{
  "best_global_step": 38000,
  "best_metric": 0.6343372808016398,
  "best_model_checkpoint": "/scratch/users/k24112973/data/dynamic_training/oversample/checkpoint-38000",
  "epoch": 974.3636363636364,
  "eval_steps": 500,
  "global_step": 38000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 2.571428571428571,
      "grad_norm": 0.9723352193832397,
      "learning_rate": 1.2692307692307692e-06,
      "loss": 0.4224576187133789,
      "step": 100
    },
    {
      "epoch": 5.12987012987013,
      "grad_norm": 0.7024934887886047,
      "learning_rate": 2.5512820512820516e-06,
      "loss": 0.22791364669799805,
      "step": 200
    },
    {
      "epoch": 7.701298701298701,
      "grad_norm": 0.2942417860031128,
      "learning_rate": 3.833333333333334e-06,
      "loss": 0.0733012866973877,
      "step": 300
    },
    {
      "epoch": 10.25974025974026,
      "grad_norm": 0.1313406378030777,
      "learning_rate": 5.115384615384616e-06,
      "loss": 0.039429664611816406,
      "step": 400
    },
    {
      "epoch": 12.831168831168831,
      "grad_norm": 0.10131426155567169,
      "learning_rate": 6.3974358974358984e-06,
      "loss": 0.03325621843338013,
      "step": 500
    },
    {
      "epoch": 12.831168831168831,
      "eval_exact_match_ratio": 0.010239344681940356,
      "eval_loss": 0.026465708389878273,
      "eval_macro_f1": 2.2791503327559487e-05,
      "eval_macro_precision": 0.017857142857142856,
      "eval_macro_recall": 1.1403028644407956e-05,
      "eval_macro_roc_auc": 0.6593257939423923,
      "eval_micro_f1": 4.855783237836263e-05,
      "eval_micro_precision": 0.02040816326530612,
      "eval_micro_recall": 2.4307834415031966e-05,
      "eval_runtime": 101.181,
      "eval_samples_per_second": 77.218,
      "eval_steps_per_second": 4.833,
      "step": 500
    },
    {
      "epoch": 15.38961038961039,
      "grad_norm": 0.10254181176424026,
      "learning_rate": 7.67948717948718e-06,
      "loss": 0.03204334259033203,
      "step": 600
    },
    {
      "epoch": 17.961038961038962,
      "grad_norm": 0.08924805372953415,
      "learning_rate": 8.96153846153846e-06,
      "loss": 0.02998211145401001,
      "step": 700
    },
    {
      "epoch": 20.51948051948052,
      "grad_norm": 0.09124752134084702,
      "learning_rate": 1.0243589743589744e-05,
      "loss": 0.028628075122833253,
      "step": 800
    },
    {
      "epoch": 23.07792207792208,
      "grad_norm": 0.09543249756097794,
      "learning_rate": 1.1525641025641026e-05,
      "loss": 0.028393406867980957,
      "step": 900
    },
    {
      "epoch": 25.649350649350648,
      "grad_norm": 0.08555900305509567,
      "learning_rate": 1.280769230769231e-05,
      "loss": 0.02726546287536621,
      "step": 1000
    },
    {
      "epoch": 25.649350649350648,
      "eval_exact_match_ratio": 0.015487008831434787,
      "eval_loss": 0.022449757903814316,
      "eval_macro_f1": 0.04015436423609551,
      "eval_macro_precision": 0.1501749391031702,
      "eval_macro_recall": 0.024922522039782197,
      "eval_macro_roc_auc": 0.8140910680524438,
      "eval_micro_f1": 0.10766036202771634,
      "eval_micro_precision": 0.7530786233028103,
      "eval_micro_recall": 0.057974185079851234,
      "eval_runtime": 101.805,
      "eval_samples_per_second": 76.745,
      "eval_steps_per_second": 4.803,
      "step": 1000
    },
    {
      "epoch": 28.207792207792206,
      "grad_norm": 0.09994173794984818,
      "learning_rate": 1.408974358974359e-05,
      "loss": 0.02647646188735962,
      "step": 1100
    },
    {
      "epoch": 30.77922077922078,
      "grad_norm": 0.09997841715812683,
      "learning_rate": 1.5371794871794872e-05,
      "loss": 0.02625159740447998,
      "step": 1200
    },
    {
      "epoch": 33.33766233766234,
      "grad_norm": 0.07699156552553177,
      "learning_rate": 1.6653846153846156e-05,
      "loss": 0.025394675731658937,
      "step": 1300
    },
    {
      "epoch": 35.90909090909091,
      "grad_norm": 0.08908795565366745,
      "learning_rate": 1.7935897435897436e-05,
      "loss": 0.02497950315475464,
      "step": 1400
    },
    {
      "epoch": 38.467532467532465,
      "grad_norm": 0.09753439575433731,
      "learning_rate": 1.921794871794872e-05,
      "loss": 0.02420916795730591,
      "step": 1500
    },
    {
      "epoch": 38.467532467532465,
      "eval_exact_match_ratio": 0.04134135415333419,
      "eval_loss": 0.019969576969742775,
      "eval_macro_f1": 0.1500658433419304,
      "eval_macro_precision": 0.45900795060760735,
      "eval_macro_recall": 0.10204770730152682,
      "eval_macro_roc_auc": 0.8592341139831198,
      "eval_micro_f1": 0.2797057867360208,
      "eval_micro_precision": 0.8521728364491766,
      "eval_micro_recall": 0.167310824278665,
      "eval_runtime": 102.1239,
      "eval_samples_per_second": 76.505,
      "eval_steps_per_second": 4.788,
      "step": 1500
    },
    {
      "epoch": 41.02597402597402,
      "grad_norm": 0.09155359119176865,
      "learning_rate": 2.05e-05,
      "loss": 0.024083244800567626,
      "step": 1600
    },
    {
      "epoch": 43.5974025974026,
      "grad_norm": 0.09570193290710449,
      "learning_rate": 2.1782051282051284e-05,
      "loss": 0.023498144149780274,
      "step": 1700
    },
    {
      "epoch": 46.15584415584416,
      "grad_norm": 0.07617457956075668,
      "learning_rate": 2.3064102564102568e-05,
      "loss": 0.02319591522216797,
      "step": 1800
    },
    {
      "epoch": 48.72727272727273,
      "grad_norm": 0.10715755820274353,
      "learning_rate": 2.4346153846153848e-05,
      "loss": 0.022931249141693116,
      "step": 1900
    },
    {
      "epoch": 51.285714285714285,
      "grad_norm": 0.07499901950359344,
      "learning_rate": 2.5628205128205128e-05,
      "loss": 0.022387499809265136,
      "step": 2000
    },
    {
      "epoch": 51.285714285714285,
      "eval_exact_match_ratio": 0.043389223089722255,
      "eval_loss": 0.019956817850470543,
      "eval_macro_f1": 0.13833827834016052,
      "eval_macro_precision": 0.5337436020047565,
      "eval_macro_recall": 0.08828918347723172,
      "eval_macro_roc_auc": 0.8762218972851165,
      "eval_micro_f1": 0.23172156038308805,
      "eval_micro_precision": 0.9166666666666666,
      "eval_micro_recall": 0.1326235445684144,
      "eval_runtime": 102.2018,
      "eval_samples_per_second": 76.447,
      "eval_steps_per_second": 4.785,
      "step": 2000
    },
    {
      "epoch": 53.857142857142854,
      "grad_norm": 0.07609694451093674,
      "learning_rate": 2.6910256410256412e-05,
      "loss": 0.022320070266723634,
      "step": 2100
    },
    {
      "epoch": 56.41558441558441,
      "grad_norm": 0.07854855805635452,
      "learning_rate": 2.8192307692307696e-05,
      "loss": 0.02212502956390381,
      "step": 2200
    },
    {
      "epoch": 58.98701298701299,
      "grad_norm": 0.06473389267921448,
      "learning_rate": 2.9474358974358972e-05,
      "loss": 0.021405723094940186,
      "step": 2300
    },
    {
      "epoch": 61.54545454545455,
      "grad_norm": 0.08168220520019531,
      "learning_rate": 3.075641025641026e-05,
      "loss": 0.021646018028259276,
      "step": 2400
    },
    {
      "epoch": 64.1038961038961,
      "grad_norm": 0.07714404910802841,
      "learning_rate": 3.2038461538461536e-05,
      "loss": 0.021127231121063232,
      "step": 2500
    },
    {
      "epoch": 64.1038961038961,
      "eval_exact_match_ratio": 0.06015615000639959,
      "eval_loss": 0.01814069040119648,
      "eval_macro_f1": 0.24566480305372065,
      "eval_macro_precision": 0.5808284865248478,
      "eval_macro_recall": 0.17255970450010616,
      "eval_macro_roc_auc": 0.8893594530608899,
      "eval_micro_f1": 0.4059803849273652,
      "eval_micro_precision": 0.8402136589088134,
      "eval_micro_recall": 0.26765356474391694,
      "eval_runtime": 102.1616,
      "eval_samples_per_second": 76.477,
      "eval_steps_per_second": 4.787,
      "step": 2500
    },
    {
      "epoch": 66.67532467532467,
      "grad_norm": 0.0774487778544426,
      "learning_rate": 3.332051282051282e-05,
      "loss": 0.02119396924972534,
      "step": 2600
    },
    {
      "epoch": 69.23376623376623,
      "grad_norm": 0.08460704237222672,
      "learning_rate": 3.4602564102564104e-05,
      "loss": 0.02077857732772827,
      "step": 2700
    },
    {
      "epoch": 71.8051948051948,
      "grad_norm": 0.06910490989685059,
      "learning_rate": 3.588461538461539e-05,
      "loss": 0.020965633392333986,
      "step": 2800
    },
    {
      "epoch": 74.36363636363636,
      "grad_norm": 0.060772985219955444,
      "learning_rate": 3.7166666666666664e-05,
      "loss": 0.020291900634765624,
      "step": 2900
    },
    {
      "epoch": 76.93506493506493,
      "grad_norm": 0.06566117703914642,
      "learning_rate": 3.844871794871795e-05,
      "loss": 0.020529248714447022,
      "step": 3000
    },
    {
      "epoch": 76.93506493506493,
      "eval_exact_match_ratio": 0.0674516830922821,
      "eval_loss": 0.017710894346237183,
      "eval_macro_f1": 0.25788730715836283,
      "eval_macro_precision": 0.6595957697827093,
      "eval_macro_recall": 0.1863560199250396,
      "eval_macro_roc_auc": 0.8968546657274837,
      "eval_micro_f1": 0.42316413107080164,
      "eval_micro_precision": 0.8540113661524836,
      "eval_micro_recall": 0.28126595201633486,
      "eval_runtime": 102.1271,
      "eval_samples_per_second": 76.503,
      "eval_steps_per_second": 4.788,
      "step": 3000
    },
    {
      "epoch": 79.49350649350649,
      "grad_norm": 0.06747350096702576,
      "learning_rate": 3.973076923076923e-05,
      "loss": 0.019568560123443605,
      "step": 3100
    },
    {
      "epoch": 82.05194805194805,
      "grad_norm": 0.08599278330802917,
      "learning_rate": 4.1012820512820515e-05,
      "loss": 0.019356379508972166,
      "step": 3200
    },
    {
      "epoch": 84.62337662337663,
      "grad_norm": 0.07699308544397354,
      "learning_rate": 4.22948717948718e-05,
      "loss": 0.01968835711479187,
      "step": 3300
    },
    {
      "epoch": 87.18181818181819,
      "grad_norm": 0.07464466989040375,
      "learning_rate": 4.3576923076923076e-05,
      "loss": 0.019133061170578003,
      "step": 3400
    },
    {
      "epoch": 89.75324675324676,
      "grad_norm": 0.057159002870321274,
      "learning_rate": 4.485897435897436e-05,
      "loss": 0.019086666107177734,
      "step": 3500
    },
    {
      "epoch": 89.75324675324676,
      "eval_exact_match_ratio": 0.06693971585818508,
      "eval_loss": 0.017398711293935776,
      "eval_macro_f1": 0.3135355629530234,
      "eval_macro_precision": 0.7128723183711959,
      "eval_macro_recall": 0.23280118458092305,
      "eval_macro_roc_auc": 0.9029103776493033,
      "eval_micro_f1": 0.4553707578282918,
      "eval_micro_precision": 0.8415361426043646,
      "eval_micro_recall": 0.31213690172342545,
      "eval_runtime": 102.1495,
      "eval_samples_per_second": 76.486,
      "eval_steps_per_second": 4.787,
      "step": 3500
    },
    {
      "epoch": 92.31168831168831,
      "grad_norm": 0.0754333958029747,
      "learning_rate": 4.614102564102564e-05,
      "loss": 0.018581944704055785,
      "step": 3600
    },
    {
      "epoch": 94.88311688311688,
      "grad_norm": 0.07137007266283035,
      "learning_rate": 4.742307692307693e-05,
      "loss": 0.01879796862602234,
      "step": 3700
    },
    {
      "epoch": 97.44155844155844,
      "grad_norm": 0.07455797493457794,
      "learning_rate": 4.870512820512821e-05,
      "loss": 0.018829100131988526,
      "step": 3800
    },
    {
      "epoch": 100.0,
      "grad_norm": 0.08882606029510498,
      "learning_rate": 4.9987179487179494e-05,
      "loss": 0.018090782165527345,
      "step": 3900
    },
    {
      "epoch": 102.57142857142857,
      "grad_norm": 0.07004310935735703,
      "learning_rate": 4.999901856153945e-05,
      "loss": 0.018008456230163575,
      "step": 4000
    },
    {
      "epoch": 102.57142857142857,
      "eval_exact_match_ratio": 0.05913221553820555,
      "eval_loss": 0.017673490568995476,
      "eval_macro_f1": 0.28538834333600155,
      "eval_macro_precision": 0.7215161598314275,
      "eval_macro_recall": 0.20060156364179246,
      "eval_macro_roc_auc": 0.9039128340841802,
      "eval_micro_f1": 0.4328613547846977,
      "eval_micro_precision": 0.8613022379952198,
      "eval_micro_recall": 0.28906876686356014,
      "eval_runtime": 102.1514,
      "eval_samples_per_second": 76.485,
      "eval_steps_per_second": 4.787,
      "step": 4000
    },
    {
      "epoch": 105.12987012987013,
      "grad_norm": 0.071014404296875,
      "learning_rate": 4.999603457083103e-05,
      "loss": 0.017830185890197754,
      "step": 4100
    },
    {
      "epoch": 107.7012987012987,
      "grad_norm": 0.07118625193834305,
      "learning_rate": 4.999104816678274e-05,
      "loss": 0.017918087244033813,
      "step": 4200
    },
    {
      "epoch": 110.25974025974025,
      "grad_norm": 0.06566045433282852,
      "learning_rate": 4.998405974885163e-05,
      "loss": 0.01788578987121582,
      "step": 4300
    },
    {
      "epoch": 112.83116883116882,
      "grad_norm": 0.0653366893529892,
      "learning_rate": 4.9975069876874545e-05,
      "loss": 0.017887649536132814,
      "step": 4400
    },
    {
      "epoch": 115.3896103896104,
      "grad_norm": 0.07393347471952438,
      "learning_rate": 4.996407927102331e-05,
      "loss": 0.01751647353172302,
      "step": 4500
    },
    {
      "epoch": 115.3896103896104,
      "eval_exact_match_ratio": 0.07013951107129143,
      "eval_loss": 0.017509067431092262,
      "eval_macro_f1": 0.37014410895371136,
      "eval_macro_precision": 0.697715256791381,
      "eval_macro_recall": 0.2842852948568902,
      "eval_macro_roc_auc": 0.9079131324403187,
      "eval_micro_f1": 0.5182345456648545,
      "eval_micro_precision": 0.8092166374238827,
      "eval_micro_recall": 0.3811711514621162,
      "eval_runtime": 102.1935,
      "eval_samples_per_second": 76.453,
      "eval_steps_per_second": 4.785,
      "step": 4500
    },
    {
      "epoch": 117.96103896103897,
      "grad_norm": 0.05196866765618324,
      "learning_rate": 4.9951088811747e-05,
      "loss": 0.017132682800292967,
      "step": 4600
    },
    {
      "epoch": 120.51948051948052,
      "grad_norm": 0.0642334371805191,
      "learning_rate": 4.9936099539701444e-05,
      "loss": 0.01745768427848816,
      "step": 4700
    },
    {
      "epoch": 123.07792207792208,
      "grad_norm": 0.06535433232784271,
      "learning_rate": 4.991911265566582e-05,
      "loss": 0.017694165706634523,
      "step": 4800
    },
    {
      "epoch": 125.64935064935065,
      "grad_norm": 0.11530792713165283,
      "learning_rate": 4.990012952044654e-05,
      "loss": 0.017267401218414306,
      "step": 4900
    },
    {
      "epoch": 128.2077922077922,
      "grad_norm": 0.06359517574310303,
      "learning_rate": 4.987915165476811e-05,
      "loss": 0.016948271989822388,
      "step": 5000
    },
    {
      "epoch": 128.2077922077922,
      "eval_exact_match_ratio": 0.06783565851785486,
      "eval_loss": 0.016847912222146988,
      "eval_macro_f1": 0.36597600181513823,
      "eval_macro_precision": 0.7139592509552216,
      "eval_macro_recall": 0.2715970879612518,
      "eval_macro_roc_auc": 0.9133004873427402,
      "eval_micro_f1": 0.5028703029481991,
      "eval_micro_precision": 0.8288489449592498,
      "eval_micro_recall": 0.36092272539439463,
      "eval_runtime": 102.1597,
      "eval_samples_per_second": 76.478,
      "eval_steps_per_second": 4.787,
      "step": 5000
    },
    {
      "epoch": 130.7792207792208,
      "grad_norm": 0.0596756786108017,
      "learning_rate": 4.985618073915143e-05,
      "loss": 0.01628790855407715,
      "step": 5100
    },
    {
      "epoch": 133.33766233766235,
      "grad_norm": 0.058088693767786026,
      "learning_rate": 4.9831218613779095e-05,
      "loss": 0.016356936693191527,
      "step": 5200
    },
    {
      "epoch": 135.9090909090909,
      "grad_norm": 0.05349066108465195,
      "learning_rate": 4.9804267278348006e-05,
      "loss": 0.015819222927093507,
      "step": 5300
    },
    {
      "epoch": 138.46753246753246,
      "grad_norm": 0.058746371418237686,
      "learning_rate": 4.977532889190916e-05,
      "loss": 0.016046864986419676,
      "step": 5400
    },
    {
      "epoch": 141.02597402597402,
      "grad_norm": 0.06296154856681824,
      "learning_rate": 4.9744405772694725e-05,
      "loss": 0.016026583909988405,
      "step": 5500
    },
    {
      "epoch": 141.02597402597402,
      "eval_exact_match_ratio": 0.07321131447587355,
      "eval_loss": 0.01681182160973549,
      "eval_macro_f1": 0.3796766306924299,
      "eval_macro_precision": 0.7387063979509644,
      "eval_macro_recall": 0.2871505002687625,
      "eval_macro_roc_auc": 0.916403901192674,
      "eval_micro_f1": 0.5243835525331294,
      "eval_micro_precision": 0.8297726794359083,
      "eval_micro_recall": 0.38331024089063903,
      "eval_runtime": 102.1895,
      "eval_samples_per_second": 76.456,
      "eval_steps_per_second": 4.785,
      "step": 5500
    },
    {
      "epoch": 143.5974025974026,
      "grad_norm": 0.06563801318407059,
      "learning_rate": 4.9711500397932284e-05,
      "loss": 0.015465517044067383,
      "step": 5600
    },
    {
      "epoch": 146.15584415584416,
      "grad_norm": 0.07564526796340942,
      "learning_rate": 4.96766154036464e-05,
      "loss": 0.015505937337875366,
      "step": 5700
    },
    {
      "epoch": 148.72727272727272,
      "grad_norm": 0.10228395462036133,
      "learning_rate": 4.963975358444748e-05,
      "loss": 0.015380208492279052,
      "step": 5800
    },
    {
      "epoch": 151.28571428571428,
      "grad_norm": 0.06292247027158737,
      "learning_rate": 4.960091789330784e-05,
      "loss": 0.015107660293579102,
      "step": 5900
    },
    {
      "epoch": 153.85714285714286,
      "grad_norm": 0.06817900389432907,
      "learning_rate": 4.956011144132521e-05,
      "loss": 0.01492759346961975,
      "step": 6000
    },
    {
      "epoch": 153.85714285714286,
      "eval_exact_match_ratio": 0.07679508511455267,
      "eval_loss": 0.016874490305781364,
      "eval_macro_f1": 0.3877334139086099,
      "eval_macro_precision": 0.7358665653332759,
      "eval_macro_recall": 0.28940207330254175,
      "eval_macro_roc_auc": 0.9184438464756707,
      "eval_micro_f1": 0.52094626246403,
      "eval_micro_precision": 0.8355605645897064,
      "eval_micro_recall": 0.37844867400763266,
      "eval_runtime": 102.1468,
      "eval_samples_per_second": 76.488,
      "eval_steps_per_second": 4.787,
      "step": 6000
    },
    {
      "epoch": 156.41558441558442,
      "grad_norm": 0.05961983650922775,
      "learning_rate": 4.951733749747345e-05,
      "loss": 0.014758473634719849,
      "step": 6100
    },
    {
      "epoch": 158.98701298701297,
      "grad_norm": 0.06107480823993683,
      "learning_rate": 4.947259948834069e-05,
      "loss": 0.0149067223072052,
      "step": 6200
    },
    {
      "epoch": 161.54545454545453,
      "grad_norm": 0.060411397367715836,
      "learning_rate": 4.942590099785488e-05,
      "loss": 0.01451791763305664,
      "step": 6300
    },
    {
      "epoch": 164.1038961038961,
      "grad_norm": 0.054140854626894,
      "learning_rate": 4.93772457669966e-05,
      "loss": 0.014490524530410767,
      "step": 6400
    },
    {
      "epoch": 166.67532467532467,
      "grad_norm": 0.055879805237054825,
      "learning_rate": 4.932663769349945e-05,
      "loss": 0.014141499996185303,
      "step": 6500
    },
    {
      "epoch": 166.67532467532467,
      "eval_exact_match_ratio": 0.07589914245488288,
      "eval_loss": 0.016887707635760307,
      "eval_macro_f1": 0.4082256470374264,
      "eval_macro_precision": 0.7508155552966105,
      "eval_macro_recall": 0.31539244031288544,
      "eval_macro_roc_auc": 0.9191473709383688,
      "eval_micro_f1": 0.5334341011402096,
      "eval_micro_precision": 0.8229330359405891,
      "eval_micro_recall": 0.39461338389362893,
      "eval_runtime": 102.1626,
      "eval_samples_per_second": 76.476,
      "eval_steps_per_second": 4.786,
      "step": 6500
    },
    {
      "epoch": 169.23376623376623,
      "grad_norm": 0.06160331889986992,
      "learning_rate": 4.9274080831537714e-05,
      "loss": 0.014069569110870362,
      "step": 6600
    },
    {
      "epoch": 171.80519480519482,
      "grad_norm": 0.06609829515218735,
      "learning_rate": 4.92195793914017e-05,
      "loss": 0.014194674491882324,
      "step": 6700
    },
    {
      "epoch": 174.36363636363637,
      "grad_norm": 0.06778760254383087,
      "learning_rate": 4.9163137739160386e-05,
      "loss": 0.01400719404220581,
      "step": 6800
    },
    {
      "epoch": 176.93506493506493,
      "grad_norm": 0.060218971222639084,
      "learning_rate": 4.9104760396311634e-05,
      "loss": 0.014022926092147827,
      "step": 6900
    },
    {
      "epoch": 179.4935064935065,
      "grad_norm": 0.06643606722354889,
      "learning_rate": 4.904445203942002e-05,
      "loss": 0.013471884727478027,
      "step": 7000
    },
    {
      "epoch": 179.4935064935065,
      "eval_exact_match_ratio": 0.07730705234864968,
      "eval_loss": 0.016866693273186684,
      "eval_macro_f1": 0.41576847181651655,
      "eval_macro_precision": 0.744571222614412,
      "eval_macro_recall": 0.31318365475605586,
      "eval_macro_roc_auc": 0.9178895140330184,
      "eval_micro_f1": 0.5388735541535226,
      "eval_micro_precision": 0.8313814955640051,
      "eval_micro_recall": 0.3986241765721092,
      "eval_runtime": 102.1754,
      "eval_samples_per_second": 76.467,
      "eval_steps_per_second": 4.786,
      "step": 7000
    },
    {
      "epoch": 182.05194805194805,
      "grad_norm": 0.05420941859483719,
      "learning_rate": 4.898221749974223e-05,
      "loss": 0.013712399005889893,
      "step": 7100
    },
    {
      "epoch": 184.62337662337663,
      "grad_norm": 0.05109863728284836,
      "learning_rate": 4.891806176283993e-05,
      "loss": 0.013760663270950317,
      "step": 7200
    },
    {
      "epoch": 187.1818181818182,
      "grad_norm": 0.05963105708360672,
      "learning_rate": 4.885198996818048e-05,
      "loss": 0.013506799936294556,
      "step": 7300
    },
    {
      "epoch": 189.75324675324674,
      "grad_norm": 0.07721959054470062,
      "learning_rate": 4.878400740872514e-05,
      "loss": 0.012987656593322754,
      "step": 7400
    },
    {
      "epoch": 192.3116883116883,
      "grad_norm": 0.06577788293361664,
      "learning_rate": 4.871411953050512e-05,
      "loss": 0.013001085519790649,
      "step": 7500
    },
    {
      "epoch": 192.3116883116883,
      "eval_exact_match_ratio": 0.07513119160373737,
      "eval_loss": 0.016872141510248184,
      "eval_macro_f1": 0.4409711297909289,
      "eval_macro_precision": 0.7129979657202209,
      "eval_macro_recall": 0.3480125684191662,
      "eval_macro_roc_auc": 0.9192576870565512,
      "eval_micro_f1": 0.5707223977724244,
      "eval_micro_precision": 0.8005441699214464,
      "eval_micro_recall": 0.4434235153990131,
      "eval_runtime": 102.1701,
      "eval_samples_per_second": 76.471,
      "eval_steps_per_second": 4.786,
      "step": 7500
    },
    {
      "epoch": 194.88311688311688,
      "grad_norm": 0.05922481790184975,
      "learning_rate": 4.864233193218523e-05,
      "loss": 0.013138995170593262,
      "step": 7600
    },
    {
      "epoch": 197.44155844155844,
      "grad_norm": 0.06037190929055214,
      "learning_rate": 4.856865036461545e-05,
      "loss": 0.012665022611618042,
      "step": 7700
    },
    {
      "epoch": 200.0,
      "grad_norm": 0.08931905776262283,
      "learning_rate": 4.8493080730370175e-05,
      "loss": 0.013038263320922852,
      "step": 7800
    },
    {
      "epoch": 202.57142857142858,
      "grad_norm": 0.06864232569932938,
      "learning_rate": 4.841562908327539e-05,
      "loss": 0.013083732128143311,
      "step": 7900
    },
    {
      "epoch": 205.12987012987014,
      "grad_norm": 0.0662744864821434,
      "learning_rate": 4.833630162792372e-05,
      "loss": 0.012569264173507691,
      "step": 8000
    },
    {
      "epoch": 205.12987012987014,
      "eval_exact_match_ratio": 0.07500319979521311,
      "eval_loss": 0.017042480409145355,
      "eval_macro_f1": 0.40625272508360205,
      "eval_macro_precision": 0.7716337720371749,
      "eval_macro_recall": 0.3067554564431893,
      "eval_macro_roc_auc": 0.9207542392535188,
      "eval_micro_f1": 0.5226256842951132,
      "eval_micro_precision": 0.8394643433365602,
      "eval_micro_recall": 0.37942098738423397,
      "eval_runtime": 102.1628,
      "eval_samples_per_second": 76.476,
      "eval_steps_per_second": 4.786,
      "step": 8000
    },
    {
      "epoch": 207.7012987012987,
      "grad_norm": 0.08280864357948303,
      "learning_rate": 4.8255104719177324e-05,
      "loss": 0.013807666301727296,
      "step": 8100
    },
    {
      "epoch": 210.25974025974025,
      "grad_norm": 0.05355716869235039,
      "learning_rate": 4.8172044861658895e-05,
      "loss": 0.012576370239257813,
      "step": 8200
    },
    {
      "epoch": 212.83116883116884,
      "grad_norm": 0.06488946080207825,
      "learning_rate": 4.808712870923051e-05,
      "loss": 0.012673401832580566,
      "step": 8300
    },
    {
      "epoch": 215.3896103896104,
      "grad_norm": 0.06996672600507736,
      "learning_rate": 4.800036306446063e-05,
      "loss": 0.01278065800666809,
      "step": 8400
    },
    {
      "epoch": 217.96103896103895,
      "grad_norm": 0.06019515171647072,
      "learning_rate": 4.791175487807914e-05,
      "loss": 0.012741817235946655,
      "step": 8500
    },
    {
      "epoch": 217.96103896103895,
      "eval_exact_match_ratio": 0.07961090490208626,
      "eval_loss": 0.0170586034655571,
      "eval_macro_f1": 0.45372991858744083,
      "eval_macro_precision": 0.7275439363935846,
      "eval_macro_recall": 0.35349195212720724,
      "eval_macro_roc_auc": 0.9195610154086726,
      "eval_micro_f1": 0.5620295067878344,
      "eval_micro_precision": 0.8139116806490274,
      "eval_micro_recall": 0.4292034322662194,
      "eval_runtime": 102.1456,
      "eval_samples_per_second": 76.489,
      "eval_steps_per_second": 4.787,
      "step": 8500
    },
    {
      "epoch": 220.5194805194805,
      "grad_norm": 0.05731276795268059,
      "learning_rate": 4.782131124842054e-05,
      "loss": 0.012283004522323608,
      "step": 8600
    },
    {
      "epoch": 223.07792207792207,
      "grad_norm": 0.06507930159568787,
      "learning_rate": 4.772903942085528e-05,
      "loss": 0.012043085098266602,
      "step": 8700
    },
    {
      "epoch": 225.64935064935065,
      "grad_norm": 0.0741071030497551,
      "learning_rate": 4.763494678720938e-05,
      "loss": 0.015712366104125977,
      "step": 8800
    },
    {
      "epoch": 228.2077922077922,
      "grad_norm": 0.07785498350858688,
      "learning_rate": 4.753904088517223e-05,
      "loss": 0.013810333013534546,
      "step": 8900
    },
    {
      "epoch": 230.7792207792208,
      "grad_norm": 0.08087580651044846,
      "learning_rate": 4.744132939769278e-05,
      "loss": 0.012514785528182984,
      "step": 9000
    },
    {
      "epoch": 230.7792207792208,
      "eval_exact_match_ratio": 0.07564315883783437,
      "eval_loss": 0.01717931404709816,
      "eval_macro_f1": 0.45775823344233985,
      "eval_macro_precision": 0.7157794582662921,
      "eval_macro_recall": 0.3572054692302405,
      "eval_macro_roc_auc": 0.9205136928314973,
      "eval_micro_f1": 0.5658782718385368,
      "eval_micro_precision": 0.8053498496476819,
      "eval_micro_recall": 0.4361797807433336,
      "eval_runtime": 102.141,
      "eval_samples_per_second": 76.492,
      "eval_steps_per_second": 4.788,
      "step": 9000
    },
    {
      "epoch": 233.33766233766235,
      "grad_norm": 0.06029883772134781,
      "learning_rate": 4.734182015236407e-05,
      "loss": 0.01256476879119873,
      "step": 9100
    },
    {
      "epoch": 235.9090909090909,
      "grad_norm": 0.06095489487051964,
      "learning_rate": 4.724052112079614e-05,
      "loss": 0.012196015119552612,
      "step": 9200
    },
    {
      "epoch": 238.46753246753246,
      "grad_norm": 0.05542588606476784,
      "learning_rate": 4.7137440417977464e-05,
      "loss": 0.011755993366241455,
      "step": 9300
    },
    {
      "epoch": 241.02597402597402,
      "grad_norm": 0.0759962648153305,
      "learning_rate": 4.70325863016248e-05,
      "loss": 0.013117395639419556,
      "step": 9400
    },
    {
      "epoch": 243.5974025974026,
      "grad_norm": 2.043687105178833,
      "learning_rate": 4.692596717152179e-05,
      "loss": 0.012369465827941895,
      "step": 9500
    },
    {
      "epoch": 243.5974025974026,
      "eval_exact_match_ratio": 0.07679508511455267,
      "eval_loss": 0.01774309016764164,
      "eval_macro_f1": 0.4646990659068998,
      "eval_macro_precision": 0.7077453391599539,
      "eval_macro_recall": 0.36856962101064555,
      "eval_macro_roc_auc": 0.9200068792026982,
      "eval_micro_f1": 0.572256677158974,
      "eval_micro_precision": 0.7971347948773605,
      "eval_micro_recall": 0.44634045552881696,
      "eval_runtime": 102.1434,
      "eval_samples_per_second": 76.49,
      "eval_steps_per_second": 4.787,
      "step": 9500
    },
    {
      "epoch": 246.15584415584416,
      "grad_norm": 0.2821160852909088,
      "learning_rate": 4.6817591568845934e-05,
      "loss": 0.012548868656158447,
      "step": 9600
    },
    {
      "epoch": 248.72727272727272,
      "grad_norm": 0.06456050276756287,
      "learning_rate": 4.6707468175484415e-05,
      "loss": 0.012746795415878295,
      "step": 9700
    },
    {
      "epoch": 251.28571428571428,
      "grad_norm": 0.06523428112268448,
      "learning_rate": 4.6595605813338626e-05,
      "loss": 0.013266347646713257,
      "step": 9800
    },
    {
      "epoch": 253.85714285714286,
      "grad_norm": 0.07367172092199326,
      "learning_rate": 4.6482013443617415e-05,
      "loss": 0.011969189643859863,
      "step": 9900
    },
    {
      "epoch": 256.4155844155844,
      "grad_norm": 0.05717799812555313,
      "learning_rate": 4.6366700166119225e-05,
      "loss": 0.0123068368434906,
      "step": 10000
    },
    {
      "epoch": 256.4155844155844,
      "eval_exact_match_ratio": 0.07820299500831947,
      "eval_loss": 0.016695819795131683,
      "eval_macro_f1": 0.4386405320492238,
      "eval_macro_precision": 0.7362820141564432,
      "eval_macro_recall": 0.3376546549355771,
      "eval_macro_roc_auc": 0.9195907346871441,
      "eval_micro_f1": 0.5427499592236177,
      "eval_micro_precision": 0.8248475534182738,
      "eval_micro_recall": 0.40443374899730183,
      "eval_runtime": 102.1679,
      "eval_samples_per_second": 76.472,
      "eval_steps_per_second": 4.786,
      "step": 10000
    },
    {
      "epoch": 258.987012987013,
      "grad_norm": 0.05555659160017967,
      "learning_rate": 4.624967521850311e-05,
      "loss": 0.0128944993019104,
      "step": 10100
    },
    {
      "epoch": 261.54545454545456,
      "grad_norm": 0.06749725341796875,
      "learning_rate": 4.6130947975548714e-05,
      "loss": 0.012046520709991454,
      "step": 10200
    },
    {
      "epoch": 264.1038961038961,
      "grad_norm": 0.05790464207530022,
      "learning_rate": 4.6010527948405284e-05,
      "loss": 0.01183195948600769,
      "step": 10300
    },
    {
      "epoch": 266.6753246753247,
      "grad_norm": 0.07628806680440903,
      "learning_rate": 4.588842478382971e-05,
      "loss": 0.011336252689361573,
      "step": 10400
    },
    {
      "epoch": 269.23376623376623,
      "grad_norm": 0.058772485703229904,
      "learning_rate": 4.576464826341373e-05,
      "loss": 0.01166085124015808,
      "step": 10500
    },
    {
      "epoch": 269.23376623376623,
      "eval_exact_match_ratio": 0.07986688851913477,
      "eval_loss": 0.017322411760687828,
      "eval_macro_f1": 0.45475743745987046,
      "eval_macro_precision": 0.7238330693213869,
      "eval_macro_recall": 0.35233480253030286,
      "eval_macro_roc_auc": 0.9197818030099011,
      "eval_micro_f1": 0.5656597673460332,
      "eval_micro_precision": 0.812696388724441,
      "eval_micro_recall": 0.43379761297066044,
      "eval_runtime": 102.243,
      "eval_samples_per_second": 76.416,
      "eval_steps_per_second": 4.783,
      "step": 10500
    },
    {
      "epoch": 271.8051948051948,
      "grad_norm": 0.048091158270835876,
      "learning_rate": 4.563920830280037e-05,
      "loss": 0.011261138916015625,
      "step": 10600
    },
    {
      "epoch": 274.3636363636364,
      "grad_norm": 0.07051370292901993,
      "learning_rate": 4.551211495088958e-05,
      "loss": 0.010775072574615479,
      "step": 10700
    },
    {
      "epoch": 276.93506493506493,
      "grad_norm": 0.06992930173873901,
      "learning_rate": 4.538337838903322e-05,
      "loss": 0.011018173694610596,
      "step": 10800
    },
    {
      "epoch": 279.4935064935065,
      "grad_norm": 0.6913730502128601,
      "learning_rate": 4.5253008930219474e-05,
      "loss": 0.013334927558898925,
      "step": 10900
    },
    {
      "epoch": 282.05194805194805,
      "grad_norm": 0.08289528638124466,
      "learning_rate": 4.5121017018246625e-05,
      "loss": 0.019828026294708253,
      "step": 11000
    },
    {
      "epoch": 282.05194805194805,
      "eval_exact_match_ratio": 0.06937156022014591,
      "eval_loss": 0.017677050083875656,
      "eval_macro_f1": 0.417881040712503,
      "eval_macro_precision": 0.6891737957828016,
      "eval_macro_recall": 0.32455667445623465,
      "eval_macro_roc_auc": 0.9120942872370977,
      "eval_micro_f1": 0.5276260062720248,
      "eval_micro_precision": 0.7875307629204266,
      "eval_micro_recall": 0.39670385765332167,
      "eval_runtime": 102.2474,
      "eval_samples_per_second": 76.413,
      "eval_steps_per_second": 4.783,
      "step": 11000
    },
    {
      "epoch": 284.6233766233766,
      "grad_norm": 0.10242991149425507,
      "learning_rate": 4.498741322688648e-05,
      "loss": 0.013998864889144898,
      "step": 11100
    },
    {
      "epoch": 287.1818181818182,
      "grad_norm": 0.08067472279071808,
      "learning_rate": 4.4852208259037264e-05,
      "loss": 0.013458923101425171,
      "step": 11200
    },
    {
      "epoch": 289.75324675324674,
      "grad_norm": 0.05978621542453766,
      "learning_rate": 4.471541294586623e-05,
      "loss": 0.012362684011459351,
      "step": 11300
    },
    {
      "epoch": 292.31168831168833,
      "grad_norm": 0.06926928460597992,
      "learning_rate": 4.457703824594201e-05,
      "loss": 0.011729782819747925,
      "step": 11400
    },
    {
      "epoch": 294.8831168831169,
      "grad_norm": 0.0719408169388771,
      "learning_rate": 4.4437095244356654e-05,
      "loss": 0.011548184156417847,
      "step": 11500
    },
    {
      "epoch": 294.8831168831169,
      "eval_exact_match_ratio": 0.07999488032765903,
      "eval_loss": 0.017668213695287704,
      "eval_macro_f1": 0.45782690727302355,
      "eval_macro_precision": 0.7119229416605312,
      "eval_macro_recall": 0.36388761766657174,
      "eval_macro_roc_auc": 0.919269940531085,
      "eval_micro_f1": 0.5720337663146746,
      "eval_micro_precision": 0.7961156630684527,
      "eval_micro_recall": 0.446389071197647,
      "eval_runtime": 102.2175,
      "eval_samples_per_second": 76.435,
      "eval_steps_per_second": 4.784,
      "step": 11500
    },
    {
      "epoch": 297.4415584415584,
      "grad_norm": 0.06696569919586182,
      "learning_rate": 4.429559515183774e-05,
      "loss": 0.011518007516860962,
      "step": 11600
    },
    {
      "epoch": 300.0,
      "grad_norm": 0.10853864997625351,
      "learning_rate": 4.415254930385016e-05,
      "loss": 0.011066470146179199,
      "step": 11700
    },
    {
      "epoch": 302.57142857142856,
      "grad_norm": 0.05707867071032524,
      "learning_rate": 4.400796915968813e-05,
      "loss": 0.010919314622879029,
      "step": 11800
    },
    {
      "epoch": 305.12987012987014,
      "grad_norm": 0.06344405561685562,
      "learning_rate": 4.386186630155717e-05,
      "loss": 0.011025784015655517,
      "step": 11900
    },
    {
      "epoch": 307.7012987012987,
      "grad_norm": 0.18356481194496155,
      "learning_rate": 4.3714252433646233e-05,
      "loss": 0.011042345762252808,
      "step": 12000
    },
    {
      "epoch": 307.7012987012987,
      "eval_exact_match_ratio": 0.07730705234864968,
      "eval_loss": 0.01683417335152626,
      "eval_macro_f1": 0.45021063147768897,
      "eval_macro_precision": 0.7296708808497332,
      "eval_macro_recall": 0.34357324527190675,
      "eval_macro_roc_auc": 0.9216874338801263,
      "eval_micro_f1": 0.5344574417185,
      "eval_micro_precision": 0.8318097781429745,
      "eval_micro_recall": 0.39371399402027274,
      "eval_runtime": 102.2777,
      "eval_samples_per_second": 76.39,
      "eval_steps_per_second": 4.781,
      "step": 12000
    },
    {
      "epoch": 310.2597402597403,
      "grad_norm": 0.0740647241473198,
      "learning_rate": 4.356513938119015e-05,
      "loss": 0.01079399585723877,
      "step": 12100
    },
    {
      "epoch": 312.83116883116884,
      "grad_norm": 0.07097247987985611,
      "learning_rate": 4.341453908952226e-05,
      "loss": 0.01074166178703308,
      "step": 12200
    },
    {
      "epoch": 315.38961038961037,
      "grad_norm": 0.12918981909751892,
      "learning_rate": 4.3262463623117524e-05,
      "loss": 0.010779845714569091,
      "step": 12300
    },
    {
      "epoch": 317.961038961039,
      "grad_norm": 0.07301953434944153,
      "learning_rate": 4.3108925164626015e-05,
      "loss": 0.01032239556312561,
      "step": 12400
    },
    {
      "epoch": 320.5194805194805,
      "grad_norm": 0.062331970781087875,
      "learning_rate": 4.295393601389701e-05,
      "loss": 0.010224409103393554,
      "step": 12500
    },
    {
      "epoch": 320.5194805194805,
      "eval_exact_match_ratio": 0.07973889671061052,
      "eval_loss": 0.01777857169508934,
      "eval_macro_f1": 0.47315968038788886,
      "eval_macro_precision": 0.715080594586631,
      "eval_macro_recall": 0.3776558621230981,
      "eval_macro_roc_auc": 0.918684819210908,
      "eval_micro_f1": 0.5652235583363541,
      "eval_micro_precision": 0.8036930799569738,
      "eval_micro_recall": 0.4358880867303532,
      "eval_runtime": 102.2201,
      "eval_samples_per_second": 76.433,
      "eval_steps_per_second": 4.784,
      "step": 12500
    },
    {
      "epoch": 323.0779220779221,
      "grad_norm": 0.06769008934497833,
      "learning_rate": 4.2797508586993616e-05,
      "loss": 0.010095107555389404,
      "step": 12600
    },
    {
      "epoch": 325.64935064935065,
      "grad_norm": 0.06683795154094696,
      "learning_rate": 4.263965541519818e-05,
      "loss": 0.009790812730789184,
      "step": 12700
    },
    {
      "epoch": 328.2077922077922,
      "grad_norm": 0.06232920661568642,
      "learning_rate": 4.248038914400835e-05,
      "loss": 0.009567870497703553,
      "step": 12800
    },
    {
      "epoch": 330.7792207792208,
      "grad_norm": 0.17745979130268097,
      "learning_rate": 4.231972253212414e-05,
      "loss": 0.009850963354110717,
      "step": 12900
    },
    {
      "epoch": 333.3376623376623,
      "grad_norm": 0.05588870495557785,
      "learning_rate": 4.2157668450425745e-05,
      "loss": 0.00982754111289978,
      "step": 13000
    },
    {
      "epoch": 333.3376623376623,
      "eval_exact_match_ratio": 0.0792269294765135,
      "eval_loss": 0.01711001619696617,
      "eval_macro_f1": 0.47405529136163205,
      "eval_macro_precision": 0.7128848133401314,
      "eval_macro_recall": 0.37381816799438905,
      "eval_macro_roc_auc": 0.9217057490188912,
      "eval_micro_f1": 0.5627246385990469,
      "eval_micro_precision": 0.8070669452266328,
      "eval_micro_recall": 0.43195021755511803,
      "eval_runtime": 102.2506,
      "eval_samples_per_second": 76.41,
      "eval_steps_per_second": 4.782,
      "step": 13000
    },
    {
      "epoch": 335.90909090909093,
      "grad_norm": 0.07357288151979446,
      "learning_rate": 4.1994239880942555e-05,
      "loss": 0.010003459453582764,
      "step": 13100
    },
    {
      "epoch": 338.46753246753246,
      "grad_norm": 0.06038432940840721,
      "learning_rate": 4.1829449915813104e-05,
      "loss": 0.009759486317634583,
      "step": 13200
    },
    {
      "epoch": 341.02597402597405,
      "grad_norm": 0.055783241987228394,
      "learning_rate": 4.166331175623631e-05,
      "loss": 0.00970048189163208,
      "step": 13300
    },
    {
      "epoch": 343.5974025974026,
      "grad_norm": 0.05792766809463501,
      "learning_rate": 4.149583871141393e-05,
      "loss": 0.009768897891044617,
      "step": 13400
    },
    {
      "epoch": 346.15584415584414,
      "grad_norm": 0.06152275204658508,
      "learning_rate": 4.1327044197484335e-05,
      "loss": 0.009740744829177856,
      "step": 13500
    },
    {
      "epoch": 346.15584415584414,
      "eval_exact_match_ratio": 0.08089082298732882,
      "eval_loss": 0.017114389687776566,
      "eval_macro_f1": 0.4625623359053038,
      "eval_macro_precision": 0.736632786768169,
      "eval_macro_recall": 0.3599164668951838,
      "eval_macro_roc_auc": 0.9221306145137217,
      "eval_micro_f1": 0.5614438221073292,
      "eval_micro_precision": 0.8171683389074693,
      "eval_micro_recall": 0.42762342302924233,
      "eval_runtime": 102.2376,
      "eval_samples_per_second": 76.42,
      "eval_steps_per_second": 4.783,
      "step": 13500
    },
    {
      "epoch": 348.72727272727275,
      "grad_norm": 0.06315884739160538,
      "learning_rate": 4.115694173644782e-05,
      "loss": 0.009266348481178284,
      "step": 13600
    },
    {
      "epoch": 351.2857142857143,
      "grad_norm": 0.08249340206384659,
      "learning_rate": 4.09855449550833e-05,
      "loss": 0.00944661557674408,
      "step": 13700
    },
    {
      "epoch": 353.85714285714283,
      "grad_norm": 0.13554084300994873,
      "learning_rate": 4.081286758385672e-05,
      "loss": 0.009609118103981018,
      "step": 13800
    },
    {
      "epoch": 356.4155844155844,
      "grad_norm": 0.07106807082891464,
      "learning_rate": 4.063892345582111e-05,
      "loss": 0.009459911584854127,
      "step": 13900
    },
    {
      "epoch": 358.987012987013,
      "grad_norm": 0.062474314123392105,
      "learning_rate": 4.046372650550842e-05,
      "loss": 0.009663520455360413,
      "step": 14000
    },
    {
      "epoch": 358.987012987013,
      "eval_exact_match_ratio": 0.07909893766798926,
      "eval_loss": 0.017970101907849312,
      "eval_macro_f1": 0.4944946900308322,
      "eval_macro_precision": 0.7001791029876936,
      "eval_macro_recall": 0.40705386132757987,
      "eval_macro_roc_auc": 0.9200375956528422,
      "eval_micro_f1": 0.585702507099887,
      "eval_micro_precision": 0.7875179634571956,
      "eval_micro_recall": 0.4662242640803131,
      "eval_runtime": 102.2698,
      "eval_samples_per_second": 76.396,
      "eval_steps_per_second": 4.781,
      "step": 14000
    },
    {
      "epoch": 361.54545454545456,
      "grad_norm": 0.07693453133106232,
      "learning_rate": 4.028729076781324e-05,
      "loss": 0.009516471028327943,
      "step": 14100
    },
    {
      "epoch": 364.1038961038961,
      "grad_norm": 0.07621068507432938,
      "learning_rate": 4.010963037686847e-05,
      "loss": 0.009409522414207458,
      "step": 14200
    },
    {
      "epoch": 366.6753246753247,
      "grad_norm": 0.07692443579435349,
      "learning_rate": 3.993075956491308e-05,
      "loss": 0.00964044153690338,
      "step": 14300
    },
    {
      "epoch": 369.23376623376623,
      "grad_norm": 0.05125854164361954,
      "learning_rate": 3.97506926611519e-05,
      "loss": 0.009347828626632691,
      "step": 14400
    },
    {
      "epoch": 371.8051948051948,
      "grad_norm": 0.0802665576338768,
      "learning_rate": 3.956944409060781e-05,
      "loss": 0.009300141930580139,
      "step": 14500
    },
    {
      "epoch": 371.8051948051948,
      "eval_exact_match_ratio": 0.08473057724305644,
      "eval_loss": 0.018408922478556633,
      "eval_macro_f1": 0.5184574801131602,
      "eval_macro_precision": 0.6860164338168132,
      "eval_macro_recall": 0.4402426848995402,
      "eval_macro_roc_auc": 0.9220007921525831,
      "eval_micro_f1": 0.6151445777862028,
      "eval_micro_precision": 0.768695335276968,
      "eval_micro_recall": 0.5127251513162693,
      "eval_runtime": 102.2596,
      "eval_samples_per_second": 76.404,
      "eval_steps_per_second": 4.782,
      "step": 14500
    },
    {
      "epoch": 374.3636363636364,
      "grad_norm": 0.05051404610276222,
      "learning_rate": 3.938702837296611e-05,
      "loss": 0.009183371663093567,
      "step": 14600
    },
    {
      "epoch": 376.93506493506493,
      "grad_norm": 0.074142225086689,
      "learning_rate": 3.9203460121411326e-05,
      "loss": 0.008864011168479919,
      "step": 14700
    },
    {
      "epoch": 379.4935064935065,
      "grad_norm": 0.06583409756422043,
      "learning_rate": 3.901875404145666e-05,
      "loss": 0.008919439911842346,
      "step": 14800
    },
    {
      "epoch": 382.05194805194805,
      "grad_norm": 0.06410767883062363,
      "learning_rate": 3.8832924929765845e-05,
      "loss": 0.009650479555130004,
      "step": 14900
    },
    {
      "epoch": 384.6233766233766,
      "grad_norm": 0.0898592546582222,
      "learning_rate": 3.864598767296785e-05,
      "loss": 0.009149614572525024,
      "step": 15000
    },
    {
      "epoch": 384.6233766233766,
      "eval_exact_match_ratio": 0.08101881479585306,
      "eval_loss": 0.01826385408639908,
      "eval_macro_f1": 0.5045578914046115,
      "eval_macro_precision": 0.7001430797958645,
      "eval_macro_recall": 0.41279733811801556,
      "eval_macro_roc_auc": 0.9196100077596443,
      "eval_micro_f1": 0.5993424706889046,
      "eval_micro_precision": 0.7836225170762345,
      "eval_micro_recall": 0.48523299059286806,
      "eval_runtime": 102.2662,
      "eval_samples_per_second": 76.399,
      "eval_steps_per_second": 4.782,
      "step": 15000
    },
    {
      "epoch": 387.1818181818182,
      "grad_norm": 0.07705102860927582,
      "learning_rate": 3.845795724646429e-05,
      "loss": 0.008750916123390198,
      "step": 15100
    },
    {
      "epoch": 389.75324675324674,
      "grad_norm": 0.05844540521502495,
      "learning_rate": 3.82688487132298e-05,
      "loss": 0.009145564436912536,
      "step": 15200
    },
    {
      "epoch": 392.31168831168833,
      "grad_norm": 0.05733802542090416,
      "learning_rate": 3.807867722260529e-05,
      "loss": 0.008802804946899414,
      "step": 15300
    },
    {
      "epoch": 394.8831168831169,
      "grad_norm": 0.05915013700723648,
      "learning_rate": 3.788745800908441e-05,
      "loss": 0.008625359535217285,
      "step": 15400
    },
    {
      "epoch": 397.4415584415584,
      "grad_norm": 0.05206109955906868,
      "learning_rate": 3.769520639109308e-05,
      "loss": 0.00883878767490387,
      "step": 15500
    },
    {
      "epoch": 397.4415584415584,
      "eval_exact_match_ratio": 0.08114680660437731,
      "eval_loss": 0.018691128119826317,
      "eval_macro_f1": 0.5062740939280871,
      "eval_macro_precision": 0.6900678564426678,
      "eval_macro_recall": 0.4187990596339661,
      "eval_macro_roc_auc": 0.9199016434747219,
      "eval_micro_f1": 0.6093214227899516,
      "eval_micro_precision": 0.7721169625540802,
      "eval_micro_recall": 0.5032207880599917,
      "eval_runtime": 102.254,
      "eval_samples_per_second": 76.408,
      "eval_steps_per_second": 4.782,
      "step": 15500
    },
    {
      "epoch": 400.0,
      "grad_norm": 0.07545632123947144,
      "learning_rate": 3.750193776976239e-05,
      "loss": 0.008836173415184022,
      "step": 15600
    },
    {
      "epoch": 402.57142857142856,
      "grad_norm": 0.07578064501285553,
      "learning_rate": 3.730766762769473e-05,
      "loss": 0.008473769426345826,
      "step": 15700
    },
    {
      "epoch": 405.12987012987014,
      "grad_norm": 0.3554503321647644,
      "learning_rate": 3.7112411527723614e-05,
      "loss": 0.008956490159034729,
      "step": 15800
    },
    {
      "epoch": 407.7012987012987,
      "grad_norm": 0.06172866374254227,
      "learning_rate": 3.691618511166687e-05,
      "loss": 0.008492099046707153,
      "step": 15900
    },
    {
      "epoch": 410.2597402597403,
      "grad_norm": 0.2783391773700714,
      "learning_rate": 3.67190040990736e-05,
      "loss": 0.008880165219306946,
      "step": 16000
    },
    {
      "epoch": 410.2597402597403,
      "eval_exact_match_ratio": 0.08281070011519263,
      "eval_loss": 0.018855877220630646,
      "eval_macro_f1": 0.5090366067146753,
      "eval_macro_precision": 0.6890681878375199,
      "eval_macro_recall": 0.42546098529909154,
      "eval_macro_roc_auc": 0.9196987162323282,
      "eval_micro_f1": 0.6090965146665883,
      "eval_micro_precision": 0.7690083410565338,
      "eval_micro_recall": 0.5042417171054231,
      "eval_runtime": 102.2618,
      "eval_samples_per_second": 76.402,
      "eval_steps_per_second": 4.782,
      "step": 16000
    },
    {
      "epoch": 412.83116883116884,
      "grad_norm": 0.08073753118515015,
      "learning_rate": 3.6520884285964944e-05,
      "loss": 0.008532912731170654,
      "step": 16100
    },
    {
      "epoch": 415.38961038961037,
      "grad_norm": 0.07652072608470917,
      "learning_rate": 3.6321841543568615e-05,
      "loss": 0.008614161610603332,
      "step": 16200
    },
    {
      "epoch": 417.961038961039,
      "grad_norm": 0.056339651346206665,
      "learning_rate": 3.612189181704748e-05,
      "loss": 0.008198503255844116,
      "step": 16300
    },
    {
      "epoch": 420.5194805194805,
      "grad_norm": 0.06549147516489029,
      "learning_rate": 3.592105112422224e-05,
      "loss": 0.008499912619590759,
      "step": 16400
    },
    {
      "epoch": 423.0779220779221,
      "grad_norm": 0.1966307908296585,
      "learning_rate": 3.571933555428821e-05,
      "loss": 0.008244459629058837,
      "step": 16500
    },
    {
      "epoch": 423.0779220779221,
      "eval_exact_match_ratio": 0.08524254447715346,
      "eval_loss": 0.018479451537132263,
      "eval_macro_f1": 0.509428996063012,
      "eval_macro_precision": 0.7075950863965321,
      "eval_macro_recall": 0.41157342120431734,
      "eval_macro_roc_auc": 0.9189355817154282,
      "eval_micro_f1": 0.5950503956938747,
      "eval_micro_precision": 0.7982980116193438,
      "eval_micro_recall": 0.4742944651061037,
      "eval_runtime": 102.1981,
      "eval_samples_per_second": 76.45,
      "eval_steps_per_second": 4.785,
      "step": 16500
    },
    {
      "epoch": 425.64935064935065,
      "grad_norm": 0.0854429304599762,
      "learning_rate": 3.551676126652646e-05,
      "loss": 0.008234282732009887,
      "step": 16600
    },
    {
      "epoch": 428.2077922077922,
      "grad_norm": 0.05865246802568436,
      "learning_rate": 3.531334448900928e-05,
      "loss": 0.008386944532394408,
      "step": 16700
    },
    {
      "epoch": 430.7792207792208,
      "grad_norm": 0.05798732489347458,
      "learning_rate": 3.510910151730017e-05,
      "loss": 0.008281524777412414,
      "step": 16800
    },
    {
      "epoch": 433.3376623376623,
      "grad_norm": 0.04840199276804924,
      "learning_rate": 3.4904048713148454e-05,
      "loss": 0.008205247521400451,
      "step": 16900
    },
    {
      "epoch": 435.90909090909093,
      "grad_norm": 0.0655435249209404,
      "learning_rate": 3.4698202503178484e-05,
      "loss": 0.008291175961494446,
      "step": 17000
    },
    {
      "epoch": 435.90909090909093,
      "eval_exact_match_ratio": 0.08114680660437731,
      "eval_loss": 0.01775345392525196,
      "eval_macro_f1": 0.4722220915320051,
      "eval_macro_precision": 0.7329299559313631,
      "eval_macro_recall": 0.3677824490997907,
      "eval_macro_roc_auc": 0.9207820883969813,
      "eval_micro_f1": 0.570677311030856,
      "eval_micro_precision": 0.8190787977992998,
      "eval_micro_recall": 0.4378813291523858,
      "eval_runtime": 102.5227,
      "eval_samples_per_second": 76.208,
      "eval_steps_per_second": 4.77,
      "step": 17000
    },
    {
      "epoch": 438.46753246753246,
      "grad_norm": 0.07557027041912079,
      "learning_rate": 3.449157937757375e-05,
      "loss": 0.007943571805953979,
      "step": 17100
    },
    {
      "epoch": 441.02597402597405,
      "grad_norm": 0.06562158465385437,
      "learning_rate": 3.428419588875588e-05,
      "loss": 0.008024969696998596,
      "step": 17200
    },
    {
      "epoch": 443.5974025974026,
      "grad_norm": 0.045727718621492386,
      "learning_rate": 3.4076068650058596e-05,
      "loss": 0.007892493605613709,
      "step": 17300
    },
    {
      "epoch": 446.15584415584414,
      "grad_norm": 0.062242183834314346,
      "learning_rate": 3.3867214334396896e-05,
      "loss": 0.007983540892601013,
      "step": 17400
    },
    {
      "epoch": 448.72727272727275,
      "grad_norm": 0.09176430106163025,
      "learning_rate": 3.365764967293134e-05,
      "loss": 0.008140393495559693,
      "step": 17500
    },
    {
      "epoch": 448.72727272727275,
      "eval_exact_match_ratio": 0.08050684756175605,
      "eval_loss": 0.018921300768852234,
      "eval_macro_f1": 0.5247225649992565,
      "eval_macro_precision": 0.6819721460003585,
      "eval_macro_recall": 0.4462178509471472,
      "eval_macro_roc_auc": 0.9197569641974261,
      "eval_micro_f1": 0.61399416909621,
      "eval_micro_precision": 0.7669057936710244,
      "eval_micro_recall": 0.5119229927805732,
      "eval_runtime": 102.4537,
      "eval_samples_per_second": 76.259,
      "eval_steps_per_second": 4.773,
      "step": 17500
    },
    {
      "epoch": 451.2857142857143,
      "grad_norm": 0.069572314620018,
      "learning_rate": 3.3447391453727736e-05,
      "loss": 0.008110844492912293,
      "step": 17600
    },
    {
      "epoch": 453.85714285714283,
      "grad_norm": 0.08083420991897583,
      "learning_rate": 3.32364565204123e-05,
      "loss": 0.012586702108383179,
      "step": 17700
    },
    {
      "epoch": 456.4155844155844,
      "grad_norm": 0.08073645085096359,
      "learning_rate": 3.302486177082228e-05,
      "loss": 0.008544116020202637,
      "step": 17800
    },
    {
      "epoch": 458.987012987013,
      "grad_norm": 0.0571538545191288,
      "learning_rate": 3.281262415565233e-05,
      "loss": 0.008015545010566711,
      "step": 17900
    },
    {
      "epoch": 461.54545454545456,
      "grad_norm": 0.10560425370931625,
      "learning_rate": 3.2599760677096566e-05,
      "loss": 0.008030588626861573,
      "step": 18000
    },
    {
      "epoch": 461.54545454545456,
      "eval_exact_match_ratio": 0.08140279022142582,
      "eval_loss": 0.0184748824685812,
      "eval_macro_f1": 0.5029128457480906,
      "eval_macro_precision": 0.7041952411473982,
      "eval_macro_recall": 0.4086470727913426,
      "eval_macro_roc_auc": 0.9201859910354003,
      "eval_micro_f1": 0.600189419564335,
      "eval_micro_precision": 0.7865248226950354,
      "eval_micro_recall": 0.48523299059286806,
      "eval_runtime": 102.3897,
      "eval_samples_per_second": 76.306,
      "eval_steps_per_second": 4.776,
      "step": 18000
    },
    {
      "epoch": 464.1038961038961,
      "grad_norm": 0.20402133464813232,
      "learning_rate": 3.238628838748653e-05,
      "loss": 0.00802460491657257,
      "step": 18100
    },
    {
      "epoch": 466.6753246753247,
      "grad_norm": 0.07640907913446426,
      "learning_rate": 3.217222438792517e-05,
      "loss": 0.008065465092658996,
      "step": 18200
    },
    {
      "epoch": 469.23376623376623,
      "grad_norm": 0.05053655430674553,
      "learning_rate": 3.1957585826916856e-05,
      "loss": 0.00797804892063141,
      "step": 18300
    },
    {
      "epoch": 471.8051948051948,
      "grad_norm": 0.07274461537599564,
      "learning_rate": 3.1742389898993675e-05,
      "loss": 0.00811425805091858,
      "step": 18400
    },
    {
      "epoch": 474.3636363636364,
      "grad_norm": 0.08433771133422852,
      "learning_rate": 3.152665384333791e-05,
      "loss": 0.007987116575241088,
      "step": 18500
    },
    {
      "epoch": 474.3636363636364,
      "eval_exact_match_ratio": 0.08217074107257136,
      "eval_loss": 0.019883858039975166,
      "eval_macro_f1": 0.5164657636910949,
      "eval_macro_precision": 0.6930716865996135,
      "eval_macro_recall": 0.4274164596148955,
      "eval_macro_roc_auc": 0.9170862775179656,
      "eval_micro_f1": 0.5956786403085824,
      "eval_micro_precision": 0.7835031114986721,
      "eval_micro_recall": 0.4804929628819368,
      "eval_runtime": 102.446,
      "eval_samples_per_second": 76.265,
      "eval_steps_per_second": 4.773,
      "step": 18500
    },
    {
      "epoch": 476.93506493506493,
      "grad_norm": 0.06414658576250076,
      "learning_rate": 3.131039494240107e-05,
      "loss": 0.007713909745216369,
      "step": 18600
    },
    {
      "epoch": 479.4935064935065,
      "grad_norm": 0.05157611891627312,
      "learning_rate": 3.109363052051944e-05,
      "loss": 0.007647395133972168,
      "step": 18700
    },
    {
      "epoch": 482.05194805194805,
      "grad_norm": 0.07206937670707703,
      "learning_rate": 3.0876377942526166e-05,
      "loss": 0.007665642499923706,
      "step": 18800
    },
    {
      "epoch": 484.6233766233766,
      "grad_norm": 0.062469009310007095,
      "learning_rate": 3.0658654612360246e-05,
      "loss": 0.007545911073684693,
      "step": 18900
    },
    {
      "epoch": 487.1818181818182,
      "grad_norm": 0.07884809374809265,
      "learning_rate": 3.044047797167222e-05,
      "loss": 0.007437369227409363,
      "step": 19000
    },
    {
      "epoch": 487.1818181818182,
      "eval_exact_match_ratio": 0.078970945859465,
      "eval_loss": 0.018954811617732048,
      "eval_macro_f1": 0.5000744829757097,
      "eval_macro_precision": 0.7138213385513621,
      "eval_macro_recall": 0.4066385603558665,
      "eval_macro_roc_auc": 0.9175250522097761,
      "eval_micro_f1": 0.5868171640646567,
      "eval_micro_precision": 0.7949474384011301,
      "eval_micro_recall": 0.46505748802839153,
      "eval_runtime": 102.4305,
      "eval_samples_per_second": 76.276,
      "eval_steps_per_second": 4.774,
      "step": 19000
    },
    {
      "epoch": 489.75324675324674,
      "grad_norm": 0.04569287598133087,
      "learning_rate": 3.0221865498427038e-05,
      "loss": 0.007479221820831299,
      "step": 19100
    },
    {
      "epoch": 492.31168831168833,
      "grad_norm": 0.06724037230014801,
      "learning_rate": 3.0002834705503862e-05,
      "loss": 0.007417883276939392,
      "step": 19200
    },
    {
      "epoch": 494.8831168831169,
      "grad_norm": 0.06905155628919601,
      "learning_rate": 2.9783403139293103e-05,
      "loss": 0.007773596048355103,
      "step": 19300
    },
    {
      "epoch": 497.4415584415584,
      "grad_norm": 0.06041852384805679,
      "learning_rate": 2.956358837829083e-05,
      "loss": 0.007952988743782044,
      "step": 19400
    },
    {
      "epoch": 500.0,
      "grad_norm": 0.11891217529773712,
      "learning_rate": 2.934340803169055e-05,
      "loss": 0.007487051486968994,
      "step": 19500
    },
    {
      "epoch": 500.0,
      "eval_exact_match_ratio": 0.08293869192371688,
      "eval_loss": 0.01937742345035076,
      "eval_macro_f1": 0.5192784158682981,
      "eval_macro_precision": 0.6897390500461071,
      "eval_macro_recall": 0.43222927870173006,
      "eval_macro_roc_auc": 0.9210025526507667,
      "eval_micro_f1": 0.5999253341297692,
      "eval_micro_precision": 0.7777820800743437,
      "eval_micro_recall": 0.48827146989474707,
      "eval_runtime": 102.4352,
      "eval_samples_per_second": 76.273,
      "eval_steps_per_second": 4.774,
      "step": 19500
    },
    {
      "epoch": 502.57142857142856,
      "grad_norm": 0.06379847973585129,
      "learning_rate": 2.9122879737972557e-05,
      "loss": 0.007715587615966797,
      "step": 19600
    },
    {
      "epoch": 505.12987012987014,
      "grad_norm": 0.10253430157899857,
      "learning_rate": 2.8902021163490935e-05,
      "loss": 0.00750508964061737,
      "step": 19700
    },
    {
      "epoch": 507.7012987012987,
      "grad_norm": 0.06131616607308388,
      "learning_rate": 2.8680850001058278e-05,
      "loss": 0.007374524474143982,
      "step": 19800
    },
    {
      "epoch": 510.2597402597403,
      "grad_norm": 0.17613640427589417,
      "learning_rate": 2.845938396852841e-05,
      "loss": 0.007288593053817749,
      "step": 19900
    },
    {
      "epoch": 512.8311688311688,
      "grad_norm": 0.06226614862680435,
      "learning_rate": 2.8237640807376913e-05,
      "loss": 0.007043229341506958,
      "step": 20000
    },
    {
      "epoch": 512.8311688311688,
      "eval_exact_match_ratio": 0.08396262639191092,
      "eval_loss": 0.018526235595345497,
      "eval_macro_f1": 0.49900657541274457,
      "eval_macro_precision": 0.7015647685000564,
      "eval_macro_recall": 0.41016326992459046,
      "eval_macro_roc_auc": 0.919464958554154,
      "eval_micro_f1": 0.5897169724910695,
      "eval_micro_precision": 0.7926704149054048,
      "eval_micro_recall": 0.4695058217263424,
      "eval_runtime": 102.44,
      "eval_samples_per_second": 76.269,
      "eval_steps_per_second": 4.774,
      "step": 20000
    },
    {
      "epoch": 515.3896103896104,
      "grad_norm": 0.06631801277399063,
      "learning_rate": 2.8015638281280005e-05,
      "loss": 0.007223088741302491,
      "step": 20100
    },
    {
      "epoch": 517.961038961039,
      "grad_norm": 0.07037802785634995,
      "learning_rate": 2.7793394174691375e-05,
      "loss": 0.007282387018203735,
      "step": 20200
    },
    {
      "epoch": 520.5194805194806,
      "grad_norm": 0.04856858775019646,
      "learning_rate": 2.7570926291417564e-05,
      "loss": 0.007098578214645386,
      "step": 20300
    },
    {
      "epoch": 523.077922077922,
      "grad_norm": 0.07581399381160736,
      "learning_rate": 2.7348252453191686e-05,
      "loss": 0.006906571984291077,
      "step": 20400
    },
    {
      "epoch": 525.6493506493506,
      "grad_norm": 0.0609334260225296,
      "learning_rate": 2.7125390498245768e-05,
      "loss": 0.00665643036365509,
      "step": 20500
    },
    {
      "epoch": 525.6493506493506,
      "eval_exact_match_ratio": 0.08370664277486241,
      "eval_loss": 0.01919318363070488,
      "eval_macro_f1": 0.517190556697626,
      "eval_macro_precision": 0.7064945903569632,
      "eval_macro_recall": 0.4219813566177954,
      "eval_macro_roc_auc": 0.9193944928124468,
      "eval_micro_f1": 0.6047604087189384,
      "eval_micro_precision": 0.7863367569249922,
      "eval_micro_recall": 0.4913099491966261,
      "eval_runtime": 102.4915,
      "eval_samples_per_second": 76.231,
      "eval_steps_per_second": 4.771,
      "step": 20500
    },
    {
      "epoch": 528.2077922077922,
      "grad_norm": 0.0603305846452713,
      "learning_rate": 2.6902358279881705e-05,
      "loss": 0.0069662582874298095,
      "step": 20600
    },
    {
      "epoch": 530.7792207792207,
      "grad_norm": 0.07023122906684875,
      "learning_rate": 2.6679173665041074e-05,
      "loss": 0.006979672908782959,
      "step": 20700
    },
    {
      "epoch": 533.3376623376623,
      "grad_norm": 0.04295852407813072,
      "learning_rate": 2.6455854532873807e-05,
      "loss": 0.006755899786949158,
      "step": 20800
    },
    {
      "epoch": 535.9090909090909,
      "grad_norm": 0.09937848150730133,
      "learning_rate": 2.6232418773305935e-05,
      "loss": 0.006762274503707886,
      "step": 20900
    },
    {
      "epoch": 538.4675324675325,
      "grad_norm": 0.07118673622608185,
      "learning_rate": 2.6008884285606405e-05,
      "loss": 0.007052455544471741,
      "step": 21000
    },
    {
      "epoch": 538.4675324675325,
      "eval_exact_match_ratio": 0.08869832330730833,
      "eval_loss": 0.018636878579854965,
      "eval_macro_f1": 0.5187186366390416,
      "eval_macro_precision": 0.7058134511458013,
      "eval_macro_recall": 0.42410991039587237,
      "eval_macro_roc_auc": 0.9204623680132684,
      "eval_micro_f1": 0.611543951469014,
      "eval_micro_precision": 0.7874181567561358,
      "eval_micro_recall": 0.49989061474513236,
      "eval_runtime": 102.4549,
      "eval_samples_per_second": 76.258,
      "eval_steps_per_second": 4.773,
      "step": 21000
    },
    {
      "epoch": 541.025974025974,
      "grad_norm": 0.045981913805007935,
      "learning_rate": 2.578526897695321e-05,
      "loss": 0.006719233989715576,
      "step": 21100
    },
    {
      "epoch": 543.5974025974026,
      "grad_norm": 0.04921331629157066,
      "learning_rate": 2.5561590760998844e-05,
      "loss": 0.006752750277519226,
      "step": 21200
    },
    {
      "epoch": 546.1558441558442,
      "grad_norm": 0.05600263550877571,
      "learning_rate": 2.533786755643526e-05,
      "loss": 0.006711591482162476,
      "step": 21300
    },
    {
      "epoch": 548.7272727272727,
      "grad_norm": 0.10798894613981247,
      "learning_rate": 2.5114117285558402e-05,
      "loss": 0.006908068656921387,
      "step": 21400
    },
    {
      "epoch": 551.2857142857143,
      "grad_norm": 3.2802741527557373,
      "learning_rate": 2.4890357872832497e-05,
      "loss": 0.00682065486907959,
      "step": 21500
    },
    {
      "epoch": 551.2857142857143,
      "eval_exact_match_ratio": 0.08703442979649302,
      "eval_loss": 0.019169533625245094,
      "eval_macro_f1": 0.519388429251354,
      "eval_macro_precision": 0.6924490148287507,
      "eval_macro_recall": 0.43284218884773823,
      "eval_macro_roc_auc": 0.9204647135184783,
      "eval_micro_f1": 0.6107118083740425,
      "eval_micro_precision": 0.7772935607483658,
      "eval_micro_recall": 0.5029290940470114,
      "eval_runtime": 102.4504,
      "eval_samples_per_second": 76.261,
      "eval_steps_per_second": 4.773,
      "step": 21500
    },
    {
      "epoch": 553.8571428571429,
      "grad_norm": 0.052002616226673126,
      "learning_rate": 2.4666607243454113e-05,
      "loss": 0.006475589871406555,
      "step": 21600
    },
    {
      "epoch": 556.4155844155844,
      "grad_norm": 0.057201314717531204,
      "learning_rate": 2.4442883321916162e-05,
      "loss": 0.006451610922813416,
      "step": 21700
    },
    {
      "epoch": 558.987012987013,
      "grad_norm": 0.07654105126857758,
      "learning_rate": 2.4219204030572056e-05,
      "loss": 0.006630535125732422,
      "step": 21800
    },
    {
      "epoch": 561.5454545454545,
      "grad_norm": 0.0669059008359909,
      "learning_rate": 2.3995587288199893e-05,
      "loss": 0.006242602467536927,
      "step": 21900
    },
    {
      "epoch": 564.1038961038961,
      "grad_norm": 0.0663963183760643,
      "learning_rate": 2.3772051008566998e-05,
      "loss": 0.006227378249168396,
      "step": 22000
    },
    {
      "epoch": 564.1038961038961,
      "eval_exact_match_ratio": 0.0862664789453475,
      "eval_loss": 0.018407773226499557,
      "eval_macro_f1": 0.5033171825135012,
      "eval_macro_precision": 0.716437883141322,
      "eval_macro_recall": 0.404093693013916,
      "eval_macro_roc_auc": 0.9219270573353048,
      "eval_micro_f1": 0.5866724268279512,
      "eval_micro_precision": 0.8030751034890597,
      "eval_micro_recall": 0.4621405478985877,
      "eval_runtime": 102.4381,
      "eval_samples_per_second": 76.27,
      "eval_steps_per_second": 4.774,
      "step": 22000
    },
    {
      "epoch": 566.6753246753246,
      "grad_norm": 0.1108517125248909,
      "learning_rate": 2.3548613098994942e-05,
      "loss": 0.00636039912700653,
      "step": 22100
    },
    {
      "epoch": 569.2337662337662,
      "grad_norm": 0.26771992444992065,
      "learning_rate": 2.332529145892489e-05,
      "loss": 0.006258840560913086,
      "step": 22200
    },
    {
      "epoch": 571.8051948051948,
      "grad_norm": 0.07792654633522034,
      "learning_rate": 2.3102103978483786e-05,
      "loss": 0.0062644177675247196,
      "step": 22300
    },
    {
      "epoch": 574.3636363636364,
      "grad_norm": 0.058059047907590866,
      "learning_rate": 2.2879068537051118e-05,
      "loss": 0.006302875280380249,
      "step": 22400
    },
    {
      "epoch": 576.9350649350649,
      "grad_norm": 0.07015325874090195,
      "learning_rate": 2.2656203001826677e-05,
      "loss": 0.006340340375900268,
      "step": 22500
    },
    {
      "epoch": 576.9350649350649,
      "eval_exact_match_ratio": 0.08588250351977474,
      "eval_loss": 0.01992851123213768,
      "eval_macro_f1": 0.5326560761262493,
      "eval_macro_precision": 0.6843528286112909,
      "eval_macro_recall": 0.4536617861823306,
      "eval_macro_roc_auc": 0.9193491146852543,
      "eval_micro_f1": 0.6285494025440062,
      "eval_micro_precision": 0.7615193026151931,
      "eval_micro_recall": 0.5351126668125137,
      "eval_runtime": 102.5313,
      "eval_samples_per_second": 76.201,
      "eval_steps_per_second": 4.769,
      "step": 22500
    },
    {
      "epoch": 579.4935064935065,
      "grad_norm": 0.0467400997877121,
      "learning_rate": 2.243352522639916e-05,
      "loss": 0.006265664100646972,
      "step": 22600
    },
    {
      "epoch": 582.0519480519481,
      "grad_norm": 0.058566998690366745,
      "learning_rate": 2.221105304931599e-05,
      "loss": 0.006237353682518005,
      "step": 22700
    },
    {
      "epoch": 584.6233766233767,
      "grad_norm": 0.08225824683904648,
      "learning_rate": 2.198880429265427e-05,
      "loss": 0.006210183501243591,
      "step": 22800
    },
    {
      "epoch": 587.1818181818181,
      "grad_norm": 0.044421106576919556,
      "learning_rate": 2.1766796760593037e-05,
      "loss": 0.005878788828849793,
      "step": 22900
    },
    {
      "epoch": 589.7532467532468,
      "grad_norm": 0.060877785086631775,
      "learning_rate": 2.154504823798704e-05,
      "loss": 0.005977361798286438,
      "step": 23000
    },
    {
      "epoch": 589.7532467532468,
      "eval_exact_match_ratio": 0.0890822987328811,
      "eval_loss": 0.019213346764445305,
      "eval_macro_f1": 0.5308768421530116,
      "eval_macro_precision": 0.6984561703055269,
      "eval_macro_recall": 0.44094993012900996,
      "eval_macro_roc_auc": 0.917706661134088,
      "eval_micro_f1": 0.6160920895544386,
      "eval_micro_precision": 0.7825370675453048,
      "eval_micro_recall": 0.5080337392741681,
      "eval_runtime": 102.4297,
      "eval_samples_per_second": 76.277,
      "eval_steps_per_second": 4.774,
      "step": 23000
    },
    {
      "epoch": 592.3116883116883,
      "grad_norm": 0.06045812740921974,
      "learning_rate": 2.132357648894195e-05,
      "loss": 0.006307604312896728,
      "step": 23100
    },
    {
      "epoch": 594.8831168831168,
      "grad_norm": 0.047916702926158905,
      "learning_rate": 2.1102399255391337e-05,
      "loss": 0.006031628251075744,
      "step": 23200
    },
    {
      "epoch": 597.4415584415584,
      "grad_norm": 0.07191205769777298,
      "learning_rate": 2.0881534255675356e-05,
      "loss": 0.006112837791442871,
      "step": 23300
    },
    {
      "epoch": 600.0,
      "grad_norm": 0.08692576736211777,
      "learning_rate": 2.066099918312132e-05,
      "loss": 0.006083353161811828,
      "step": 23400
    },
    {
      "epoch": 602.5714285714286,
      "grad_norm": 0.05415201559662819,
      "learning_rate": 2.044081170462638e-05,
      "loss": 0.006193019151687622,
      "step": 23500
    },
    {
      "epoch": 602.5714285714286,
      "eval_exact_match_ratio": 0.08524254447715346,
      "eval_loss": 0.018975459039211273,
      "eval_macro_f1": 0.5187755372247677,
      "eval_macro_precision": 0.6936045861883563,
      "eval_macro_recall": 0.4258015695655271,
      "eval_macro_roc_auc": 0.9178839451406456,
      "eval_micro_f1": 0.6080152048346623,
      "eval_micro_precision": 0.7812118437118437,
      "eval_micro_recall": 0.49767860181336443,
      "eval_runtime": 102.4368,
      "eval_samples_per_second": 76.271,
      "eval_steps_per_second": 4.774,
      "step": 23500
    },
    {
      "epoch": 605.1298701298701,
      "grad_norm": 0.06566471606492996,
      "learning_rate": 2.0220989459242133e-05,
      "loss": 0.006098871827125549,
      "step": 23600
    },
    {
      "epoch": 607.7012987012987,
      "grad_norm": 1.7153992652893066,
      "learning_rate": 2.0001550056761675e-05,
      "loss": 0.0059553635120391845,
      "step": 23700
    },
    {
      "epoch": 610.2597402597403,
      "grad_norm": 0.06797078251838684,
      "learning_rate": 1.9782511076308815e-05,
      "loss": 0.0059465599060058594,
      "step": 23800
    },
    {
      "epoch": 612.8311688311688,
      "grad_norm": 0.042730364948511124,
      "learning_rate": 1.9563890064929897e-05,
      "loss": 0.005820088386535645,
      "step": 23900
    },
    {
      "epoch": 615.3896103896104,
      "grad_norm": 0.054665978997945786,
      "learning_rate": 1.934570453618803e-05,
      "loss": 0.00604570984840393,
      "step": 24000
    },
    {
      "epoch": 615.3896103896104,
      "eval_exact_match_ratio": 0.08754639703059004,
      "eval_loss": 0.01899653859436512,
      "eval_macro_f1": 0.5116703285892151,
      "eval_macro_precision": 0.7140171782293508,
      "eval_macro_recall": 0.4144373553882202,
      "eval_macro_roc_auc": 0.9184215515074319,
      "eval_micro_f1": 0.6024522625221111,
      "eval_micro_precision": 0.7968325067989122,
      "eval_micro_recall": 0.4843092928850969,
      "eval_runtime": 102.4302,
      "eval_samples_per_second": 76.276,
      "eval_steps_per_second": 4.774,
      "step": 24000
    },
    {
      "epoch": 617.961038961039,
      "grad_norm": 0.0778375044465065,
      "learning_rate": 1.912797196876017e-05,
      "loss": 0.0058346974849700925,
      "step": 24100
    },
    {
      "epoch": 620.5194805194806,
      "grad_norm": 0.08718080818653107,
      "learning_rate": 1.8910709805036894e-05,
      "loss": 0.005794444680213928,
      "step": 24200
    },
    {
      "epoch": 623.077922077922,
      "grad_norm": 0.05367223173379898,
      "learning_rate": 1.8693935449725065e-05,
      "loss": 0.005720954537391663,
      "step": 24300
    },
    {
      "epoch": 625.6493506493506,
      "grad_norm": 0.04278705269098282,
      "learning_rate": 1.847766626845362e-05,
      "loss": 0.005706824660301209,
      "step": 24400
    },
    {
      "epoch": 628.2077922077922,
      "grad_norm": 0.059389378875494,
      "learning_rate": 1.8261919586382353e-05,
      "loss": 0.005698502659797668,
      "step": 24500
    },
    {
      "epoch": 628.2077922077922,
      "eval_exact_match_ratio": 0.08677844617944451,
      "eval_loss": 0.01986958645284176,
      "eval_macro_f1": 0.5196027596854446,
      "eval_macro_precision": 0.695960295868562,
      "eval_macro_recall": 0.43010956509184517,
      "eval_macro_roc_auc": 0.9181660211732365,
      "eval_micro_f1": 0.6097643843710762,
      "eval_micro_precision": 0.7771878294923934,
      "eval_micro_recall": 0.5016893944918447,
      "eval_runtime": 102.4483,
      "eval_samples_per_second": 76.263,
      "eval_steps_per_second": 4.773,
      "step": 24500
    },
    {
      "epoch": 630.7792207792207,
      "grad_norm": 0.05704973638057709,
      "learning_rate": 1.804671268681407e-05,
      "loss": 0.005646364688873291,
      "step": 24600
    },
    {
      "epoch": 633.3376623376623,
      "grad_norm": 0.08267273008823395,
      "learning_rate": 1.7832062809810014e-05,
      "loss": 0.005629185438156128,
      "step": 24700
    },
    {
      "epoch": 635.9090909090909,
      "grad_norm": 0.7053768634796143,
      "learning_rate": 1.7617987150808732e-05,
      "loss": 0.005555727481842041,
      "step": 24800
    },
    {
      "epoch": 638.4675324675325,
      "grad_norm": 0.07147101312875748,
      "learning_rate": 1.7404502859248654e-05,
      "loss": 0.005533997416496277,
      "step": 24900
    },
    {
      "epoch": 641.025974025974,
      "grad_norm": 0.05342170596122742,
      "learning_rate": 1.7191627037194186e-05,
      "loss": 0.005432799458503723,
      "step": 25000
    },
    {
      "epoch": 641.025974025974,
      "eval_exact_match_ratio": 0.09036221681812363,
      "eval_loss": 0.020724816247820854,
      "eval_macro_f1": 0.5382663740140988,
      "eval_macro_precision": 0.683825789003131,
      "eval_macro_recall": 0.45486292848211296,
      "eval_macro_roc_auc": 0.9167133000856893,
      "eval_micro_f1": 0.6248647400845465,
      "eval_micro_precision": 0.7686710208717876,
      "eval_micro_recall": 0.5263861542575172,
      "eval_runtime": 102.4518,
      "eval_samples_per_second": 76.26,
      "eval_steps_per_second": 4.773,
      "step": 25000
    },
    {
      "epoch": 643.5974025974026,
      "grad_norm": 0.0691411942243576,
      "learning_rate": 1.6979376737965715e-05,
      "loss": 0.005394455790519714,
      "step": 25100
    },
    {
      "epoch": 646.1558441558442,
      "grad_norm": 0.05541299283504486,
      "learning_rate": 1.6767768964773462e-05,
      "loss": 0.00530026376247406,
      "step": 25200
    },
    {
      "epoch": 648.7272727272727,
      "grad_norm": 0.06575076282024384,
      "learning_rate": 1.6556820669355368e-05,
      "loss": 0.005507146716117859,
      "step": 25300
    },
    {
      "epoch": 651.2857142857143,
      "grad_norm": 0.05441686883568764,
      "learning_rate": 1.6346548750619132e-05,
      "loss": 0.005455752611160278,
      "step": 25400
    },
    {
      "epoch": 653.8571428571429,
      "grad_norm": 0.06322501599788666,
      "learning_rate": 1.6136970053288407e-05,
      "loss": 0.005336692333221435,
      "step": 25500
    },
    {
      "epoch": 653.8571428571429,
      "eval_exact_match_ratio": 0.09164213490336619,
      "eval_loss": 0.020510081201791763,
      "eval_macro_f1": 0.5326481503388606,
      "eval_macro_precision": 0.6863773093457161,
      "eval_macro_recall": 0.44768434106788085,
      "eval_macro_roc_auc": 0.9177996942670225,
      "eval_micro_f1": 0.6191676991472924,
      "eval_micro_precision": 0.7732833321197117,
      "eval_micro_recall": 0.5162740951408639,
      "eval_runtime": 102.4821,
      "eval_samples_per_second": 76.238,
      "eval_steps_per_second": 4.772,
      "step": 25500
    },
    {
      "epoch": 656.4155844155844,
      "grad_norm": 0.05284987390041351,
      "learning_rate": 1.5928101366553444e-05,
      "loss": 0.005241130590438843,
      "step": 25600
    },
    {
      "epoch": 658.987012987013,
      "grad_norm": 0.0750289261341095,
      "learning_rate": 1.5719959422726026e-05,
      "loss": 0.005260994434356689,
      "step": 25700
    },
    {
      "epoch": 661.5454545454545,
      "grad_norm": 0.06609339267015457,
      "learning_rate": 1.5512560895899178e-05,
      "loss": 0.005315477848052978,
      "step": 25800
    },
    {
      "epoch": 664.1038961038961,
      "grad_norm": 0.06626884639263153,
      "learning_rate": 1.53059224006113e-05,
      "loss": 0.005291938185691834,
      "step": 25900
    },
    {
      "epoch": 666.6753246753246,
      "grad_norm": 0.0454958900809288,
      "learning_rate": 1.5100060490515267e-05,
      "loss": 0.005286111831665039,
      "step": 26000
    },
    {
      "epoch": 666.6753246753246,
      "eval_exact_match_ratio": 0.0862664789453475,
      "eval_loss": 0.01993655227124691,
      "eval_macro_f1": 0.5263185100891884,
      "eval_macro_precision": 0.703115935221635,
      "eval_macro_recall": 0.43452028856146985,
      "eval_macro_roc_auc": 0.9170057409055367,
      "eval_micro_f1": 0.6061583972031733,
      "eval_micro_precision": 0.7864619678995115,
      "eval_micro_recall": 0.49310872894333846,
      "eval_runtime": 102.5114,
      "eval_samples_per_second": 76.216,
      "eval_steps_per_second": 4.77,
      "step": 26000
    },
    {
      "epoch": 669.2337662337662,
      "grad_norm": 0.07089071720838547,
      "learning_rate": 1.4894991657052318e-05,
      "loss": 0.005330728888511657,
      "step": 26100
    },
    {
      "epoch": 671.8051948051948,
      "grad_norm": 0.20719651877880096,
      "learning_rate": 1.4690732328130891e-05,
      "loss": 0.005397461652755738,
      "step": 26200
    },
    {
      "epoch": 674.3636363636364,
      "grad_norm": 0.05946618318557739,
      "learning_rate": 1.4487298866810664e-05,
      "loss": 0.005473027229309082,
      "step": 26300
    },
    {
      "epoch": 676.9350649350649,
      "grad_norm": 0.06262410432100296,
      "learning_rate": 1.4284707569991662e-05,
      "loss": 0.005509186387062073,
      "step": 26400
    },
    {
      "epoch": 679.4935064935065,
      "grad_norm": 0.06768012046813965,
      "learning_rate": 1.4082974667108768e-05,
      "loss": 0.005361970663070679,
      "step": 26500
    },
    {
      "epoch": 679.4935064935065,
      "eval_exact_match_ratio": 0.08869832330730833,
      "eval_loss": 0.02036813087761402,
      "eval_macro_f1": 0.5428305872455483,
      "eval_macro_precision": 0.6891954433620443,
      "eval_macro_recall": 0.457988531399064,
      "eval_macro_roc_auc": 0.9174813538600485,
      "eval_micro_f1": 0.6258097952837522,
      "eval_micro_precision": 0.7673975214489991,
      "eval_micro_recall": 0.5283307810107197,
      "eval_runtime": 102.4605,
      "eval_samples_per_second": 76.254,
      "eval_steps_per_second": 4.773,
      "step": 26500
    },
    {
      "epoch": 682.0519480519481,
      "grad_norm": 0.05607151612639427,
      "learning_rate": 1.3882116318831584e-05,
      "loss": 0.005440906286239624,
      "step": 26600
    },
    {
      "epoch": 684.6233766233767,
      "grad_norm": 0.05418792739510536,
      "learning_rate": 1.3682148615769783e-05,
      "loss": 0.005189067125320435,
      "step": 26700
    },
    {
      "epoch": 687.1818181818181,
      "grad_norm": 0.11867716163396835,
      "learning_rate": 1.3483087577184155e-05,
      "loss": 0.005343307852745056,
      "step": 26800
    },
    {
      "epoch": 689.7532467532468,
      "grad_norm": 0.06363189965486526,
      "learning_rate": 1.3284949149703258e-05,
      "loss": 0.004987302422523499,
      "step": 26900
    },
    {
      "epoch": 692.3116883116883,
      "grad_norm": 0.06163704767823219,
      "learning_rate": 1.3087749206046007e-05,
      "loss": 0.00512208878993988,
      "step": 27000
    },
    {
      "epoch": 692.3116883116883,
      "eval_exact_match_ratio": 0.0893382823499296,
      "eval_loss": 0.020353108644485474,
      "eval_macro_f1": 0.5497736284039709,
      "eval_macro_precision": 0.6856352520657081,
      "eval_macro_recall": 0.4696166403332373,
      "eval_macro_roc_auc": 0.9185865366187935,
      "eval_micro_f1": 0.6307762478976026,
      "eval_micro_precision": 0.7625004307522658,
      "eval_micro_recall": 0.5378594521014123,
      "eval_runtime": 102.4631,
      "eval_samples_per_second": 76.252,
      "eval_steps_per_second": 4.772,
      "step": 27000
    },
    {
      "epoch": 694.8831168831168,
      "grad_norm": 0.05955934524536133,
      "learning_rate": 1.2891503543750056e-05,
      "loss": 0.0048657068610191345,
      "step": 27100
    },
    {
      "epoch": 697.4415584415584,
      "grad_norm": 0.06009102240204811,
      "learning_rate": 1.2696227883906342e-05,
      "loss": 0.004967238008975983,
      "step": 27200
    },
    {
      "epoch": 700.0,
      "grad_norm": 0.07863937318325043,
      "learning_rate": 1.2501937869899616e-05,
      "loss": 0.0048710906505584715,
      "step": 27300
    },
    {
      "epoch": 702.5714285714286,
      "grad_norm": 0.06519953161478043,
      "learning_rate": 1.2308649066155284e-05,
      "loss": 0.004968412816524505,
      "step": 27400
    },
    {
      "epoch": 705.1298701298701,
      "grad_norm": 0.04750102758407593,
      "learning_rate": 1.2116376956892586e-05,
      "loss": 0.004930488765239716,
      "step": 27500
    },
    {
      "epoch": 705.1298701298701,
      "eval_exact_match_ratio": 0.09228209394598745,
      "eval_loss": 0.02042371965944767,
      "eval_macro_f1": 0.5315062283942905,
      "eval_macro_precision": 0.7033915733815073,
      "eval_macro_recall": 0.43964185993927263,
      "eval_macro_roc_auc": 0.9162567343563148,
      "eval_micro_f1": 0.6133586146621832,
      "eval_micro_precision": 0.78515625,
      "eval_micro_recall": 0.5032450958944068,
      "eval_runtime": 102.4426,
      "eval_samples_per_second": 76.267,
      "eval_steps_per_second": 4.773,
      "step": 27500
    },
    {
      "epoch": 707.7012987012987,
      "grad_norm": 0.0600716769695282,
      "learning_rate": 1.1925136944884118e-05,
      "loss": 0.005037544369697571,
      "step": 27600
    },
    {
      "epoch": 710.2597402597403,
      "grad_norm": 0.08386489003896713,
      "learning_rate": 1.1734944350221971e-05,
      "loss": 0.004885256588459015,
      "step": 27700
    },
    {
      "epoch": 712.8311688311688,
      "grad_norm": 0.050300367176532745,
      "learning_rate": 1.1545814409090393e-05,
      "loss": 0.004709890782833099,
      "step": 27800
    },
    {
      "epoch": 715.3896103896104,
      "grad_norm": 0.051148027181625366,
      "learning_rate": 1.1357762272545305e-05,
      "loss": 0.004758725762367248,
      "step": 27900
    },
    {
      "epoch": 717.961038961039,
      "grad_norm": 0.04437381029129028,
      "learning_rate": 1.117080300530049e-05,
      "loss": 0.004740619957447052,
      "step": 28000
    },
    {
      "epoch": 717.961038961039,
      "eval_exact_match_ratio": 0.09087418405222066,
      "eval_loss": 0.020683255046606064,
      "eval_macro_f1": 0.5378589945081652,
      "eval_macro_precision": 0.6895489930482531,
      "eval_macro_recall": 0.45068156231120365,
      "eval_macro_roc_auc": 0.9166536224563916,
      "eval_micro_f1": 0.6207560097380352,
      "eval_micro_precision": 0.775402432806468,
      "eval_micro_recall": 0.5175381025304455,
      "eval_runtime": 102.4462,
      "eval_samples_per_second": 76.264,
      "eval_steps_per_second": 4.773,
      "step": 28000
    },
    {
      "epoch": 720.5194805194806,
      "grad_norm": 0.04835651069879532,
      "learning_rate": 1.0984951584520803e-05,
      "loss": 0.00483533650636673,
      "step": 28100
    },
    {
      "epoch": 723.077922077922,
      "grad_norm": 0.05378466844558716,
      "learning_rate": 1.0800222898622389e-05,
      "loss": 0.004820087552070617,
      "step": 28200
    },
    {
      "epoch": 725.6493506493506,
      "grad_norm": 0.05804194137454033,
      "learning_rate": 1.0616631746079923e-05,
      "loss": 0.004568048119544983,
      "step": 28300
    },
    {
      "epoch": 728.2077922077922,
      "grad_norm": 0.05501865595579147,
      "learning_rate": 1.0434192834241163e-05,
      "loss": 0.004445096552371979,
      "step": 28400
    },
    {
      "epoch": 730.7792207792207,
      "grad_norm": 0.06859777122735977,
      "learning_rate": 1.025292077814875e-05,
      "loss": 0.004875591993331909,
      "step": 28500
    },
    {
      "epoch": 730.7792207792207,
      "eval_exact_match_ratio": 0.09356201203123,
      "eval_loss": 0.020116720348596573,
      "eval_macro_f1": 0.5353128036191978,
      "eval_macro_precision": 0.7085658267938172,
      "eval_macro_recall": 0.44253946802379573,
      "eval_macro_roc_auc": 0.9177891899462797,
      "eval_micro_f1": 0.618904593639576,
      "eval_micro_precision": 0.7848101265822784,
      "eval_micro_recall": 0.5109020637351418,
      "eval_runtime": 102.4904,
      "eval_samples_per_second": 76.232,
      "eval_steps_per_second": 4.771,
      "step": 28500
    },
    {
      "epoch": 733.3376623376623,
      "grad_norm": 0.07353351265192032,
      "learning_rate": 1.0072830099369362e-05,
      "loss": 0.004735169410705567,
      "step": 28600
    },
    {
      "epoch": 735.9090909090909,
      "grad_norm": 0.07899585366249084,
      "learning_rate": 9.893935224830484e-06,
      "loss": 0.004543424546718597,
      "step": 28700
    },
    {
      "epoch": 738.4675324675325,
      "grad_norm": 0.0518750436604023,
      "learning_rate": 9.716250485664586e-06,
      "loss": 0.004658754169940949,
      "step": 28800
    },
    {
      "epoch": 741.025974025974,
      "grad_norm": 0.039970334619283676,
      "learning_rate": 9.539790116061151e-06,
      "loss": 0.004515608549118042,
      "step": 28900
    },
    {
      "epoch": 743.5974025974026,
      "grad_norm": 0.05125045403838158,
      "learning_rate": 9.364568252126319e-06,
      "loss": 0.004613224565982819,
      "step": 29000
    },
    {
      "epoch": 743.5974025974026,
      "eval_exact_match_ratio": 0.0924100857545117,
      "eval_loss": 0.020941568538546562,
      "eval_macro_f1": 0.5390058053687804,
      "eval_macro_precision": 0.6973278001560494,
      "eval_macro_recall": 0.45079511207887135,
      "eval_macro_roc_auc": 0.9154630758415354,
      "eval_micro_f1": 0.6170613880051844,
      "eval_micro_precision": 0.7828979332511119,
      "eval_micro_recall": 0.5092005153260896,
      "eval_runtime": 102.4716,
      "eval_samples_per_second": 76.246,
      "eval_steps_per_second": 4.772,
      "step": 29000
    },
    {
      "epoch": 746.1558441558442,
      "grad_norm": 0.11053027212619781,
      "learning_rate": 9.190598930750502e-06,
      "loss": 0.0046744984388351445,
      "step": 29100
    },
    {
      "epoch": 748.7272727272727,
      "grad_norm": 0.05686669796705246,
      "learning_rate": 9.01789608848386e-06,
      "loss": 0.004727273285388947,
      "step": 29200
    },
    {
      "epoch": 751.2857142857143,
      "grad_norm": 0.055179938673973083,
      "learning_rate": 8.846473560419888e-06,
      "loss": 0.004479790329933167,
      "step": 29300
    },
    {
      "epoch": 753.8571428571429,
      "grad_norm": 0.05653301998972893,
      "learning_rate": 8.676345079087087e-06,
      "loss": 0.004562745094299317,
      "step": 29400
    },
    {
      "epoch": 756.4155844155844,
      "grad_norm": 0.0768360048532486,
      "learning_rate": 8.507524273348835e-06,
      "loss": 0.004600732326507568,
      "step": 29500
    },
    {
      "epoch": 756.4155844155844,
      "eval_exact_match_ratio": 0.09343402022270575,
      "eval_loss": 0.021207865327596664,
      "eval_macro_f1": 0.5393111862830572,
      "eval_macro_precision": 0.6917878939386617,
      "eval_macro_recall": 0.4536591544852183,
      "eval_macro_roc_auc": 0.9165133828176482,
      "eval_micro_f1": 0.6232494624905572,
      "eval_micro_precision": 0.7744882117196809,
      "eval_micro_recall": 0.5214273560368506,
      "eval_runtime": 102.4626,
      "eval_samples_per_second": 76.252,
      "eval_steps_per_second": 4.772,
      "step": 29500
    },
    {
      "epoch": 758.987012987013,
      "grad_norm": 0.0622091069817543,
      "learning_rate": 8.340024667311643e-06,
      "loss": 0.004532504975795746,
      "step": 29600
    },
    {
      "epoch": 761.5454545454545,
      "grad_norm": 0.07209206372499466,
      "learning_rate": 8.173859679241683e-06,
      "loss": 0.004609873294830322,
      "step": 29700
    },
    {
      "epoch": 764.1038961038961,
      "grad_norm": 0.06777409464120865,
      "learning_rate": 8.009042620489925e-06,
      "loss": 0.004413497447967529,
      "step": 29800
    },
    {
      "epoch": 766.6753246753246,
      "grad_norm": 0.07615849375724792,
      "learning_rate": 7.84558669442573e-06,
      "loss": 0.004251837432384491,
      "step": 29900
    },
    {
      "epoch": 769.2337662337662,
      "grad_norm": 0.09144994616508484,
      "learning_rate": 7.683504995379148e-06,
      "loss": 0.004310681521892548,
      "step": 30000
    },
    {
      "epoch": 769.2337662337662,
      "eval_exact_match_ratio": 0.09023422500959939,
      "eval_loss": 0.021334001794457436,
      "eval_macro_f1": 0.5402881748890329,
      "eval_macro_precision": 0.6899468496407114,
      "eval_macro_recall": 0.45428226732850757,
      "eval_macro_roc_auc": 0.9167386011721886,
      "eval_micro_f1": 0.6215267153327214,
      "eval_micro_precision": 0.7757893205296086,
      "eval_micro_recall": 0.5184374924038018,
      "eval_runtime": 102.4009,
      "eval_samples_per_second": 76.298,
      "eval_steps_per_second": 4.775,
      "step": 30000
    },
    {
      "epoch": 771.8051948051948,
      "grad_norm": 0.0747210830450058,
      "learning_rate": 7.522810507591963e-06,
      "loss": 0.004427394568920135,
      "step": 30100
    },
    {
      "epoch": 774.3636363636364,
      "grad_norm": 0.0587286576628685,
      "learning_rate": 7.363516104177512e-06,
      "loss": 0.004328975975513458,
      "step": 30200
    },
    {
      "epoch": 776.9350649350649,
      "grad_norm": 0.060415104031562805,
      "learning_rate": 7.205634546089441e-06,
      "loss": 0.004391759634017944,
      "step": 30300
    },
    {
      "epoch": 779.4935064935065,
      "grad_norm": 0.06651648879051208,
      "learning_rate": 7.049178481099411e-06,
      "loss": 0.0043424275517463685,
      "step": 30400
    },
    {
      "epoch": 782.0519480519481,
      "grad_norm": 0.054684337228536606,
      "learning_rate": 6.894160442783945e-06,
      "loss": 0.004217715859413147,
      "step": 30500
    },
    {
      "epoch": 782.0519480519481,
      "eval_exact_match_ratio": 0.09023422500959939,
      "eval_loss": 0.021332407370209694,
      "eval_macro_f1": 0.5454786648621834,
      "eval_macro_precision": 0.6855032840659904,
      "eval_macro_recall": 0.46273205862347355,
      "eval_macro_roc_auc": 0.9170952435656144,
      "eval_micro_f1": 0.6272312102920202,
      "eval_micro_precision": 0.7699299709980901,
      "eval_micro_recall": 0.5291572473808308,
      "eval_runtime": 102.4798,
      "eval_samples_per_second": 76.239,
      "eval_steps_per_second": 4.772,
      "step": 30500
    },
    {
      "epoch": 784.6233766233767,
      "grad_norm": 0.03413702920079231,
      "learning_rate": 6.740592849520314e-06,
      "loss": 0.004243727922439575,
      "step": 30600
    },
    {
      "epoch": 787.1818181818181,
      "grad_norm": 0.08396663516759872,
      "learning_rate": 6.588488003491744e-06,
      "loss": 0.004365263879299164,
      "step": 30700
    },
    {
      "epoch": 789.7532467532468,
      "grad_norm": 0.0551944337785244,
      "learning_rate": 6.437858089701906e-06,
      "loss": 0.004189511835575104,
      "step": 30800
    },
    {
      "epoch": 792.3116883116883,
      "grad_norm": 0.05307900533080101,
      "learning_rate": 6.288715174998749e-06,
      "loss": 0.004200064539909363,
      "step": 30900
    },
    {
      "epoch": 794.8831168831168,
      "grad_norm": 0.06540679931640625,
      "learning_rate": 6.141071207107874e-06,
      "loss": 0.004175088405609131,
      "step": 31000
    },
    {
      "epoch": 794.8831168831168,
      "eval_exact_match_ratio": 0.09010623320107514,
      "eval_loss": 0.022123005241155624,
      "eval_macro_f1": 0.552493587052672,
      "eval_macro_precision": 0.6803594261423801,
      "eval_macro_recall": 0.474371105087053,
      "eval_macro_roc_auc": 0.917284771118801,
      "eval_micro_f1": 0.6317771406564252,
      "eval_micro_precision": 0.7644987572493787,
      "eval_micro_recall": 0.5383213009552978,
      "eval_runtime": 102.4636,
      "eval_samples_per_second": 76.251,
      "eval_steps_per_second": 4.772,
      "step": 31000
    },
    {
      "epoch": 797.4415584415584,
      "grad_norm": 0.08182214945554733,
      "learning_rate": 5.994938013675375e-06,
      "loss": 0.004129273295402527,
      "step": 31100
    },
    {
      "epoch": 800.0,
      "grad_norm": 0.09082362800836563,
      "learning_rate": 5.8503273013203745e-06,
      "loss": 0.004135433137416839,
      "step": 31200
    },
    {
      "epoch": 802.5714285714286,
      "grad_norm": 0.047166865319013596,
      "learning_rate": 5.707250654697169e-06,
      "loss": 0.0041842591762542725,
      "step": 31300
    },
    {
      "epoch": 805.1298701298701,
      "grad_norm": 0.05579902231693268,
      "learning_rate": 5.565719535567212e-06,
      "loss": 0.004283812642097473,
      "step": 31400
    },
    {
      "epoch": 807.7012987012987,
      "grad_norm": 0.04604563117027283,
      "learning_rate": 5.4257452818809564e-06,
      "loss": 0.004151110947132111,
      "step": 31500
    },
    {
      "epoch": 807.7012987012987,
      "eval_exact_match_ratio": 0.09036221681812363,
      "eval_loss": 0.02227086015045643,
      "eval_macro_f1": 0.55002505268058,
      "eval_macro_precision": 0.6809654082076066,
      "eval_macro_recall": 0.472064243270022,
      "eval_macro_roc_auc": 0.9168277014000817,
      "eval_micro_f1": 0.6328406710089364,
      "eval_micro_precision": 0.7649186879823594,
      "eval_micro_recall": 0.5396582318481247,
      "eval_runtime": 102.4628,
      "eval_samples_per_second": 76.252,
      "eval_steps_per_second": 4.772,
      "step": 31500
    },
    {
      "epoch": 810.2597402597403,
      "grad_norm": 0.04115508869290352,
      "learning_rate": 5.287339106869507e-06,
      "loss": 0.004166992008686065,
      "step": 31600
    },
    {
      "epoch": 812.8311688311688,
      "grad_norm": 0.045943696051836014,
      "learning_rate": 5.150512098146409e-06,
      "loss": 0.0041005188226699825,
      "step": 31700
    },
    {
      "epoch": 815.3896103896104,
      "grad_norm": 0.08061197400093079,
      "learning_rate": 5.0152752168193764e-06,
      "loss": 0.004030687510967254,
      "step": 31800
    },
    {
      "epoch": 817.961038961039,
      "grad_norm": 0.08528733253479004,
      "learning_rate": 4.8816392966122306e-06,
      "loss": 0.004062847495079041,
      "step": 31900
    },
    {
      "epoch": 820.5194805194806,
      "grad_norm": 0.05935657024383545,
      "learning_rate": 4.749615042997027e-06,
      "loss": 0.003978443443775177,
      "step": 32000
    },
    {
      "epoch": 820.5194805194806,
      "eval_exact_match_ratio": 0.09407397926532701,
      "eval_loss": 0.022233232855796814,
      "eval_macro_f1": 0.5488654165676495,
      "eval_macro_precision": 0.681222886000314,
      "eval_macro_recall": 0.4695009123125078,
      "eval_macro_roc_auc": 0.9160235788327328,
      "eval_micro_f1": 0.6300307374105886,
      "eval_micro_precision": 0.7700031597795176,
      "eval_micro_recall": 0.5331194243904811,
      "eval_runtime": 102.4751,
      "eval_samples_per_second": 76.243,
      "eval_steps_per_second": 4.772,
      "step": 32000
    },
    {
      "epoch": 823.077922077922,
      "grad_norm": 0.5270804762840271,
      "learning_rate": 4.619213032336409e-06,
      "loss": 0.003936229050159454,
      "step": 32100
    },
    {
      "epoch": 825.6493506493506,
      "grad_norm": 0.04487597569823265,
      "learning_rate": 4.490443711036396e-06,
      "loss": 0.004148940443992615,
      "step": 32200
    },
    {
      "epoch": 828.2077922077922,
      "grad_norm": 0.0743882954120636,
      "learning_rate": 4.3633173947094795e-06,
      "loss": 0.004046813249588012,
      "step": 32300
    },
    {
      "epoch": 830.7792207792207,
      "grad_norm": 0.04291708767414093,
      "learning_rate": 4.23784426734829e-06,
      "loss": 0.003959880769252777,
      "step": 32400
    },
    {
      "epoch": 833.3376623376623,
      "grad_norm": 0.049565643072128296,
      "learning_rate": 4.114034380509737e-06,
      "loss": 0.004028058350086212,
      "step": 32500
    },
    {
      "epoch": 833.3376623376623,
      "eval_exact_match_ratio": 0.09445795469089978,
      "eval_loss": 0.021341387182474136,
      "eval_macro_f1": 0.5458380783747244,
      "eval_macro_precision": 0.6917933487081671,
      "eval_macro_recall": 0.4608674332911666,
      "eval_macro_roc_auc": 0.9164873156530098,
      "eval_micro_f1": 0.6276178484072192,
      "eval_micro_precision": 0.7756313944337125,
      "eval_micro_recall": 0.5270424657867231,
      "eval_runtime": 102.4201,
      "eval_samples_per_second": 76.284,
      "eval_steps_per_second": 4.774,
      "step": 32500
    },
    {
      "epoch": 835.9090909090909,
      "grad_norm": 0.7939916253089905,
      "learning_rate": 3.99189765250979e-06,
      "loss": 0.004156913161277771,
      "step": 32600
    },
    {
      "epoch": 838.4675324675325,
      "grad_norm": 0.2271702140569687,
      "learning_rate": 3.8714438676289515e-06,
      "loss": 0.004061726033687592,
      "step": 32700
    },
    {
      "epoch": 841.025974025974,
      "grad_norm": 0.05354131758213043,
      "learning_rate": 3.752682675328406e-06,
      "loss": 0.003989363312721253,
      "step": 32800
    },
    {
      "epoch": 843.5974025974026,
      "grad_norm": 0.06650190055370331,
      "learning_rate": 3.6356235894770497e-06,
      "loss": 0.004005050361156464,
      "step": 32900
    },
    {
      "epoch": 846.1558441558442,
      "grad_norm": 0.05612694099545479,
      "learning_rate": 3.5202759875893247e-06,
      "loss": 0.004036085903644562,
      "step": 33000
    },
    {
      "epoch": 846.1558441558442,
      "eval_exact_match_ratio": 0.09266606937156022,
      "eval_loss": 0.0219939723610878,
      "eval_macro_f1": 0.5526371489281796,
      "eval_macro_precision": 0.6811432438459936,
      "eval_macro_recall": 0.4743215700661558,
      "eval_macro_roc_auc": 0.9170157989509532,
      "eval_micro_f1": 0.6313078562946218,
      "eval_micro_precision": 0.7672671628295193,
      "eval_micro_recall": 0.5362794428644352,
      "eval_runtime": 102.4551,
      "eval_samples_per_second": 76.258,
      "eval_steps_per_second": 4.773,
      "step": 33000
    },
    {
      "epoch": 848.7272727272727,
      "grad_norm": 0.04954937845468521,
      "learning_rate": 3.4066491100739923e-06,
      "loss": 0.003917406499385834,
      "step": 33100
    },
    {
      "epoch": 851.2857142857143,
      "grad_norm": 0.04812461882829666,
      "learning_rate": 3.2947520594938804e-06,
      "loss": 0.003960456252098083,
      "step": 33200
    },
    {
      "epoch": 853.8571428571429,
      "grad_norm": 0.056437257677316666,
      "learning_rate": 3.1845937998366966e-06,
      "loss": 0.004030337333679199,
      "step": 33300
    },
    {
      "epoch": 856.4155844155844,
      "grad_norm": 0.06082282215356827,
      "learning_rate": 3.076183155796947e-06,
      "loss": 0.003885360062122345,
      "step": 33400
    },
    {
      "epoch": 858.987012987013,
      "grad_norm": 0.04901837557554245,
      "learning_rate": 2.9695288120689527e-06,
      "loss": 0.0038607800006866455,
      "step": 33500
    },
    {
      "epoch": 858.987012987013,
      "eval_exact_match_ratio": 0.09458594649942403,
      "eval_loss": 0.02158532477915287,
      "eval_macro_f1": 0.5468801945713692,
      "eval_macro_precision": 0.6889413932339268,
      "eval_macro_recall": 0.46205305380055683,
      "eval_macro_roc_auc": 0.9174391135187892,
      "eval_micro_f1": 0.6256536539436518,
      "eval_micro_precision": 0.7741611700602237,
      "eval_micro_recall": 0.5249519920270304,
      "eval_runtime": 102.471,
      "eval_samples_per_second": 76.246,
      "eval_steps_per_second": 4.772,
      "step": 33500
    },
    {
      "epoch": 861.5454545454545,
      "grad_norm": 0.06145643815398216,
      "learning_rate": 2.864639312651174e-06,
      "loss": 0.004067351520061493,
      "step": 33600
    },
    {
      "epoch": 864.1038961038961,
      "grad_norm": 0.06027722731232643,
      "learning_rate": 2.7615230601617163e-06,
      "loss": 0.003923779428005219,
      "step": 33700
    },
    {
      "epoch": 866.6753246753246,
      "grad_norm": 0.19871559739112854,
      "learning_rate": 2.6601883151652324e-06,
      "loss": 0.0039792534708976745,
      "step": 33800
    },
    {
      "epoch": 869.2337662337662,
      "grad_norm": 0.06889095157384872,
      "learning_rate": 2.5606431955111672e-06,
      "loss": 0.0037630471587181092,
      "step": 33900
    },
    {
      "epoch": 871.8051948051948,
      "grad_norm": 0.05139719322323799,
      "learning_rate": 2.462895675683416e-06,
      "loss": 0.003732728362083435,
      "step": 34000
    },
    {
      "epoch": 871.8051948051948,
      "eval_exact_match_ratio": 0.09266606937156022,
      "eval_loss": 0.022047758102416992,
      "eval_macro_f1": 0.5519438408644283,
      "eval_macro_precision": 0.6881882167648339,
      "eval_macro_recall": 0.47097013584045666,
      "eval_macro_roc_auc": 0.9169544490121787,
      "eval_micro_f1": 0.6315019575786258,
      "eval_micro_precision": 0.7700944386149003,
      "eval_micro_recall": 0.5351855903157587,
      "eval_runtime": 102.4588,
      "eval_samples_per_second": 76.255,
      "eval_steps_per_second": 4.773,
      "step": 34000
    },
    {
      "epoch": 874.3636363636364,
      "grad_norm": 0.052824586629867554,
      "learning_rate": 2.3669535861615465e-06,
      "loss": 0.0039046216011047363,
      "step": 34100
    },
    {
      "epoch": 876.9350649350649,
      "grad_norm": 0.04024183005094528,
      "learning_rate": 2.2728246127934534e-06,
      "loss": 0.0038760393857955933,
      "step": 34200
    },
    {
      "epoch": 879.4935064935065,
      "grad_norm": 0.05224721133708954,
      "learning_rate": 2.1805162961796882e-06,
      "loss": 0.0037830841541290285,
      "step": 34300
    },
    {
      "epoch": 882.0519480519481,
      "grad_norm": 0.03586350008845329,
      "learning_rate": 2.0900360310693646e-06,
      "loss": 0.0037104877829551696,
      "step": 34400
    },
    {
      "epoch": 884.6233766233767,
      "grad_norm": 0.06388697773218155,
      "learning_rate": 2.0013910657677855e-06,
      "loss": 0.003926326334476471,
      "step": 34500
    },
    {
      "epoch": 884.6233766233767,
      "eval_exact_match_ratio": 0.09266606937156022,
      "eval_loss": 0.022172635421156883,
      "eval_macro_f1": 0.552790260776053,
      "eval_macro_precision": 0.6847316585204405,
      "eval_macro_recall": 0.47308025495494516,
      "eval_macro_roc_auc": 0.9172368166779918,
      "eval_micro_f1": 0.6327479044306323,
      "eval_micro_precision": 0.7650384363473405,
      "eval_micro_recall": 0.5394637691728044,
      "eval_runtime": 102.4638,
      "eval_samples_per_second": 76.251,
      "eval_steps_per_second": 4.772,
      "step": 34500
    },
    {
      "epoch": 887.1818181818181,
      "grad_norm": 0.06209539622068405,
      "learning_rate": 1.9145885015557726e-06,
      "loss": 0.0038799986243247984,
      "step": 34600
    },
    {
      "epoch": 889.7532467532468,
      "grad_norm": 0.04874173179268837,
      "learning_rate": 1.8296352921208044e-06,
      "loss": 0.003744651973247528,
      "step": 34700
    },
    {
      "epoch": 892.3116883116883,
      "grad_norm": 0.03912191465497017,
      "learning_rate": 1.7465382429999511e-06,
      "loss": 0.003766476511955261,
      "step": 34800
    },
    {
      "epoch": 894.8831168831168,
      "grad_norm": 0.057853471487760544,
      "learning_rate": 1.6653040110346824e-06,
      "loss": 0.0038407155871391296,
      "step": 34900
    },
    {
      "epoch": 897.4415584415584,
      "grad_norm": 0.058590102940797806,
      "learning_rate": 1.5859391038376165e-06,
      "loss": 0.0037665438652038576,
      "step": 35000
    },
    {
      "epoch": 897.4415584415584,
      "eval_exact_match_ratio": 0.09305004479713298,
      "eval_loss": 0.02215137891471386,
      "eval_macro_f1": 0.5528647349627238,
      "eval_macro_precision": 0.6873987077415322,
      "eval_macro_recall": 0.47146761168848705,
      "eval_macro_roc_auc": 0.9171181743901659,
      "eval_micro_f1": 0.6322933520424965,
      "eval_micro_precision": 0.7692843704271479,
      "eval_micro_recall": 0.5367169838839058,
      "eval_runtime": 102.511,
      "eval_samples_per_second": 76.216,
      "eval_steps_per_second": 4.77,
      "step": 35000
    },
    {
      "epoch": 900.0,
      "grad_norm": 0.06625393778085709,
      "learning_rate": 1.5084498792711622e-06,
      "loss": 0.003918989896774292,
      "step": 35100
    },
    {
      "epoch": 902.5714285714286,
      "grad_norm": 0.0739719569683075,
      "learning_rate": 1.4328425449382428e-06,
      "loss": 0.0036373594403266906,
      "step": 35200
    },
    {
      "epoch": 905.1298701298701,
      "grad_norm": 0.08472265303134918,
      "learning_rate": 1.3591231576849662e-06,
      "loss": 0.0038391059637069703,
      "step": 35300
    },
    {
      "epoch": 907.7012987012987,
      "grad_norm": 0.061051417142152786,
      "learning_rate": 1.2872976231154382e-06,
      "loss": 0.0037648078799247743,
      "step": 35400
    },
    {
      "epoch": 910.2597402597403,
      "grad_norm": 0.03717147558927536,
      "learning_rate": 1.2173716951186713e-06,
      "loss": 0.003802294433116913,
      "step": 35500
    },
    {
      "epoch": 910.2597402597403,
      "eval_exact_match_ratio": 0.09381799564827852,
      "eval_loss": 0.02224075049161911,
      "eval_macro_f1": 0.5538077232029822,
      "eval_macro_precision": 0.6862840426595412,
      "eval_macro_recall": 0.4740112712714454,
      "eval_macro_roc_auc": 0.9166632565114002,
      "eval_micro_f1": 0.6331097796399308,
      "eval_micro_precision": 0.7688594053903862,
      "eval_micro_recall": 0.5381025304455626,
      "eval_runtime": 102.5069,
      "eval_samples_per_second": 76.219,
      "eval_steps_per_second": 4.77,
      "step": 35500
    },
    {
      "epoch": 912.8311688311688,
      "grad_norm": 0.0466252900660038,
      "learning_rate": 1.1493509754076315e-06,
      "loss": 0.0038596293330192565,
      "step": 35600
    },
    {
      "epoch": 915.3896103896104,
      "grad_norm": 0.05678614601492882,
      "learning_rate": 1.0832409130705046e-06,
      "loss": 0.003809744119644165,
      "step": 35700
    },
    {
      "epoch": 917.961038961039,
      "grad_norm": 0.06241026520729065,
      "learning_rate": 1.019046804134155e-06,
      "loss": 0.004008056223392486,
      "step": 35800
    },
    {
      "epoch": 920.5194805194806,
      "grad_norm": 0.28280916810035706,
      "learning_rate": 9.567737911398811e-07,
      "loss": 0.0036841532588005065,
      "step": 35900
    },
    {
      "epoch": 923.077922077922,
      "grad_norm": 0.05519231781363487,
      "learning_rate": 8.964268627314476e-07,
      "loss": 0.0036885297298431398,
      "step": 36000
    },
    {
      "epoch": 923.077922077922,
      "eval_exact_match_ratio": 0.09177012671189044,
      "eval_loss": 0.02231145091354847,
      "eval_macro_f1": 0.5531246544748587,
      "eval_macro_precision": 0.6872579282528493,
      "eval_macro_recall": 0.4720735475811719,
      "eval_macro_roc_auc": 0.9165041953344719,
      "eval_micro_f1": 0.6320301635771938,
      "eval_micro_precision": 0.7703571678199482,
      "eval_micro_recall": 0.5358175940105496,
      "eval_runtime": 102.4404,
      "eval_samples_per_second": 76.269,
      "eval_steps_per_second": 4.774,
      "step": 36000
    },
    {
      "epoch": 925.6493506493506,
      "grad_norm": 0.10396624356508255,
      "learning_rate": 8.380108532554415e-07,
      "loss": 0.0036322930455207824,
      "step": 36100
    },
    {
      "epoch": 928.2077922077922,
      "grad_norm": 0.06927410513162613,
      "learning_rate": 7.815304423740039e-07,
      "loss": 0.003674095273017883,
      "step": 36200
    },
    {
      "epoch": 930.7792207792207,
      "grad_norm": 0.05493837594985962,
      "learning_rate": 7.269901546899382e-07,
      "loss": 0.003674429655075073,
      "step": 36300
    },
    {
      "epoch": 933.3376623376623,
      "grad_norm": 0.06781146675348282,
      "learning_rate": 6.743943593842616e-07,
      "loss": 0.0037695294618606567,
      "step": 36400
    },
    {
      "epoch": 935.9090909090909,
      "grad_norm": 0.06820249557495117,
      "learning_rate": 6.23747269866179e-07,
      "loss": 0.003663244545459747,
      "step": 36500
    },
    {
      "epoch": 935.9090909090909,
      "eval_exact_match_ratio": 0.09292205298860873,
      "eval_loss": 0.02243983931839466,
      "eval_macro_f1": 0.5548244723161484,
      "eval_macro_precision": 0.6842705622464772,
      "eval_macro_recall": 0.47528953157517456,
      "eval_macro_roc_auc": 0.9165889581029931,
      "eval_micro_f1": 0.6333647502356268,
      "eval_micro_precision": 0.7676809637553225,
      "eval_micro_recall": 0.5390505359877489,
      "eval_runtime": 102.4913,
      "eval_samples_per_second": 76.231,
      "eval_steps_per_second": 4.771,
      "step": 36500
    },
    {
      "epoch": 938.4675324675325,
      "grad_norm": 0.04225321114063263,
      "learning_rate": 5.7505294343555e-07,
      "loss": 0.0036578994989395143,
      "step": 36600
    },
    {
      "epoch": 941.025974025974,
      "grad_norm": 0.05938267707824707,
      "learning_rate": 5.283152809578751e-07,
      "loss": 0.0037998878955841065,
      "step": 36700
    },
    {
      "epoch": 943.5974025974026,
      "grad_norm": 0.07503011077642441,
      "learning_rate": 4.835380265517858e-07,
      "loss": 0.003644196391105652,
      "step": 36800
    },
    {
      "epoch": 946.1558441558442,
      "grad_norm": 0.06064372882246971,
      "learning_rate": 4.407247672891218e-07,
      "loss": 0.0037766295671463014,
      "step": 36900
    },
    {
      "epoch": 948.7272727272727,
      "grad_norm": 0.043074797838926315,
      "learning_rate": 3.998789329075492e-07,
      "loss": 0.003732178211212158,
      "step": 37000
    },
    {
      "epoch": 948.7272727272727,
      "eval_exact_match_ratio": 0.09317803660565724,
      "eval_loss": 0.02236187271773815,
      "eval_macro_f1": 0.5530787262984328,
      "eval_macro_precision": 0.6840495574925283,
      "eval_macro_recall": 0.4729763441113856,
      "eval_macro_roc_auc": 0.9165241503976501,
      "eval_micro_f1": 0.6327554418511584,
      "eval_micro_precision": 0.7678645761065631,
      "eval_micro_recall": 0.5380782226111476,
      "eval_runtime": 102.5085,
      "eval_samples_per_second": 76.218,
      "eval_steps_per_second": 4.77,
      "step": 37000
    },
    {
      "epoch": 951.2857142857143,
      "grad_norm": 0.056531645357608795,
      "learning_rate": 3.6100379553583306e-07,
      "loss": 0.0034473416209220885,
      "step": 37100
    },
    {
      "epoch": 953.8571428571429,
      "grad_norm": 0.09468604624271393,
      "learning_rate": 3.2410246943169464e-07,
      "loss": 0.0038387244939804077,
      "step": 37200
    },
    {
      "epoch": 956.4155844155844,
      "grad_norm": 0.046114273369312286,
      "learning_rate": 2.8917791073233557e-07,
      "loss": 0.00378182053565979,
      "step": 37300
    },
    {
      "epoch": 958.987012987013,
      "grad_norm": 0.04046035185456276,
      "learning_rate": 2.562329172176137e-07,
      "loss": 0.003562682569026947,
      "step": 37400
    },
    {
      "epoch": 961.5454545454545,
      "grad_norm": 0.09024065732955933,
      "learning_rate": 2.252701280859415e-07,
      "loss": 0.0037803184986114503,
      "step": 37500
    },
    {
      "epoch": 961.5454545454545,
      "eval_exact_match_ratio": 0.09279406118008447,
      "eval_loss": 0.02233230508863926,
      "eval_macro_f1": 0.5536486223383112,
      "eval_macro_precision": 0.6838886001732458,
      "eval_macro_recall": 0.4740626198061371,
      "eval_macro_roc_auc": 0.9164343709190207,
      "eval_micro_f1": 0.6332433474739684,
      "eval_micro_precision": 0.7677680798004988,
      "eval_micro_recall": 0.5388317654780136,
      "eval_runtime": 102.4487,
      "eval_samples_per_second": 76.263,
      "eval_steps_per_second": 4.773,
      "step": 37500
    },
    {
      "epoch": 964.1038961038961,
      "grad_norm": 0.035194531083106995,
      "learning_rate": 1.9629202374282786e-07,
      "loss": 0.003699219822883606,
      "step": 37600
    },
    {
      "epoch": 966.6753246753246,
      "grad_norm": 0.06676267832517624,
      "learning_rate": 1.693009256021949e-07,
      "loss": 0.0036105597019195556,
      "step": 37700
    },
    {
      "epoch": 969.2337662337662,
      "grad_norm": 0.05439554527401924,
      "learning_rate": 1.4429899590041317e-07,
      "loss": 0.0036472779512405394,
      "step": 37800
    },
    {
      "epoch": 971.8051948051948,
      "grad_norm": 0.062256623059511185,
      "learning_rate": 1.2128823752307893e-07,
      "loss": 0.0035172441601753236,
      "step": 37900
    },
    {
      "epoch": 974.3636363636364,
      "grad_norm": 0.05331027880311012,
      "learning_rate": 1.0027049384455923e-07,
      "loss": 0.003587372899055481,
      "step": 38000
    },
    {
      "epoch": 974.3636363636364,
      "eval_exact_match_ratio": 0.09266606937156022,
      "eval_loss": 0.022459441795945168,
      "eval_macro_f1": 0.5559551224668399,
      "eval_macro_precision": 0.6808657254529437,
      "eval_macro_recall": 0.47816630868888055,
      "eval_macro_roc_auc": 0.9165614485528388,
      "eval_micro_f1": 0.6343372808016398,
      "eval_micro_precision": 0.7652917539581687,
      "eval_micro_recall": 0.5416514742701573,
      "eval_runtime": 102.4988,
      "eval_samples_per_second": 76.225,
      "eval_steps_per_second": 4.771,
      "step": 38000
    }
  ],
  "logging_steps": 100,
  "max_steps": 39000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1000,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.1153453957029176e+18,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}