{
  "best_metric": 0.6471614548382768,
  "best_model_checkpoint": "/tmp/test-ner1_/checkpoint-50648",
  "epoch": 120.0,
  "global_step": 58440,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7624102478471411,
      "eval_f1": 0.47653262255186,
      "eval_loss": 0.7562578320503235,
      "eval_precision": 0.40737071527640184,
      "eval_recall": 0.5739810161920714,
      "eval_runtime": 2.7545,
      "eval_samples_per_second": 403.708,
      "eval_steps_per_second": 25.413,
      "step": 487
    },
    {
      "epoch": 1.03,
      "learning_rate": 2.9743326488706367e-05,
      "loss": 1.1928,
      "step": 500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7788459244769924,
      "eval_f1": 0.5465686274509804,
      "eval_loss": 0.6944684386253357,
      "eval_precision": 0.4871122761031018,
      "eval_recall": 0.6225572305974316,
      "eval_runtime": 2.779,
      "eval_samples_per_second": 400.142,
      "eval_steps_per_second": 25.189,
      "step": 974
    },
    {
      "epoch": 2.05,
      "learning_rate": 2.9486652977412733e-05,
      "loss": 0.6432,
      "step": 1000
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7798001001884497,
      "eval_f1": 0.5616201531242281,
      "eval_loss": 0.7379429936408997,
      "eval_precision": 0.5035429583702391,
      "eval_recall": 0.6348408710217756,
      "eval_runtime": 2.7688,
      "eval_samples_per_second": 401.624,
      "eval_steps_per_second": 25.282,
      "step": 1461
    },
    {
      "epoch": 3.08,
      "learning_rate": 2.9229979466119096e-05,
      "loss": 0.4783,
      "step": 1500
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7848572314591732,
      "eval_f1": 0.56262278978389,
      "eval_loss": 0.7334258556365967,
      "eval_precision": 0.502192021043402,
      "eval_recall": 0.6395868230039085,
      "eval_runtime": 2.7553,
      "eval_samples_per_second": 403.589,
      "eval_steps_per_second": 25.406,
      "step": 1948
    },
    {
      "epoch": 4.11,
      "learning_rate": 2.8973305954825462e-05,
      "loss": 0.3609,
      "step": 2000
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.7818754323608692,
      "eval_f1": 0.5866432777847388,
      "eval_loss": 0.8052919507026672,
      "eval_precision": 0.5321664014548761,
      "eval_recall": 0.6535455053042992,
      "eval_runtime": 2.7356,
      "eval_samples_per_second": 406.499,
      "eval_steps_per_second": 25.589,
      "step": 2435
    },
    {
      "epoch": 5.13,
      "learning_rate": 2.8716632443531828e-05,
      "loss": 0.2735,
      "step": 2500
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.7850242122086782,
      "eval_f1": 0.5788942186316586,
      "eval_loss": 0.8289278745651245,
      "eval_precision": 0.5283410138248847,
      "eval_recall": 0.6401451702959241,
      "eval_runtime": 2.794,
      "eval_samples_per_second": 397.99,
      "eval_steps_per_second": 25.053,
      "step": 2922
    },
    {
      "epoch": 6.16,
      "learning_rate": 2.8459958932238194e-05,
      "loss": 0.2243,
      "step": 3000
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.7811836549700627,
      "eval_f1": 0.5892489711934157,
      "eval_loss": 0.9322538375854492,
      "eval_precision": 0.5462565569861707,
      "eval_recall": 0.6395868230039085,
      "eval_runtime": 2.7662,
      "eval_samples_per_second": 401.992,
      "eval_steps_per_second": 25.305,
      "step": 3409
    },
    {
      "epoch": 7.19,
      "learning_rate": 2.8203285420944557e-05,
      "loss": 0.1753,
      "step": 3500
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.7798001001884497,
      "eval_f1": 0.5883828879688872,
      "eval_loss": 0.9913250803947449,
      "eval_precision": 0.5342902711323764,
      "eval_recall": 0.6546621998883305,
      "eval_runtime": 2.787,
      "eval_samples_per_second": 399.0,
      "eval_steps_per_second": 25.117,
      "step": 3896
    },
    {
      "epoch": 8.21,
      "learning_rate": 2.7946611909650923e-05,
      "loss": 0.147,
      "step": 4000
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.7737649388134825,
      "eval_f1": 0.5875266658300916,
      "eval_loss": 1.070313572883606,
      "eval_precision": 0.5336220651926146,
      "eval_recall": 0.6535455053042992,
      "eval_runtime": 2.7595,
      "eval_samples_per_second": 402.973,
      "eval_steps_per_second": 25.367,
      "step": 4383
    },
    {
      "epoch": 9.24,
      "learning_rate": 2.7689938398357292e-05,
      "loss": 0.1204,
      "step": 4500
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.785286610529329,
      "eval_f1": 0.5879922530664945,
      "eval_loss": 1.0797334909439087,
      "eval_precision": 0.5469613259668509,
      "eval_recall": 0.635678391959799,
      "eval_runtime": 2.8881,
      "eval_samples_per_second": 385.035,
      "eval_steps_per_second": 24.238,
      "step": 4870
    },
    {
      "epoch": 10.27,
      "learning_rate": 2.743326488706366e-05,
      "loss": 0.1046,
      "step": 5000
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.7803726056153241,
      "eval_f1": 0.5878064110622251,
      "eval_loss": 1.128252625465393,
      "eval_precision": 0.5346444088726275,
      "eval_recall": 0.6527079843662759,
      "eval_runtime": 2.7415,
      "eval_samples_per_second": 405.623,
      "eval_steps_per_second": 25.534,
      "step": 5357
    },
    {
      "epoch": 11.29,
      "learning_rate": 2.717659137577002e-05,
      "loss": 0.0839,
      "step": 5500
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.7841415996755803,
      "eval_f1": 0.5948770230661399,
      "eval_loss": 1.1799981594085693,
      "eval_precision": 0.5472450175849941,
      "eval_recall": 0.6515912897822446,
      "eval_runtime": 2.7908,
      "eval_samples_per_second": 398.457,
      "eval_steps_per_second": 25.083,
      "step": 5844
    },
    {
      "epoch": 12.32,
      "learning_rate": 2.6919917864476387e-05,
      "loss": 0.0702,
      "step": 6000
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.7898427995515375,
      "eval_f1": 0.6095139873662497,
      "eval_loss": 1.2153304815292358,
      "eval_precision": 0.5662275449101797,
      "eval_recall": 0.6599664991624791,
      "eval_runtime": 2.747,
      "eval_samples_per_second": 404.804,
      "eval_steps_per_second": 25.482,
      "step": 6331
    },
    {
      "epoch": 13.35,
      "learning_rate": 2.6663244353182754e-05,
      "loss": 0.0631,
      "step": 6500
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.7823763746093844,
      "eval_f1": 0.5905834186284545,
      "eval_loss": 1.2911711931228638,
      "eval_precision": 0.5451110061407652,
      "eval_recall": 0.6443327749860414,
      "eval_runtime": 2.7671,
      "eval_samples_per_second": 401.867,
      "eval_steps_per_second": 25.297,
      "step": 6818
    },
    {
      "epoch": 14.37,
      "learning_rate": 2.640657084188912e-05,
      "loss": 0.0508,
      "step": 7000
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.7855251544571933,
      "eval_f1": 0.6055880441845354,
      "eval_loss": 1.3288202285766602,
      "eval_precision": 0.5664964745927547,
      "eval_recall": 0.6504745951982133,
      "eval_runtime": 2.7475,
      "eval_samples_per_second": 404.729,
      "eval_steps_per_second": 25.478,
      "step": 7305
    },
    {
      "epoch": 15.4,
      "learning_rate": 2.6149897330595482e-05,
      "loss": 0.0463,
      "step": 7500
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.7857398439922711,
      "eval_f1": 0.6076379806434737,
      "eval_loss": 1.4110201597213745,
      "eval_precision": 0.5716043307086615,
      "eval_recall": 0.6485203796761586,
      "eval_runtime": 2.7575,
      "eval_samples_per_second": 403.266,
      "eval_steps_per_second": 25.385,
      "step": 7792
    },
    {
      "epoch": 16.43,
      "learning_rate": 2.589322381930185e-05,
      "loss": 0.0387,
      "step": 8000
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.7860499510984947,
      "eval_f1": 0.606709677419355,
      "eval_loss": 1.4022138118743896,
      "eval_precision": 0.5640595009596929,
      "eval_recall": 0.6563372417643775,
      "eval_runtime": 2.7523,
      "eval_samples_per_second": 404.029,
      "eval_steps_per_second": 25.433,
      "step": 8279
    },
    {
      "epoch": 17.45,
      "learning_rate": 2.5636550308008215e-05,
      "loss": 0.0352,
      "step": 8500
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.7891510221607309,
      "eval_f1": 0.6005115089514067,
      "eval_loss": 1.4306492805480957,
      "eval_precision": 0.5540349221330817,
      "eval_recall": 0.655499720826354,
      "eval_runtime": 2.7318,
      "eval_samples_per_second": 407.059,
      "eval_steps_per_second": 25.624,
      "step": 8766
    },
    {
      "epoch": 18.48,
      "learning_rate": 2.537987679671458e-05,
      "loss": 0.031,
      "step": 9000
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.7930869969704921,
      "eval_f1": 0.6042940793754067,
      "eval_loss": 1.4502052068710327,
      "eval_precision": 0.5659273702169144,
      "eval_recall": 0.6482412060301508,
      "eval_runtime": 2.7492,
      "eval_samples_per_second": 404.487,
      "eval_steps_per_second": 25.462,
      "step": 9253
    },
    {
      "epoch": 19.51,
      "learning_rate": 2.5123203285420943e-05,
      "loss": 0.0286,
      "step": 9500
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.7866940197037284,
      "eval_f1": 0.5970073548059852,
      "eval_loss": 1.5110787153244019,
      "eval_precision": 0.5469330855018587,
      "eval_recall": 0.6571747627024009,
      "eval_runtime": 2.7416,
      "eval_samples_per_second": 405.609,
      "eval_steps_per_second": 25.533,
      "step": 9740
    },
    {
      "epoch": 20.53,
      "learning_rate": 2.486652977412731e-05,
      "loss": 0.0262,
      "step": 10000
    },
    {
      "epoch": 21.0,
      "eval_accuracy": 0.7825672097516758,
      "eval_f1": 0.6085357846355877,
      "eval_loss": 1.6086113452911377,
      "eval_precision": 0.5745102901066204,
      "eval_recall": 0.6468453378001117,
      "eval_runtime": 2.7751,
      "eval_samples_per_second": 400.703,
      "eval_steps_per_second": 25.224,
      "step": 10227
    },
    {
      "epoch": 21.56,
      "learning_rate": 2.4609856262833676e-05,
      "loss": 0.0212,
      "step": 10500
    },
    {
      "epoch": 22.0,
      "eval_accuracy": 0.7844755611745903,
      "eval_f1": 0.6177887745800236,
      "eval_loss": 1.6133506298065186,
      "eval_precision": 0.5789602147913108,
      "eval_recall": 0.6621998883305416,
      "eval_runtime": 2.7516,
      "eval_samples_per_second": 404.131,
      "eval_steps_per_second": 25.44,
      "step": 10714
    },
    {
      "epoch": 22.59,
      "learning_rate": 2.4353182751540042e-05,
      "loss": 0.0188,
      "step": 11000
    },
    {
      "epoch": 23.0,
      "eval_accuracy": 0.7830681520001909,
      "eval_f1": 0.6066104990278678,
      "eval_loss": 1.6032028198242188,
      "eval_precision": 0.5661746915073796,
      "eval_recall": 0.6532663316582915,
      "eval_runtime": 2.7575,
      "eval_samples_per_second": 403.261,
      "eval_steps_per_second": 25.385,
      "step": 11201
    },
    {
      "epoch": 23.61,
      "learning_rate": 2.4096509240246405e-05,
      "loss": 0.0179,
      "step": 11500
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 0.7821139762887336,
      "eval_f1": 0.6070782743477138,
      "eval_loss": 1.6523734331130981,
      "eval_precision": 0.5649038461538461,
      "eval_recall": 0.6560580681183696,
      "eval_runtime": 2.7371,
      "eval_samples_per_second": 406.274,
      "eval_steps_per_second": 25.575,
      "step": 11688
    },
    {
      "epoch": 24.64,
      "learning_rate": 2.383983572895277e-05,
      "loss": 0.0156,
      "step": 12000
    },
    {
      "epoch": 25.0,
      "eval_accuracy": 0.783187423964123,
      "eval_f1": 0.6089974293059126,
      "eval_loss": 1.6467918157577515,
      "eval_precision": 0.5643163411148165,
      "eval_recall": 0.6613623673925182,
      "eval_runtime": 2.8919,
      "eval_samples_per_second": 384.523,
      "eval_steps_per_second": 24.206,
      "step": 12175
    },
    {
      "epoch": 25.67,
      "learning_rate": 2.3583162217659137e-05,
      "loss": 0.0165,
      "step": 12500
    },
    {
      "epoch": 26.0,
      "eval_accuracy": 0.7857398439922711,
      "eval_f1": 0.6114599686028258,
      "eval_loss": 1.6683323383331299,
      "eval_precision": 0.5753323485967504,
      "eval_recall": 0.652428810720268,
      "eval_runtime": 2.7393,
      "eval_samples_per_second": 405.942,
      "eval_steps_per_second": 25.554,
      "step": 12662
    },
    {
      "epoch": 26.69,
      "learning_rate": 2.3326488706365506e-05,
      "loss": 0.0141,
      "step": 13000
    },
    {
      "epoch": 27.0,
      "eval_accuracy": 0.7868371460604471,
      "eval_f1": 0.6095836606441478,
      "eval_loss": 1.688219428062439,
      "eval_precision": 0.5739644970414202,
      "eval_recall": 0.6499162479061976,
      "eval_runtime": 2.7856,
      "eval_samples_per_second": 399.201,
      "eval_steps_per_second": 25.13,
      "step": 13149
    },
    {
      "epoch": 27.72,
      "learning_rate": 2.306981519507187e-05,
      "loss": 0.013,
      "step": 13500
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.7891748765535174,
      "eval_f1": 0.6083289404322615,
      "eval_loss": 1.7041257619857788,
      "eval_precision": 0.5761357963055417,
      "eval_recall": 0.6443327749860414,
      "eval_runtime": 2.7487,
      "eval_samples_per_second": 404.554,
      "eval_steps_per_second": 25.467,
      "step": 13636
    },
    {
      "epoch": 28.75,
      "learning_rate": 2.2813141683778235e-05,
      "loss": 0.0125,
      "step": 14000
    },
    {
      "epoch": 29.0,
      "eval_accuracy": 0.7868371460604471,
      "eval_f1": 0.615344262295082,
      "eval_loss": 1.7665446996688843,
      "eval_precision": 0.5802621815483552,
      "eval_recall": 0.6549413735343383,
      "eval_runtime": 2.7723,
      "eval_samples_per_second": 401.108,
      "eval_steps_per_second": 25.25,
      "step": 14123
    },
    {
      "epoch": 29.77,
      "learning_rate": 2.25564681724846e-05,
      "loss": 0.0113,
      "step": 14500
    },
    {
      "epoch": 30.0,
      "eval_accuracy": 0.7934925216478614,
      "eval_f1": 0.6155059132720105,
      "eval_loss": 1.723854899406433,
      "eval_precision": 0.5814299900695135,
      "eval_recall": 0.653824678950307,
      "eval_runtime": 2.7827,
      "eval_samples_per_second": 399.605,
      "eval_steps_per_second": 25.155,
      "step": 14610
    },
    {
      "epoch": 30.8,
      "learning_rate": 2.2299794661190967e-05,
      "loss": 0.0115,
      "step": 15000
    },
    {
      "epoch": 31.0,
      "eval_accuracy": 0.7847618138880275,
      "eval_f1": 0.6101120667187907,
      "eval_loss": 1.8082807064056396,
      "eval_precision": 0.5720918866080157,
      "eval_recall": 0.6535455053042992,
      "eval_runtime": 2.7565,
      "eval_samples_per_second": 403.407,
      "eval_steps_per_second": 25.394,
      "step": 15097
    },
    {
      "epoch": 31.83,
      "learning_rate": 2.204312114989733e-05,
      "loss": 0.0121,
      "step": 15500
    },
    {
      "epoch": 32.0,
      "eval_accuracy": 0.7925144915436177,
      "eval_f1": 0.6105967078189302,
      "eval_loss": 1.7592214345932007,
      "eval_precision": 0.5660467334287077,
      "eval_recall": 0.6627582356225572,
      "eval_runtime": 2.7597,
      "eval_samples_per_second": 402.936,
      "eval_steps_per_second": 25.365,
      "step": 15584
    },
    {
      "epoch": 32.85,
      "learning_rate": 2.1786447638603696e-05,
      "loss": 0.0105,
      "step": 16000
    },
    {
      "epoch": 33.0,
      "eval_accuracy": 0.7882207008420601,
      "eval_f1": 0.6161497186232169,
      "eval_loss": 1.780286431312561,
      "eval_precision": 0.5799457994579946,
      "eval_recall": 0.6571747627024009,
      "eval_runtime": 2.7565,
      "eval_samples_per_second": 403.416,
      "eval_steps_per_second": 25.395,
      "step": 16071
    },
    {
      "epoch": 33.88,
      "learning_rate": 2.1529774127310062e-05,
      "loss": 0.0089,
      "step": 16500
    },
    {
      "epoch": 34.0,
      "eval_accuracy": 0.7870995443810977,
      "eval_f1": 0.6128184922511163,
      "eval_loss": 1.8191593885421753,
      "eval_precision": 0.5786210317460317,
      "eval_recall": 0.6513121161362367,
      "eval_runtime": 2.745,
      "eval_samples_per_second": 405.095,
      "eval_steps_per_second": 25.501,
      "step": 16558
    },
    {
      "epoch": 34.91,
      "learning_rate": 2.127310061601643e-05,
      "loss": 0.0107,
      "step": 17000
    },
    {
      "epoch": 35.0,
      "eval_accuracy": 0.7860260967057083,
      "eval_f1": 0.6097277770610243,
      "eval_loss": 1.8328877687454224,
      "eval_precision": 0.5668025905492924,
      "eval_recall": 0.6596873255164712,
      "eval_runtime": 2.7678,
      "eval_samples_per_second": 401.766,
      "eval_steps_per_second": 25.291,
      "step": 17045
    },
    {
      "epoch": 35.93,
      "learning_rate": 2.1016427104722795e-05,
      "loss": 0.011,
      "step": 17500
    },
    {
      "epoch": 36.0,
      "eval_accuracy": 0.7834021134992009,
      "eval_f1": 0.6102003642987249,
      "eval_loss": 1.8010284900665283,
      "eval_precision": 0.5713937621832359,
      "eval_recall": 0.6546621998883305,
      "eval_runtime": 2.7737,
      "eval_samples_per_second": 400.906,
      "eval_steps_per_second": 25.237,
      "step": 17532
    },
    {
      "epoch": 36.96,
      "learning_rate": 2.0759753593429157e-05,
      "loss": 0.0087,
      "step": 18000
    },
    {
      "epoch": 37.0,
      "eval_accuracy": 0.789818945158751,
      "eval_f1": 0.6208449212024898,
      "eval_loss": 1.8313595056533813,
      "eval_precision": 0.5905769715293525,
      "eval_recall": 0.6543830262423227,
      "eval_runtime": 2.7538,
      "eval_samples_per_second": 403.804,
      "eval_steps_per_second": 25.419,
      "step": 18019
    },
    {
      "epoch": 37.99,
      "learning_rate": 2.0503080082135523e-05,
      "loss": 0.0075,
      "step": 18500
    },
    {
      "epoch": 38.0,
      "eval_accuracy": 0.7912740631187233,
      "eval_f1": 0.6227038456455664,
      "eval_loss": 1.8428292274475098,
      "eval_precision": 0.5912170639899623,
      "eval_recall": 0.6577331099944165,
      "eval_runtime": 2.7724,
      "eval_samples_per_second": 401.093,
      "eval_steps_per_second": 25.249,
      "step": 18506
    },
    {
      "epoch": 39.0,
      "eval_accuracy": 0.7892702941246631,
      "eval_f1": 0.6217024041585445,
      "eval_loss": 1.8756595849990845,
      "eval_precision": 0.5815706297106735,
      "eval_recall": 0.6677833612506979,
      "eval_runtime": 2.7463,
      "eval_samples_per_second": 404.906,
      "eval_steps_per_second": 25.489,
      "step": 18993
    },
    {
      "epoch": 39.01,
      "learning_rate": 2.024640657084189e-05,
      "loss": 0.0079,
      "step": 19000
    },
    {
      "epoch": 40.0,
      "eval_accuracy": 0.7897235275876052,
      "eval_f1": 0.6222632550778159,
      "eval_loss": 1.8513692617416382,
      "eval_precision": 0.58975,
      "eval_recall": 0.65857063093244,
      "eval_runtime": 2.8669,
      "eval_samples_per_second": 387.875,
      "eval_steps_per_second": 24.417,
      "step": 19480
    },
    {
      "epoch": 40.04,
      "learning_rate": 1.9989733059548256e-05,
      "loss": 0.0086,
      "step": 19500
    },
    {
      "epoch": 41.0,
      "eval_accuracy": 0.7896758188020324,
      "eval_f1": 0.6242471851269966,
      "eval_loss": 1.8782603740692139,
      "eval_precision": 0.5877712031558185,
      "eval_recall": 0.6655499720826354,
      "eval_runtime": 2.7468,
      "eval_samples_per_second": 404.838,
      "eval_steps_per_second": 25.484,
      "step": 19967
    },
    {
      "epoch": 41.07,
      "learning_rate": 1.973305954825462e-05,
      "loss": 0.0075,
      "step": 20000
    },
    {
      "epoch": 42.0,
      "eval_accuracy": 0.7950669115717659,
      "eval_f1": 0.6231997905210788,
      "eval_loss": 1.8176987171173096,
      "eval_precision": 0.5867850098619329,
      "eval_recall": 0.6644332774986041,
      "eval_runtime": 2.7491,
      "eval_samples_per_second": 404.496,
      "eval_steps_per_second": 25.463,
      "step": 20454
    },
    {
      "epoch": 42.09,
      "learning_rate": 1.9476386036960984e-05,
      "loss": 0.0071,
      "step": 20500
    },
    {
      "epoch": 43.0,
      "eval_accuracy": 0.794017318289163,
      "eval_f1": 0.6329214826624153,
      "eval_loss": 1.88504159450531,
      "eval_precision": 0.6038022813688213,
      "eval_recall": 0.6649916247906198,
      "eval_runtime": 2.7727,
      "eval_samples_per_second": 401.056,
      "eval_steps_per_second": 25.246,
      "step": 20941
    },
    {
      "epoch": 43.12,
      "learning_rate": 1.921971252566735e-05,
      "loss": 0.0068,
      "step": 21000
    },
    {
      "epoch": 44.0,
      "eval_accuracy": 0.7917511509744519,
      "eval_f1": 0.6311334479566195,
      "eval_loss": 1.9209691286087036,
      "eval_precision": 0.5996481528022116,
      "eval_recall": 0.666108319374651,
      "eval_runtime": 2.8074,
      "eval_samples_per_second": 396.092,
      "eval_steps_per_second": 24.934,
      "step": 21428
    },
    {
      "epoch": 44.15,
      "learning_rate": 1.896303901437372e-05,
      "loss": 0.006,
      "step": 21500
    },
    {
      "epoch": 45.0,
      "eval_accuracy": 0.7912502087259369,
      "eval_f1": 0.6239327466176278,
      "eval_loss": 1.9289366006851196,
      "eval_precision": 0.5891838253535103,
      "eval_recall": 0.663037409268565,
      "eval_runtime": 2.7508,
      "eval_samples_per_second": 404.251,
      "eval_steps_per_second": 25.447,
      "step": 21915
    },
    {
      "epoch": 45.17,
      "learning_rate": 1.8706365503080083e-05,
      "loss": 0.0077,
      "step": 22000
    },
    {
      "epoch": 46.0,
      "eval_accuracy": 0.793802628754085,
      "eval_f1": 0.6217957144735113,
      "eval_loss": 1.901074767112732,
      "eval_precision": 0.5875776397515527,
      "eval_recall": 0.6602456728084869,
      "eval_runtime": 2.7683,
      "eval_samples_per_second": 401.688,
      "eval_steps_per_second": 25.286,
      "step": 22402
    },
    {
      "epoch": 46.2,
      "learning_rate": 1.844969199178645e-05,
      "loss": 0.0047,
      "step": 22500
    },
    {
      "epoch": 47.0,
      "eval_accuracy": 0.7934448128622886,
      "eval_f1": 0.6218905472636815,
      "eval_loss": 1.909180998802185,
      "eval_precision": 0.585552268244576,
      "eval_recall": 0.663037409268565,
      "eval_runtime": 2.7768,
      "eval_samples_per_second": 400.465,
      "eval_steps_per_second": 25.209,
      "step": 22889
    },
    {
      "epoch": 47.23,
      "learning_rate": 1.8193018480492815e-05,
      "loss": 0.0073,
      "step": 23000
    },
    {
      "epoch": 48.0,
      "eval_accuracy": 0.7884592447699244,
      "eval_f1": 0.6239832065074784,
      "eval_loss": 1.9654145240783691,
      "eval_precision": 0.5886138613861386,
      "eval_recall": 0.6638749302065885,
      "eval_runtime": 2.7606,
      "eval_samples_per_second": 402.811,
      "eval_steps_per_second": 25.357,
      "step": 23376
    },
    {
      "epoch": 48.25,
      "learning_rate": 1.793634496919918e-05,
      "loss": 0.0058,
      "step": 23500
    },
    {
      "epoch": 49.0,
      "eval_accuracy": 0.7884115359843515,
      "eval_f1": 0.6195935383011986,
      "eval_loss": 1.948256254196167,
      "eval_precision": 0.5808500244259892,
      "eval_recall": 0.6638749302065885,
      "eval_runtime": 2.7711,
      "eval_samples_per_second": 401.285,
      "eval_steps_per_second": 25.261,
      "step": 23863
    },
    {
      "epoch": 49.28,
      "learning_rate": 1.7679671457905544e-05,
      "loss": 0.0081,
      "step": 24000
    },
    {
      "epoch": 50.0,
      "eval_accuracy": 0.7899382171226832,
      "eval_f1": 0.626782145236509,
      "eval_loss": 1.943403959274292,
      "eval_precision": 0.599541167473872,
      "eval_recall": 0.6566164154103853,
      "eval_runtime": 2.7621,
      "eval_samples_per_second": 402.586,
      "eval_steps_per_second": 25.343,
      "step": 24350
    },
    {
      "epoch": 50.31,
      "learning_rate": 1.742299794661191e-05,
      "loss": 0.0063,
      "step": 24500
    },
    {
      "epoch": 51.0,
      "eval_accuracy": 0.7902960330144796,
      "eval_f1": 0.626861737182022,
      "eval_loss": 1.9489706754684448,
      "eval_precision": 0.5937578027465668,
      "eval_recall": 0.6638749302065885,
      "eval_runtime": 2.7691,
      "eval_samples_per_second": 401.574,
      "eval_steps_per_second": 25.279,
      "step": 24837
    },
    {
      "epoch": 51.33,
      "learning_rate": 1.7166324435318276e-05,
      "loss": 0.0052,
      "step": 25000
    },
    {
      "epoch": 52.0,
      "eval_accuracy": 0.78788673934305,
      "eval_f1": 0.6303209346045388,
      "eval_loss": 1.9653586149215698,
      "eval_precision": 0.6072445019404916,
      "eval_recall": 0.6552205471803462,
      "eval_runtime": 2.7625,
      "eval_samples_per_second": 402.535,
      "eval_steps_per_second": 25.339,
      "step": 25324
    },
    {
      "epoch": 52.36,
      "learning_rate": 1.6909650924024642e-05,
      "loss": 0.007,
      "step": 25500
    },
    {
      "epoch": 53.0,
      "eval_accuracy": 0.7880298656997686,
      "eval_f1": 0.6263430163151613,
      "eval_loss": 1.9698514938354492,
      "eval_precision": 0.5966641394996209,
      "eval_recall": 0.6591289782244556,
      "eval_runtime": 2.7557,
      "eval_samples_per_second": 403.531,
      "eval_steps_per_second": 25.402,
      "step": 25811
    },
    {
      "epoch": 53.39,
      "learning_rate": 1.6652977412731005e-05,
      "loss": 0.0047,
      "step": 26000
    },
    {
      "epoch": 54.0,
      "eval_accuracy": 0.7909401016197133,
      "eval_f1": 0.6273834745762713,
      "eval_loss": 1.971285343170166,
      "eval_precision": 0.5967254408060454,
      "eval_recall": 0.6613623673925182,
      "eval_runtime": 2.7453,
      "eval_samples_per_second": 405.054,
      "eval_steps_per_second": 25.498,
      "step": 26298
    },
    {
      "epoch": 54.41,
      "learning_rate": 1.639630390143737e-05,
      "loss": 0.0041,
      "step": 26500
    },
    {
      "epoch": 55.0,
      "eval_accuracy": 0.7894611292669546,
      "eval_f1": 0.6249177739771082,
      "eval_loss": 1.9533612728118896,
      "eval_precision": 0.5909430206519034,
      "eval_recall": 0.663037409268565,
      "eval_runtime": 2.8696,
      "eval_samples_per_second": 387.509,
      "eval_steps_per_second": 24.394,
      "step": 26785
    },
    {
      "epoch": 55.44,
      "learning_rate": 1.6139630390143737e-05,
      "loss": 0.0042,
      "step": 27000
    },
    {
      "epoch": 56.0,
      "eval_accuracy": 0.7940650270747358,
      "eval_f1": 0.631480989098644,
      "eval_loss": 1.9982216358184814,
      "eval_precision": 0.6027918781725888,
      "eval_recall": 0.663037409268565,
      "eval_runtime": 2.7416,
      "eval_samples_per_second": 405.607,
      "eval_steps_per_second": 25.533,
      "step": 27272
    },
    {
      "epoch": 56.47,
      "learning_rate": 1.5882956878850103e-05,
      "loss": 0.0045,
      "step": 27500
    },
    {
      "epoch": 57.0,
      "eval_accuracy": 0.792085112473462,
      "eval_f1": 0.6291772916387062,
      "eval_loss": 1.996762990951538,
      "eval_precision": 0.6058413026621866,
      "eval_recall": 0.6543830262423227,
      "eval_runtime": 2.7399,
      "eval_samples_per_second": 405.854,
      "eval_steps_per_second": 25.548,
      "step": 27759
    },
    {
      "epoch": 57.49,
      "learning_rate": 1.5626283367556466e-05,
      "loss": 0.0045,
      "step": 28000
    },
    {
      "epoch": 58.0,
      "eval_accuracy": 0.7904630137639846,
      "eval_f1": 0.6297929191716767,
      "eval_loss": 1.985123634338379,
      "eval_precision": 0.603894440174225,
      "eval_recall": 0.6580122836404243,
      "eval_runtime": 2.7713,
      "eval_samples_per_second": 401.262,
      "eval_steps_per_second": 25.259,
      "step": 28246
    },
    {
      "epoch": 58.52,
      "learning_rate": 1.5369609856262832e-05,
      "loss": 0.0039,
      "step": 28500
    },
    {
      "epoch": 59.0,
      "eval_accuracy": 0.789103313375158,
      "eval_f1": 0.6346205059920107,
      "eval_loss": 2.043133497238159,
      "eval_precision": 0.6066700610997964,
      "eval_recall": 0.6652707984366276,
      "eval_runtime": 2.7416,
      "eval_samples_per_second": 405.599,
      "eval_steps_per_second": 25.532,
      "step": 28733
    },
    {
      "epoch": 59.55,
      "learning_rate": 1.5112936344969198e-05,
      "loss": 0.0048,
      "step": 29000
    },
    {
      "epoch": 60.0,
      "eval_accuracy": 0.7877913217719043,
      "eval_f1": 0.6211777273334224,
      "eval_loss": 2.0035526752471924,
      "eval_precision": 0.5953416943946762,
      "eval_recall": 0.649357900614182,
      "eval_runtime": 2.7429,
      "eval_samples_per_second": 405.406,
      "eval_steps_per_second": 25.52,
      "step": 29220
    },
    {
      "epoch": 60.57,
      "learning_rate": 1.4856262833675564e-05,
      "loss": 0.004,
      "step": 29500
    },
    {
      "epoch": 61.0,
      "eval_accuracy": 0.791369480689869,
      "eval_f1": 0.6329315141078288,
      "eval_loss": 1.9970779418945312,
      "eval_precision": 0.6022183009831107,
      "eval_recall": 0.6669458403126745,
      "eval_runtime": 2.7486,
      "eval_samples_per_second": 404.568,
      "eval_steps_per_second": 25.467,
      "step": 29707
    },
    {
      "epoch": 61.6,
      "learning_rate": 1.459958932238193e-05,
      "loss": 0.0032,
      "step": 30000
    },
    {
      "epoch": 62.0,
      "eval_accuracy": 0.7912024999403641,
      "eval_f1": 0.6301771207883874,
      "eval_loss": 2.007345199584961,
      "eval_precision": 0.6024955436720143,
      "eval_recall": 0.6605248464544947,
      "eval_runtime": 2.7495,
      "eval_samples_per_second": 404.438,
      "eval_steps_per_second": 25.459,
      "step": 30194
    },
    {
      "epoch": 62.63,
      "learning_rate": 1.4342915811088295e-05,
      "loss": 0.0033,
      "step": 30500
    },
    {
      "epoch": 63.0,
      "eval_accuracy": 0.7917511509744519,
      "eval_f1": 0.626853813559322,
      "eval_loss": 2.0134215354919434,
      "eval_precision": 0.5962216624685138,
      "eval_recall": 0.6608040201005025,
      "eval_runtime": 2.7805,
      "eval_samples_per_second": 399.935,
      "eval_steps_per_second": 25.176,
      "step": 30681
    },
    {
      "epoch": 63.66,
      "learning_rate": 1.4086242299794661e-05,
      "loss": 0.0035,
      "step": 31000
    },
    {
      "epoch": 64.0,
      "eval_accuracy": 0.7937072111829393,
      "eval_f1": 0.628412403922608,
      "eval_loss": 2.0015182495117188,
      "eval_precision": 0.5981331987891019,
      "eval_recall": 0.6619207146845337,
      "eval_runtime": 2.7746,
      "eval_samples_per_second": 400.774,
      "eval_steps_per_second": 25.229,
      "step": 31168
    },
    {
      "epoch": 64.68,
      "learning_rate": 1.3829568788501027e-05,
      "loss": 0.0032,
      "step": 31500
    },
    {
      "epoch": 65.0,
      "eval_accuracy": 0.7939934638963765,
      "eval_f1": 0.6255252100840337,
      "eval_loss": 1.9973812103271484,
      "eval_precision": 0.5904809122459098,
      "eval_recall": 0.6649916247906198,
      "eval_runtime": 2.7988,
      "eval_samples_per_second": 397.311,
      "eval_steps_per_second": 25.011,
      "step": 31655
    },
    {
      "epoch": 65.71,
      "learning_rate": 1.3572895277207393e-05,
      "loss": 0.0036,
      "step": 32000
    },
    {
      "epoch": 66.0,
      "eval_accuracy": 0.7891987309463038,
      "eval_f1": 0.6282034432908398,
      "eval_loss": 2.052255392074585,
      "eval_precision": 0.5934939160665508,
      "eval_recall": 0.6672250139586823,
      "eval_runtime": 2.7831,
      "eval_samples_per_second": 399.549,
      "eval_steps_per_second": 25.151,
      "step": 32142
    },
    {
      "epoch": 66.74,
      "learning_rate": 1.331622176591376e-05,
      "loss": 0.0027,
      "step": 32500
    },
    {
      "epoch": 67.0,
      "eval_accuracy": 0.7900813434794017,
      "eval_f1": 0.6333861595351294,
      "eval_loss": 2.068305730819702,
      "eval_precision": 0.6010025062656642,
      "eval_recall": 0.6694584031267449,
      "eval_runtime": 2.801,
      "eval_samples_per_second": 397.005,
      "eval_steps_per_second": 24.991,
      "step": 32629
    },
    {
      "epoch": 67.76,
      "learning_rate": 1.3059548254620124e-05,
      "loss": 0.0039,
      "step": 33000
    },
    {
      "epoch": 68.0,
      "eval_accuracy": 0.7876481954151857,
      "eval_f1": 0.6244558765334389,
      "eval_loss": 2.1081135272979736,
      "eval_precision": 0.5918979744936234,
      "eval_recall": 0.6608040201005025,
      "eval_runtime": 2.7627,
      "eval_samples_per_second": 402.499,
      "eval_steps_per_second": 25.337,
      "step": 33116
    },
    {
      "epoch": 68.79,
      "learning_rate": 1.280287474332649e-05,
      "loss": 0.0027,
      "step": 33500
    },
    {
      "epoch": 69.0,
      "eval_accuracy": 0.7922759476157535,
      "eval_f1": 0.6296051762841676,
      "eval_loss": 2.0554866790771484,
      "eval_precision": 0.5973440240541218,
      "eval_recall": 0.6655499720826354,
      "eval_runtime": 2.8053,
      "eval_samples_per_second": 396.39,
      "eval_steps_per_second": 24.953,
      "step": 33603
    },
    {
      "epoch": 69.82,
      "learning_rate": 1.2546201232032854e-05,
      "loss": 0.003,
      "step": 34000
    },
    {
      "epoch": 70.0,
      "eval_accuracy": 0.7879821569141957,
      "eval_f1": 0.6243246804585585,
      "eval_loss": 2.100715398788452,
      "eval_precision": 0.59121537309708,
      "eval_recall": 0.6613623673925182,
      "eval_runtime": 2.8873,
      "eval_samples_per_second": 385.136,
      "eval_steps_per_second": 24.244,
      "step": 34090
    },
    {
      "epoch": 70.84,
      "learning_rate": 1.228952772073922e-05,
      "loss": 0.0023,
      "step": 34500
    },
    {
      "epoch": 71.0,
      "eval_accuracy": 0.7936595023973665,
      "eval_f1": 0.6381622626477227,
      "eval_loss": 2.091642379760742,
      "eval_precision": 0.6085084831602937,
      "eval_recall": 0.6708542713567839,
      "eval_runtime": 2.7949,
      "eval_samples_per_second": 397.869,
      "eval_steps_per_second": 25.046,
      "step": 34577
    },
    {
      "epoch": 71.87,
      "learning_rate": 1.2032854209445585e-05,
      "loss": 0.0016,
      "step": 35000
    },
    {
      "epoch": 72.0,
      "eval_accuracy": 0.7908208296557812,
      "eval_f1": 0.6252314202591908,
      "eval_loss": 2.1563851833343506,
      "eval_precision": 0.5939698492462312,
      "eval_recall": 0.6599664991624791,
      "eval_runtime": 2.7897,
      "eval_samples_per_second": 398.603,
      "eval_steps_per_second": 25.092,
      "step": 35064
    },
    {
      "epoch": 72.9,
      "learning_rate": 1.1776180698151951e-05,
      "loss": 0.0028,
      "step": 35500
    },
    {
      "epoch": 73.0,
      "eval_accuracy": 0.7862646406335727,
      "eval_f1": 0.6271611455721261,
      "eval_loss": 2.161975383758545,
      "eval_precision": 0.5947434292866083,
      "eval_recall": 0.6633165829145728,
      "eval_runtime": 2.7929,
      "eval_samples_per_second": 398.158,
      "eval_steps_per_second": 25.064,
      "step": 35551
    },
    {
      "epoch": 73.92,
      "learning_rate": 1.1519507186858315e-05,
      "loss": 0.0028,
      "step": 36000
    },
    {
      "epoch": 74.0,
      "eval_accuracy": 0.7891510221607309,
      "eval_f1": 0.6318289786223278,
      "eval_loss": 2.1389882564544678,
      "eval_precision": 0.5990990990990991,
      "eval_recall": 0.6683417085427136,
      "eval_runtime": 2.7738,
      "eval_samples_per_second": 400.894,
      "eval_steps_per_second": 25.236,
      "step": 36038
    },
    {
      "epoch": 74.95,
      "learning_rate": 1.1262833675564683e-05,
      "loss": 0.0025,
      "step": 36500
    },
    {
      "epoch": 75.0,
      "eval_accuracy": 0.7925144915436177,
      "eval_f1": 0.6336555011253806,
      "eval_loss": 2.120399236679077,
      "eval_precision": 0.6026189876605389,
      "eval_recall": 0.6680625348967058,
      "eval_runtime": 2.7787,
      "eval_samples_per_second": 400.19,
      "eval_steps_per_second": 25.192,
      "step": 36525
    },
    {
      "epoch": 75.98,
      "learning_rate": 1.1006160164271048e-05,
      "loss": 0.0026,
      "step": 37000
    },
    {
      "epoch": 76.0,
      "eval_accuracy": 0.7884115359843515,
      "eval_f1": 0.6298019407151402,
      "eval_loss": 2.1699509620666504,
      "eval_precision": 0.6011164679015478,
      "eval_recall": 0.6613623673925182,
      "eval_runtime": 2.772,
      "eval_samples_per_second": 401.15,
      "eval_steps_per_second": 25.252,
      "step": 37012
    },
    {
      "epoch": 77.0,
      "eval_accuracy": 0.7923713651868992,
      "eval_f1": 0.6300172208239503,
      "eval_loss": 2.1477670669555664,
      "eval_precision": 0.5994454247542224,
      "eval_recall": 0.6638749302065885,
      "eval_runtime": 2.7809,
      "eval_samples_per_second": 399.866,
      "eval_steps_per_second": 25.171,
      "step": 37499
    },
    {
      "epoch": 77.0,
      "learning_rate": 1.0749486652977414e-05,
      "loss": 0.0022,
      "step": 37500
    },
    {
      "epoch": 78.0,
      "eval_accuracy": 0.7879344481286229,
      "eval_f1": 0.6282473954899117,
      "eval_loss": 2.15474009513855,
      "eval_precision": 0.5953511622094476,
      "eval_recall": 0.6649916247906198,
      "eval_runtime": 2.7719,
      "eval_samples_per_second": 401.163,
      "eval_steps_per_second": 25.253,
      "step": 37986
    },
    {
      "epoch": 78.03,
      "learning_rate": 1.0492813141683778e-05,
      "loss": 0.0026,
      "step": 38000
    },
    {
      "epoch": 79.0,
      "eval_accuracy": 0.78788673934305,
      "eval_f1": 0.6241042345276874,
      "eval_loss": 2.1489484310150146,
      "eval_precision": 0.5851453701441486,
      "eval_recall": 0.6686208821887214,
      "eval_runtime": 2.7999,
      "eval_samples_per_second": 397.156,
      "eval_steps_per_second": 25.001,
      "step": 38473
    },
    {
      "epoch": 79.06,
      "learning_rate": 1.0236139630390144e-05,
      "loss": 0.0017,
      "step": 38500
    },
    {
      "epoch": 80.0,
      "eval_accuracy": 0.7869564180243792,
      "eval_f1": 0.6278917788524376,
      "eval_loss": 2.1788620948791504,
      "eval_precision": 0.5903170312116,
      "eval_recall": 0.6705750977107761,
      "eval_runtime": 2.7799,
      "eval_samples_per_second": 400.01,
      "eval_steps_per_second": 25.18,
      "step": 38960
    },
    {
      "epoch": 80.08,
      "learning_rate": 9.979466119096509e-06,
      "loss": 0.0016,
      "step": 39000
    },
    {
      "epoch": 81.0,
      "eval_accuracy": 0.7877436129863314,
      "eval_f1": 0.6317747077577046,
      "eval_loss": 2.188213586807251,
      "eval_precision": 0.6026355803345159,
      "eval_recall": 0.6638749302065885,
      "eval_runtime": 2.8389,
      "eval_samples_per_second": 391.698,
      "eval_steps_per_second": 24.657,
      "step": 39447
    },
    {
      "epoch": 81.11,
      "learning_rate": 9.722792607802875e-06,
      "loss": 0.0014,
      "step": 39500
    },
    {
      "epoch": 82.0,
      "eval_accuracy": 0.7879821569141957,
      "eval_f1": 0.6343844834410872,
      "eval_loss": 2.1824533939361572,
      "eval_precision": 0.6014510883162372,
      "eval_recall": 0.6711334450027917,
      "eval_runtime": 2.7717,
      "eval_samples_per_second": 401.192,
      "eval_steps_per_second": 25.255,
      "step": 39934
    },
    {
      "epoch": 82.14,
      "learning_rate": 9.46611909650924e-06,
      "loss": 0.0019,
      "step": 40000
    },
    {
      "epoch": 83.0,
      "eval_accuracy": 0.7902721786216932,
      "eval_f1": 0.6320529801324504,
      "eval_loss": 2.1753077507019043,
      "eval_precision": 0.6013104838709677,
      "eval_recall": 0.666108319374651,
      "eval_runtime": 2.7935,
      "eval_samples_per_second": 398.061,
      "eval_steps_per_second": 25.058,
      "step": 40421
    },
    {
      "epoch": 83.16,
      "learning_rate": 9.209445585215607e-06,
      "loss": 0.0014,
      "step": 40500
    },
    {
      "epoch": 84.0,
      "eval_accuracy": 0.7911070823692183,
      "eval_f1": 0.6313839640116432,
      "eval_loss": 2.1886837482452393,
      "eval_precision": 0.6001006036217303,
      "eval_recall": 0.666108319374651,
      "eval_runtime": 2.7541,
      "eval_samples_per_second": 403.769,
      "eval_steps_per_second": 25.417,
      "step": 40908
    },
    {
      "epoch": 84.19,
      "learning_rate": 8.952772073921972e-06,
      "loss": 0.0011,
      "step": 41000
    },
    {
      "epoch": 85.0,
      "eval_accuracy": 0.7912979175115098,
      "eval_f1": 0.6346000531490833,
      "eval_loss": 2.1973717212677,
      "eval_precision": 0.6054766734279919,
      "eval_recall": 0.6666666666666666,
      "eval_runtime": 2.8017,
      "eval_samples_per_second": 396.905,
      "eval_steps_per_second": 24.985,
      "step": 41395
    },
    {
      "epoch": 85.22,
      "learning_rate": 8.696098562628338e-06,
      "loss": 0.0019,
      "step": 41500
    },
    {
      "epoch": 86.0,
      "eval_accuracy": 0.7913217719042962,
      "eval_f1": 0.6334745762711864,
      "eval_loss": 2.1918320655822754,
      "eval_precision": 0.6025188916876574,
      "eval_recall": 0.6677833612506979,
      "eval_runtime": 2.8875,
      "eval_samples_per_second": 385.106,
      "eval_steps_per_second": 24.242,
      "step": 41882
    },
    {
      "epoch": 86.24,
      "learning_rate": 8.439425051334702e-06,
      "loss": 0.0014,
      "step": 42000
    },
    {
      "epoch": 87.0,
      "eval_accuracy": 0.7900813434794017,
      "eval_f1": 0.6352624495289367,
      "eval_loss": 2.196187734603882,
      "eval_precision": 0.6133056133056133,
      "eval_recall": 0.6588498045784478,
      "eval_runtime": 2.7929,
      "eval_samples_per_second": 398.148,
      "eval_steps_per_second": 25.063,
      "step": 42369
    },
    {
      "epoch": 87.27,
      "learning_rate": 8.182751540041068e-06,
      "loss": 0.0019,
      "step": 42500
    },
    {
      "epoch": 88.0,
      "eval_accuracy": 0.7902006154433339,
      "eval_f1": 0.6272126816380448,
      "eval_loss": 2.197391986846924,
      "eval_precision": 0.5952858575727181,
      "eval_recall": 0.6627582356225572,
      "eval_runtime": 2.7744,
      "eval_samples_per_second": 400.806,
      "eval_steps_per_second": 25.231,
      "step": 42856
    },
    {
      "epoch": 88.3,
      "learning_rate": 7.926078028747433e-06,
      "loss": 0.0009,
      "step": 43000
    },
    {
      "epoch": 89.0,
      "eval_accuracy": 0.7918227141528112,
      "eval_f1": 0.6316624785081338,
      "eval_loss": 2.181831121444702,
      "eval_precision": 0.600150791656195,
      "eval_recall": 0.6666666666666666,
      "eval_runtime": 2.7914,
      "eval_samples_per_second": 398.372,
      "eval_steps_per_second": 25.077,
      "step": 43343
    },
    {
      "epoch": 89.32,
      "learning_rate": 7.669404517453799e-06,
      "loss": 0.0016,
      "step": 43500
    },
    {
      "epoch": 90.0,
      "eval_accuracy": 0.7944705517521051,
      "eval_f1": 0.6410461702695489,
      "eval_loss": 2.205850124359131,
      "eval_precision": 0.6140081799591002,
      "eval_recall": 0.6705750977107761,
      "eval_runtime": 2.7843,
      "eval_samples_per_second": 399.377,
      "eval_steps_per_second": 25.141,
      "step": 43830
    },
    {
      "epoch": 90.35,
      "learning_rate": 7.412731006160165e-06,
      "loss": 0.0013,
      "step": 44000
    },
    {
      "epoch": 91.0,
      "eval_accuracy": 0.7921805300446078,
      "eval_f1": 0.6387156693644687,
      "eval_loss": 2.2012627124786377,
      "eval_precision": 0.6085967130214918,
      "eval_recall": 0.6719709659408152,
      "eval_runtime": 2.7984,
      "eval_samples_per_second": 397.373,
      "eval_steps_per_second": 25.014,
      "step": 44317
    },
    {
      "epoch": 91.38,
      "learning_rate": 7.15605749486653e-06,
      "loss": 0.001,
      "step": 44500
    },
    {
      "epoch": 92.0,
      "eval_accuracy": 0.7944944061448915,
      "eval_f1": 0.6372340425531915,
      "eval_loss": 2.172316551208496,
      "eval_precision": 0.6084306754697816,
      "eval_recall": 0.6689000558347292,
      "eval_runtime": 2.7639,
      "eval_samples_per_second": 402.326,
      "eval_steps_per_second": 25.326,
      "step": 44804
    },
    {
      "epoch": 92.4,
      "learning_rate": 6.899383983572895e-06,
      "loss": 0.0012,
      "step": 45000
    },
    {
      "epoch": 93.0,
      "eval_accuracy": 0.7965935927100976,
      "eval_f1": 0.6390847412531595,
      "eval_loss": 2.1967203617095947,
      "eval_precision": 0.6104193138500635,
      "eval_recall": 0.6705750977107761,
      "eval_runtime": 2.8557,
      "eval_samples_per_second": 389.403,
      "eval_steps_per_second": 24.513,
      "step": 45291
    },
    {
      "epoch": 93.43,
      "learning_rate": 6.642710472279261e-06,
      "loss": 0.0023,
      "step": 45500
    },
    {
      "epoch": 94.0,
      "eval_accuracy": 0.7938741919324444,
      "eval_f1": 0.6414337301056573,
      "eval_loss": 2.202401876449585,
      "eval_precision": 0.6156611039794608,
      "eval_recall": 0.6694584031267449,
      "eval_runtime": 2.795,
      "eval_samples_per_second": 397.859,
      "eval_steps_per_second": 25.045,
      "step": 45778
    },
    {
      "epoch": 94.46,
      "learning_rate": 6.386036960985627e-06,
      "loss": 0.0012,
      "step": 46000
    },
    {
      "epoch": 95.0,
      "eval_accuracy": 0.7929200162209871,
      "eval_f1": 0.6406042936655181,
      "eval_loss": 2.225001096725464,
      "eval_precision": 0.6097376387487387,
      "eval_recall": 0.6747627024008933,
      "eval_runtime": 2.7893,
      "eval_samples_per_second": 398.662,
      "eval_steps_per_second": 25.096,
      "step": 46265
    },
    {
      "epoch": 95.48,
      "learning_rate": 6.129363449691992e-06,
      "loss": 0.0015,
      "step": 46500
    },
    {
      "epoch": 96.0,
      "eval_accuracy": 0.7913933350826555,
      "eval_f1": 0.6457831325301205,
      "eval_loss": 2.193800210952759,
      "eval_precision": 0.6203703703703703,
      "eval_recall": 0.6733668341708543,
      "eval_runtime": 2.8031,
      "eval_samples_per_second": 396.697,
      "eval_steps_per_second": 24.972,
      "step": 46752
    },
    {
      "epoch": 96.51,
      "learning_rate": 5.872689938398357e-06,
      "loss": 0.0012,
      "step": 47000
    },
    {
      "epoch": 97.0,
      "eval_accuracy": 0.7897473819803917,
      "eval_f1": 0.638197537332984,
      "eval_loss": 2.1853880882263184,
      "eval_precision": 0.6011846001974334,
      "eval_recall": 0.6800670016750419,
      "eval_runtime": 2.7798,
      "eval_samples_per_second": 400.033,
      "eval_steps_per_second": 25.182,
      "step": 47239
    },
    {
      "epoch": 97.54,
      "learning_rate": 5.6160164271047226e-06,
      "loss": 0.0008,
      "step": 47500
    },
    {
      "epoch": 98.0,
      "eval_accuracy": 0.7930392881849192,
      "eval_f1": 0.6455238859895625,
      "eval_loss": 2.200537919998169,
      "eval_precision": 0.6198920585967618,
      "eval_recall": 0.6733668341708543,
      "eval_runtime": 2.7585,
      "eval_samples_per_second": 403.112,
      "eval_steps_per_second": 25.376,
      "step": 47726
    },
    {
      "epoch": 98.56,
      "learning_rate": 5.359342915811089e-06,
      "loss": 0.0008,
      "step": 48000
    },
    {
      "epoch": 99.0,
      "eval_accuracy": 0.7896042556236731,
      "eval_f1": 0.6393529567753912,
      "eval_loss": 2.1999213695526123,
      "eval_precision": 0.6088383838383838,
      "eval_recall": 0.6730876605248465,
      "eval_runtime": 2.8138,
      "eval_samples_per_second": 395.198,
      "eval_steps_per_second": 24.878,
      "step": 48213
    },
    {
      "epoch": 99.59,
      "learning_rate": 5.102669404517454e-06,
      "loss": 0.0011,
      "step": 48500
    },
    {
      "epoch": 100.0,
      "eval_accuracy": 0.7930631425777057,
      "eval_f1": 0.6375963839404414,
      "eval_loss": 2.2227866649627686,
      "eval_precision": 0.6086294416243655,
      "eval_recall": 0.6694584031267449,
      "eval_runtime": 2.7694,
      "eval_samples_per_second": 401.531,
      "eval_steps_per_second": 25.276,
      "step": 48700
    },
    {
      "epoch": 100.62,
      "learning_rate": 4.845995893223819e-06,
      "loss": 0.0006,
      "step": 49000
    },
    {
      "epoch": 101.0,
      "eval_accuracy": 0.7925144915436177,
      "eval_f1": 0.6429421881201217,
      "eval_loss": 2.2299931049346924,
      "eval_precision": 0.6110133266281117,
      "eval_recall": 0.678391959798995,
      "eval_runtime": 2.9094,
      "eval_samples_per_second": 382.209,
      "eval_steps_per_second": 24.06,
      "step": 49187
    },
    {
      "epoch": 101.64,
      "learning_rate": 4.5893223819301845e-06,
      "loss": 0.001,
      "step": 49500
    },
    {
      "epoch": 102.0,
      "eval_accuracy": 0.7917034421888791,
      "eval_f1": 0.6384889710738344,
      "eval_loss": 2.2193572521209717,
      "eval_precision": 0.6059162697417899,
      "eval_recall": 0.6747627024008933,
      "eval_runtime": 2.7802,
      "eval_samples_per_second": 399.973,
      "eval_steps_per_second": 25.178,
      "step": 49674
    },
    {
      "epoch": 102.67,
      "learning_rate": 4.332648870636551e-06,
      "loss": 0.0007,
      "step": 50000
    },
    {
      "epoch": 103.0,
      "eval_accuracy": 0.794685241287183,
      "eval_f1": 0.6422018348623854,
      "eval_loss": 2.2047975063323975,
      "eval_precision": 0.6130997715156131,
      "eval_recall": 0.6742043551088778,
      "eval_runtime": 2.8095,
      "eval_samples_per_second": 395.8,
      "eval_steps_per_second": 24.915,
      "step": 50161
    },
    {
      "epoch": 103.7,
      "learning_rate": 4.075975359342916e-06,
      "loss": 0.0003,
      "step": 50500
    },
    {
      "epoch": 104.0,
      "eval_accuracy": 0.7937549199685122,
      "eval_f1": 0.6471614548382768,
      "eval_loss": 2.2270121574401855,
      "eval_precision": 0.6231584388730939,
      "eval_recall": 0.6730876605248465,
      "eval_runtime": 2.7614,
      "eval_samples_per_second": 402.698,
      "eval_steps_per_second": 25.35,
      "step": 50648
    },
    {
      "epoch": 104.72,
      "learning_rate": 3.819301848049281e-06,
      "loss": 0.0008,
      "step": 51000
    },
    {
      "epoch": 105.0,
      "eval_accuracy": 0.7951861835356981,
      "eval_f1": 0.6451182048884735,
      "eval_loss": 2.228421211242676,
      "eval_precision": 0.618437900128041,
      "eval_recall": 0.6742043551088778,
      "eval_runtime": 2.7968,
      "eval_samples_per_second": 397.592,
      "eval_steps_per_second": 25.028,
      "step": 51135
    },
    {
      "epoch": 105.75,
      "learning_rate": 3.5626283367556473e-06,
      "loss": 0.0005,
      "step": 51500
    },
    {
      "epoch": 106.0,
      "eval_accuracy": 0.7920612580806755,
      "eval_f1": 0.6393963463065925,
      "eval_loss": 2.2278361320495605,
      "eval_precision": 0.6080060422960725,
      "eval_recall": 0.6742043551088778,
      "eval_runtime": 2.7765,
      "eval_samples_per_second": 400.508,
      "eval_steps_per_second": 25.212,
      "step": 51622
    },
    {
      "epoch": 106.78,
      "learning_rate": 3.3059548254620125e-06,
      "loss": 0.0004,
      "step": 52000
    },
    {
      "epoch": 107.0,
      "eval_accuracy": 0.7925622003291907,
      "eval_f1": 0.6443971253659835,
      "eval_loss": 2.257086992263794,
      "eval_precision": 0.6157171922685656,
      "eval_recall": 0.6758793969849246,
      "eval_runtime": 2.7854,
      "eval_samples_per_second": 399.221,
      "eval_steps_per_second": 25.131,
      "step": 52109
    },
    {
      "epoch": 107.8,
      "learning_rate": 3.049281314168378e-06,
      "loss": 0.0006,
      "step": 52500
    },
    {
      "epoch": 108.0,
      "eval_accuracy": 0.7924906371508313,
      "eval_f1": 0.6378807947019867,
      "eval_loss": 2.256169319152832,
      "eval_precision": 0.6068548387096774,
      "eval_recall": 0.672250139586823,
      "eval_runtime": 2.7803,
      "eval_samples_per_second": 399.96,
      "eval_steps_per_second": 25.177,
      "step": 52596
    },
    {
      "epoch": 108.83,
      "learning_rate": 2.7926078028747435e-06,
      "loss": 0.0005,
      "step": 53000
    },
    {
      "epoch": 109.0,
      "eval_accuracy": 0.7949714940006202,
      "eval_f1": 0.6433155080213904,
      "eval_loss": 2.225450277328491,
      "eval_precision": 0.6172396100564392,
      "eval_recall": 0.6716917922948074,
      "eval_runtime": 2.7966,
      "eval_samples_per_second": 397.621,
      "eval_steps_per_second": 25.03,
      "step": 53083
    },
    {
      "epoch": 109.86,
      "learning_rate": 2.5359342915811088e-06,
      "loss": 0.0006,
      "step": 53500
    },
    {
      "epoch": 110.0,
      "eval_accuracy": 0.7930869969704921,
      "eval_f1": 0.6414944356120826,
      "eval_loss": 2.242906332015991,
      "eval_precision": 0.6104387291981845,
      "eval_recall": 0.6758793969849246,
      "eval_runtime": 2.7792,
      "eval_samples_per_second": 400.111,
      "eval_steps_per_second": 25.187,
      "step": 53570
    },
    {
      "epoch": 110.88,
      "learning_rate": 2.2792607802874745e-06,
      "loss": 0.0004,
      "step": 54000
    },
    {
      "epoch": 111.0,
      "eval_accuracy": 0.7927291810786956,
      "eval_f1": 0.6417751793781558,
      "eval_loss": 2.241628885269165,
      "eval_precision": 0.6123225152129818,
      "eval_recall": 0.6742043551088778,
      "eval_runtime": 2.7871,
      "eval_samples_per_second": 398.979,
      "eval_steps_per_second": 25.116,
      "step": 54057
    },
    {
      "epoch": 111.91,
      "learning_rate": 2.0225872689938397e-06,
      "loss": 0.0004,
      "step": 54500
    },
    {
      "epoch": 112.0,
      "eval_accuracy": 0.7938980463252308,
      "eval_f1": 0.6393595730486992,
      "eval_loss": 2.262876033782959,
      "eval_precision": 0.612317914643496,
      "eval_recall": 0.6689000558347292,
      "eval_runtime": 2.7846,
      "eval_samples_per_second": 399.344,
      "eval_steps_per_second": 25.139,
      "step": 54544
    },
    {
      "epoch": 112.94,
      "learning_rate": 1.7659137577002054e-06,
      "loss": 0.0004,
      "step": 55000
    },
    {
      "epoch": 113.0,
      "eval_accuracy": 0.7931824145416378,
      "eval_f1": 0.6427336790320436,
      "eval_loss": 2.2645206451416016,
      "eval_precision": 0.6136075145976136,
      "eval_recall": 0.6747627024008933,
      "eval_runtime": 2.7845,
      "eval_samples_per_second": 399.355,
      "eval_steps_per_second": 25.139,
      "step": 55031
    },
    {
      "epoch": 113.96,
      "learning_rate": 1.509240246406571e-06,
      "loss": 0.0003,
      "step": 55500
    },
    {
      "epoch": 114.0,
      "eval_accuracy": 0.7943751341809594,
      "eval_f1": 0.6461373677868524,
      "eval_loss": 2.2760562896728516,
      "eval_precision": 0.6207872395163365,
      "eval_recall": 0.6736460078168621,
      "eval_runtime": 2.7804,
      "eval_samples_per_second": 399.941,
      "eval_steps_per_second": 25.176,
      "step": 55518
    },
    {
      "epoch": 114.99,
      "learning_rate": 1.2525667351129362e-06,
      "loss": 0.0004,
      "step": 56000
    },
    {
      "epoch": 115.0,
      "eval_accuracy": 0.7936595023973665,
      "eval_f1": 0.6438374417055296,
      "eval_loss": 2.2683961391448975,
      "eval_precision": 0.6158552128473107,
      "eval_recall": 0.6744835287548855,
      "eval_runtime": 2.8007,
      "eval_samples_per_second": 397.04,
      "eval_steps_per_second": 24.994,
      "step": 56005
    },
    {
      "epoch": 116.0,
      "eval_accuracy": 0.7926337635075499,
      "eval_f1": 0.6435524736631552,
      "eval_loss": 2.274099349975586,
      "eval_precision": 0.6160326780699515,
      "eval_recall": 0.6736460078168621,
      "eval_runtime": 2.8997,
      "eval_samples_per_second": 383.493,
      "eval_steps_per_second": 24.141,
      "step": 56492
    },
    {
      "epoch": 116.02,
      "learning_rate": 9.958932238193019e-07,
      "loss": 0.0003,
      "step": 56500
    },
    {
      "epoch": 117.0,
      "eval_accuracy": 0.7939457551108037,
      "eval_f1": 0.6435524736631552,
      "eval_loss": 2.257563591003418,
      "eval_precision": 0.6160326780699515,
      "eval_recall": 0.6736460078168621,
      "eval_runtime": 2.7639,
      "eval_samples_per_second": 402.33,
      "eval_steps_per_second": 25.327,
      "step": 56979
    },
    {
      "epoch": 117.04,
      "learning_rate": 7.392197125256674e-07,
      "loss": 0.001,
      "step": 57000
    },
    {
      "epoch": 118.0,
      "eval_accuracy": 0.7942558622170273,
      "eval_f1": 0.643104827954121,
      "eval_loss": 2.2542884349823,
      "eval_precision": 0.615679264555669,
      "eval_recall": 0.6730876605248465,
      "eval_runtime": 2.7671,
      "eval_samples_per_second": 401.861,
      "eval_steps_per_second": 25.297,
      "step": 57466
    },
    {
      "epoch": 118.07,
      "learning_rate": 4.82546201232033e-07,
      "loss": 0.0003,
      "step": 57500
    },
    {
      "epoch": 119.0,
      "eval_accuracy": 0.794685241287183,
      "eval_f1": 0.6438191758901186,
      "eval_loss": 2.2540647983551025,
      "eval_precision": 0.6162879754914475,
      "eval_recall": 0.67392518146287,
      "eval_runtime": 2.7786,
      "eval_samples_per_second": 400.208,
      "eval_steps_per_second": 25.193,
      "step": 57953
    },
    {
      "epoch": 119.1,
      "learning_rate": 2.2587268993839835e-07,
      "loss": 0.0005,
      "step": 58000
    },
    {
      "epoch": 120.0,
      "eval_accuracy": 0.7944228429665323,
      "eval_f1": 0.6433529097704217,
      "eval_loss": 2.2551677227020264,
      "eval_precision": 0.6163682864450127,
      "eval_recall": 0.6728084868788387,
      "eval_runtime": 2.7762,
      "eval_samples_per_second": 400.553,
      "eval_steps_per_second": 25.215,
      "step": 58440
    },
    {
      "epoch": 120.0,
      "step": 58440,
      "total_flos": 1.220726808511488e+17,
      "train_loss": 0.039664112105659635,
      "train_runtime": 7090.7344,
      "train_samples_per_second": 131.749,
      "train_steps_per_second": 8.242
    }
  ],
  "max_steps": 58440,
  "num_train_epochs": 120,
  "total_flos": 1.220726808511488e+17,
  "trial_name": null,
  "trial_params": null
}