{ "best_metric": null, "best_model_checkpoint": null, "epoch": 47.0, "eval_steps": 500, "global_step": 3807, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 1.0, "eval_B": { "f1-score": 0.7905866302864938, "precision": 0.7768096514745308, "recall": 0.8048611111111111, "support": 1440.0 }, "eval_I": { "f1-score": 0.9005331227169513, "precision": 0.9637593111099372, "recall": 0.8450919534905267, "support": 21587.0 }, "eval_O": { "f1-score": 0.8288892663043479, "precision": 0.7463108800367001, "recall": 0.9320156593144275, "support": 10473.0 }, "eval_accuracy": 0.8705373134328358, "eval_loss": 0.32930415868759155, "eval_macro avg": { "f1-score": 0.8400030064359308, "precision": 0.8289599475403894, "recall": 0.8606562413053552, "support": 33500.0 }, "eval_runtime": 4.7349, "eval_samples_per_second": 17.107, "eval_steps_per_second": 2.323, "eval_weighted avg": { "f1-score": 0.8734092702599646, "precision": 0.8877430445874, "recall": 0.8705373134328358, "support": 33500.0 }, "step": 81 }, { "epoch": 2.0, "eval_B": { "f1-score": 0.8656218739579861, "precision": 0.8325849903784477, "recall": 0.9013888888888889, "support": 1440.0 }, "eval_I": { "f1-score": 0.9547168948727912, "precision": 0.9423736462093862, "recall": 0.9673877796822161, "support": 21587.0 }, "eval_O": { "f1-score": 0.9013528191962081, "precision": 0.9332379102341274, "recall": 0.8715745249689678, "support": 10473.0 }, "eval_accuracy": 0.9345970149253732, "eval_loss": 0.2162298709154129, "eval_macro avg": { "f1-score": 0.9072305293423284, "precision": 0.902732182273987, "recall": 0.9134503978466909, "support": 33500.0 }, "eval_runtime": 4.753, "eval_samples_per_second": 17.042, "eval_steps_per_second": 2.314, "eval_weighted avg": { "f1-score": 0.9342040950316517, "precision": 0.9347982961417612, "recall": 0.9345970149253732, "support": 33500.0 }, "step": 162 }, { "epoch": 3.0, "eval_B": { "f1-score": 0.8603587759409076, "precision": 0.8717034925160371, "recall": 0.8493055555555555, "support": 1440.0 }, "eval_I": { "f1-score": 0.9434195218082959, "precision": 0.9677057963955188, "recall": 0.9203224162690509, "support": 21587.0 }, "eval_O": { "f1-score": 0.8879310344827586, "precision": 0.8459410391631365, "recall": 0.9343072663038289, "support": 10473.0 }, "eval_accuracy": 0.9216417910447762, "eval_loss": 0.24082203209400177, "eval_macro avg": { "f1-score": 0.8972364440773206, "precision": 0.8951167760248975, "recall": 0.9013117460428117, "support": 33500.0 }, "eval_runtime": 4.7554, "eval_samples_per_second": 17.033, "eval_steps_per_second": 2.313, "eval_weighted avg": { "f1-score": 0.9225019575751797, "precision": 0.9255121957960801, "recall": 0.9216417910447762, "support": 33500.0 }, "step": 243 }, { "epoch": 4.0, "eval_B": { "f1-score": 0.8804020100502512, "precision": 0.8504854368932039, "recall": 0.9125, "support": 1440.0 }, "eval_I": { "f1-score": 0.9589184992350025, "precision": 0.945595388218339, "recall": 0.9726224116366332, "support": 21587.0 }, "eval_O": { "f1-score": 0.9091178797468354, "precision": 0.9427750999897446, "recall": 0.877780960565263, "support": 10473.0 }, "eval_accuracy": 0.9403880597014925, "eval_loss": 0.22083505988121033, "eval_macro avg": { "f1-score": 0.9161461296773631, "precision": 0.9129519750337626, "recall": 0.9209677907339654, "support": 33500.0 }, "eval_runtime": 4.7449, "eval_samples_per_second": 17.071, "eval_steps_per_second": 2.318, "eval_weighted avg": { "f1-score": 0.9399744505088647, "precision": 0.9406253819936745, "recall": 0.9403880597014925, "support": 33500.0 }, "step": 324 }, { "epoch": 5.0, "eval_B": { "f1-score": 0.8879781420765027, "precision": 0.8736559139784946, "recall": 0.9027777777777778, "support": 1440.0 }, "eval_I": { "f1-score": 0.9613201259451634, "precision": 0.9539317642765919, "recall": 0.9688238291564367, "support": 21587.0 }, "eval_O": { "f1-score": 0.916006030835076, "precision": 0.9334853291038858, "recall": 0.899169292466342, "support": 10473.0 }, "eval_accuracy": 0.9442089552238806, "eval_loss": 0.2257838249206543, "eval_macro avg": { "f1-score": 0.921768099618914, "precision": 0.9203576691196576, "recall": 0.9235902998001855, "support": 33500.0 }, "eval_runtime": 4.7845, "eval_samples_per_second": 16.93, "eval_steps_per_second": 2.299, "eval_weighted avg": { "f1-score": 0.9440011117702735, "precision": 0.9440889959185916, "recall": 0.9442089552238806, "support": 33500.0 }, "step": 405 }, { "epoch": 6.0, "eval_B": { "f1-score": 0.8873579056148813, "precision": 0.8803827751196173, "recall": 0.8944444444444445, "support": 1440.0 }, "eval_I": { "f1-score": 0.9622060984570168, "precision": 0.9539267015706806, "recall": 0.9706304720433594, "support": 21587.0 }, "eval_O": { "f1-score": 0.9182769530299342, "precision": 0.9365567911040509, "recall": 0.9006970304592762, "support": 10473.0 }, "eval_accuracy": 0.9454925373134329, "eval_loss": 0.2242102175951004, "eval_macro avg": { "f1-score": 0.9226136523672773, "precision": 0.923622089264783, "recall": 0.9219239823156933, "support": 33500.0 }, "eval_runtime": 4.7715, "eval_samples_per_second": 16.976, "eval_steps_per_second": 2.305, "eval_weighted avg": { "f1-score": 0.945255312255509, "precision": 0.9453351097376493, "recall": 0.9454925373134329, "support": 33500.0 }, "step": 486 }, { "epoch": 6.17, "grad_norm": 1.8705081939697266, "learning_rate": 1.7530864197530865e-05, "loss": 0.1601, "step": 500 }, { "epoch": 7.0, "eval_B": { "f1-score": 0.8876749743939911, "precision": 0.8730691739422431, "recall": 0.9027777777777778, "support": 1440.0 }, "eval_I": { "f1-score": 0.9611415113943341, "precision": 0.9581108451482232, "recall": 0.9641914114976606, "support": 21587.0 }, "eval_O": { "f1-score": 0.9164739884393064, "precision": 0.9247594050743657, "recall": 0.9083357204239473, "support": 10473.0 }, "eval_accuracy": 0.944089552238806, "eval_loss": 0.22517083585262299, "eval_macro avg": { "f1-score": 0.9217634914092105, "precision": 0.9186464747216107, "recall": 0.9251016365664619, "support": 33500.0 }, "eval_runtime": 4.7926, "eval_samples_per_second": 16.901, "eval_steps_per_second": 2.295, "eval_weighted avg": { "f1-score": 0.9440192791200506, "precision": 0.9440287663891153, "recall": 0.944089552238806, "support": 33500.0 }, "step": 567 }, { "epoch": 8.0, "eval_B": { "f1-score": 0.894540522210919, "precision": 0.8740888005301524, "recall": 0.9159722222222222, "support": 1440.0 }, "eval_I": { "f1-score": 0.9638018385639943, "precision": 0.9540255968049378, "recall": 0.9737805160513272, "support": 21587.0 }, "eval_O": { "f1-score": 0.9203132648066569, "precision": 0.9441598875163202, "recall": 0.8976415544734078, "support": 10473.0 }, "eval_accuracy": 0.9474925373134329, "eval_loss": 0.24869142472743988, "eval_macro avg": { "f1-score": 0.9262185418605234, "precision": 0.9240914282838034, "recall": 0.9291314309156524, "support": 33500.0 }, "eval_runtime": 4.7743, "eval_samples_per_second": 16.966, "eval_steps_per_second": 2.304, "eval_weighted avg": { "f1-score": 0.947228939205516, "precision": 0.9475052218791054, "recall": 0.9474925373134329, "support": 33500.0 }, "step": 648 }, { "epoch": 9.0, "eval_B": { "f1-score": 0.8851236868858015, "precision": 0.8643282594308405, "recall": 0.9069444444444444, "support": 1440.0 }, "eval_I": { "f1-score": 0.9617001723147618, "precision": 0.9540067462849849, "recall": 0.9695186918052532, "support": 21587.0 }, "eval_O": { "f1-score": 0.9165854609237966, "precision": 0.9358272808675754, "recall": 0.8981189725961998, "support": 10473.0 }, "eval_accuracy": 0.9445074626865672, "eval_loss": 0.3045330345630646, "eval_macro avg": { "f1-score": 0.9211364400414532, "precision": 0.9180540955278002, "recall": 0.9248607029486324, "support": 33500.0 }, "eval_runtime": 4.7599, "eval_samples_per_second": 17.017, "eval_steps_per_second": 2.311, "eval_weighted avg": { "f1-score": 0.9443044555560967, "precision": 0.9444685205421044, "recall": 0.9445074626865672, "support": 33500.0 }, "step": 729 }, { "epoch": 10.0, "eval_B": { "f1-score": 0.8815068493150685, "precision": 0.8695945945945946, "recall": 0.89375, "support": 1440.0 }, "eval_I": { "f1-score": 0.956297959947595, "precision": 0.9660159757999717, "recall": 0.9467735211006625, "support": 21587.0 }, "eval_O": { "f1-score": 0.9101049868766404, "precision": 0.8937678357728068, "recall": 0.9270505108373914, "support": 10473.0 }, "eval_accuracy": 0.9383283582089552, "eval_loss": 0.32627996802330017, "eval_macro avg": { "f1-score": 0.9159699320464346, "precision": 0.909792802055791, "recall": 0.9225246773126846, "support": 33500.0 }, "eval_runtime": 4.7534, "eval_samples_per_second": 17.04, "eval_steps_per_second": 2.314, "eval_weighted avg": { "f1-score": 0.9386418940884026, "precision": 0.9392845859659347, "recall": 0.9383283582089552, "support": 33500.0 }, "step": 810 }, { "epoch": 11.0, "eval_B": { "f1-score": 0.8901136755080951, "precision": 0.8831168831168831, "recall": 0.8972222222222223, "support": 1440.0 }, "eval_I": { "f1-score": 0.9620073001078947, "precision": 0.9535359970874671, "recall": 0.9706304720433594, "support": 21587.0 }, "eval_O": { "f1-score": 0.917413322945072, "precision": 0.9361025539103647, "recall": 0.8994557433400172, "support": 10473.0 }, "eval_accuracy": 0.9452238805970149, "eval_loss": 0.3123624622821808, "eval_macro avg": { "f1-score": 0.9231780995203539, "precision": 0.9242518113715716, "recall": 0.922436145868533, "support": 33500.0 }, "eval_runtime": 4.7539, "eval_samples_per_second": 17.039, "eval_steps_per_second": 2.314, "eval_weighted avg": { "f1-score": 0.9449756719810304, "precision": 0.9450588635199616, "recall": 0.9452238805970149, "support": 33500.0 }, "step": 891 }, { "epoch": 12.0, "eval_B": { "f1-score": 0.8800819952169457, "precision": 0.8661735036987223, "recall": 0.8944444444444445, "support": 1440.0 }, "eval_I": { "f1-score": 0.9595463311598861, "precision": 0.9512039692293687, "recall": 0.9680363181544448, "support": 21587.0 }, "eval_O": { "f1-score": 0.9115367743822197, "precision": 0.9310035842293907, "recall": 0.8928673732454884, "support": 10473.0 }, "eval_accuracy": 0.9413731343283582, "eval_loss": 0.3427491784095764, "eval_macro avg": { "f1-score": 0.9170550335863505, "precision": 0.9161270190524938, "recall": 0.9184493786147926, "support": 33500.0 }, "eval_runtime": 4.7455, "eval_samples_per_second": 17.069, "eval_steps_per_second": 2.318, "eval_weighted avg": { "f1-score": 0.9411214734915179, "precision": 0.9412337452750731, "recall": 0.9413731343283582, "support": 33500.0 }, "step": 972 }, { "epoch": 12.35, "grad_norm": 0.47551417350769043, "learning_rate": 1.506172839506173e-05, "loss": 0.0222, "step": 1000 }, { "epoch": 13.0, "eval_B": { "f1-score": 0.8896434634974534, "precision": 0.8704318936877077, "recall": 0.9097222222222222, "support": 1440.0 }, "eval_I": { "f1-score": 0.9620136667970457, "precision": 0.9556591698665204, "recall": 0.9684532357437347, "support": 21587.0 }, "eval_O": { "f1-score": 0.9183177933177933, "precision": 0.9343808676746714, "recall": 0.9027976701995608, "support": 10473.0 }, "eval_accuracy": 0.9454029850746268, "eval_loss": 0.3816670775413513, "eval_macro avg": { "f1-score": 0.9233249745374308, "precision": 0.9201573104096331, "recall": 0.9269910427218392, "support": 33500.0 }, "eval_runtime": 4.7247, "eval_samples_per_second": 17.144, "eval_steps_per_second": 2.328, "eval_weighted avg": { "f1-score": 0.945242324238848, "precision": 0.9453435001186778, "recall": 0.9454029850746268, "support": 33500.0 }, "step": 1053 }, { "epoch": 14.0, "eval_B": { "f1-score": 0.8865272540281112, "precision": 0.8754231550440081, "recall": 0.8979166666666667, "support": 1440.0 }, "eval_I": { "f1-score": 0.9599206715247671, "precision": 0.9557331129172981, "recall": 0.9641450873210728, "support": 21587.0 }, "eval_O": { "f1-score": 0.9141367826632559, "precision": 0.9242631270739801, "recall": 0.9042299245679366, "support": 10473.0 }, "eval_accuracy": 0.9425671641791045, "eval_loss": 0.3827010989189148, "eval_macro avg": { "f1-score": 0.9201949027387114, "precision": 0.9184731316784287, "recall": 0.9220972261852255, "support": 33500.0 }, "eval_runtime": 4.7319, "eval_samples_per_second": 17.118, "eval_steps_per_second": 2.325, "eval_weighted avg": { "f1-score": 0.9424525763235196, "precision": 0.9424426203479068, "recall": 0.9425671641791045, "support": 33500.0 }, "step": 1134 }, { "epoch": 15.0, "eval_B": { "f1-score": 0.8871919472405415, "precision": 0.8868841082581541, "recall": 0.8875, "support": 1440.0 }, "eval_I": { "f1-score": 0.9616856471022177, "precision": 0.9571841585975862, "recall": 0.9662296752675221, "support": 21587.0 }, "eval_O": { "f1-score": 0.9181813798756087, "precision": 0.9273470977795092, "recall": 0.9091950730449728, "support": 10473.0 }, "eval_accuracy": 0.9450149253731344, "eval_loss": 0.39016807079315186, "eval_macro avg": { "f1-score": 0.9223529914061226, "precision": 0.9238051215450831, "recall": 0.920974916104165, "support": 33500.0 }, "eval_runtime": 4.7376, "eval_samples_per_second": 17.097, "eval_steps_per_second": 2.322, "eval_weighted avg": { "f1-score": 0.944882927148036, "precision": 0.9448344388830637, "recall": 0.9450149253731344, "support": 33500.0 }, "step": 1215 }, { "epoch": 16.0, "eval_B": { "f1-score": 0.8897338403041825, "precision": 0.8857536132140399, "recall": 0.89375, "support": 1440.0 }, "eval_I": { "f1-score": 0.9608010520729992, "precision": 0.9570693142121713, "recall": 0.9645620049103627, "support": 21587.0 }, "eval_O": { "f1-score": 0.9159121556540166, "precision": 0.9240112719852298, "recall": 0.9079537859257137, "support": 10473.0 }, "eval_accuracy": 0.943820895522388, "eval_loss": 0.40087994933128357, "eval_macro avg": { "f1-score": 0.9221490160103993, "precision": 0.9222780664704805, "recall": 0.9220885969453588, "support": 33500.0 }, "eval_runtime": 4.8145, "eval_samples_per_second": 16.824, "eval_steps_per_second": 2.285, "eval_weighted avg": { "f1-score": 0.9437127476806678, "precision": 0.9436689713560499, "recall": 0.943820895522388, "support": 33500.0 }, "step": 1296 }, { "epoch": 17.0, "eval_B": { "f1-score": 0.8904347826086957, "precision": 0.89198606271777, "recall": 0.8888888888888888, "support": 1440.0 }, "eval_I": { "f1-score": 0.9615171394252979, "precision": 0.9589034784611841, "recall": 0.9641450873210728, "support": 21587.0 }, "eval_O": { "f1-score": 0.9182546920750733, "precision": 0.9232625482625483, "recall": 0.9133008689009835, "support": 10473.0 }, "eval_accuracy": 0.9450149253731344, "eval_loss": 0.42787015438079834, "eval_macro avg": { "f1-score": 0.9234022047030224, "precision": 0.9247173631471673, "recall": 0.9221116150369818, "support": 33500.0 }, "eval_runtime": 4.7453, "eval_samples_per_second": 17.07, "eval_steps_per_second": 2.318, "eval_weighted avg": { "f1-score": 0.9449366556964975, "precision": 0.9448847160539952, "recall": 0.9450149253731344, "support": 33500.0 }, "step": 1377 }, { "epoch": 18.0, "eval_B": { "f1-score": 0.8895027624309393, "precision": 0.8846153846153846, "recall": 0.8944444444444445, "support": 1440.0 }, "eval_I": { "f1-score": 0.9607206210979795, "precision": 0.9555912007332722, "recall": 0.9659054060314078, "support": 21587.0 }, "eval_O": { "f1-score": 0.9159781610861478, "precision": 0.9271322378716745, "recall": 0.905089277188962, "support": 10473.0 }, "eval_accuracy": 0.943820895522388, "eval_loss": 0.422620564699173, "eval_macro avg": { "f1-score": 0.9220671815383555, "precision": 0.9224462744067772, "recall": 0.9218130425549381, "support": 33500.0 }, "eval_runtime": 4.7703, "eval_samples_per_second": 16.98, "eval_steps_per_second": 2.306, "eval_weighted avg": { "f1-score": 0.9436716210924735, "precision": 0.9436432636210552, "recall": 0.943820895522388, "support": 33500.0 }, "step": 1458 }, { "epoch": 18.52, "grad_norm": 0.1027584820985794, "learning_rate": 1.2592592592592593e-05, "loss": 0.0054, "step": 1500 }, { "epoch": 19.0, "eval_B": { "f1-score": 0.8924247665167762, "precision": 0.8890420399724328, "recall": 0.8958333333333334, "support": 1440.0 }, "eval_I": { "f1-score": 0.9599648245857633, "precision": 0.959121387283237, "recall": 0.9608097466067541, "support": 21587.0 }, "eval_O": { "f1-score": 0.9158252380724505, "precision": 0.9179777436684574, "recall": 0.913682803399217, "support": 10473.0 }, "eval_accuracy": 0.9432835820895522, "eval_loss": 0.4375981390476227, "eval_macro avg": { "f1-score": 0.9227382763916633, "precision": 0.9220470569747091, "recall": 0.9234419611131015, "support": 33500.0 }, "eval_runtime": 4.7487, "eval_samples_per_second": 17.057, "eval_steps_per_second": 2.316, "eval_weighted avg": { "f1-score": 0.9432623895656658, "precision": 0.9432464129636805, "recall": 0.9432835820895522, "support": 33500.0 }, "step": 1539 }, { "epoch": 20.0, "eval_B": { "f1-score": 0.8907273354015857, "precision": 0.8843258042436687, "recall": 0.8972222222222223, "support": 1440.0 }, "eval_I": { "f1-score": 0.9620682542229149, "precision": 0.9520958083832335, "recall": 0.972251818223931, "support": 21587.0 }, "eval_O": { "f1-score": 0.9177252296267345, "precision": 0.9396698349174587, "recall": 0.8967822018523823, "support": 10473.0 }, "eval_accuracy": 0.9454328358208955, "eval_loss": 0.44444534182548523, "eval_macro avg": { "f1-score": 0.9235069397504118, "precision": 0.9253638158481202, "recall": 0.922085414099512, "support": 33500.0 }, "eval_runtime": 4.7631, "eval_samples_per_second": 17.006, "eval_steps_per_second": 2.309, "eval_weighted avg": { "f1-score": 0.9451388387095264, "precision": 0.9452980165603072, "recall": 0.9454328358208955, "support": 33500.0 }, "step": 1620 }, { "epoch": 21.0, "eval_B": { "f1-score": 0.890790825059911, "precision": 0.8784604996623903, "recall": 0.9034722222222222, "support": 1440.0 }, "eval_I": { "f1-score": 0.9615375783600082, "precision": 0.9533284764593388, "recall": 0.9698892852179553, "support": 21587.0 }, "eval_O": { "f1-score": 0.9162201656113005, "precision": 0.9351695336581486, "recall": 0.8980234889716414, "support": 10473.0 }, "eval_accuracy": 0.9445671641791045, "eval_loss": 0.4363422691822052, "eval_macro avg": { "f1-score": 0.9228495230104065, "precision": 0.9223195032599593, "recall": 0.9237949988039397, "support": 33500.0 }, "eval_runtime": 4.7902, "eval_samples_per_second": 16.91, "eval_steps_per_second": 2.296, "eval_weighted avg": { "f1-score": 0.9443290831818184, "precision": 0.9444332975177129, "recall": 0.9445671641791045, "support": 33500.0 }, "step": 1701 }, { "epoch": 22.0, "eval_B": { "f1-score": 0.8907216494845361, "precision": 0.8816326530612245, "recall": 0.9, "support": 1440.0 }, "eval_I": { "f1-score": 0.961758728985775, "precision": 0.9588801399825022, "recall": 0.9646546532635383, "support": 21587.0 }, "eval_O": { "f1-score": 0.9186952756663139, "precision": 0.9258217783380199, "recall": 0.9116776472834909, "support": 10473.0 }, "eval_accuracy": 0.9453134328358209, "eval_loss": 0.3998025953769684, "eval_macro avg": { "f1-score": 0.9237252180455416, "precision": 0.9221115237939155, "recall": 0.925444100182343, "support": 33500.0 }, "eval_runtime": 4.7313, "eval_samples_per_second": 17.12, "eval_steps_per_second": 2.325, "eval_weighted avg": { "f1-score": 0.9452424023858794, "precision": 0.9452247190072991, "recall": 0.9453134328358209, "support": 33500.0 }, "step": 1782 }, { "epoch": 23.0, "eval_B": { "f1-score": 0.8938053097345132, "precision": 0.8765020026702269, "recall": 0.9118055555555555, "support": 1440.0 }, "eval_I": { "f1-score": 0.9620159689850926, "precision": 0.9584770313146641, "recall": 0.9655811367952934, "support": 21587.0 }, "eval_O": { "f1-score": 0.918660748745658, "precision": 0.9284251584592882, "recall": 0.9090995894204144, "support": 10473.0 }, "eval_accuracy": 0.9456119402985075, "eval_loss": 0.4420088231563568, "eval_macro avg": { "f1-score": 0.9248273424884212, "precision": 0.9211347308147264, "recall": 0.9288287605904211, "support": 33500.0 }, "eval_runtime": 4.7375, "eval_samples_per_second": 17.098, "eval_steps_per_second": 2.322, "eval_weighted avg": { "f1-score": 0.9455299220929005, "precision": 0.9455583057725045, "recall": 0.9456119402985075, "support": 33500.0 }, "step": 1863 }, { "epoch": 24.0, "eval_B": { "f1-score": 0.89482699554642, "precision": 0.8830290736984449, "recall": 0.9069444444444444, "support": 1440.0 }, "eval_I": { "f1-score": 0.962135833103575, "precision": 0.9544605005242285, "recall": 0.969935609394543, "support": 21587.0 }, "eval_O": { "f1-score": 0.9185192391885975, "precision": 0.9362356207854026, "recall": 0.9014608994557434, "support": 10473.0 }, "eval_accuracy": 0.945820895522388, "eval_loss": 0.4822453558444977, "eval_macro avg": { "f1-score": 0.9251606892795309, "precision": 0.924575065002692, "recall": 0.9261136510982436, "support": 33500.0 }, "eval_runtime": 4.7345, "eval_samples_per_second": 17.108, "eval_steps_per_second": 2.323, "eval_weighted avg": { "f1-score": 0.9456068386512209, "precision": 0.945692428281427, "recall": 0.945820895522388, "support": 33500.0 }, "step": 1944 }, { "epoch": 24.69, "grad_norm": 0.10262901335954666, "learning_rate": 1.0123456790123458e-05, "loss": 0.0037, "step": 2000 }, { "epoch": 25.0, "eval_B": { "f1-score": 0.8864342423196411, "precision": 0.8812628689087165, "recall": 0.8916666666666667, "support": 1440.0 }, "eval_I": { "f1-score": 0.9595285234126801, "precision": 0.9537320946409775, "recall": 0.9653958400889424, "support": 21587.0 }, "eval_O": { "f1-score": 0.9129445923058311, "precision": 0.9255298273155416, "recall": 0.9006970304592762, "support": 10473.0 }, "eval_accuracy": 0.942, "eval_loss": 0.46240004897117615, "eval_macro avg": { "f1-score": 0.9196357860127174, "precision": 0.9201749302884119, "recall": 0.9192531790716284, "support": 33500.0 }, "eval_runtime": 4.7408, "eval_samples_per_second": 17.086, "eval_steps_per_second": 2.32, "eval_weighted avg": { "f1-score": 0.941823171912501, "precision": 0.9418002131259403, "recall": 0.942, "support": 33500.0 }, "step": 2025 }, { "epoch": 26.0, "eval_B": { "f1-score": 0.892392049348869, "precision": 0.8809201623815968, "recall": 0.9041666666666667, "support": 1440.0 }, "eval_I": { "f1-score": 0.9613188432492171, "precision": 0.9556420233463035, "recall": 0.9670635104461018, "support": 21587.0 }, "eval_O": { "f1-score": 0.9172881355932202, "precision": 0.9306278864105335, "recall": 0.904325408192495, "support": 10473.0 }, "eval_accuracy": 0.9447462686567164, "eval_loss": 0.4685353934764862, "eval_macro avg": { "f1-score": 0.9236663427304355, "precision": 0.9223966907128114, "recall": 0.9251851951017546, "support": 33500.0 }, "eval_runtime": 4.7522, "eval_samples_per_second": 17.045, "eval_steps_per_second": 2.315, "eval_weighted avg": { "f1-score": 0.9445908377418215, "precision": 0.9446100073487664, "recall": 0.9447462686567164, "support": 33500.0 }, "step": 2106 }, { "epoch": 27.0, "eval_B": { "f1-score": 0.8878923766816142, "precision": 0.882111034955449, "recall": 0.89375, "support": 1440.0 }, "eval_I": { "f1-score": 0.9603469923524712, "precision": 0.9467098748762265, "recall": 0.9743827303469681, "support": 21587.0 }, "eval_O": { "f1-score": 0.9131848640126133, "precision": 0.943398147205538, "recall": 0.8848467487825837, "support": 10473.0 }, "eval_accuracy": 0.9429253731343283, "eval_loss": 0.5011060833930969, "eval_macro avg": { "f1-score": 0.9204747443488995, "precision": 0.9240730190124045, "recall": 0.9176598263765173, "support": 33500.0 }, "eval_runtime": 4.7877, "eval_samples_per_second": 16.918, "eval_steps_per_second": 2.298, "eval_weighted avg": { "f1-score": 0.942488376929505, "precision": 0.9428977538797776, "recall": 0.9429253731343283, "support": 33500.0 }, "step": 2187 }, { "epoch": 28.0, "eval_B": { "f1-score": 0.8914835164835164, "precision": 0.8817934782608695, "recall": 0.9013888888888889, "support": 1440.0 }, "eval_I": { "f1-score": 0.9599201652393882, "precision": 0.9618175053483397, "recall": 0.9580302960114884, "support": 21587.0 }, "eval_O": { "f1-score": 0.9164245916472212, "precision": 0.9141174235227056, "recall": 0.9187434355008116, "support": 10473.0 }, "eval_accuracy": 0.9433134328358209, "eval_loss": 0.4657354950904846, "eval_macro avg": { "f1-score": 0.9226094244567086, "precision": 0.919242802377305, "recall": 0.9260542068003962, "support": 33500.0 }, "eval_runtime": 4.7715, "eval_samples_per_second": 16.976, "eval_steps_per_second": 2.305, "eval_weighted avg": { "f1-score": 0.9433805259426951, "precision": 0.9434653394986137, "recall": 0.9433134328358209, "support": 33500.0 }, "step": 2268 }, { "epoch": 29.0, "eval_B": { "f1-score": 0.8932038834951457, "precision": 0.8623141564318035, "recall": 0.9263888888888889, "support": 1440.0 }, "eval_I": { "f1-score": 0.9610192139335576, "precision": 0.9559099867088318, "recall": 0.9661833510909343, "support": 21587.0 }, "eval_O": { "f1-score": 0.9164846896685592, "precision": 0.9318136964673377, "recall": 0.9016518667048601, "support": 10473.0 }, "eval_accuracy": 0.9442985074626866, "eval_loss": 0.45431333780288696, "eval_macro avg": { "f1-score": 0.9235692623657542, "precision": 0.9166792798693243, "recall": 0.931408035561561, "support": 33500.0 }, "eval_runtime": 4.7759, "eval_samples_per_second": 16.96, "eval_steps_per_second": 2.303, "eval_weighted avg": { "f1-score": 0.9441814781586727, "precision": 0.9443536152670977, "recall": 0.9442985074626866, "support": 33500.0 }, "step": 2349 }, { "epoch": 30.0, "eval_B": { "f1-score": 0.8953290146607569, "precision": 0.8794373744139317, "recall": 0.9118055555555555, "support": 1440.0 }, "eval_I": { "f1-score": 0.9620589989448088, "precision": 0.9528786295269687, "recall": 0.9714179830453513, "support": 21587.0 }, "eval_O": { "f1-score": 0.9176964782884776, "precision": 0.9394, "recall": 0.896973169101499, "support": 10473.0 }, "eval_accuracy": 0.9455820895522388, "eval_loss": 0.4872027635574341, "eval_macro avg": { "f1-score": 0.9250281639646811, "precision": 0.9239053346469669, "recall": 0.926732235900802, "support": 33500.0 }, "eval_runtime": 4.7685, "eval_samples_per_second": 16.987, "eval_steps_per_second": 2.307, "eval_weighted avg": { "f1-score": 0.945321689804427, "precision": 0.9455079699926786, "recall": 0.9455820895522388, "support": 33500.0 }, "step": 2430 }, { "epoch": 30.86, "grad_norm": 8.135757446289062, "learning_rate": 7.654320987654322e-06, "loss": 0.0025, "step": 2500 }, { "epoch": 31.0, "eval_B": { "f1-score": 0.8891177480260899, "precision": 0.8791581805838425, "recall": 0.8993055555555556, "support": 1440.0 }, "eval_I": { "f1-score": 0.9598950783461034, "precision": 0.9535978787601719, "recall": 0.9662759994441099, "support": 21587.0 }, "eval_O": { "f1-score": 0.9137011538834482, "precision": 0.9281000689451394, "recall": 0.8997421942136924, "support": 10473.0 }, "eval_accuracy": 0.9425970149253732, "eval_loss": 0.4678420424461365, "eval_macro avg": { "f1-score": 0.9209046600852139, "precision": 0.9202853760963846, "recall": 0.9217745830711194, "support": 33500.0 }, "eval_runtime": 4.7618, "eval_samples_per_second": 17.01, "eval_steps_per_second": 2.31, "eval_weighted avg": { "f1-score": 0.9424112477025748, "precision": 0.9424267824148957, "recall": 0.9425970149253732, "support": 33500.0 }, "step": 2511 }, { "epoch": 32.0, "eval_B": { "f1-score": 0.8871355060034306, "precision": 0.8766101694915254, "recall": 0.8979166666666667, "support": 1440.0 }, "eval_I": { "f1-score": 0.9603147215128719, "precision": 0.9538848263254114, "recall": 0.966831889563163, "support": 21587.0 }, "eval_O": { "f1-score": 0.9144436899796294, "precision": 0.9292262198127156, "recall": 0.9001241287119259, "support": 10473.0 }, "eval_accuracy": 0.9430149253731344, "eval_loss": 0.4877029359340668, "eval_macro avg": { "f1-score": 0.9206313058319773, "precision": 0.9199070718765507, "recall": 0.9216242283139185, "support": 33500.0 }, "eval_runtime": 4.759, "eval_samples_per_second": 17.02, "eval_steps_per_second": 2.311, "eval_weighted avg": { "f1-score": 0.9428285906597003, "precision": 0.942854226568747, "recall": 0.9430149253731344, "support": 33500.0 }, "step": 2592 }, { "epoch": 33.0, "eval_B": { "f1-score": 0.8910756701730572, "precision": 0.8712674187126742, "recall": 0.9118055555555555, "support": 1440.0 }, "eval_I": { "f1-score": 0.9623726096415894, "precision": 0.9516713470423046, "recall": 0.9733172742854496, "support": 21587.0 }, "eval_O": { "f1-score": 0.9176966843241121, "precision": 0.9435199193141705, "recall": 0.8932493077437219, "support": 10473.0 }, "eval_accuracy": 0.9456417910447761, "eval_loss": 0.49512049555778503, "eval_macro avg": { "f1-score": 0.9237149880462528, "precision": 0.9221528950230496, "recall": 0.9261240458615757, "support": 33500.0 }, "eval_runtime": 4.765, "eval_samples_per_second": 16.999, "eval_steps_per_second": 2.308, "eval_weighted avg": { "f1-score": 0.9453410108748842, "precision": 0.9456668228813668, "recall": 0.9456417910447761, "support": 33500.0 }, "step": 2673 }, { "epoch": 34.0, "eval_B": { "f1-score": 0.8933423913043478, "precision": 0.8743351063829787, "recall": 0.9131944444444444, "support": 1440.0 }, "eval_I": { "f1-score": 0.9620096067661051, "precision": 0.9546159459952563, "recall": 0.9695186918052532, "support": 21587.0 }, "eval_O": { "f1-score": 0.9179849111706012, "precision": 0.9362589356632248, "recall": 0.9004105795856011, "support": 10473.0 }, "eval_accuracy": 0.9454925373134329, "eval_loss": 0.47959965467453003, "eval_macro avg": { "f1-score": 0.9244456364136847, "precision": 0.9217366626804866, "recall": 0.9277079052784329, "support": 33500.0 }, "eval_runtime": 4.7437, "eval_samples_per_second": 17.075, "eval_steps_per_second": 2.319, "eval_weighted avg": { "f1-score": 0.9452946387888919, "precision": 0.9454261735102102, "recall": 0.9454925373134329, "support": 33500.0 }, "step": 2754 }, { "epoch": 35.0, "eval_B": { "f1-score": 0.8896504455106238, "precision": 0.878213802435724, "recall": 0.9013888888888889, "support": 1440.0 }, "eval_I": { "f1-score": 0.9619017054832203, "precision": 0.9520805917320869, "recall": 0.9719275489878167, "support": 21587.0 }, "eval_O": { "f1-score": 0.917587251930785, "precision": 0.9400100150225338, "recall": 0.896209300105032, "support": 10473.0 }, "eval_accuracy": 0.9452238805970149, "eval_loss": 0.4806292951107025, "eval_macro avg": { "f1-score": 0.923046467641543, "precision": 0.9234348030634482, "recall": 0.9231752459939125, "support": 33500.0 }, "eval_runtime": 4.73, "eval_samples_per_second": 17.125, "eval_steps_per_second": 2.326, "eval_weighted avg": { "f1-score": 0.9449420909633638, "precision": 0.9451318357181789, "recall": 0.9452238805970149, "support": 33500.0 }, "step": 2835 }, { "epoch": 36.0, "eval_B": { "f1-score": 0.8926979773740144, "precision": 0.8815165876777251, "recall": 0.9041666666666667, "support": 1440.0 }, "eval_I": { "f1-score": 0.9623375129414473, "precision": 0.9559374714324893, "recall": 0.9688238291564367, "support": 21587.0 }, "eval_O": { "f1-score": 0.9190998156950239, "precision": 0.9339576145884673, "recall": 0.9047073426907285, "support": 10473.0 }, "eval_accuracy": 0.946, "eval_loss": 0.4832863509654999, "eval_macro avg": { "f1-score": 0.9247117686701619, "precision": 0.923803891232894, "recall": 0.9258992795046107, "support": 33500.0 }, "eval_runtime": 4.7435, "eval_samples_per_second": 17.076, "eval_steps_per_second": 2.319, "eval_weighted avg": { "f1-score": 0.9458267865390921, "precision": 0.9458669904374355, "recall": 0.946, "support": 33500.0 }, "step": 2916 }, { "epoch": 37.0, "eval_B": { "f1-score": 0.8904958677685951, "precision": 0.8831967213114754, "recall": 0.8979166666666667, "support": 1440.0 }, "eval_I": { "f1-score": 0.9614206416841329, "precision": 0.9560696289509849, "recall": 0.966831889563163, "support": 21587.0 }, "eval_O": { "f1-score": 0.9176459209826394, "precision": 0.9296492259455222, "recall": 0.9059486298099876, "support": 10473.0 }, "eval_accuracy": 0.9448358208955224, "eval_loss": 0.4824429750442505, "eval_macro avg": { "f1-score": 0.9231874768117891, "precision": 0.9229718587359942, "recall": 0.9235657286799391, "support": 33500.0 }, "eval_runtime": 4.7583, "eval_samples_per_second": 17.023, "eval_steps_per_second": 2.312, "eval_weighted avg": { "f1-score": 0.9446867812559204, "precision": 0.9446774537964444, "recall": 0.9448358208955224, "support": 33500.0 }, "step": 2997 }, { "epoch": 37.04, "grad_norm": 0.2996862828731537, "learning_rate": 5.185185185185185e-06, "loss": 0.0016, "step": 3000 }, { "epoch": 38.0, "eval_B": { "f1-score": 0.8921302578018996, "precision": 0.8720159151193634, "recall": 0.9131944444444444, "support": 1440.0 }, "eval_I": { "f1-score": 0.9625111778598124, "precision": 0.9529192772178334, "recall": 0.9722981424005188, "support": 21587.0 }, "eval_O": { "f1-score": 0.9186359410930084, "precision": 0.9420028095524784, "recall": 0.8964002673541488, "support": 10473.0 }, "eval_accuracy": 0.9460298507462687, "eval_loss": 0.49525222182273865, "eval_macro avg": { "f1-score": 0.9244257922515735, "precision": 0.9223126672965584, "recall": 0.9272976180663707, "support": 33500.0 }, "eval_runtime": 4.7487, "eval_samples_per_second": 17.057, "eval_steps_per_second": 2.316, "eval_weighted avg": { "f1-score": 0.9457692710078085, "precision": 0.9460288590900406, "recall": 0.9460298507462687, "support": 33500.0 }, "step": 3078 }, { "epoch": 39.0, "eval_B": { "f1-score": 0.8914011647824598, "precision": 0.8796484110885734, "recall": 0.9034722222222222, "support": 1440.0 }, "eval_I": { "f1-score": 0.9630648600316389, "precision": 0.9533817521561507, "recall": 0.9729466808727475, "support": 21587.0 }, "eval_O": { "f1-score": 0.9199569976544176, "precision": 0.9421479331398258, "recall": 0.8987873579681085, "support": 10473.0 }, "eval_accuracy": 0.9467761194029851, "eval_loss": 0.4956269860267639, "eval_macro avg": { "f1-score": 0.9248076741561722, "precision": 0.9250593654615167, "recall": 0.9250687536876927, "support": 33500.0 }, "eval_runtime": 4.7588, "eval_samples_per_second": 17.021, "eval_steps_per_second": 2.312, "eval_weighted avg": { "f1-score": 0.9465077148425208, "precision": 0.9467003253592766, "recall": 0.9467761194029851, "support": 33500.0 }, "step": 3159 }, { "epoch": 40.0, "eval_B": { "f1-score": 0.8935439560439561, "precision": 0.8838315217391305, "recall": 0.9034722222222222, "support": 1440.0 }, "eval_I": { "f1-score": 0.962365961745999, "precision": 0.9541067200874158, "recall": 0.9707694445731228, "support": 21587.0 }, "eval_O": { "f1-score": 0.9186346593952379, "precision": 0.9373012718600954, "recall": 0.9006970304592762, "support": 10473.0 }, "eval_accuracy": 0.9459701492537314, "eval_loss": 0.48434382677078247, "eval_macro avg": { "f1-score": 0.9248481923950643, "precision": 0.9250798378955473, "recall": 0.9249795657515404, "support": 33500.0 }, "eval_runtime": 4.7973, "eval_samples_per_second": 16.884, "eval_steps_per_second": 2.293, "eval_weighted avg": { "f1-score": 0.9457360627092687, "precision": 0.9458321008364828, "recall": 0.9459701492537314, "support": 33500.0 }, "step": 3240 }, { "epoch": 41.0, "eval_B": { "f1-score": 0.8901893287435455, "precision": 0.8825938566552901, "recall": 0.8979166666666667, "support": 1440.0 }, "eval_I": { "f1-score": 0.9599889178768498, "precision": 0.9569179784589892, "recall": 0.9630796312595543, "support": 21587.0 }, "eval_O": { "f1-score": 0.915215089981715, "precision": 0.9224949073624988, "recall": 0.9080492695502721, "support": 10473.0 }, "eval_accuracy": 0.9430746268656717, "eval_loss": 0.4890231490135193, "eval_macro avg": { "f1-score": 0.92179777886737, "precision": 0.920668914158926, "recall": 0.9230151891588311, "support": 33500.0 }, "eval_runtime": 4.7614, "eval_samples_per_second": 17.012, "eval_steps_per_second": 2.31, "eval_weighted avg": { "f1-score": 0.9429910758500526, "precision": 0.942961573712993, "recall": 0.9430746268656717, "support": 33500.0 }, "step": 3321 }, { "epoch": 42.0, "eval_B": { "f1-score": 0.8924028875902373, "precision": 0.8835942818243703, "recall": 0.9013888888888889, "support": 1440.0 }, "eval_I": { "f1-score": 0.9607680236314964, "precision": 0.9572775350655323, "recall": 0.9642840598508362, "support": 21587.0 }, "eval_O": { "f1-score": 0.9167108242208198, "precision": 0.9250437487847559, "recall": 0.9085266876730641, "support": 10473.0 }, "eval_accuracy": 0.9441492537313433, "eval_loss": 0.493211567401886, "eval_macro avg": { "f1-score": 0.9232939118141844, "precision": 0.9219718552248861, "recall": 0.9247332121375963, "support": 33500.0 }, "eval_runtime": 4.7587, "eval_samples_per_second": 17.022, "eval_steps_per_second": 2.312, "eval_weighted avg": { "f1-score": 0.9440558789948568, "precision": 0.9440331073525219, "recall": 0.9441492537313433, "support": 33500.0 }, "step": 3402 }, { "epoch": 43.0, "eval_B": { "f1-score": 0.8943758573388203, "precision": 0.8834688346883469, "recall": 0.9055555555555556, "support": 1440.0 }, "eval_I": { "f1-score": 0.9620124858900229, "precision": 0.956832554302997, "recall": 0.9672488071524529, "support": 21587.0 }, "eval_O": { "f1-score": 0.9189842805320435, "precision": 0.9311899627524015, "recall": 0.9070944333046882, "support": 10473.0 }, "eval_accuracy": 0.9457910447761194, "eval_loss": 0.48890525102615356, "eval_macro avg": { "f1-score": 0.9251242079202955, "precision": 0.9238304505812485, "recall": 0.9266329320042322, "support": 33500.0 }, "eval_runtime": 4.7372, "eval_samples_per_second": 17.099, "eval_steps_per_second": 2.322, "eval_weighted avg": { "f1-score": 0.945653347387699, "precision": 0.945662446316296, "recall": 0.9457910447761194, "support": 33500.0 }, "step": 3483 }, { "epoch": 43.21, "grad_norm": 0.36121609807014465, "learning_rate": 2.7160493827160496e-06, "loss": 0.0015, "step": 3500 }, { "epoch": 44.0, "eval_B": { "f1-score": 0.8921703296703297, "precision": 0.8824728260869565, "recall": 0.9020833333333333, "support": 1440.0 }, "eval_I": { "f1-score": 0.9620328986776021, "precision": 0.956918282231083, "recall": 0.9672024829758651, "support": 21587.0 }, "eval_O": { "f1-score": 0.9190600522193212, "precision": 0.9309432853364679, "recall": 0.9074763678029218, "support": 10473.0 }, "eval_accuracy": 0.9457313432835821, "eval_loss": 0.49400997161865234, "eval_macro avg": { "f1-score": 0.9244210935224176, "precision": 0.9234447978848358, "recall": 0.9255873947040402, "support": 33500.0 }, "eval_runtime": 4.7323, "eval_samples_per_second": 17.116, "eval_steps_per_second": 2.324, "eval_weighted avg": { "f1-score": 0.9455953846379588, "precision": 0.9455977568781024, "recall": 0.9457313432835821, "support": 33500.0 }, "step": 3564 }, { "epoch": 45.0, "eval_B": { "f1-score": 0.8950808393532852, "precision": 0.8868438991138378, "recall": 0.9034722222222222, "support": 1440.0 }, "eval_I": { "f1-score": 0.9623735050597976, "precision": 0.9556479239939707, "recall": 0.9691944225691388, "support": 21587.0 }, "eval_O": { "f1-score": 0.9191287051860476, "precision": 0.9342209072978304, "recall": 0.9045163754416118, "support": 10473.0 }, "eval_accuracy": 0.9461492537313433, "eval_loss": 0.49951982498168945, "eval_macro avg": { "f1-score": 0.92552768319971, "precision": 0.925570910135213, "recall": 0.9257276734109908, "support": 33500.0 }, "eval_runtime": 4.7702, "eval_samples_per_second": 16.98, "eval_steps_per_second": 2.306, "eval_weighted avg": { "f1-score": 0.9459614385614344, "precision": 0.9459917167794611, "recall": 0.9461492537313433, "support": 33500.0 }, "step": 3645 }, { "epoch": 46.0, "eval_B": { "f1-score": 0.8926862611073137, "precision": 0.8788694481830417, "recall": 0.9069444444444444, "support": 1440.0 }, "eval_I": { "f1-score": 0.9626744265976216, "precision": 0.9524392455567646, "recall": 0.9731319775790985, "support": 21587.0 }, "eval_O": { "f1-score": 0.918604082032206, "precision": 0.9423579031934123, "recall": 0.8960183328559153, "support": 10473.0 }, "eval_accuracy": 0.9461791044776119, "eval_loss": 0.5005789995193481, "eval_macro avg": { "f1-score": 0.9246549232457139, "precision": 0.9245555323110729, "recall": 0.9253649182931527, "support": 33500.0 }, "eval_runtime": 4.7467, "eval_samples_per_second": 17.064, "eval_steps_per_second": 2.317, "eval_weighted avg": { "f1-score": 0.9458884063904681, "precision": 0.9461251438615541, "recall": 0.9461791044776119, "support": 33500.0 }, "step": 3726 }, { "epoch": 47.0, "eval_B": { "f1-score": 0.8930774503084304, "precision": 0.881596752368065, "recall": 0.9048611111111111, "support": 1440.0 }, "eval_I": { "f1-score": 0.9616694275722437, "precision": 0.9552955158385519, "recall": 0.9681289665076204, "support": 21587.0 }, "eval_O": { "f1-score": 0.9176447764089631, "precision": 0.9324790537210449, "recall": 0.9032750883223527, "support": 10473.0 }, "eval_accuracy": 0.945134328358209, "eval_loss": 0.4959483742713928, "eval_macro avg": { "f1-score": 0.9241305514298791, "precision": 0.9231237739758873, "recall": 0.9254217219803614, "support": 33500.0 }, "eval_runtime": 4.7667, "eval_samples_per_second": 16.993, "eval_steps_per_second": 2.308, "eval_weighted avg": { "f1-score": 0.9449577076052906, "precision": 0.9449945299533534, "recall": 0.945134328358209, "support": 33500.0 }, "step": 3807 } ], "logging_steps": 500, "max_steps": 4050, "num_input_tokens_seen": 0, "num_train_epochs": 50, "save_steps": 500, "total_flos": 6736985185782600.0, "train_batch_size": 4, "trial_name": null, "trial_params": null }