "learning_rate": 0.00022768048249248665, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9203820766839513, "recall": { "identifier_lower": 0.9875311720698254, "identifier_upper": 0.8636363636363636, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 396, 4, 1 ], [ 1, 19, 2 ], [ 1, 1, 42 ] ], "ece": 0.013505464233618508 } }, { "epoch": 42, "training_loss": 0.041365806391399314, "learning_rate": 0.00020610737385376354, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9227902073474655, "recall": { "identifier_lower": 0.9850374064837906, "identifier_upper": 0.9090909090909091, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 395, 5, 1 ], [ 0, 20, 2 ], [ 1, 1, 42 ] ], "ece": 0.017064912102465496 } }, { "epoch": 43, "training_loss": 0.0394672335816889, "learning_rate": 0.0001853398044750814, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9203820766839513, "recall": { "identifier_lower": 0.9875311720698254, "identifier_upper": 0.8636363636363636, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 396, 4, 1 ], [ 1, 19, 2 ], [ 1, 1, 42 ] ], "ece": 0.014093965084634147 } }, { "epoch": 44, "training_loss": 0.03919070995379999, "learning_rate": 0.0001654346968205711, "validation": { "samples": 467, "accuracy": 0.978586733341217, "macro_f1": 0.9203820766839513, "recall": { "identifier_lower": 0.9875311720698254, "identifier_upper": 0.8636363636363636, "multiply_operator": 0.9545454545454546 }, "confusion": [ [ 396, 4, 1 ], [ 1, 19, 2 ], [ 1, 1, 42 ] ], "ece": 0.014121455602647742 } } ], "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." }