[
  {
    "loss": 1.6348,
    "grad_norm": 12.6875,
    "learning_rate": 9.849246231155778e-07,
    "epoch": 0.25125628140703515,
    "step": 50
  },
  {
    "loss": 1.6242,
    "grad_norm": 13.625,
    "learning_rate": 1.989949748743719e-06,
    "epoch": 0.5025125628140703,
    "step": 100
  },
  {
    "loss": 1.6156,
    "grad_norm": 13.0,
    "learning_rate": 2.994974874371859e-06,
    "epoch": 0.7537688442211056,
    "step": 150
  },
  {
    "eval_loss": 1.6190789937973022,
    "eval_accuracy": 0.21059268600252207,
    "eval_top3": 0.6633039092055486,
    "eval_macro_f1": 0.2037528618679592,
    "eval_kappa": 0.03531167592000617,
    "eval_runtime": 64.558,
    "eval_samples_per_second": 12.284,
    "eval_steps_per_second": 0.387,
    "epoch": 1.0,
    "step": 199
  },
  {
    "loss": 1.6267,
    "grad_norm": 12.5625,
    "learning_rate": 4.000000000000001e-06,
    "epoch": 1.0050251256281406,
    "step": 200
  },
  {
    "loss": 1.6181,
    "grad_norm": 10.9375,
    "learning_rate": 5.005025125628141e-06,
    "epoch": 1.2562814070351758,
    "step": 250
  },
  {
    "loss": 1.5978,
    "grad_norm": 14.4375,
    "learning_rate": 6.010050251256282e-06,
    "epoch": 1.507537688442211,
    "step": 300
  },
  {
    "loss": 1.5929,
    "grad_norm": 11.125,
    "learning_rate": 7.015075376884422e-06,
    "epoch": 1.758793969849246,
    "step": 350
  },
  {
    "eval_loss": 1.5563229322433472,
    "eval_accuracy": 0.2938209331651955,
    "eval_top3": 0.7364438839848676,
    "eval_macro_f1": 0.26941324280219475,
    "eval_kappa": 0.11484951166035484,
    "eval_runtime": 1.348,
    "eval_samples_per_second": 588.28,
    "eval_steps_per_second": 18.546,
    "epoch": 2.0,
    "step": 398
  },
  {
    "loss": 1.5629,
    "grad_norm": 16.125,
    "learning_rate": 8.020100502512563e-06,
    "epoch": 2.0100502512562812,
    "step": 400
  },
  {
    "loss": 1.5411,
    "grad_norm": 12.125,
    "learning_rate": 9.025125628140704e-06,
    "epoch": 2.2613065326633164,
    "step": 450
  },
  {
    "loss": 1.5202,
    "grad_norm": 10.375,
    "learning_rate": 1.0030150753768846e-05,
    "epoch": 2.5125628140703515,
    "step": 500
  },
  {
    "loss": 1.4924,
    "grad_norm": 34.25,
    "learning_rate": 1.1035175879396986e-05,
    "epoch": 2.7638190954773867,
    "step": 550
  },
  {
    "eval_loss": 1.4240866899490356,
    "eval_accuracy": 0.46532156368221944,
    "eval_top3": 0.8348045397225725,
    "eval_macro_f1": 0.39144397784367796,
    "eval_kappa": 0.2981436602019366,
    "eval_runtime": 1.339,
    "eval_samples_per_second": 592.233,
    "eval_steps_per_second": 18.671,
    "epoch": 3.0,
    "step": 597
  },
  {
    "loss": 1.4388,
    "grad_norm": 12.375,
    "learning_rate": 1.2040201005025128e-05,
    "epoch": 3.0150753768844223,
    "step": 600
  },
  {
    "loss": 1.3922,
    "grad_norm": 10.125,
    "learning_rate": 1.3045226130653268e-05,
    "epoch": 3.2663316582914574,
    "step": 650
  },
  {
    "loss": 1.3391,
    "grad_norm": 13.0,
    "learning_rate": 1.4050251256281408e-05,
    "epoch": 3.5175879396984926,
    "step": 700
  },
  {
    "loss": 1.2838,
    "grad_norm": 11.625,
    "learning_rate": 1.5055276381909548e-05,
    "epoch": 3.7688442211055277,
    "step": 750
  },
  {
    "eval_loss": 1.202721118927002,
    "eval_accuracy": 0.5800756620428752,
    "eval_top3": 0.9205548549810845,
    "eval_macro_f1": 0.497103403115527,
    "eval_kappa": 0.4472502653133419,
    "eval_runtime": 1.337,
    "eval_samples_per_second": 593.118,
    "eval_steps_per_second": 18.699,
    "epoch": 4.0,
    "step": 796
  },
  {
    "loss": 1.2487,
    "grad_norm": 14.4375,
    "learning_rate": 1.606030150753769e-05,
    "epoch": 4.0201005025125625,
    "step": 800
  },
  {
    "loss": 1.1559,
    "grad_norm": 10.4375,
    "learning_rate": 1.706532663316583e-05,
    "epoch": 4.271356783919598,
    "step": 850
  },
  {
    "loss": 1.0407,
    "grad_norm": 20.25,
    "learning_rate": 1.8070351758793973e-05,
    "epoch": 4.522613065326633,
    "step": 900
  },
  {
    "loss": 0.9483,
    "grad_norm": 10.125,
    "learning_rate": 1.907537688442211e-05,
    "epoch": 4.773869346733669,
    "step": 950
  },
  {
    "eval_loss": 0.8165158629417419,
    "eval_accuracy": 0.7452711223203027,
    "eval_top3": 0.9672131147540983,
    "eval_macro_f1": 0.6978370232443252,
    "eval_kappa": 0.6680433777984089,
    "eval_runtime": 1.3351,
    "eval_samples_per_second": 593.948,
    "eval_steps_per_second": 18.725,
    "epoch": 5.0,
    "step": 995
  },
  {
    "loss": 0.8677,
    "grad_norm": 12.625,
    "learning_rate": 1.9999990154030625e-05,
    "epoch": 5.025125628140704,
    "step": 1000
  },
  {
    "loss": 0.7965,
    "grad_norm": 11.4375,
    "learning_rate": 1.9998205625451942e-05,
    "epoch": 5.276381909547739,
    "step": 1050
  },
  {
    "loss": 0.6202,
    "grad_norm": 10.5,
    "learning_rate": 1.9993344861920247e-05,
    "epoch": 5.527638190954773,
    "step": 1100
  },
  {
    "loss": 0.655,
    "grad_norm": 14.8125,
    "learning_rate": 1.9985409358988966e-05,
    "epoch": 5.778894472361809,
    "step": 1150
  },
  {
    "eval_loss": 0.5597819685935974,
    "eval_accuracy": 0.8007566204287516,
    "eval_top3": 0.9810844892812106,
    "eval_macro_f1": 0.7631654003718922,
    "eval_kappa": 0.7409659352859331,
    "eval_runtime": 1.388,
    "eval_samples_per_second": 571.326,
    "eval_steps_per_second": 18.012,
    "epoch": 6.0,
    "step": 1194
  },
  {
    "loss": 0.5711,
    "grad_norm": 14.375,
    "learning_rate": 1.9974401558243355e-05,
    "epoch": 6.030150753768845,
    "step": 1200
  },
  {
    "loss": 0.5362,
    "grad_norm": 13.25,
    "learning_rate": 1.9960324846549286e-05,
    "epoch": 6.28140703517588,
    "step": 1250
  },
  {
    "loss": 0.5104,
    "grad_norm": 12.9375,
    "learning_rate": 1.994318355501118e-05,
    "epoch": 6.532663316582915,
    "step": 1300
  },
  {
    "loss": 0.4477,
    "grad_norm": 12.0625,
    "learning_rate": 1.99229829576394e-05,
    "epoch": 6.78391959798995,
    "step": 1350
  },
  {
    "eval_loss": 0.46245822310447693,
    "eval_accuracy": 0.8310214375788146,
    "eval_top3": 0.987389659520807,
    "eval_macro_f1": 0.8040762608401479,
    "eval_kappa": 0.7808619761646,
    "eval_runtime": 1.334,
    "eval_samples_per_second": 594.453,
    "eval_steps_per_second": 18.741,
    "epoch": 7.0,
    "step": 1393
  },
  {
    "loss": 0.496,
    "grad_norm": 8.9375,
    "learning_rate": 1.9899729269727584e-05,
    "epoch": 7.035175879396985,
    "step": 1400
  },
  {
    "loss": 0.4427,
    "grad_norm": 8.5625,
    "learning_rate": 1.9873429645940293e-05,
    "epoch": 7.28643216080402,
    "step": 1450
  },
  {
    "loss": 0.4144,
    "grad_norm": 10.75,
    "learning_rate": 1.9844092178111703e-05,
    "epoch": 7.5376884422110555,
    "step": 1500
  },
  {
    "loss": 0.4074,
    "grad_norm": 11.0625,
    "learning_rate": 1.9811725892755905e-05,
    "epoch": 7.788944723618091,
    "step": 1550
  },
  {
    "eval_loss": 0.4191744029521942,
    "eval_accuracy": 0.8423707440100883,
    "eval_top3": 0.9886506935687264,
    "eval_macro_f1": 0.8179574437589323,
    "eval_kappa": 0.7955173694199191,
    "eval_runtime": 1.325,
    "eval_samples_per_second": 598.49,
    "eval_steps_per_second": 18.868,
    "epoch": 8.0,
    "step": 1592
  },
  {
    "loss": 0.402,
    "grad_norm": 11.875,
    "learning_rate": 1.9776340748289643e-05,
    "epoch": 8.040201005025125,
    "step": 1600
  },
  {
    "loss": 0.359,
    "grad_norm": 12.5,
    "learning_rate": 1.973794763196832e-05,
    "epoch": 8.291457286432161,
    "step": 1650
  },
  {
    "loss": 0.326,
    "grad_norm": 12.75,
    "learning_rate": 1.9696558356536224e-05,
    "epoch": 8.542713567839195,
    "step": 1700
  },
  {
    "loss": 0.3792,
    "grad_norm": 13.3125,
    "learning_rate": 1.9652185656591992e-05,
    "epoch": 8.793969849246231,
    "step": 1750
  },
  {
    "eval_loss": 0.39110326766967773,
    "eval_accuracy": 0.8537200504413619,
    "eval_top3": 0.9911727616645649,
    "eval_macro_f1": 0.8292392112258031,
    "eval_kappa": 0.8104571257842844,
    "eval_runtime": 1.37,
    "eval_samples_per_second": 578.832,
    "eval_steps_per_second": 18.248,
    "epoch": 9.0,
    "step": 1791
  },
  {
    "loss": 0.3652,
    "grad_norm": 12.0625,
    "learning_rate": 1.9604843184670462e-05,
    "epoch": 9.045226130653266,
    "step": 1800
  },
  {
    "loss": 0.3435,
    "grad_norm": 19.25,
    "learning_rate": 1.9554545507042047e-05,
    "epoch": 9.296482412060302,
    "step": 1850
  },
  {
    "loss": 0.3287,
    "grad_norm": 6.09375,
    "learning_rate": 1.950130809923104e-05,
    "epoch": 9.547738693467338,
    "step": 1900
  },
  {
    "loss": 0.2922,
    "grad_norm": 11.4375,
    "learning_rate": 1.9445147341254094e-05,
    "epoch": 9.798994974874372,
    "step": 1950
  },
  {
    "eval_loss": 0.3803386092185974,
    "eval_accuracy": 0.8511979823455234,
    "eval_top3": 0.9911727616645649,
    "eval_macro_f1": 0.8273026210900127,
    "eval_kappa": 0.8074331693865371,
    "eval_runtime": 1.391,
    "eval_samples_per_second": 570.092,
    "eval_steps_per_second": 17.973,
    "epoch": 10.0,
    "step": 1990
  },
  {
    "loss": 0.3334,
    "grad_norm": 10.8125,
    "learning_rate": 1.938608051258047e-05,
    "epoch": 10.050251256281408,
    "step": 2000
  },
  {
    "loss": 0.2973,
    "grad_norm": 12.0,
    "learning_rate": 1.932412578681548e-05,
    "epoch": 10.301507537688442,
    "step": 2050
  },
  {
    "loss": 0.2925,
    "grad_norm": 7.53125,
    "learning_rate": 1.9259302226108878e-05,
    "epoch": 10.552763819095478,
    "step": 2100
  },
  {
    "loss": 0.3054,
    "grad_norm": 44.0,
    "learning_rate": 1.9191629775289836e-05,
    "epoch": 10.804020100502512,
    "step": 2150
  },
  {
    "eval_loss": 0.36623987555503845,
    "eval_accuracy": 0.8638083228247163,
    "eval_top3": 0.9911727616645649,
    "eval_macro_f1": 0.8409607591489069,
    "eval_kappa": 0.823617727196901,
    "eval_runtime": 1.364,
    "eval_samples_per_second": 581.379,
    "eval_steps_per_second": 18.328,
    "epoch": 11.0,
    "step": 2189
  },
  {
    "loss": 0.2989,
    "grad_norm": 6.65625,
    "learning_rate": 1.912112925573035e-05,
    "epoch": 11.055276381909549,
    "step": 2200
  },
  {
    "loss": 0.2812,
    "grad_norm": 15.125,
    "learning_rate": 1.9047822358938933e-05,
    "epoch": 11.306532663316583,
    "step": 2250
  },
  {
    "loss": 0.2535,
    "grad_norm": 9.25,
    "learning_rate": 1.897173163988661e-05,
    "epoch": 11.557788944723619,
    "step": 2300
  },
  {
    "loss": 0.2797,
    "grad_norm": 22.0,
    "learning_rate": 1.8892880510067234e-05,
    "epoch": 11.809045226130653,
    "step": 2350
  },
  {
    "eval_loss": 0.3603597581386566,
    "eval_accuracy": 0.8688524590163934,
    "eval_top3": 0.9886506935687264,
    "eval_macro_f1": 0.8466482549906356,
    "eval_kappa": 0.8301133168401472,
    "eval_runtime": 1.368,
    "eval_samples_per_second": 579.678,
    "eval_steps_per_second": 18.275,
    "epoch": 12.0,
    "step": 2388
  },
  {
    "loss": 0.283,
    "grad_norm": 14.0625,
    "learning_rate": 1.881129323029427e-05,
    "epoch": 12.06030150753769,
    "step": 2400
  },
  {
    "loss": 0.2537,
    "grad_norm": 15.0,
    "learning_rate": 1.872699490323626e-05,
    "epoch": 12.311557788944723,
    "step": 2450
  },
  {
    "loss": 0.2988,
    "grad_norm": 14.25,
    "learning_rate": 1.8640011465693237e-05,
    "epoch": 12.56281407035176,
    "step": 2500
  },
  {
    "loss": 0.2403,
    "grad_norm": 8.3125,
    "learning_rate": 1.8550369680616584e-05,
    "epoch": 12.814070351758794,
    "step": 2550
  },
  {
    "eval_loss": 0.35605388879776,
    "eval_accuracy": 0.8738965952080706,
    "eval_top3": 0.9911727616645649,
    "eval_macro_f1": 0.8539434147627019,
    "eval_kappa": 0.8367116653351014,
    "eval_runtime": 1.361,
    "eval_samples_per_second": 582.66,
    "eval_steps_per_second": 18.369,
    "epoch": 13.0,
    "step": 2587
  },
  {
    "loss": 0.2475,
    "grad_norm": 15.5,
    "learning_rate": 1.8458097128874583e-05,
    "epoch": 13.06532663316583,
    "step": 2600
  },
  {
    "loss": 0.2492,
    "grad_norm": 5.5625,
    "learning_rate": 1.8363222200766406e-05,
    "epoch": 13.316582914572864,
    "step": 2650
  },
  {
    "loss": 0.2392,
    "grad_norm": 43.75,
    "learning_rate": 1.8265774087287016e-05,
    "epoch": 13.5678391959799,
    "step": 2700
  },
  {
    "loss": 0.2525,
    "grad_norm": 6.09375,
    "learning_rate": 1.816578277114571e-05,
    "epoch": 13.819095477386934,
    "step": 2750
  },
  {
    "eval_loss": 0.348567396402359,
    "eval_accuracy": 0.8764186633039092,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8541541677819424,
    "eval_kappa": 0.8398197730261395,
    "eval_runtime": 1.403,
    "eval_samples_per_second": 565.218,
    "eval_steps_per_second": 17.819,
    "epoch": 14.0,
    "step": 2786
  },
  {
    "loss": 0.2064,
    "grad_norm": 8.5,
    "learning_rate": 1.8063279017541086e-05,
    "epoch": 14.07035175879397,
    "step": 2800
  },
  {
    "loss": 0.2228,
    "grad_norm": 29.5,
    "learning_rate": 1.7958294364695254e-05,
    "epoch": 14.321608040201005,
    "step": 2850
  },
  {
    "loss": 0.23,
    "grad_norm": 9.5,
    "learning_rate": 1.7850861114150174e-05,
    "epoch": 14.57286432160804,
    "step": 2900
  },
  {
    "loss": 0.2341,
    "grad_norm": 18.75,
    "learning_rate": 1.7741012320829187e-05,
    "epoch": 14.824120603015075,
    "step": 2950
  },
  {
    "eval_loss": 0.3470554053783417,
    "eval_accuracy": 0.8764186633039092,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8554054945028179,
    "eval_kappa": 0.8397983100426509,
    "eval_runtime": 1.396,
    "eval_samples_per_second": 568.052,
    "eval_steps_per_second": 17.908,
    "epoch": 15.0,
    "step": 2985
  },
  {
    "loss": 0.2557,
    "grad_norm": 12.875,
    "learning_rate": 1.7628781782866696e-05,
    "epoch": 15.075376884422111,
    "step": 3000
  },
  {
    "loss": 0.2163,
    "grad_norm": 9.4375,
    "learning_rate": 1.751420403120922e-05,
    "epoch": 15.326633165829145,
    "step": 3050
  },
  {
    "loss": 0.2159,
    "grad_norm": 28.75,
    "learning_rate": 1.7397314318990933e-05,
    "epoch": 15.577889447236181,
    "step": 3100
  },
  {
    "loss": 0.2219,
    "grad_norm": 10.4375,
    "learning_rate": 1.7278148610687053e-05,
    "epoch": 15.829145728643216,
    "step": 3150
  },
  {
    "eval_loss": 0.3462725579738617,
    "eval_accuracy": 0.8764186633039092,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8554054945028179,
    "eval_kappa": 0.8397983100426509,
    "eval_runtime": 1.355,
    "eval_samples_per_second": 585.241,
    "eval_steps_per_second": 18.45,
    "epoch": 16.0,
    "step": 3184
  },
  {
    "loss": 0.2271,
    "grad_norm": 13.8125,
    "learning_rate": 1.7156743571048314e-05,
    "epoch": 16.08040201005025,
    "step": 3200
  },
  {
    "loss": 0.2162,
    "grad_norm": 11.125,
    "learning_rate": 1.703313655382002e-05,
    "epoch": 16.331658291457288,
    "step": 3250
  },
  {
    "loss": 0.2019,
    "grad_norm": 12.75,
    "learning_rate": 1.6907365590249083e-05,
    "epoch": 16.582914572864322,
    "step": 3300
  },
  {
    "loss": 0.1926,
    "grad_norm": 3.6875,
    "learning_rate": 1.6779469377382624e-05,
    "epoch": 16.834170854271356,
    "step": 3350
  },
  {
    "eval_loss": 0.34653961658477783,
    "eval_accuracy": 0.8751576292559899,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.853208113072801,
    "eval_kappa": 0.8382066236630051,
    "eval_runtime": 1.3501,
    "eval_samples_per_second": 587.384,
    "eval_steps_per_second": 18.518,
    "epoch": 17.0,
    "step": 3383
  },
  {
    "loss": 0.2128,
    "grad_norm": 16.5,
    "learning_rate": 1.6649487266161718e-05,
    "epoch": 17.08542713567839,
    "step": 3400
  },
  {
    "loss": 0.1929,
    "grad_norm": 10.5625,
    "learning_rate": 1.6517459249313926e-05,
    "epoch": 17.33668341708543,
    "step": 3450
  },
  {
    "loss": 0.2036,
    "grad_norm": 13.4375,
    "learning_rate": 1.638342594904843e-05,
    "epoch": 17.587939698492463,
    "step": 3500
  },
  {
    "loss": 0.1969,
    "grad_norm": 11.1875,
    "learning_rate": 1.6247428604557392e-05,
    "epoch": 17.839195979899497,
    "step": 3550
  },
  {
    "eval_loss": 0.3459188640117645,
    "eval_accuracy": 0.880201765447667,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8600443723197964,
    "eval_kappa": 0.8447552476373783,
    "eval_runtime": 1.344,
    "eval_samples_per_second": 590.028,
    "eval_steps_per_second": 18.601,
    "epoch": 18.0,
    "step": 3582
  },
  {
    "loss": 0.2207,
    "grad_norm": 18.125,
    "learning_rate": 1.610950905932756e-05,
    "epoch": 18.09045226130653,
    "step": 3600
  },
  {
    "loss": 0.2066,
    "grad_norm": 18.25,
    "learning_rate": 1.5969709748265933e-05,
    "epoch": 18.34170854271357,
    "step": 3650
  },
  {
    "loss": 0.1658,
    "grad_norm": 11.5625,
    "learning_rate": 1.58280736846434e-05,
    "epoch": 18.592964824120603,
    "step": 3700
  },
  {
    "loss": 0.2012,
    "grad_norm": 2.890625,
    "learning_rate": 1.5684644446860518e-05,
    "epoch": 18.844221105527637,
    "step": 3750
  },
  {
    "eval_loss": 0.3451780676841736,
    "eval_accuracy": 0.8751576292559899,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8549175688845306,
    "eval_kappa": 0.8383209115828104,
    "eval_runtime": 1.351,
    "eval_samples_per_second": 586.973,
    "eval_steps_per_second": 18.505,
    "epoch": 19.0,
    "step": 3781
  },
  {
    "loss": 0.2053,
    "grad_norm": 10.75,
    "learning_rate": 1.5539466165039338e-05,
    "epoch": 19.09547738693467,
    "step": 3800
  },
  {
    "loss": 0.1818,
    "grad_norm": 8.375,
    "learning_rate": 1.5392583507445508e-05,
    "epoch": 19.34673366834171,
    "step": 3850
  },
  {
    "loss": 0.2039,
    "grad_norm": 10.0625,
    "learning_rate": 1.524404166674482e-05,
    "epoch": 19.597989949748744,
    "step": 3900
  },
  {
    "loss": 0.1937,
    "grad_norm": 9.5,
    "learning_rate": 1.5093886346098372e-05,
    "epoch": 19.849246231155778,
    "step": 3950
  },
  {
    "eval_loss": 0.3440295159816742,
    "eval_accuracy": 0.880201765447667,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8590029071176492,
    "eval_kappa": 0.844723249362074,
    "eval_runtime": 1.345,
    "eval_samples_per_second": 589.592,
    "eval_steps_per_second": 18.587,
    "epoch": 20.0,
    "step": 3980
  },
  {
    "loss": 0.2018,
    "grad_norm": 17.375,
    "learning_rate": 1.4942163745100668e-05,
    "epoch": 20.100502512562816,
    "step": 4000
  },
  {
    "loss": 0.1673,
    "grad_norm": 7.625,
    "learning_rate": 1.4788920545565014e-05,
    "epoch": 20.35175879396985,
    "step": 4050
  },
  {
    "loss": 0.1789,
    "grad_norm": 20.75,
    "learning_rate": 1.4634203897160497e-05,
    "epoch": 20.603015075376884,
    "step": 4100
  },
  {
    "loss": 0.2001,
    "grad_norm": 3.453125,
    "learning_rate": 1.4478061402905048e-05,
    "epoch": 20.85427135678392,
    "step": 4150
  },
  {
    "eval_loss": 0.3432489037513733,
    "eval_accuracy": 0.8814627994955864,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8622259247649311,
    "eval_kappa": 0.8463894029254059,
    "eval_runtime": 1.365,
    "eval_samples_per_second": 580.952,
    "eval_steps_per_second": 18.315,
    "epoch": 21.0,
    "step": 4179
  },
  {
    "loss": 0.1866,
    "grad_norm": 3.546875,
    "learning_rate": 1.4320541104518992e-05,
    "epoch": 21.105527638190956,
    "step": 4200
  },
  {
    "loss": 0.184,
    "grad_norm": 15.375,
    "learning_rate": 1.4161691467643659e-05,
    "epoch": 21.35678391959799,
    "step": 4250
  },
  {
    "loss": 0.1923,
    "grad_norm": 8.25,
    "learning_rate": 1.400156136692951e-05,
    "epoch": 21.608040201005025,
    "step": 4300
  },
  {
    "loss": 0.1762,
    "grad_norm": 4.46875,
    "learning_rate": 1.3840200070998466e-05,
    "epoch": 21.85929648241206,
    "step": 4350
  },
  {
    "eval_loss": 0.34197643399238586,
    "eval_accuracy": 0.8839848675914249,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8649545688082612,
    "eval_kappa": 0.8496555447709369,
    "eval_runtime": 1.366,
    "eval_samples_per_second": 580.527,
    "eval_steps_per_second": 18.302,
    "epoch": 22.0,
    "step": 4378
  },
  {
    "loss": 0.1991,
    "grad_norm": 13.625,
    "learning_rate": 1.3677657227285e-05,
    "epoch": 22.110552763819097,
    "step": 4400
  },
  {
    "loss": 0.2095,
    "grad_norm": 23.875,
    "learning_rate": 1.3513982846760669e-05,
    "epoch": 22.36180904522613,
    "step": 4450
  },
  {
    "loss": 0.1788,
    "grad_norm": 13.5,
    "learning_rate": 1.3349227288546815e-05,
    "epoch": 22.613065326633166,
    "step": 4500
  },
  {
    "loss": 0.1537,
    "grad_norm": 16.375,
    "learning_rate": 1.3183441244420129e-05,
    "epoch": 22.8643216080402,
    "step": 4550
  },
  {
    "eval_loss": 0.34272438287734985,
    "eval_accuracy": 0.880201765447667,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8598827006510683,
    "eval_kappa": 0.8447574870227686,
    "eval_runtime": 1.407,
    "eval_samples_per_second": 563.611,
    "eval_steps_per_second": 17.768,
    "epoch": 23.0,
    "step": 4577
  },
  {
    "loss": 0.1774,
    "grad_norm": 9.3125,
    "learning_rate": 1.3016675723215879e-05,
    "epoch": 23.115577889447238,
    "step": 4600
  },
  {
    "loss": 0.1776,
    "grad_norm": 7.78125,
    "learning_rate": 1.2848982035133555e-05,
    "epoch": 23.366834170854272,
    "step": 4650
  },
  {
    "loss": 0.1666,
    "grad_norm": 8.5625,
    "learning_rate": 1.2680411775949847e-05,
    "epoch": 23.618090452261306,
    "step": 4700
  },
  {
    "loss": 0.1787,
    "grad_norm": 8.0625,
    "learning_rate": 1.2511016811143697e-05,
    "epoch": 23.86934673366834,
    "step": 4750
  },
  {
    "eval_loss": 0.34232452511787415,
    "eval_accuracy": 0.8827238335435057,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8619418239031577,
    "eval_kappa": 0.8479944267302592,
    "eval_runtime": 1.395,
    "eval_samples_per_second": 568.458,
    "eval_steps_per_second": 17.921,
    "epoch": 24.0,
    "step": 4776
  },
  {
    "loss": 0.1751,
    "grad_norm": 18.25,
    "learning_rate": 1.2340849259938404e-05,
    "epoch": 24.12060301507538,
    "step": 4800
  },
  {
    "loss": 0.1651,
    "grad_norm": 10.625,
    "learning_rate": 1.2169961479265653e-05,
    "epoch": 24.371859296482413,
    "step": 4850
  },
  {
    "loss": 0.1746,
    "grad_norm": 5.28125,
    "learning_rate": 1.1998406047656396e-05,
    "epoch": 24.623115577889447,
    "step": 4900
  },
  {
    "loss": 0.1905,
    "grad_norm": 12.5625,
    "learning_rate": 1.182623574906356e-05,
    "epoch": 24.87437185929648,
    "step": 4950
  },
  {
    "eval_loss": 0.34067824482917786,
    "eval_accuracy": 0.8852459016393442,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8661537754454564,
    "eval_kappa": 0.8512633637898235,
    "eval_runtime": 1.358,
    "eval_samples_per_second": 583.949,
    "eval_steps_per_second": 18.409,
    "epoch": 25.0,
    "step": 4975
  },
  {
    "loss": 0.1701,
    "grad_norm": 6.125,
    "learning_rate": 1.1653503556621553e-05,
    "epoch": 25.12562814070352,
    "step": 5000
  },
  {
    "loss": 0.1833,
    "grad_norm": 16.0,
    "learning_rate": 1.1480262616347544e-05,
    "epoch": 25.376884422110553,
    "step": 5050
  },
  {
    "loss": 0.167,
    "grad_norm": 6.3125,
    "learning_rate": 1.1306566230789593e-05,
    "epoch": 25.628140703517587,
    "step": 5100
  },
  {
    "loss": 0.167,
    "grad_norm": 25.75,
    "learning_rate": 1.1132467842626555e-05,
    "epoch": 25.87939698492462,
    "step": 5150
  },
  {
    "eval_loss": 0.34115472435951233,
    "eval_accuracy": 0.8852459016393442,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8661537754454564,
    "eval_kappa": 0.8512633637898235,
    "eval_runtime": 1.347,
    "eval_samples_per_second": 588.717,
    "eval_steps_per_second": 18.56,
    "epoch": 26.0,
    "step": 5174
  },
  {
    "loss": 0.1756,
    "grad_norm": 6.8125,
    "learning_rate": 1.0958021018224942e-05,
    "epoch": 26.13065326633166,
    "step": 5200
  },
  {
    "loss": 0.1735,
    "grad_norm": 4.59375,
    "learning_rate": 1.0783279431157668e-05,
    "epoch": 26.381909547738694,
    "step": 5250
  },
  {
    "loss": 0.1778,
    "grad_norm": 12.375,
    "learning_rate": 1.0608296845689843e-05,
    "epoch": 26.633165829145728,
    "step": 5300
  },
  {
    "loss": 0.1699,
    "grad_norm": 10.5,
    "learning_rate": 1.0433127100236655e-05,
    "epoch": 26.884422110552762,
    "step": 5350
  },
  {
    "eval_loss": 0.3404243290424347,
    "eval_accuracy": 0.8839848675914249,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8649838454025487,
    "eval_kappa": 0.8496422211963331,
    "eval_runtime": 1.365,
    "eval_samples_per_second": 580.951,
    "eval_steps_per_second": 18.315,
    "epoch": 27.0,
    "step": 5373
  },
  {
    "loss": 0.1706,
    "grad_norm": 5.9375,
    "learning_rate": 1.0257824090798431e-05,
    "epoch": 27.1356783919598,
    "step": 5400
  },
  {
    "loss": 0.1549,
    "grad_norm": 4.5625,
    "learning_rate": 1.0082441754377996e-05,
    "epoch": 27.386934673366834,
    "step": 5450
  },
  {
    "loss": 0.1816,
    "grad_norm": 10.75,
    "learning_rate": 9.907034052385383e-06,
    "epoch": 27.63819095477387,
    "step": 5500
  },
  {
    "loss": 0.1777,
    "grad_norm": 14.0625,
    "learning_rate": 9.731654954035082e-06,
    "epoch": 27.889447236180903,
    "step": 5550
  },
  {
    "eval_loss": 0.3424096405506134,
    "eval_accuracy": 0.8827238335435057,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8628805718895952,
    "eval_kappa": 0.8480078975136847,
    "eval_runtime": 1.368,
    "eval_samples_per_second": 579.679,
    "eval_steps_per_second": 18.275,
    "epoch": 28.0,
    "step": 5572
  },
  {
    "loss": 0.1808,
    "grad_norm": 3.65625,
    "learning_rate": 9.556358419740852e-06,
    "epoch": 28.14070351758794,
    "step": 5600
  },
  {
    "loss": 0.1799,
    "grad_norm": 9.25,
    "learning_rate": 9.381198384513253e-06,
    "epoch": 28.391959798994975,
    "step": 5650
  },
  {
    "loss": 0.155,
    "grad_norm": 7.25,
    "learning_rate": 9.206228741364972e-06,
    "epoch": 28.64321608040201,
    "step": 5700
  },
  {
    "loss": 0.17,
    "grad_norm": 20.375,
    "learning_rate": 9.031503324729128e-06,
    "epoch": 28.894472361809044,
    "step": 5750
  },
  {
    "eval_loss": 0.3412993550300598,
    "eval_accuracy": 0.8827238335435057,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8622758895354596,
    "eval_kappa": 0.8479787600258079,
    "eval_runtime": 1.356,
    "eval_samples_per_second": 584.808,
    "eval_steps_per_second": 18.437,
    "epoch": 29.0,
    "step": 5771
  },
  {
    "loss": 0.1891,
    "grad_norm": 13.625,
    "learning_rate": 8.857075893895537e-06,
    "epoch": 29.14572864321608,
    "step": 5800
  },
  {
    "loss": 0.1624,
    "grad_norm": 9.0625,
    "learning_rate": 8.683000116470117e-06,
    "epoch": 29.396984924623116,
    "step": 5850
  },
  {
    "loss": 0.1562,
    "grad_norm": 9.4375,
    "learning_rate": 8.509329551862535e-06,
    "epoch": 29.64824120603015,
    "step": 5900
  },
  {
    "loss": 0.1926,
    "grad_norm": 10.625,
    "learning_rate": 8.336117634807107e-06,
    "epoch": 29.899497487437188,
    "step": 5950
  },
  {
    "eval_loss": 0.3421500027179718,
    "eval_accuracy": 0.8827238335435057,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8628805718895952,
    "eval_kappa": 0.8480078975136847,
    "eval_runtime": 1.352,
    "eval_samples_per_second": 586.538,
    "eval_steps_per_second": 18.491,
    "epoch": 30.0,
    "step": 5970
  },
  {
    "loss": 0.1803,
    "grad_norm": 12.3125,
    "learning_rate": 8.163417658922048e-06,
    "epoch": 30.150753768844222,
    "step": 6000
  },
  {
    "loss": 0.1612,
    "grad_norm": 19.625,
    "learning_rate": 7.991282760312188e-06,
    "epoch": 30.402010050251256,
    "step": 6050
  },
  {
    "loss": 0.1568,
    "grad_norm": 8.125,
    "learning_rate": 7.819765901220113e-06,
    "epoch": 30.65326633165829,
    "step": 6100
  },
  {
    "loss": 0.1639,
    "grad_norm": 10.6875,
    "learning_rate": 7.648919853730804e-06,
    "epoch": 30.90452261306533,
    "step": 6150
  },
  {
    "eval_loss": 0.34186428785324097,
    "eval_accuracy": 0.8827238335435057,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8628805718895952,
    "eval_kappa": 0.8480078975136847,
    "eval_runtime": 1.391,
    "eval_samples_per_second": 570.094,
    "eval_steps_per_second": 17.973,
    "epoch": 31.0,
    "step": 6169
  },
  {
    "loss": 0.1836,
    "grad_norm": 15.1875,
    "learning_rate": 7.478797183534803e-06,
    "epoch": 31.155778894472363,
    "step": 6200
  },
  {
    "loss": 0.1477,
    "grad_norm": 13.8125,
    "learning_rate": 7.309450233754869e-06,
    "epoch": 31.407035175879397,
    "step": 6250
  },
  {
    "loss": 0.1907,
    "grad_norm": 12.875,
    "learning_rate": 7.140931108841094e-06,
    "epoch": 31.65829145728643,
    "step": 6300
  },
  {
    "loss": 0.1552,
    "grad_norm": 8.25,
    "learning_rate": 6.973291658539496e-06,
    "epoch": 31.90954773869347,
    "step": 6350
  },
  {
    "eval_loss": 0.3415157198905945,
    "eval_accuracy": 0.8839848675914249,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8640505019325028,
    "eval_kappa": 0.8496288952600413,
    "eval_runtime": 1.327,
    "eval_samples_per_second": 597.588,
    "eval_steps_per_second": 18.839,
    "epoch": 32.0,
    "step": 6368
  },
  {
    "loss": 0.1516,
    "grad_norm": 9.0625,
    "learning_rate": 6.806583461938956e-06,
    "epoch": 32.1608040201005,
    "step": 6400
  },
  {
    "loss": 0.187,
    "grad_norm": 4.09375,
    "learning_rate": 6.6408578116014025e-06,
    "epoch": 32.41206030150754,
    "step": 6450
  },
  {
    "loss": 0.1809,
    "grad_norm": 6.09375,
    "learning_rate": 6.4761656977802236e-06,
    "epoch": 32.663316582914575,
    "step": 6500
  },
  {
    "loss": 0.1543,
    "grad_norm": 28.75,
    "learning_rate": 6.312557792731619e-06,
    "epoch": 32.914572864321606,
    "step": 6550
  },
  {
    "eval_loss": 0.342151015996933,
    "eval_accuracy": 0.8839848675914249,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8640505019325028,
    "eval_kappa": 0.8496288952600413,
    "eval_runtime": 1.349,
    "eval_samples_per_second": 587.844,
    "eval_steps_per_second": 18.532,
    "epoch": 33.0,
    "step": 6567
  },
  {
    "loss": 0.1549,
    "grad_norm": 8.3125,
    "learning_rate": 6.150084435123842e-06,
    "epoch": 33.165829145728644,
    "step": 6600
  },
  {
    "loss": 0.1766,
    "grad_norm": 10.75,
    "learning_rate": 5.988795614549054e-06,
    "epoch": 33.41708542713568,
    "step": 6650
  },
  {
    "loss": 0.1597,
    "grad_norm": 5.9375,
    "learning_rate": 5.828740956142611e-06,
    "epoch": 33.66834170854271,
    "step": 6700
  },
  {
    "loss": 0.1688,
    "grad_norm": 6.59375,
    "learning_rate": 5.669969705314432e-06,
    "epoch": 33.91959798994975,
    "step": 6750
  },
  {
    "eval_loss": 0.3422100245952606,
    "eval_accuracy": 0.8839848675914249,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8640505019325028,
    "eval_kappa": 0.8496288952600413,
    "eval_runtime": 1.363,
    "eval_samples_per_second": 581.803,
    "eval_steps_per_second": 18.342,
    "epoch": 34.0,
    "step": 6766
  },
  {
    "loss": 0.1756,
    "grad_norm": 7.875,
    "learning_rate": 5.512530712597268e-06,
    "epoch": 34.17085427135678,
    "step": 6800
  },
  {
    "loss": 0.1617,
    "grad_norm": 5.34375,
    "learning_rate": 5.356472418616425e-06,
    "epoch": 34.42211055276382,
    "step": 6850
  },
  {
    "loss": 0.162,
    "grad_norm": 10.75,
    "learning_rate": 5.201842839185598e-06,
    "epoch": 34.67336683417086,
    "step": 6900
  },
  {
    "loss": 0.1638,
    "grad_norm": 3.078125,
    "learning_rate": 5.048689550533444e-06,
    "epoch": 34.92462311557789,
    "step": 6950
  },
  {
    "eval_loss": 0.34207883477211,
    "eval_accuracy": 0.8839848675914249,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8640505019325028,
    "eval_kappa": 0.8496288952600413,
    "eval_runtime": 1.365,
    "eval_samples_per_second": 580.953,
    "eval_steps_per_second": 18.315,
    "epoch": 35.0,
    "step": 6965
  },
  {
    "loss": 0.1858,
    "grad_norm": 21.625,
    "learning_rate": 4.897059674665391e-06,
    "epoch": 35.175879396984925,
    "step": 7000
  },
  {
    "loss": 0.1714,
    "grad_norm": 6.875,
    "learning_rate": 4.746999864865188e-06,
    "epoch": 35.42713567839196,
    "step": 7050
  },
  {
    "loss": 0.1725,
    "grad_norm": 6.5625,
    "learning_rate": 4.5985562913407015e-06,
    "epoch": 35.678391959798994,
    "step": 7100
  },
  {
    "loss": 0.1641,
    "grad_norm": 23.75,
    "learning_rate": 4.451774627018303e-06,
    "epoch": 35.92964824120603,
    "step": 7150
  },
  {
    "eval_loss": 0.34272658824920654,
    "eval_accuracy": 0.8827238335435057,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8628805718895952,
    "eval_kappa": 0.8480078975136847,
    "eval_runtime": 1.403,
    "eval_samples_per_second": 565.216,
    "eval_steps_per_second": 17.819,
    "epoch": 36.0,
    "step": 7164
  },
  {
    "loss": 0.1798,
    "grad_norm": 23.375,
    "learning_rate": 4.3067000334902985e-06,
    "epoch": 36.18090452261306,
    "step": 7200
  },
  {
    "loss": 0.1658,
    "grad_norm": 15.625,
    "learning_rate": 4.163377147119664e-06,
    "epoch": 36.4321608040201,
    "step": 7250
  },
  {
    "loss": 0.1755,
    "grad_norm": 5.0,
    "learning_rate": 4.0218500653063905e-06,
    "epoch": 36.68341708542714,
    "step": 7300
  },
  {
    "loss": 0.1623,
    "grad_norm": 5.5625,
    "learning_rate": 3.882162332919645e-06,
    "epoch": 36.93467336683417,
    "step": 7350
  },
  {
    "eval_loss": 0.3424283564090729,
    "eval_accuracy": 0.8827238335435057,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8628805718895952,
    "eval_kappa": 0.8480078975136847,
    "eval_runtime": 1.372,
    "eval_samples_per_second": 577.988,
    "eval_steps_per_second": 18.222,
    "epoch": 37.0,
    "step": 7363
  },
  {
    "loss": 0.1677,
    "grad_norm": 7.96875,
    "learning_rate": 3.744356928899939e-06,
    "epoch": 37.185929648241206,
    "step": 7400
  },
  {
    "loss": 0.1683,
    "grad_norm": 14.375,
    "learning_rate": 3.6084762530354487e-06,
    "epoch": 37.437185929648244,
    "step": 7450
  },
  {
    "loss": 0.1778,
    "grad_norm": 8.375,
    "learning_rate": 3.4745621129164676e-06,
    "epoch": 37.688442211055275,
    "step": 7500
  },
  {
    "loss": 0.1565,
    "grad_norm": 7.84375,
    "learning_rate": 3.3426557110721326e-06,
    "epoch": 37.93969849246231,
    "step": 7550
  },
  {
    "eval_loss": 0.3421717584133148,
    "eval_accuracy": 0.8839848675914249,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8640505019325028,
    "eval_kappa": 0.8496288952600413,
    "eval_runtime": 1.341,
    "eval_samples_per_second": 591.348,
    "eval_steps_per_second": 18.643,
    "epoch": 38.0,
    "step": 7562
  },
  {
    "loss": 0.1708,
    "grad_norm": 40.75,
    "learning_rate": 3.2127976322932596e-06,
    "epoch": 38.19095477386934,
    "step": 7600
  },
  {
    "loss": 0.1841,
    "grad_norm": 9.125,
    "learning_rate": 3.0850278311452797e-06,
    "epoch": 38.44221105527638,
    "step": 7650
  },
  {
    "loss": 0.1649,
    "grad_norm": 6.9375,
    "learning_rate": 2.9593856196750705e-06,
    "epoch": 38.69346733668342,
    "step": 7700
  },
  {
    "loss": 0.1606,
    "grad_norm": 22.125,
    "learning_rate": 2.8359096553154806e-06,
    "epoch": 38.94472361809045,
    "step": 7750
  },
  {
    "eval_loss": 0.34265828132629395,
    "eval_accuracy": 0.8827238335435057,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8628805718895952,
    "eval_kappa": 0.8480078975136847,
    "eval_runtime": 1.358,
    "eval_samples_per_second": 583.946,
    "eval_steps_per_second": 18.409,
    "epoch": 39.0,
    "step": 7761
  },
  {
    "loss": 0.176,
    "grad_norm": 12.875,
    "learning_rate": 2.7146379289912337e-06,
    "epoch": 39.19597989949749,
    "step": 7800
  },
  {
    "loss": 0.1633,
    "grad_norm": 5.71875,
    "learning_rate": 2.5956077534299715e-06,
    "epoch": 39.447236180904525,
    "step": 7850
  },
  {
    "loss": 0.1517,
    "grad_norm": 9.5,
    "learning_rate": 2.4788557516818935e-06,
    "epoch": 39.698492462311556,
    "step": 7900
  },
  {
    "loss": 0.1839,
    "grad_norm": 4.5625,
    "learning_rate": 2.3644178458516353e-06,
    "epoch": 39.949748743718594,
    "step": 7950
  },
  {
    "eval_loss": 0.3417166769504547,
    "eval_accuracy": 0.8839848675914249,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8640505019325028,
    "eval_kappa": 0.8496288952600413,
    "eval_runtime": 1.347,
    "eval_samples_per_second": 588.716,
    "eval_steps_per_second": 18.56,
    "epoch": 40.0,
    "step": 7960
  },
  {
    "loss": 0.1759,
    "grad_norm": 9.0625,
    "learning_rate": 2.252329246045787e-06,
    "epoch": 40.20100502512563,
    "step": 8000
  },
  {
    "loss": 0.1724,
    "grad_norm": 10.0625,
    "learning_rate": 2.142624439539516e-06,
    "epoch": 40.45226130653266,
    "step": 8050
  },
  {
    "loss": 0.1762,
    "grad_norm": 5.875,
    "learning_rate": 2.0353371801655564e-06,
    "epoch": 40.7035175879397,
    "step": 8100
  },
  {
    "loss": 0.1538,
    "grad_norm": 17.625,
    "learning_rate": 1.9305004779288673e-06,
    "epoch": 40.95477386934673,
    "step": 8150
  },
  {
    "eval_loss": 0.3423360288143158,
    "eval_accuracy": 0.8827238335435057,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8628805718895952,
    "eval_kappa": 0.8480078975136847,
    "eval_runtime": 1.323,
    "eval_samples_per_second": 599.394,
    "eval_steps_per_second": 18.896,
    "epoch": 41.0,
    "step": 8159
  },
  {
    "loss": 0.1696,
    "grad_norm": 11.875,
    "learning_rate": 1.8281465888501627e-06,
    "epoch": 41.20603015075377,
    "step": 8200
  },
  {
    "loss": 0.1642,
    "grad_norm": 12.875,
    "learning_rate": 1.7283070050414275e-06,
    "epoch": 41.45728643216081,
    "step": 8250
  },
  {
    "loss": 0.1711,
    "grad_norm": 5.75,
    "learning_rate": 1.6310124450164488e-06,
    "epoch": 41.70854271356784,
    "step": 8300
  },
  {
    "loss": 0.1544,
    "grad_norm": 11.25,
    "learning_rate": 1.5362928442394055e-06,
    "epoch": 41.959798994974875,
    "step": 8350
  },
  {
    "eval_loss": 0.34329381585121155,
    "eval_accuracy": 0.8827238335435057,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8628805718895952,
    "eval_kappa": 0.8480078975136847,
    "eval_runtime": 1.344,
    "eval_samples_per_second": 590.03,
    "eval_steps_per_second": 18.601,
    "epoch": 42.0,
    "step": 8358
  },
  {
    "loss": 0.1794,
    "grad_norm": 14.5,
    "learning_rate": 1.4441773459143405e-06,
    "epoch": 42.21105527638191,
    "step": 8400
  },
  {
    "loss": 0.146,
    "grad_norm": 7.28125,
    "learning_rate": 1.3546942920184303e-06,
    "epoch": 42.462311557788944,
    "step": 8450
  },
  {
    "loss": 0.1634,
    "grad_norm": 15.875,
    "learning_rate": 1.267871214581764e-06,
    "epoch": 42.71356783919598,
    "step": 8500
  },
  {
    "loss": 0.1804,
    "grad_norm": 10.25,
    "learning_rate": 1.1837348272163208e-06,
    "epoch": 42.96482412060301,
    "step": 8550
  },
  {
    "eval_loss": 0.3423879146575928,
    "eval_accuracy": 0.8827238335435057,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8628805718895952,
    "eval_kappa": 0.8480078975136847,
    "eval_runtime": 1.341,
    "eval_samples_per_second": 591.35,
    "eval_steps_per_second": 18.643,
    "epoch": 43.0,
    "step": 8557
  },
  {
    "loss": 0.162,
    "grad_norm": 18.375,
    "learning_rate": 1.1023110168967554e-06,
    "epoch": 43.21608040201005,
    "step": 8600
  },
  {
    "loss": 0.1698,
    "grad_norm": 106.0,
    "learning_rate": 1.0236248359955293e-06,
    "epoch": 43.46733668341709,
    "step": 8650
  },
  {
    "loss": 0.1629,
    "grad_norm": 16.875,
    "learning_rate": 9.477004945748414e-07,
    "epoch": 43.71859296482412,
    "step": 8700
  },
  {
    "loss": 0.1766,
    "grad_norm": 7.0,
    "learning_rate": 8.745613529376851e-07,
    "epoch": 43.969849246231156,
    "step": 8750
  },
  {
    "eval_loss": 0.34215831756591797,
    "eval_accuracy": 0.8827238335435057,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8628805718895952,
    "eval_kappa": 0.8480078975136847,
    "eval_runtime": 1.36,
    "eval_samples_per_second": 583.089,
    "eval_steps_per_second": 18.382,
    "epoch": 44.0,
    "step": 8756
  },
  {
    "loss": 0.158,
    "grad_norm": 6.1875,
    "learning_rate": 8.042299144404031e-07,
    "epoch": 44.221105527638194,
    "step": 8800
  },
  {
    "loss": 0.1792,
    "grad_norm": 3.859375,
    "learning_rate": 7.367278185688697e-07,
    "epoch": 44.472361809045225,
    "step": 8850
  },
  {
    "loss": 0.1887,
    "grad_norm": 5.71875,
    "learning_rate": 6.720758342804834e-07,
    "epoch": 44.72361809045226,
    "step": 8900
  },
  {
    "loss": 0.1489,
    "grad_norm": 21.25,
    "learning_rate": 6.10293853614008e-07,
    "epoch": 44.97487437185929,
    "step": 8950
  },
  {
    "eval_loss": 0.3427422344684601,
    "eval_accuracy": 0.8827238335435057,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8628805718895952,
    "eval_kappa": 0.8480078975136847,
    "eval_runtime": 1.31,
    "eval_samples_per_second": 605.344,
    "eval_steps_per_second": 19.084,
    "epoch": 45.0,
    "step": 8955
  },
  {
    "loss": 0.1917,
    "grad_norm": 8.5625,
    "learning_rate": 5.514008855692087e-07,
    "epoch": 45.22613065326633,
    "step": 9000
  },
  {
    "loss": 0.1846,
    "grad_norm": 4.875,
    "learning_rate": 4.954150502581923e-07,
    "epoch": 45.47738693467337,
    "step": 9050
  },
  {
    "loss": 0.1375,
    "grad_norm": 7.34375,
    "learning_rate": 4.4235357333022864e-07,
    "epoch": 45.7286432160804,
    "step": 9100
  },
  {
    "loss": 0.1672,
    "grad_norm": 11.6875,
    "learning_rate": 3.9223278067179404e-07,
    "epoch": 45.97989949748744,
    "step": 9150
  },
  {
    "eval_loss": 0.34274017810821533,
    "eval_accuracy": 0.8827238335435057,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8628805718895952,
    "eval_kappa": 0.8480078975136847,
    "eval_runtime": 1.335,
    "eval_samples_per_second": 594.008,
    "eval_steps_per_second": 18.727,
    "epoch": 46.0,
    "step": 9154
  },
  {
    "loss": 0.1705,
    "grad_norm": 3.21875,
    "learning_rate": 3.4506809338343495e-07,
    "epoch": 46.231155778894475,
    "step": 9200
  },
  {
    "loss": 0.1555,
    "grad_norm": 16.25,
    "learning_rate": 3.008740230350293e-07,
    "epoch": 46.482412060301506,
    "step": 9250
  },
  {
    "loss": 0.1823,
    "grad_norm": 14.375,
    "learning_rate": 2.596641672008837e-07,
    "epoch": 46.733668341708544,
    "step": 9300
  },
  {
    "loss": 0.1669,
    "grad_norm": 10.0,
    "learning_rate": 2.214512052760487e-07,
    "epoch": 46.984924623115575,
    "step": 9350
  },
  {
    "eval_loss": 0.3425142168998718,
    "eval_accuracy": 0.8827238335435057,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8628805718895952,
    "eval_kappa": 0.8480078975136847,
    "eval_runtime": 1.372,
    "eval_samples_per_second": 577.988,
    "eval_steps_per_second": 18.222,
    "epoch": 47.0,
    "step": 9353
  },
  {
    "loss": 0.1545,
    "grad_norm": 13.375,
    "learning_rate": 1.8624689457513945e-07,
    "epoch": 47.23618090452261,
    "step": 9400
  },
  {
    "loss": 0.1768,
    "grad_norm": 11.125,
    "learning_rate": 1.54062066714864e-07,
    "epoch": 47.48743718592965,
    "step": 9450
  },
  {
    "loss": 0.1526,
    "grad_norm": 3.765625,
    "learning_rate": 1.2490662428135569e-07,
    "epoch": 47.73869346733668,
    "step": 9500
  },
  {
    "loss": 0.179,
    "grad_norm": 20.0,
    "learning_rate": 9.87895377833692e-08,
    "epoch": 47.98994974874372,
    "step": 9550
  },
  {
    "eval_loss": 0.3430207669734955,
    "eval_accuracy": 0.8827238335435057,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8628805718895952,
    "eval_kappa": 0.8480078975136847,
    "eval_runtime": 1.3929,
    "eval_samples_per_second": 569.323,
    "eval_steps_per_second": 17.948,
    "epoch": 48.0,
    "step": 9552
  },
  {
    "loss": 0.1604,
    "grad_norm": 10.875,
    "learning_rate": 7.571884289224174e-08,
    "epoch": 48.24120603015076,
    "step": 9600
  },
  {
    "loss": 0.1837,
    "grad_norm": 33.0,
    "learning_rate": 5.5701637969486355e-08,
    "epoch": 48.49246231155779,
    "step": 9650
  },
  {
    "loss": 0.1591,
    "grad_norm": 5.5625,
    "learning_rate": 3.87440818827789e-08,
    "epoch": 48.743718592964825,
    "step": 9700
  },
  {
    "loss": 0.178,
    "grad_norm": 10.5,
    "learning_rate": 2.4851392111004956e-08,
    "epoch": 48.994974874371856,
    "step": 9750
  },
  {
    "eval_loss": 0.34273195266723633,
    "eval_accuracy": 0.8839848675914249,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8640505019325028,
    "eval_kappa": 0.8496288952600413,
    "eval_runtime": 1.438,
    "eval_samples_per_second": 551.459,
    "eval_steps_per_second": 17.385,
    "epoch": 49.0,
    "step": 9751
  },
  {
    "loss": 0.1834,
    "grad_norm": 11.5,
    "learning_rate": 1.4027843138954001e-08,
    "epoch": 49.246231155778894,
    "step": 9800
  },
  {
    "loss": 0.18,
    "grad_norm": 14.6875,
    "learning_rate": 6.276765142152475e-09,
    "epoch": 49.49748743718593,
    "step": 9850
  },
  {
    "loss": 0.1612,
    "grad_norm": 10.125,
    "learning_rate": 1.6005429622334423e-09,
    "epoch": 49.74874371859296,
    "step": 9900
  },
  {
    "loss": 0.1675,
    "grad_norm": 24.5,
    "learning_rate": 6.153731813007824e-13,
    "epoch": 50.0,
    "step": 9950
  },
  {
    "eval_loss": 0.34301358461380005,
    "eval_accuracy": 0.8839848675914249,
    "eval_top3": 0.9899117276166457,
    "eval_macro_f1": 0.8640505019325028,
    "eval_kappa": 0.8496288952600413,
    "eval_runtime": 1.334,
    "eval_samples_per_second": 594.454,
    "eval_steps_per_second": 18.741,
    "epoch": 50.0,
    "step": 9950
  },
  {
    "train_runtime": 1027.9074,
    "train_samples_per_second": 154.197,
    "train_steps_per_second": 9.68,
    "total_flos": 1.2418210759836672e+19,
    "train_loss": 0.335881779062089,
    "epoch": 50.0,
    "step": 9950
  }
]