{
  "best_metric": 0.8730279414559969,
  "best_model_checkpoint": "./MDEBERTA_MLCR-outputs/checkpoint-77000",
  "epoch": 14.238165680473372,
  "eval_steps": 1000,
  "global_step": 77000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.1849112426035503,
      "grad_norm": 2.0331010818481445,
      "learning_rate": 1.849112426035503e-07,
      "loss": 0.7187,
      "step": 1000
    },
    {
      "epoch": 0.1849112426035503,
      "eval_accuracy": 0.4890784272935303,
      "eval_f1": 0.0,
      "eval_loss": 0.7106410264968872,
      "eval_precision": 0.0,
      "eval_recall": 0.0,
      "eval_runtime": 71.3076,
      "eval_samples_per_second": 134.824,
      "eval_steps_per_second": 8.428,
      "step": 1000
    },
    {
      "epoch": 0.3698224852071006,
      "grad_norm": 3.5357959270477295,
      "learning_rate": 3.698224852071006e-07,
      "loss": 0.6978,
      "step": 2000
    },
    {
      "epoch": 0.3698224852071006,
      "eval_accuracy": 0.5269398793426253,
      "eval_f1": 0.6158134820070958,
      "eval_loss": 0.6897135972976685,
      "eval_precision": 0.526277793820387,
      "eval_recall": 0.7420602605863192,
      "eval_runtime": 71.4436,
      "eval_samples_per_second": 134.568,
      "eval_steps_per_second": 8.412,
      "step": 2000
    },
    {
      "epoch": 0.5547337278106509,
      "grad_norm": 5.24600887298584,
      "learning_rate": 5.547337278106509e-07,
      "loss": 0.6881,
      "step": 3000
    },
    {
      "epoch": 0.5547337278106509,
      "eval_accuracy": 0.5542958185978781,
      "eval_f1": 0.6489143793527243,
      "eval_loss": 0.6817396283149719,
      "eval_precision": 0.5429864253393665,
      "eval_recall": 0.8061889250814332,
      "eval_runtime": 71.3024,
      "eval_samples_per_second": 134.834,
      "eval_steps_per_second": 8.429,
      "step": 3000
    },
    {
      "epoch": 0.7396449704142012,
      "grad_norm": 4.388635158538818,
      "learning_rate": 7.396449704142012e-07,
      "loss": 0.6831,
      "step": 4000
    },
    {
      "epoch": 0.7396449704142012,
      "eval_accuracy": 0.5620969419596422,
      "eval_f1": 0.666561064470141,
      "eval_loss": 0.6732956171035767,
      "eval_precision": 0.5455016852476018,
      "eval_recall": 0.8566775244299675,
      "eval_runtime": 71.4449,
      "eval_samples_per_second": 134.565,
      "eval_steps_per_second": 8.412,
      "step": 4000
    },
    {
      "epoch": 0.9245562130177515,
      "grad_norm": 6.712058067321777,
      "learning_rate": 9.245562130177515e-07,
      "loss": 0.6758,
      "step": 5000
    },
    {
      "epoch": 0.9245562130177515,
      "eval_accuracy": 0.5701060952777199,
      "eval_f1": 0.47927428499433034,
      "eval_loss": 0.6640639901161194,
      "eval_precision": 0.6287603305785124,
      "eval_recall": 0.38721498371335505,
      "eval_runtime": 71.1928,
      "eval_samples_per_second": 135.042,
      "eval_steps_per_second": 8.442,
      "step": 5000
    },
    {
      "epoch": 1.1094674556213018,
      "grad_norm": 5.304254055023193,
      "learning_rate": 1.1094674556213018e-06,
      "loss": 0.6636,
      "step": 6000
    },
    {
      "epoch": 1.1094674556213018,
      "eval_accuracy": 0.6098398169336384,
      "eval_f1": 0.6513616507110326,
      "eval_loss": 0.6411445736885071,
      "eval_precision": 0.599281682914315,
      "eval_recall": 0.7133550488599348,
      "eval_runtime": 71.3706,
      "eval_samples_per_second": 134.705,
      "eval_steps_per_second": 8.421,
      "step": 6000
    },
    {
      "epoch": 1.2943786982248522,
      "grad_norm": 17.199308395385742,
      "learning_rate": 1.294378698224852e-06,
      "loss": 0.6359,
      "step": 7000
    },
    {
      "epoch": 1.2943786982248522,
      "eval_accuracy": 0.6711046390680258,
      "eval_f1": 0.7002275312855518,
      "eval_loss": 0.5785636901855469,
      "eval_precision": 0.6552519517388219,
      "eval_recall": 0.7518322475570033,
      "eval_runtime": 71.2372,
      "eval_samples_per_second": 134.958,
      "eval_steps_per_second": 8.437,
      "step": 7000
    },
    {
      "epoch": 1.4792899408284024,
      "grad_norm": 17.191123962402344,
      "learning_rate": 1.4792899408284024e-06,
      "loss": 0.5651,
      "step": 8000
    },
    {
      "epoch": 1.4792899408284024,
      "eval_accuracy": 0.7214478884959434,
      "eval_f1": 0.6973327305605787,
      "eval_loss": 0.5006158947944641,
      "eval_precision": 0.7837906504065041,
      "eval_recall": 0.6280537459283387,
      "eval_runtime": 71.414,
      "eval_samples_per_second": 134.624,
      "eval_steps_per_second": 8.416,
      "step": 8000
    },
    {
      "epoch": 1.6642011834319526,
      "grad_norm": 17.50662612915039,
      "learning_rate": 1.6642011834319525e-06,
      "loss": 0.5189,
      "step": 9000
    },
    {
      "epoch": 1.6642011834319526,
      "eval_accuracy": 0.7335136259621385,
      "eval_f1": 0.7180897887323944,
      "eval_loss": 0.4555279016494751,
      "eval_precision": 0.7813697318007663,
      "eval_recall": 0.6642915309446255,
      "eval_runtime": 71.347,
      "eval_samples_per_second": 134.75,
      "eval_steps_per_second": 8.424,
      "step": 9000
    },
    {
      "epoch": 1.849112426035503,
      "grad_norm": 11.771262168884277,
      "learning_rate": 1.849112426035503e-06,
      "loss": 0.4782,
      "step": 10000
    },
    {
      "epoch": 1.849112426035503,
      "eval_accuracy": 0.7384023299355107,
      "eval_f1": 0.7066371165286364,
      "eval_loss": 0.43298760056495667,
      "eval_precision": 0.827369571155422,
      "eval_recall": 0.6166530944625407,
      "eval_runtime": 71.4997,
      "eval_samples_per_second": 134.462,
      "eval_steps_per_second": 8.406,
      "step": 10000
    },
    {
      "epoch": 2.0340236686390534,
      "grad_norm": 11.091800689697266,
      "learning_rate": 1.9962195923734386e-06,
      "loss": 0.4474,
      "step": 11000
    },
    {
      "epoch": 2.0340236686390534,
      "eval_accuracy": 0.7443311836904514,
      "eval_f1": 0.754739572939533,
      "eval_loss": 0.41377708315849304,
      "eval_precision": 0.7401174168297456,
      "eval_recall": 0.7699511400651465,
      "eval_runtime": 71.4016,
      "eval_samples_per_second": 134.647,
      "eval_steps_per_second": 8.417,
      "step": 11000
    },
    {
      "epoch": 2.2189349112426036,
      "grad_norm": 31.02802276611328,
      "learning_rate": 1.975673898750822e-06,
      "loss": 0.4275,
      "step": 12000
    },
    {
      "epoch": 2.2189349112426036,
      "eval_accuracy": 0.7448512585812357,
      "eval_f1": 0.769822651778174,
      "eval_loss": 0.40316781401634216,
      "eval_precision": 0.7140121845082681,
      "eval_recall": 0.8350977198697068,
      "eval_runtime": 71.3228,
      "eval_samples_per_second": 134.796,
      "eval_steps_per_second": 8.426,
      "step": 12000
    },
    {
      "epoch": 2.4038461538461537,
      "grad_norm": 30.21622085571289,
      "learning_rate": 1.955128205128205e-06,
      "loss": 0.425,
      "step": 13000
    },
    {
      "epoch": 2.4038461538461537,
      "eval_accuracy": 0.7466195132099023,
      "eval_f1": 0.792504258943782,
      "eval_loss": 0.40604618191719055,
      "eval_precision": 0.6813122437024018,
      "eval_recall": 0.9470684039087948,
      "eval_runtime": 71.2177,
      "eval_samples_per_second": 134.995,
      "eval_steps_per_second": 8.439,
      "step": 13000
    },
    {
      "epoch": 2.5887573964497044,
      "grad_norm": 7.608345031738281,
      "learning_rate": 1.9345825115055883e-06,
      "loss": 0.4093,
      "step": 14000
    },
    {
      "epoch": 2.5887573964497044,
      "eval_accuracy": 0.7550447264406075,
      "eval_f1": 0.7994208329784516,
      "eval_loss": 0.38763341307640076,
      "eval_precision": 0.6872162834968517,
      "eval_recall": 0.9554153094462541,
      "eval_runtime": 71.3556,
      "eval_samples_per_second": 134.734,
      "eval_steps_per_second": 8.423,
      "step": 14000
    },
    {
      "epoch": 2.7736686390532546,
      "grad_norm": 18.13289451599121,
      "learning_rate": 1.9140368178829715e-06,
      "loss": 0.4136,
      "step": 15000
    },
    {
      "epoch": 2.7736686390532546,
      "eval_accuracy": 0.7645100894528812,
      "eval_f1": 0.7836391437308868,
      "eval_loss": 0.3780820667743683,
      "eval_precision": 0.7384726224783862,
      "eval_recall": 0.8346905537459284,
      "eval_runtime": 71.3682,
      "eval_samples_per_second": 134.71,
      "eval_steps_per_second": 8.421,
      "step": 15000
    },
    {
      "epoch": 2.9585798816568047,
      "grad_norm": 18.237754821777344,
      "learning_rate": 1.8934911242603548e-06,
      "loss": 0.4026,
      "step": 16000
    },
    {
      "epoch": 2.9585798816568047,
      "eval_accuracy": 0.7601414603702933,
      "eval_f1": 0.8055649241146712,
      "eval_loss": 0.3755676746368408,
      "eval_precision": 0.6875359815774323,
      "eval_recall": 0.9725162866449512,
      "eval_runtime": 71.4592,
      "eval_samples_per_second": 134.538,
      "eval_steps_per_second": 8.41,
      "step": 16000
    },
    {
      "epoch": 3.143491124260355,
      "grad_norm": 7.552541255950928,
      "learning_rate": 1.8729454306377385e-06,
      "loss": 0.3895,
      "step": 17000
    },
    {
      "epoch": 3.143491124260355,
      "eval_accuracy": 0.7738714374869982,
      "eval_f1": 0.7812877263581489,
      "eval_loss": 0.3771662712097168,
      "eval_precision": 0.7722752585521082,
      "eval_recall": 0.7905130293159609,
      "eval_runtime": 71.6042,
      "eval_samples_per_second": 134.266,
      "eval_steps_per_second": 8.393,
      "step": 17000
    },
    {
      "epoch": 3.328402366863905,
      "grad_norm": 22.454090118408203,
      "learning_rate": 1.8523997370151217e-06,
      "loss": 0.3916,
      "step": 18000
    },
    {
      "epoch": 3.328402366863905,
      "eval_accuracy": 0.7645100894528812,
      "eval_f1": 0.7874178403755868,
      "eval_loss": 0.3725959360599518,
      "eval_precision": 0.7307424189613105,
      "eval_recall": 0.8536237785016286,
      "eval_runtime": 71.5401,
      "eval_samples_per_second": 134.386,
      "eval_steps_per_second": 8.401,
      "step": 18000
    },
    {
      "epoch": 3.5133136094674557,
      "grad_norm": 8.693994522094727,
      "learning_rate": 1.831854043392505e-06,
      "loss": 0.3882,
      "step": 19000
    },
    {
      "epoch": 3.5133136094674557,
      "eval_accuracy": 0.7733513625962138,
      "eval_f1": 0.8056026407351236,
      "eval_loss": 0.3690403401851654,
      "eval_precision": 0.717008099094807,
      "eval_recall": 0.9191775244299675,
      "eval_runtime": 71.609,
      "eval_samples_per_second": 134.257,
      "eval_steps_per_second": 8.393,
      "step": 19000
    },
    {
      "epoch": 3.698224852071006,
      "grad_norm": 5.8600358963012695,
      "learning_rate": 1.8113083497698882e-06,
      "loss": 0.3901,
      "step": 20000
    },
    {
      "epoch": 3.698224852071006,
      "eval_accuracy": 0.783960890368213,
      "eval_f1": 0.780188379722722,
      "eval_loss": 0.3576684594154358,
      "eval_precision": 0.8124311218867093,
      "eval_recall": 0.7504071661237784,
      "eval_runtime": 71.3088,
      "eval_samples_per_second": 134.822,
      "eval_steps_per_second": 8.428,
      "step": 20000
    },
    {
      "epoch": 3.883136094674556,
      "grad_norm": 19.961711883544922,
      "learning_rate": 1.7907626561472714e-06,
      "loss": 0.3776,
      "step": 21000
    },
    {
      "epoch": 3.883136094674556,
      "eval_accuracy": 0.7951945080091534,
      "eval_f1": 0.8153081324453616,
      "eval_loss": 0.3502472937107086,
      "eval_precision": 0.7559575578361454,
      "eval_recall": 0.884771986970684,
      "eval_runtime": 71.2812,
      "eval_samples_per_second": 134.874,
      "eval_steps_per_second": 8.431,
      "step": 21000
    },
    {
      "epoch": 4.068047337278107,
      "grad_norm": 5.484897136688232,
      "learning_rate": 1.7702169625246549e-06,
      "loss": 0.3687,
      "step": 22000
    },
    {
      "epoch": 4.068047337278107,
      "eval_accuracy": 0.7998751820262118,
      "eval_f1": 0.8182848507744617,
      "eval_loss": 0.34667038917541504,
      "eval_precision": 0.7632135306553911,
      "eval_recall": 0.8819218241042345,
      "eval_runtime": 71.5944,
      "eval_samples_per_second": 134.284,
      "eval_steps_per_second": 8.395,
      "step": 22000
    },
    {
      "epoch": 4.252958579881657,
      "grad_norm": 14.385085105895996,
      "learning_rate": 1.7496712689020381e-06,
      "loss": 0.354,
      "step": 23000
    },
    {
      "epoch": 4.252958579881657,
      "eval_accuracy": 0.8051799459122113,
      "eval_f1": 0.8130178696216432,
      "eval_loss": 0.3367505669593811,
      "eval_precision": 0.7976493633692459,
      "eval_recall": 0.8289902280130294,
      "eval_runtime": 71.182,
      "eval_samples_per_second": 135.062,
      "eval_steps_per_second": 8.443,
      "step": 23000
    },
    {
      "epoch": 4.437869822485207,
      "grad_norm": 7.935754776000977,
      "learning_rate": 1.7291255752794214e-06,
      "loss": 0.3498,
      "step": 24000
    },
    {
      "epoch": 4.437869822485207,
      "eval_accuracy": 0.8109007697108384,
      "eval_f1": 0.8206039076376554,
      "eval_loss": 0.33399322628974915,
      "eval_precision": 0.7962466487935657,
      "eval_recall": 0.8464983713355049,
      "eval_runtime": 71.7203,
      "eval_samples_per_second": 134.048,
      "eval_steps_per_second": 8.38,
      "step": 24000
    },
    {
      "epoch": 4.622781065088757,
      "grad_norm": 5.013232707977295,
      "learning_rate": 1.7085798816568046e-06,
      "loss": 0.349,
      "step": 25000
    },
    {
      "epoch": 4.622781065088757,
      "eval_accuracy": 0.8079883503224464,
      "eval_f1": 0.8264058679706602,
      "eval_loss": 0.3555152416229248,
      "eval_precision": 0.7679133170220203,
      "eval_recall": 0.8945439739413681,
      "eval_runtime": 71.3751,
      "eval_samples_per_second": 134.697,
      "eval_steps_per_second": 8.42,
      "step": 25000
    },
    {
      "epoch": 4.8076923076923075,
      "grad_norm": 15.721752166748047,
      "learning_rate": 1.688034188034188e-06,
      "loss": 0.3341,
      "step": 26000
    },
    {
      "epoch": 4.8076923076923075,
      "eval_accuracy": 0.8115248595797795,
      "eval_f1": 0.8270662340141248,
      "eval_loss": 0.3554248809814453,
      "eval_precision": 0.7784764642472153,
      "eval_recall": 0.8821254071661238,
      "eval_runtime": 71.5461,
      "eval_samples_per_second": 134.375,
      "eval_steps_per_second": 8.4,
      "step": 26000
    },
    {
      "epoch": 4.992603550295858,
      "grad_norm": 21.64997100830078,
      "learning_rate": 1.6674884944115713e-06,
      "loss": 0.3348,
      "step": 27000
    },
    {
      "epoch": 4.992603550295858,
      "eval_accuracy": 0.819742042854171,
      "eval_f1": 0.8303807379857101,
      "eval_loss": 0.3212674856185913,
      "eval_precision": 0.7996229971724788,
      "eval_recall": 0.863599348534202,
      "eval_runtime": 71.2415,
      "eval_samples_per_second": 134.949,
      "eval_steps_per_second": 8.436,
      "step": 27000
    },
    {
      "epoch": 5.177514792899408,
      "grad_norm": 15.086274147033691,
      "learning_rate": 1.6469428007889545e-06,
      "loss": 0.3122,
      "step": 28000
    },
    {
      "epoch": 5.177514792899408,
      "eval_accuracy": 0.8222384023299355,
      "eval_f1": 0.8343510710477852,
      "eval_loss": 0.3286217749118805,
      "eval_precision": 0.7962997224791859,
      "eval_recall": 0.8762214983713354,
      "eval_runtime": 71.5509,
      "eval_samples_per_second": 134.366,
      "eval_steps_per_second": 8.4,
      "step": 28000
    },
    {
      "epoch": 5.362426035502959,
      "grad_norm": 8.424311637878418,
      "learning_rate": 1.6263971071663378e-06,
      "loss": 0.3024,
      "step": 29000
    },
    {
      "epoch": 5.362426035502959,
      "eval_accuracy": 0.8286873309756605,
      "eval_f1": 0.835019533206451,
      "eval_loss": 0.32026207447052,
      "eval_precision": 0.8219286136856636,
      "eval_recall": 0.8485342019543974,
      "eval_runtime": 71.2045,
      "eval_samples_per_second": 135.02,
      "eval_steps_per_second": 8.44,
      "step": 29000
    },
    {
      "epoch": 5.547337278106509,
      "grad_norm": 31.459623336791992,
      "learning_rate": 1.605851413543721e-06,
      "loss": 0.3023,
      "step": 30000
    },
    {
      "epoch": 5.547337278106509,
      "eval_accuracy": 0.829831495735386,
      "eval_f1": 0.8446638815039879,
      "eval_loss": 0.32447922229766846,
      "eval_precision": 0.7914590747330961,
      "eval_recall": 0.9055374592833876,
      "eval_runtime": 71.4681,
      "eval_samples_per_second": 134.522,
      "eval_steps_per_second": 8.409,
      "step": 30000
    },
    {
      "epoch": 5.732248520710059,
      "grad_norm": 6.890359878540039,
      "learning_rate": 1.5853057199211045e-06,
      "loss": 0.3049,
      "step": 31000
    },
    {
      "epoch": 5.732248520710059,
      "eval_accuracy": 0.832639900145621,
      "eval_f1": 0.8481072406306052,
      "eval_loss": 0.31333744525909424,
      "eval_precision": 0.7907058616440767,
      "eval_recall": 0.9144951140065146,
      "eval_runtime": 71.3131,
      "eval_samples_per_second": 134.814,
      "eval_steps_per_second": 8.428,
      "step": 31000
    },
    {
      "epoch": 5.9171597633136095,
      "grad_norm": 23.060514450073242,
      "learning_rate": 1.5647600262984877e-06,
      "loss": 0.3039,
      "step": 32000
    },
    {
      "epoch": 5.9171597633136095,
      "eval_accuracy": 0.8346161847306012,
      "eval_f1": 0.8510957108072673,
      "eval_loss": 0.29815933108329773,
      "eval_precision": 0.78806798473812,
      "eval_recall": 0.9250814332247557,
      "eval_runtime": 71.51,
      "eval_samples_per_second": 134.443,
      "eval_steps_per_second": 8.404,
      "step": 32000
    },
    {
      "epoch": 6.10207100591716,
      "grad_norm": 13.75260066986084,
      "learning_rate": 1.544214332675871e-06,
      "loss": 0.2884,
      "step": 33000
    },
    {
      "epoch": 6.10207100591716,
      "eval_accuracy": 0.8325358851674641,
      "eval_f1": 0.8447743925954493,
      "eval_loss": 0.3285083770751953,
      "eval_precision": 0.8023809523809524,
      "eval_recall": 0.8918973941368078,
      "eval_runtime": 71.4708,
      "eval_samples_per_second": 134.516,
      "eval_steps_per_second": 8.409,
      "step": 33000
    },
    {
      "epoch": 6.28698224852071,
      "grad_norm": 9.07077407836914,
      "learning_rate": 1.5236686390532542e-06,
      "loss": 0.2803,
      "step": 34000
    },
    {
      "epoch": 6.28698224852071,
      "eval_accuracy": 0.8398169336384439,
      "eval_f1": 0.849078792630341,
      "eval_loss": 0.30692258477211,
      "eval_precision": 0.81859410430839,
      "eval_recall": 0.8819218241042345,
      "eval_runtime": 71.54,
      "eval_samples_per_second": 134.386,
      "eval_steps_per_second": 8.401,
      "step": 34000
    },
    {
      "epoch": 6.47189349112426,
      "grad_norm": 31.427412033081055,
      "learning_rate": 1.5031229454306379e-06,
      "loss": 0.2731,
      "step": 35000
    },
    {
      "epoch": 6.47189349112426,
      "eval_accuracy": 0.8372165591845225,
      "eval_f1": 0.8451261751608115,
      "eval_loss": 0.3197310268878937,
      "eval_precision": 0.822260735605623,
      "eval_recall": 0.869299674267101,
      "eval_runtime": 71.1627,
      "eval_samples_per_second": 135.099,
      "eval_steps_per_second": 8.445,
      "step": 35000
    },
    {
      "epoch": 6.65680473372781,
      "grad_norm": 18.06778335571289,
      "learning_rate": 1.4825772518080211e-06,
      "loss": 0.2717,
      "step": 36000
    },
    {
      "epoch": 6.65680473372781,
      "eval_accuracy": 0.8398169336384439,
      "eval_f1": 0.8547717842323651,
      "eval_loss": 0.29738789796829224,
      "eval_precision": 0.7962052002810963,
      "eval_recall": 0.9226384364820847,
      "eval_runtime": 71.4155,
      "eval_samples_per_second": 134.621,
      "eval_steps_per_second": 8.416,
      "step": 36000
    },
    {
      "epoch": 6.841715976331361,
      "grad_norm": 10.752886772155762,
      "learning_rate": 1.4620315581854044e-06,
      "loss": 0.2764,
      "step": 37000
    },
    {
      "epoch": 6.841715976331361,
      "eval_accuracy": 0.8452257125026004,
      "eval_f1": 0.8565644881434356,
      "eval_loss": 0.2894780933856964,
      "eval_precision": 0.8134383009886489,
      "eval_recall": 0.9045195439739414,
      "eval_runtime": 71.1003,
      "eval_samples_per_second": 135.218,
      "eval_steps_per_second": 8.453,
      "step": 37000
    },
    {
      "epoch": 7.0266272189349115,
      "grad_norm": 9.670276641845703,
      "learning_rate": 1.4414858645627876e-06,
      "loss": 0.2689,
      "step": 38000
    },
    {
      "epoch": 7.0266272189349115,
      "eval_accuracy": 0.8449136675681298,
      "eval_f1": 0.8576882695428081,
      "eval_loss": 0.30610039830207825,
      "eval_precision": 0.8073674752920036,
      "eval_recall": 0.9146986970684039,
      "eval_runtime": 71.4568,
      "eval_samples_per_second": 134.543,
      "eval_steps_per_second": 8.411,
      "step": 38000
    },
    {
      "epoch": 7.211538461538462,
      "grad_norm": 24.961204528808594,
      "learning_rate": 1.4209401709401708e-06,
      "loss": 0.256,
      "step": 39000
    },
    {
      "epoch": 7.211538461538462,
      "eval_accuracy": 0.8429373829831496,
      "eval_f1": 0.8588785046728972,
      "eval_loss": 0.3088281452655792,
      "eval_precision": 0.7938838977194195,
      "eval_recall": 0.9354641693811075,
      "eval_runtime": 71.3164,
      "eval_samples_per_second": 134.808,
      "eval_steps_per_second": 8.427,
      "step": 39000
    },
    {
      "epoch": 7.396449704142012,
      "grad_norm": 11.780752182006836,
      "learning_rate": 1.4003944773175543e-06,
      "loss": 0.2584,
      "step": 40000
    },
    {
      "epoch": 7.396449704142012,
      "eval_accuracy": 0.8485541918036197,
      "eval_f1": 0.858448376433988,
      "eval_loss": 0.28954440355300903,
      "eval_precision": 0.8215481950130257,
      "eval_recall": 0.8988192182410424,
      "eval_runtime": 72.3859,
      "eval_samples_per_second": 132.816,
      "eval_steps_per_second": 8.303,
      "step": 40000
    },
    {
      "epoch": 7.581360946745562,
      "grad_norm": 5.480474472045898,
      "learning_rate": 1.3798487836949375e-06,
      "loss": 0.2559,
      "step": 41000
    },
    {
      "epoch": 7.581360946745562,
      "eval_accuracy": 0.8439775327647181,
      "eval_f1": 0.8532002348796242,
      "eval_loss": 0.3007795214653015,
      "eval_precision": 0.8215228043724085,
      "eval_recall": 0.8874185667752443,
      "eval_runtime": 71.4351,
      "eval_samples_per_second": 134.584,
      "eval_steps_per_second": 8.413,
      "step": 41000
    },
    {
      "epoch": 7.766272189349112,
      "grad_norm": 35.79705047607422,
      "learning_rate": 1.3593030900723208e-06,
      "loss": 0.2583,
      "step": 42000
    },
    {
      "epoch": 7.766272189349112,
      "eval_accuracy": 0.8460578323278553,
      "eval_f1": 0.8545597484276729,
      "eval_loss": 0.30222100019454956,
      "eval_precision": 0.8259878419452887,
      "eval_recall": 0.8851791530944625,
      "eval_runtime": 71.1634,
      "eval_samples_per_second": 135.098,
      "eval_steps_per_second": 8.445,
      "step": 42000
    },
    {
      "epoch": 7.951183431952662,
      "grad_norm": 7.736880779266357,
      "learning_rate": 1.338757396449704e-06,
      "loss": 0.2572,
      "step": 43000
    },
    {
      "epoch": 7.951183431952662,
      "eval_accuracy": 0.8488662367380903,
      "eval_f1": 0.8615003336192927,
      "eval_loss": 0.29688963294029236,
      "eval_precision": 0.8100017924359204,
      "eval_recall": 0.9199918566775245,
      "eval_runtime": 71.4477,
      "eval_samples_per_second": 134.56,
      "eval_steps_per_second": 8.412,
      "step": 43000
    },
    {
      "epoch": 8.136094674556213,
      "grad_norm": 31.98334503173828,
      "learning_rate": 1.3182117028270875e-06,
      "loss": 0.25,
      "step": 44000
    },
    {
      "epoch": 8.136094674556213,
      "eval_accuracy": 0.8475140420220512,
      "eval_f1": 0.8559072144682524,
      "eval_loss": 0.2971686124801636,
      "eval_precision": 0.8274420372481947,
      "eval_recall": 0.886400651465798,
      "eval_runtime": 71.3134,
      "eval_samples_per_second": 134.813,
      "eval_steps_per_second": 8.428,
      "step": 44000
    },
    {
      "epoch": 8.321005917159763,
      "grad_norm": 15.46840763092041,
      "learning_rate": 1.2976660092044707e-06,
      "loss": 0.2384,
      "step": 45000
    },
    {
      "epoch": 8.321005917159763,
      "eval_accuracy": 0.8495943415851883,
      "eval_f1": 0.8559760956175299,
      "eval_loss": 0.29278796911239624,
      "eval_precision": 0.8379485179407177,
      "eval_recall": 0.8747964169381107,
      "eval_runtime": 71.4398,
      "eval_samples_per_second": 134.575,
      "eval_steps_per_second": 8.413,
      "step": 45000
    },
    {
      "epoch": 8.505917159763314,
      "grad_norm": 9.12899112701416,
      "learning_rate": 1.277120315581854e-06,
      "loss": 0.2408,
      "step": 46000
    },
    {
      "epoch": 8.505917159763314,
      "eval_accuracy": 0.8494903266070314,
      "eval_f1": 0.8632454399395142,
      "eval_loss": 0.3010319471359253,
      "eval_precision": 0.8056094549303228,
      "eval_recall": 0.9297638436482085,
      "eval_runtime": 71.1731,
      "eval_samples_per_second": 135.079,
      "eval_steps_per_second": 8.444,
      "step": 46000
    },
    {
      "epoch": 8.690828402366863,
      "grad_norm": 6.4189934730529785,
      "learning_rate": 1.2565746219592372e-06,
      "loss": 0.251,
      "step": 47000
    },
    {
      "epoch": 8.690828402366863,
      "eval_accuracy": 0.853338880798835,
      "eval_f1": 0.8613569321533924,
      "eval_loss": 0.2985897958278656,
      "eval_precision": 0.8330163560289083,
      "eval_recall": 0.8916938110749185,
      "eval_runtime": 71.3021,
      "eval_samples_per_second": 134.835,
      "eval_steps_per_second": 8.429,
      "step": 47000
    },
    {
      "epoch": 8.875739644970414,
      "grad_norm": 28.648107528686523,
      "learning_rate": 1.2360289283366204e-06,
      "loss": 0.2436,
      "step": 48000
    },
    {
      "epoch": 8.875739644970414,
      "eval_accuracy": 0.8517786561264822,
      "eval_f1": 0.8599783826274934,
      "eval_loss": 0.29831263422966003,
      "eval_precision": 0.8311490978157645,
      "eval_recall": 0.8908794788273615,
      "eval_runtime": 71.1941,
      "eval_samples_per_second": 135.039,
      "eval_steps_per_second": 8.442,
      "step": 48000
    },
    {
      "epoch": 9.060650887573965,
      "grad_norm": 15.021607398986816,
      "learning_rate": 1.2154832347140039e-06,
      "loss": 0.2386,
      "step": 49000
    },
    {
      "epoch": 9.060650887573965,
      "eval_accuracy": 0.8550031204493447,
      "eval_f1": 0.864870104691741,
      "eval_loss": 0.30714085698127747,
      "eval_precision": 0.825499629903775,
      "eval_recall": 0.9081840390879479,
      "eval_runtime": 71.3285,
      "eval_samples_per_second": 134.785,
      "eval_steps_per_second": 8.426,
      "step": 49000
    },
    {
      "epoch": 9.245562130177515,
      "grad_norm": 16.410472869873047,
      "learning_rate": 1.1949375410913871e-06,
      "loss": 0.2245,
      "step": 50000
    },
    {
      "epoch": 9.245562130177515,
      "eval_accuracy": 0.8564593301435407,
      "eval_f1": 0.8665893271461717,
      "eval_loss": 0.2869988977909088,
      "eval_precision": 0.8251104565537555,
      "eval_recall": 0.9124592833876222,
      "eval_runtime": 71.7556,
      "eval_samples_per_second": 133.983,
      "eval_steps_per_second": 8.376,
      "step": 50000
    },
    {
      "epoch": 9.430473372781066,
      "grad_norm": 3.089381694793701,
      "learning_rate": 1.1743918474687704e-06,
      "loss": 0.2318,
      "step": 51000
    },
    {
      "epoch": 9.430473372781066,
      "eval_accuracy": 0.8566673600998543,
      "eval_f1": 0.8679570716749713,
      "eval_loss": 0.2949957549571991,
      "eval_precision": 0.8198769007965243,
      "eval_recall": 0.9220276872964169,
      "eval_runtime": 71.3514,
      "eval_samples_per_second": 134.742,
      "eval_steps_per_second": 8.423,
      "step": 51000
    },
    {
      "epoch": 9.615384615384615,
      "grad_norm": 19.565244674682617,
      "learning_rate": 1.1538461538461536e-06,
      "loss": 0.2348,
      "step": 52000
    },
    {
      "epoch": 9.615384615384615,
      "eval_accuracy": 0.8584356147285209,
      "eval_f1": 0.8695485478769289,
      "eval_loss": 0.30221354961395264,
      "eval_precision": 0.8215902916138381,
      "eval_recall": 0.9234527687296417,
      "eval_runtime": 71.2576,
      "eval_samples_per_second": 134.919,
      "eval_steps_per_second": 8.434,
      "step": 52000
    },
    {
      "epoch": 9.800295857988166,
      "grad_norm": 6.412939548492432,
      "learning_rate": 1.1333004602235373e-06,
      "loss": 0.2356,
      "step": 53000
    },
    {
      "epoch": 9.800295857988166,
      "eval_accuracy": 0.8553151653838152,
      "eval_f1": 0.8600744391912283,
      "eval_loss": 0.2939301133155823,
      "eval_precision": 0.8500695963412209,
      "eval_recall": 0.8703175895765473,
      "eval_runtime": 71.2887,
      "eval_samples_per_second": 134.86,
      "eval_steps_per_second": 8.431,
      "step": 53000
    },
    {
      "epoch": 9.985207100591715,
      "grad_norm": 24.278528213500977,
      "learning_rate": 1.1127547666009205e-06,
      "loss": 0.2274,
      "step": 54000
    },
    {
      "epoch": 9.985207100591715,
      "eval_accuracy": 0.8551071354275016,
      "eval_f1": 0.8684732319894249,
      "eval_loss": 0.31603407859802246,
      "eval_precision": 0.8098256735340729,
      "eval_recall": 0.9362785016286646,
      "eval_runtime": 71.2443,
      "eval_samples_per_second": 134.944,
      "eval_steps_per_second": 8.436,
      "step": 54000
    },
    {
      "epoch": 10.170118343195266,
      "grad_norm": 7.386682987213135,
      "learning_rate": 1.0922090729783038e-06,
      "loss": 0.2226,
      "step": 55000
    },
    {
      "epoch": 10.170118343195266,
      "eval_accuracy": 0.8546910755148741,
      "eval_f1": 0.8655567317871234,
      "eval_loss": 0.3250740170478821,
      "eval_precision": 0.8207702135426173,
      "eval_recall": 0.9155130293159609,
      "eval_runtime": 71.3288,
      "eval_samples_per_second": 134.784,
      "eval_steps_per_second": 8.426,
      "step": 55000
    },
    {
      "epoch": 10.355029585798816,
      "grad_norm": 26.67705535888672,
      "learning_rate": 1.071663379355687e-06,
      "loss": 0.225,
      "step": 56000
    },
    {
      "epoch": 10.355029585798816,
      "eval_accuracy": 0.8517786561264822,
      "eval_f1": 0.8643503093764874,
      "eval_loss": 0.3132619559764862,
      "eval_precision": 0.8117289468979081,
      "eval_recall": 0.9242671009771987,
      "eval_runtime": 71.3488,
      "eval_samples_per_second": 134.746,
      "eval_steps_per_second": 8.423,
      "step": 56000
    },
    {
      "epoch": 10.539940828402367,
      "grad_norm": 6.816524028778076,
      "learning_rate": 1.0511176857330705e-06,
      "loss": 0.2257,
      "step": 57000
    },
    {
      "epoch": 10.539940828402367,
      "eval_accuracy": 0.8545870605367173,
      "eval_f1": 0.8601720344068814,
      "eval_loss": 0.30009618401527405,
      "eval_precision": 0.8454581203303185,
      "eval_recall": 0.8754071661237784,
      "eval_runtime": 71.2356,
      "eval_samples_per_second": 134.961,
      "eval_steps_per_second": 8.437,
      "step": 57000
    },
    {
      "epoch": 10.724852071005918,
      "grad_norm": 20.463912963867188,
      "learning_rate": 1.0305719921104537e-06,
      "loss": 0.2199,
      "step": 58000
    },
    {
      "epoch": 10.724852071005918,
      "eval_accuracy": 0.8569794050343249,
      "eval_f1": 0.8605900841528946,
      "eval_loss": 0.3028598725795746,
      "eval_precision": 0.8572005655423147,
      "eval_recall": 0.8640065146579805,
      "eval_runtime": 71.2275,
      "eval_samples_per_second": 134.976,
      "eval_steps_per_second": 8.438,
      "step": 58000
    },
    {
      "epoch": 10.909763313609467,
      "grad_norm": 2.745788335800171,
      "learning_rate": 1.010026298487837e-06,
      "loss": 0.2235,
      "step": 59000
    },
    {
      "epoch": 10.909763313609467,
      "eval_accuracy": 0.8582275847722072,
      "eval_f1": 0.8604484488584008,
      "eval_loss": 0.3084582984447479,
      "eval_precision": 0.8654994850669413,
      "eval_recall": 0.8554560260586319,
      "eval_runtime": 71.3104,
      "eval_samples_per_second": 134.819,
      "eval_steps_per_second": 8.428,
      "step": 59000
    },
    {
      "epoch": 11.094674556213018,
      "grad_norm": 34.805477142333984,
      "learning_rate": 9.894806048652202e-07,
      "loss": 0.2149,
      "step": 60000
    },
    {
      "epoch": 11.094674556213018,
      "eval_accuracy": 0.856251300187227,
      "eval_f1": 0.8660073686251697,
      "eval_loss": 0.3204440772533417,
      "eval_precision": 0.8267308404294705,
      "eval_recall": 0.9092019543973942,
      "eval_runtime": 71.3919,
      "eval_samples_per_second": 134.665,
      "eval_steps_per_second": 8.418,
      "step": 60000
    },
    {
      "epoch": 11.279585798816568,
      "grad_norm": 0.9099503755569458,
      "learning_rate": 9.689349112426034e-07,
      "loss": 0.2105,
      "step": 61000
    },
    {
      "epoch": 11.279585798816568,
      "eval_accuracy": 0.8560432702309132,
      "eval_f1": 0.8661249758173728,
      "eval_loss": 0.3223714828491211,
      "eval_precision": 0.8251013638039071,
      "eval_recall": 0.9114413680781759,
      "eval_runtime": 71.3086,
      "eval_samples_per_second": 134.822,
      "eval_steps_per_second": 8.428,
      "step": 61000
    },
    {
      "epoch": 11.464497041420119,
      "grad_norm": 44.76268005371094,
      "learning_rate": 9.483892176199868e-07,
      "loss": 0.2196,
      "step": 62000
    },
    {
      "epoch": 11.464497041420119,
      "eval_accuracy": 0.853338880798835,
      "eval_f1": 0.8583199356913184,
      "eval_loss": 0.3199668824672699,
      "eval_precision": 0.8474206349206349,
      "eval_recall": 0.8695032573289903,
      "eval_runtime": 71.2065,
      "eval_samples_per_second": 135.016,
      "eval_steps_per_second": 8.44,
      "step": 62000
    },
    {
      "epoch": 11.649408284023668,
      "grad_norm": 5.310791969299316,
      "learning_rate": 9.278435239973701e-07,
      "loss": 0.2112,
      "step": 63000
    },
    {
      "epoch": 11.649408284023668,
      "eval_accuracy": 0.8580195548158935,
      "eval_f1": 0.8656363815336156,
      "eval_loss": 0.3171244263648987,
      "eval_precision": 0.8380026681913474,
      "eval_recall": 0.8951547231270358,
      "eval_runtime": 71.347,
      "eval_samples_per_second": 134.75,
      "eval_steps_per_second": 8.424,
      "step": 63000
    },
    {
      "epoch": 11.834319526627219,
      "grad_norm": 23.944461822509766,
      "learning_rate": 9.072978303747534e-07,
      "loss": 0.2157,
      "step": 64000
    },
    {
      "epoch": 11.834319526627219,
      "eval_accuracy": 0.857707509881423,
      "eval_f1": 0.8635275339185954,
      "eval_loss": 0.3211972117424011,
      "eval_precision": 0.8466353677621283,
      "eval_recall": 0.8811074918566775,
      "eval_runtime": 71.2358,
      "eval_samples_per_second": 134.96,
      "eval_steps_per_second": 8.437,
      "step": 64000
    },
    {
      "epoch": 12.01923076923077,
      "grad_norm": 10.31908893585205,
      "learning_rate": 8.867521367521367e-07,
      "loss": 0.2221,
      "step": 65000
    },
    {
      "epoch": 12.01923076923077,
      "eval_accuracy": 0.8616600790513834,
      "eval_f1": 0.8634777253130774,
      "eval_loss": 0.3090285360813141,
      "eval_precision": 0.8708074534161491,
      "eval_recall": 0.8562703583061889,
      "eval_runtime": 71.285,
      "eval_samples_per_second": 134.867,
      "eval_steps_per_second": 8.431,
      "step": 65000
    },
    {
      "epoch": 12.20414201183432,
      "grad_norm": 10.808996200561523,
      "learning_rate": 8.6620644312952e-07,
      "loss": 0.2122,
      "step": 66000
    },
    {
      "epoch": 12.20414201183432,
      "eval_accuracy": 0.8554191803619722,
      "eval_f1": 0.8607214428857716,
      "eval_loss": 0.32709065079689026,
      "eval_precision": 0.8474743488555643,
      "eval_recall": 0.8743892508143323,
      "eval_runtime": 71.3543,
      "eval_samples_per_second": 134.736,
      "eval_steps_per_second": 8.423,
      "step": 66000
    },
    {
      "epoch": 12.38905325443787,
      "grad_norm": 36.5963020324707,
      "learning_rate": 8.456607495069034e-07,
      "loss": 0.2087,
      "step": 67000
    },
    {
      "epoch": 12.38905325443787,
      "eval_accuracy": 0.8594757645100894,
      "eval_f1": 0.8617619973396091,
      "eval_loss": 0.31520143151283264,
      "eval_precision": 0.8662826578893232,
      "eval_recall": 0.8572882736156352,
      "eval_runtime": 71.4929,
      "eval_samples_per_second": 134.475,
      "eval_steps_per_second": 8.406,
      "step": 67000
    },
    {
      "epoch": 12.57396449704142,
      "grad_norm": 69.17304229736328,
      "learning_rate": 8.251150558842866e-07,
      "loss": 0.2053,
      "step": 68000
    },
    {
      "epoch": 12.57396449704142,
      "eval_accuracy": 0.857603494903266,
      "eval_f1": 0.8651630060080764,
      "eval_loss": 0.32987838983535767,
      "eval_precision": 0.8380080137378363,
      "eval_recall": 0.8941368078175895,
      "eval_runtime": 71.5274,
      "eval_samples_per_second": 134.41,
      "eval_steps_per_second": 8.402,
      "step": 68000
    },
    {
      "epoch": 12.75887573964497,
      "grad_norm": 22.708887100219727,
      "learning_rate": 8.045693622616699e-07,
      "loss": 0.2124,
      "step": 69000
    },
    {
      "epoch": 12.75887573964497,
      "eval_accuracy": 0.8570834200124818,
      "eval_f1": 0.8667054714784633,
      "eval_loss": 0.33354252576828003,
      "eval_precision": 0.8278354336545589,
      "eval_recall": 0.9094055374592834,
      "eval_runtime": 71.3908,
      "eval_samples_per_second": 134.667,
      "eval_steps_per_second": 8.418,
      "step": 69000
    },
    {
      "epoch": 12.94378698224852,
      "grad_norm": 20.74620246887207,
      "learning_rate": 7.840236686390532e-07,
      "loss": 0.2086,
      "step": 70000
    },
    {
      "epoch": 12.94378698224852,
      "eval_accuracy": 0.8591637195756189,
      "eval_f1": 0.8646,
      "eval_loss": 0.32121601700782776,
      "eval_precision": 0.8496462264150944,
      "eval_recall": 0.8800895765472313,
      "eval_runtime": 71.264,
      "eval_samples_per_second": 134.907,
      "eval_steps_per_second": 8.433,
      "step": 70000
    },
    {
      "epoch": 13.128698224852071,
      "grad_norm": 191.69613647460938,
      "learning_rate": 7.634779750164365e-07,
      "loss": 0.2016,
      "step": 71000
    },
    {
      "epoch": 13.128698224852071,
      "eval_accuracy": 0.8614520490950697,
      "eval_f1": 0.8714781937475878,
      "eval_loss": 0.32725852727890015,
      "eval_precision": 0.8283198826118855,
      "eval_recall": 0.9193811074918566,
      "eval_runtime": 71.3109,
      "eval_samples_per_second": 134.818,
      "eval_steps_per_second": 8.428,
      "step": 71000
    },
    {
      "epoch": 13.31360946745562,
      "grad_norm": 9.116983413696289,
      "learning_rate": 7.429322813938198e-07,
      "loss": 0.2026,
      "step": 72000
    },
    {
      "epoch": 13.31360946745562,
      "eval_accuracy": 0.8582275847722072,
      "eval_f1": 0.8682455292411793,
      "eval_loss": 0.3508372902870178,
      "eval_precision": 0.8266151297625621,
      "eval_recall": 0.9142915309446255,
      "eval_runtime": 71.2132,
      "eval_samples_per_second": 135.003,
      "eval_steps_per_second": 8.439,
      "step": 72000
    },
    {
      "epoch": 13.498520710059172,
      "grad_norm": 4.702348232269287,
      "learning_rate": 7.223865877712031e-07,
      "loss": 0.2006,
      "step": 73000
    },
    {
      "epoch": 13.498520710059172,
      "eval_accuracy": 0.8582275847722072,
      "eval_f1": 0.8654624420096733,
      "eval_loss": 0.34033337235450745,
      "eval_precision": 0.8400076643035064,
      "eval_recall": 0.8925081433224755,
      "eval_runtime": 71.1462,
      "eval_samples_per_second": 135.13,
      "eval_steps_per_second": 8.447,
      "step": 73000
    },
    {
      "epoch": 13.683431952662723,
      "grad_norm": 29.501680374145508,
      "learning_rate": 7.018408941485864e-07,
      "loss": 0.2064,
      "step": 74000
    },
    {
      "epoch": 13.683431952662723,
      "eval_accuracy": 0.8606199292698149,
      "eval_f1": 0.863876472978464,
      "eval_loss": 0.3424365222454071,
      "eval_precision": 0.862124898621249,
      "eval_recall": 0.8656351791530945,
      "eval_runtime": 71.1345,
      "eval_samples_per_second": 135.152,
      "eval_steps_per_second": 8.449,
      "step": 74000
    },
    {
      "epoch": 13.868343195266272,
      "grad_norm": 2.9182920455932617,
      "learning_rate": 6.812952005259697e-07,
      "loss": 0.2071,
      "step": 75000
    },
    {
      "epoch": 13.868343195266272,
      "eval_accuracy": 0.8584356147285209,
      "eval_f1": 0.8673101296675441,
      "eval_loss": 0.33476874232292175,
      "eval_precision": 0.832179607109448,
      "eval_recall": 0.9055374592833876,
      "eval_runtime": 71.1759,
      "eval_samples_per_second": 135.074,
      "eval_steps_per_second": 8.444,
      "step": 75000
    },
    {
      "epoch": 14.053254437869823,
      "grad_norm": 27.45977020263672,
      "learning_rate": 6.607495069033531e-07,
      "loss": 0.1996,
      "step": 76000
    },
    {
      "epoch": 14.053254437869823,
      "eval_accuracy": 0.8590597045974621,
      "eval_f1": 0.86776617546599,
      "eval_loss": 0.32783225178718567,
      "eval_precision": 0.8333645735707591,
      "eval_recall": 0.9051302931596091,
      "eval_runtime": 71.2136,
      "eval_samples_per_second": 135.002,
      "eval_steps_per_second": 8.439,
      "step": 76000
    },
    {
      "epoch": 14.238165680473372,
      "grad_norm": 28.44346046447754,
      "learning_rate": 6.402038132807364e-07,
      "loss": 0.1946,
      "step": 77000
    },
    {
      "epoch": 14.238165680473372,
      "eval_accuracy": 0.8610359891824423,
      "eval_f1": 0.8730279414559969,
      "eval_loss": 0.33918145298957825,
      "eval_precision": 0.818716577540107,
      "eval_recall": 0.935057003257329,
      "eval_runtime": 71.1883,
      "eval_samples_per_second": 135.05,
      "eval_steps_per_second": 8.442,
      "step": 77000
    }
  ],
  "logging_steps": 1000,
  "max_steps": 108160,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "save_steps": 1000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2.5466830173511814e+17,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}