{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 19.99770484278173,
  "eval_steps": 500,
  "global_step": 21780,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.09180628873077806,
      "grad_norm": 0.26081690192222595,
      "learning_rate": 0.0001,
      "loss": 1.6526,
      "step": 100
    },
    {
      "epoch": 0.18361257746155613,
      "grad_norm": 0.25705486536026,
      "learning_rate": 0.0001,
      "loss": 1.5908,
      "step": 200
    },
    {
      "epoch": 0.2754188661923342,
      "grad_norm": 0.2516353726387024,
      "learning_rate": 0.0001,
      "loss": 1.5712,
      "step": 300
    },
    {
      "epoch": 0.36722515492311225,
      "grad_norm": 0.22883176803588867,
      "learning_rate": 0.0001,
      "loss": 1.5768,
      "step": 400
    },
    {
      "epoch": 0.4590314436538903,
      "grad_norm": 0.236283078789711,
      "learning_rate": 0.0001,
      "loss": 1.5744,
      "step": 500
    },
    {
      "epoch": 0.5508377323846684,
      "grad_norm": 0.231370747089386,
      "learning_rate": 0.0001,
      "loss": 1.5722,
      "step": 600
    },
    {
      "epoch": 0.6426440211154464,
      "grad_norm": 0.24125300347805023,
      "learning_rate": 0.0001,
      "loss": 1.5736,
      "step": 700
    },
    {
      "epoch": 0.7344503098462245,
      "grad_norm": 0.2657057046890259,
      "learning_rate": 0.0001,
      "loss": 1.5831,
      "step": 800
    },
    {
      "epoch": 0.8262565985770025,
      "grad_norm": 0.27335691452026367,
      "learning_rate": 0.0001,
      "loss": 1.5521,
      "step": 900
    },
    {
      "epoch": 0.9180628873077806,
      "grad_norm": 0.2826825976371765,
      "learning_rate": 0.0001,
      "loss": 1.5635,
      "step": 1000
    },
    {
      "epoch": 0.999770484278173,
      "eval_accuracy": 0.6795836972343523,
      "eval_loss": 1.4614675045013428,
      "eval_runtime": 9.1641,
      "eval_samples_per_second": 54.561,
      "eval_steps_per_second": 6.875,
      "step": 1089
    },
    {
      "epoch": 1.0098691760385587,
      "grad_norm": 0.24073347449302673,
      "learning_rate": 0.0001,
      "loss": 1.5475,
      "step": 1100
    },
    {
      "epoch": 1.1016754647693368,
      "grad_norm": 0.3140055239200592,
      "learning_rate": 0.0001,
      "loss": 1.4788,
      "step": 1200
    },
    {
      "epoch": 1.1934817535001148,
      "grad_norm": 0.3724195659160614,
      "learning_rate": 0.0001,
      "loss": 1.4698,
      "step": 1300
    },
    {
      "epoch": 1.2852880422308928,
      "grad_norm": 0.34302350878715515,
      "learning_rate": 0.0001,
      "loss": 1.4629,
      "step": 1400
    },
    {
      "epoch": 1.377094330961671,
      "grad_norm": 0.35881391167640686,
      "learning_rate": 0.0001,
      "loss": 1.4596,
      "step": 1500
    },
    {
      "epoch": 1.468900619692449,
      "grad_norm": 0.3676307797431946,
      "learning_rate": 0.0001,
      "loss": 1.4718,
      "step": 1600
    },
    {
      "epoch": 1.560706908423227,
      "grad_norm": 0.3709953725337982,
      "learning_rate": 0.0001,
      "loss": 1.435,
      "step": 1700
    },
    {
      "epoch": 1.652513197154005,
      "grad_norm": 0.38531753420829773,
      "learning_rate": 0.0001,
      "loss": 1.4553,
      "step": 1800
    },
    {
      "epoch": 1.744319485884783,
      "grad_norm": 0.40058839321136475,
      "learning_rate": 0.0001,
      "loss": 1.4444,
      "step": 1900
    },
    {
      "epoch": 1.836125774615561,
      "grad_norm": 0.4059107303619385,
      "learning_rate": 0.0001,
      "loss": 1.4381,
      "step": 2000
    },
    {
      "epoch": 1.9279320633463393,
      "grad_norm": 0.4201526939868927,
      "learning_rate": 0.0001,
      "loss": 1.4521,
      "step": 2100
    },
    {
      "epoch": 1.999540968556346,
      "eval_accuracy": 0.6873740902474527,
      "eval_loss": 1.362550973892212,
      "eval_runtime": 9.1565,
      "eval_samples_per_second": 54.606,
      "eval_steps_per_second": 6.88,
      "step": 2178
    },
    {
      "epoch": 2.0197383520771175,
      "grad_norm": 0.397616446018219,
      "learning_rate": 0.0001,
      "loss": 1.4163,
      "step": 2200
    },
    {
      "epoch": 2.1115446408078955,
      "grad_norm": 0.4666147232055664,
      "learning_rate": 0.0001,
      "loss": 1.3015,
      "step": 2300
    },
    {
      "epoch": 2.2033509295386735,
      "grad_norm": 0.5091608762741089,
      "learning_rate": 0.0001,
      "loss": 1.3117,
      "step": 2400
    },
    {
      "epoch": 2.2951572182694515,
      "grad_norm": 0.44425851106643677,
      "learning_rate": 0.0001,
      "loss": 1.3042,
      "step": 2500
    },
    {
      "epoch": 2.3869635070002295,
      "grad_norm": 0.4947376251220703,
      "learning_rate": 0.0001,
      "loss": 1.3091,
      "step": 2600
    },
    {
      "epoch": 2.4787697957310075,
      "grad_norm": 0.49756625294685364,
      "learning_rate": 0.0001,
      "loss": 1.2878,
      "step": 2700
    },
    {
      "epoch": 2.5705760844617855,
      "grad_norm": 0.48819100856781006,
      "learning_rate": 0.0001,
      "loss": 1.3036,
      "step": 2800
    },
    {
      "epoch": 2.6623823731925635,
      "grad_norm": 0.49992629885673523,
      "learning_rate": 0.0001,
      "loss": 1.29,
      "step": 2900
    },
    {
      "epoch": 2.754188661923342,
      "grad_norm": 0.5537226796150208,
      "learning_rate": 0.0001,
      "loss": 1.3049,
      "step": 3000
    },
    {
      "epoch": 2.84599495065412,
      "grad_norm": 0.5161275267601013,
      "learning_rate": 0.0001,
      "loss": 1.2796,
      "step": 3100
    },
    {
      "epoch": 2.937801239384898,
      "grad_norm": 0.5615408420562744,
      "learning_rate": 0.0001,
      "loss": 1.2848,
      "step": 3200
    },
    {
      "epoch": 2.9993114528345193,
      "eval_accuracy": 0.6957729257641921,
      "eval_loss": 1.257521390914917,
      "eval_runtime": 9.264,
      "eval_samples_per_second": 53.972,
      "eval_steps_per_second": 6.801,
      "step": 3267
    },
    {
      "epoch": 3.029607528115676,
      "grad_norm": 0.6211069226264954,
      "learning_rate": 0.0001,
      "loss": 1.2349,
      "step": 3300
    },
    {
      "epoch": 3.121413816846454,
      "grad_norm": 0.6274811029434204,
      "learning_rate": 0.0001,
      "loss": 1.1362,
      "step": 3400
    },
    {
      "epoch": 3.213220105577232,
      "grad_norm": 0.7168062925338745,
      "learning_rate": 0.0001,
      "loss": 1.1299,
      "step": 3500
    },
    {
      "epoch": 3.30502639430801,
      "grad_norm": 0.6573987603187561,
      "learning_rate": 0.0001,
      "loss": 1.153,
      "step": 3600
    },
    {
      "epoch": 3.396832683038788,
      "grad_norm": 0.702870786190033,
      "learning_rate": 0.0001,
      "loss": 1.1402,
      "step": 3700
    },
    {
      "epoch": 3.488638971769566,
      "grad_norm": 0.6937388181686401,
      "learning_rate": 0.0001,
      "loss": 1.1344,
      "step": 3800
    },
    {
      "epoch": 3.580445260500344,
      "grad_norm": 0.705838680267334,
      "learning_rate": 0.0001,
      "loss": 1.125,
      "step": 3900
    },
    {
      "epoch": 3.672251549231122,
      "grad_norm": 0.8442272543907166,
      "learning_rate": 0.0001,
      "loss": 1.1423,
      "step": 4000
    },
    {
      "epoch": 3.7640578379619005,
      "grad_norm": 0.9211050868034363,
      "learning_rate": 0.0001,
      "loss": 1.1227,
      "step": 4100
    },
    {
      "epoch": 3.8558641266926785,
      "grad_norm": 0.6930621862411499,
      "learning_rate": 0.0001,
      "loss": 1.1286,
      "step": 4200
    },
    {
      "epoch": 3.9476704154234565,
      "grad_norm": 0.6763383746147156,
      "learning_rate": 0.0001,
      "loss": 1.1197,
      "step": 4300
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7054410480349345,
      "eval_loss": 1.1526687145233154,
      "eval_runtime": 8.1013,
      "eval_samples_per_second": 61.718,
      "eval_steps_per_second": 7.777,
      "step": 4357
    },
    {
      "epoch": 4.039476704154235,
      "grad_norm": 0.6891958713531494,
      "learning_rate": 0.0001,
      "loss": 1.0562,
      "step": 4400
    },
    {
      "epoch": 4.131282992885013,
      "grad_norm": 0.7408663630485535,
      "learning_rate": 0.0001,
      "loss": 0.9585,
      "step": 4500
    },
    {
      "epoch": 4.223089281615791,
      "grad_norm": 0.8520354628562927,
      "learning_rate": 0.0001,
      "loss": 0.9652,
      "step": 4600
    },
    {
      "epoch": 4.314895570346569,
      "grad_norm": 0.8522772789001465,
      "learning_rate": 0.0001,
      "loss": 0.9819,
      "step": 4700
    },
    {
      "epoch": 4.406701859077347,
      "grad_norm": 0.8211854696273804,
      "learning_rate": 0.0001,
      "loss": 0.9777,
      "step": 4800
    },
    {
      "epoch": 4.498508147808125,
      "grad_norm": 0.8455579280853271,
      "learning_rate": 0.0001,
      "loss": 0.9748,
      "step": 4900
    },
    {
      "epoch": 4.590314436538903,
      "grad_norm": 0.9336457848548889,
      "learning_rate": 0.0001,
      "loss": 0.9806,
      "step": 5000
    },
    {
      "epoch": 4.682120725269681,
      "grad_norm": 0.8030388355255127,
      "learning_rate": 0.0001,
      "loss": 0.977,
      "step": 5100
    },
    {
      "epoch": 4.773927014000459,
      "grad_norm": 0.8392836451530457,
      "learning_rate": 0.0001,
      "loss": 0.9773,
      "step": 5200
    },
    {
      "epoch": 4.865733302731237,
      "grad_norm": 0.823242723941803,
      "learning_rate": 0.0001,
      "loss": 0.9776,
      "step": 5300
    },
    {
      "epoch": 4.957539591462015,
      "grad_norm": 0.9073436260223389,
      "learning_rate": 0.0001,
      "loss": 0.9756,
      "step": 5400
    },
    {
      "epoch": 4.999770484278173,
      "eval_accuracy": 0.7142823871906842,
      "eval_loss": 1.0531576871871948,
      "eval_runtime": 9.2013,
      "eval_samples_per_second": 54.34,
      "eval_steps_per_second": 6.847,
      "step": 5446
    },
    {
      "epoch": 5.049345880192793,
      "grad_norm": 1.1068471670150757,
      "learning_rate": 0.0001,
      "loss": 0.887,
      "step": 5500
    },
    {
      "epoch": 5.141152168923571,
      "grad_norm": 0.9577746987342834,
      "learning_rate": 0.0001,
      "loss": 0.8052,
      "step": 5600
    },
    {
      "epoch": 5.232958457654349,
      "grad_norm": 0.9153982996940613,
      "learning_rate": 0.0001,
      "loss": 0.8118,
      "step": 5700
    },
    {
      "epoch": 5.324764746385127,
      "grad_norm": 1.0308676958084106,
      "learning_rate": 0.0001,
      "loss": 0.8332,
      "step": 5800
    },
    {
      "epoch": 5.416571035115905,
      "grad_norm": 0.9150503873825073,
      "learning_rate": 0.0001,
      "loss": 0.8242,
      "step": 5900
    },
    {
      "epoch": 5.508377323846684,
      "grad_norm": 1.0191842317581177,
      "learning_rate": 0.0001,
      "loss": 0.8323,
      "step": 6000
    },
    {
      "epoch": 5.600183612577462,
      "grad_norm": 1.1198716163635254,
      "learning_rate": 0.0001,
      "loss": 0.8301,
      "step": 6100
    },
    {
      "epoch": 5.69198990130824,
      "grad_norm": 0.996842622756958,
      "learning_rate": 0.0001,
      "loss": 0.84,
      "step": 6200
    },
    {
      "epoch": 5.783796190039018,
      "grad_norm": 1.086377739906311,
      "learning_rate": 0.0001,
      "loss": 0.8414,
      "step": 6300
    },
    {
      "epoch": 5.875602478769796,
      "grad_norm": 0.9792770147323608,
      "learning_rate": 0.0001,
      "loss": 0.8277,
      "step": 6400
    },
    {
      "epoch": 5.967408767500574,
      "grad_norm": 1.0763967037200928,
      "learning_rate": 0.0001,
      "loss": 0.8393,
      "step": 6500
    },
    {
      "epoch": 5.999540968556346,
      "eval_accuracy": 0.7241455604075692,
      "eval_loss": 0.9538469314575195,
      "eval_runtime": 8.1157,
      "eval_samples_per_second": 61.609,
      "eval_steps_per_second": 7.763,
      "step": 6535
    },
    {
      "epoch": 6.059215056231352,
      "grad_norm": 1.08652663230896,
      "learning_rate": 0.0001,
      "loss": 0.729,
      "step": 6600
    },
    {
      "epoch": 6.15102134496213,
      "grad_norm": 0.9392278790473938,
      "learning_rate": 0.0001,
      "loss": 0.6772,
      "step": 6700
    },
    {
      "epoch": 6.242827633692908,
      "grad_norm": 1.126567006111145,
      "learning_rate": 0.0001,
      "loss": 0.6904,
      "step": 6800
    },
    {
      "epoch": 6.334633922423686,
      "grad_norm": 1.0995755195617676,
      "learning_rate": 0.0001,
      "loss": 0.69,
      "step": 6900
    },
    {
      "epoch": 6.426440211154464,
      "grad_norm": 0.983116090297699,
      "learning_rate": 0.0001,
      "loss": 0.6962,
      "step": 7000
    },
    {
      "epoch": 6.518246499885242,
      "grad_norm": 1.2054848670959473,
      "learning_rate": 0.0001,
      "loss": 0.7061,
      "step": 7100
    },
    {
      "epoch": 6.61005278861602,
      "grad_norm": 1.2262558937072754,
      "learning_rate": 0.0001,
      "loss": 0.71,
      "step": 7200
    },
    {
      "epoch": 6.701859077346798,
      "grad_norm": 1.069161295890808,
      "learning_rate": 0.0001,
      "loss": 0.7003,
      "step": 7300
    },
    {
      "epoch": 6.793665366077576,
      "grad_norm": 1.2181183099746704,
      "learning_rate": 0.0001,
      "loss": 0.7024,
      "step": 7400
    },
    {
      "epoch": 6.885471654808354,
      "grad_norm": 1.0989437103271484,
      "learning_rate": 0.0001,
      "loss": 0.7002,
      "step": 7500
    },
    {
      "epoch": 6.977277943539132,
      "grad_norm": 1.1168180704116821,
      "learning_rate": 0.0001,
      "loss": 0.7125,
      "step": 7600
    },
    {
      "epoch": 6.999311452834519,
      "eval_accuracy": 0.7324425036390102,
      "eval_loss": 0.8674135208129883,
      "eval_runtime": 9.2169,
      "eval_samples_per_second": 54.248,
      "eval_steps_per_second": 6.835,
      "step": 7624
    },
    {
      "epoch": 7.06908423226991,
      "grad_norm": 1.0699574947357178,
      "learning_rate": 0.0001,
      "loss": 0.6029,
      "step": 7700
    },
    {
      "epoch": 7.160890521000688,
      "grad_norm": 0.9937657713890076,
      "learning_rate": 0.0001,
      "loss": 0.5805,
      "step": 7800
    },
    {
      "epoch": 7.252696809731467,
      "grad_norm": 1.178791880607605,
      "learning_rate": 0.0001,
      "loss": 0.5752,
      "step": 7900
    },
    {
      "epoch": 7.344503098462245,
      "grad_norm": 1.2159409523010254,
      "learning_rate": 0.0001,
      "loss": 0.5912,
      "step": 8000
    },
    {
      "epoch": 7.436309387193023,
      "grad_norm": 1.0133622884750366,
      "learning_rate": 0.0001,
      "loss": 0.5932,
      "step": 8100
    },
    {
      "epoch": 7.528115675923801,
      "grad_norm": 1.0923631191253662,
      "learning_rate": 0.0001,
      "loss": 0.6071,
      "step": 8200
    },
    {
      "epoch": 7.619921964654579,
      "grad_norm": 1.3819491863250732,
      "learning_rate": 0.0001,
      "loss": 0.5956,
      "step": 8300
    },
    {
      "epoch": 7.711728253385357,
      "grad_norm": 1.182358980178833,
      "learning_rate": 0.0001,
      "loss": 0.5922,
      "step": 8400
    },
    {
      "epoch": 7.803534542116135,
      "grad_norm": 1.1674267053604126,
      "learning_rate": 0.0001,
      "loss": 0.5912,
      "step": 8500
    },
    {
      "epoch": 7.895340830846913,
      "grad_norm": 1.1732617616653442,
      "learning_rate": 0.0001,
      "loss": 0.5969,
      "step": 8600
    },
    {
      "epoch": 7.987147119577691,
      "grad_norm": 1.2167391777038574,
      "learning_rate": 0.0001,
      "loss": 0.6144,
      "step": 8700
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.7404046579330422,
      "eval_loss": 0.7907233834266663,
      "eval_runtime": 9.1451,
      "eval_samples_per_second": 54.674,
      "eval_steps_per_second": 6.889,
      "step": 8714
    },
    {
      "epoch": 8.07895340830847,
      "grad_norm": 1.1496325731277466,
      "learning_rate": 0.0001,
      "loss": 0.5043,
      "step": 8800
    },
    {
      "epoch": 8.170759697039248,
      "grad_norm": 1.4953396320343018,
      "learning_rate": 0.0001,
      "loss": 0.4848,
      "step": 8900
    },
    {
      "epoch": 8.262565985770026,
      "grad_norm": 1.2796908617019653,
      "learning_rate": 0.0001,
      "loss": 0.5007,
      "step": 9000
    },
    {
      "epoch": 8.354372274500804,
      "grad_norm": 1.2108944654464722,
      "learning_rate": 0.0001,
      "loss": 0.4987,
      "step": 9100
    },
    {
      "epoch": 8.446178563231582,
      "grad_norm": 0.9534372687339783,
      "learning_rate": 0.0001,
      "loss": 0.5068,
      "step": 9200
    },
    {
      "epoch": 8.53798485196236,
      "grad_norm": 1.1545357704162598,
      "learning_rate": 0.0001,
      "loss": 0.5072,
      "step": 9300
    },
    {
      "epoch": 8.629791140693138,
      "grad_norm": 1.2086093425750732,
      "learning_rate": 0.0001,
      "loss": 0.5173,
      "step": 9400
    },
    {
      "epoch": 8.721597429423916,
      "grad_norm": 1.20607328414917,
      "learning_rate": 0.0001,
      "loss": 0.519,
      "step": 9500
    },
    {
      "epoch": 8.813403718154694,
      "grad_norm": 1.2534675598144531,
      "learning_rate": 0.0001,
      "loss": 0.5261,
      "step": 9600
    },
    {
      "epoch": 8.905210006885472,
      "grad_norm": 1.2726677656173706,
      "learning_rate": 0.0001,
      "loss": 0.5213,
      "step": 9700
    },
    {
      "epoch": 8.99701629561625,
      "grad_norm": 1.28297758102417,
      "learning_rate": 0.0001,
      "loss": 0.5355,
      "step": 9800
    },
    {
      "epoch": 8.999770484278173,
      "eval_accuracy": 0.7468762736535662,
      "eval_loss": 0.7288308143615723,
      "eval_runtime": 9.1574,
      "eval_samples_per_second": 54.601,
      "eval_steps_per_second": 6.88,
      "step": 9803
    },
    {
      "epoch": 9.088822584347028,
      "grad_norm": 1.278200626373291,
      "learning_rate": 0.0001,
      "loss": 0.4247,
      "step": 9900
    },
    {
      "epoch": 9.180628873077806,
      "grad_norm": 1.3318266868591309,
      "learning_rate": 0.0001,
      "loss": 0.4284,
      "step": 10000
    },
    {
      "epoch": 9.272435161808584,
      "grad_norm": 1.209088683128357,
      "learning_rate": 0.0001,
      "loss": 0.4381,
      "step": 10100
    },
    {
      "epoch": 9.364241450539362,
      "grad_norm": 1.0169490575790405,
      "learning_rate": 0.0001,
      "loss": 0.443,
      "step": 10200
    },
    {
      "epoch": 9.45604773927014,
      "grad_norm": 1.4842835664749146,
      "learning_rate": 0.0001,
      "loss": 0.4424,
      "step": 10300
    },
    {
      "epoch": 9.547854028000918,
      "grad_norm": 1.1761025190353394,
      "learning_rate": 0.0001,
      "loss": 0.4463,
      "step": 10400
    },
    {
      "epoch": 9.639660316731696,
      "grad_norm": 1.270493984222412,
      "learning_rate": 0.0001,
      "loss": 0.4541,
      "step": 10500
    },
    {
      "epoch": 9.731466605462474,
      "grad_norm": 1.346306562423706,
      "learning_rate": 0.0001,
      "loss": 0.4551,
      "step": 10600
    },
    {
      "epoch": 9.823272894193252,
      "grad_norm": 1.2559789419174194,
      "learning_rate": 0.0001,
      "loss": 0.4611,
      "step": 10700
    },
    {
      "epoch": 9.91507918292403,
      "grad_norm": 1.4359290599822998,
      "learning_rate": 0.0001,
      "loss": 0.4584,
      "step": 10800
    },
    {
      "epoch": 9.997704842781731,
      "eval_accuracy": 0.75309461426492,
      "eval_loss": 0.6794138550758362,
      "eval_runtime": 8.1091,
      "eval_samples_per_second": 61.659,
      "eval_steps_per_second": 7.769,
      "step": 10890
    },
    {
      "epoch": 10.009180628873079,
      "grad_norm": 1.3423351049423218,
      "learning_rate": 0.0001,
      "loss": 0.386,
      "step": 10900
    },
    {
      "epoch": 10.100986917603857,
      "grad_norm": 1.132673740386963,
      "learning_rate": 0.0001,
      "loss": 0.3726,
      "step": 11000
    },
    {
      "epoch": 10.192793206334635,
      "grad_norm": 1.614931344985962,
      "learning_rate": 0.0001,
      "loss": 0.3819,
      "step": 11100
    },
    {
      "epoch": 10.284599495065413,
      "grad_norm": 1.3352196216583252,
      "learning_rate": 0.0001,
      "loss": 0.3876,
      "step": 11200
    },
    {
      "epoch": 10.37640578379619,
      "grad_norm": 1.0957690477371216,
      "learning_rate": 0.0001,
      "loss": 0.3973,
      "step": 11300
    },
    {
      "epoch": 10.468212072526969,
      "grad_norm": 1.142330527305603,
      "learning_rate": 0.0001,
      "loss": 0.396,
      "step": 11400
    },
    {
      "epoch": 10.560018361257747,
      "grad_norm": 1.4076579809188843,
      "learning_rate": 0.0001,
      "loss": 0.402,
      "step": 11500
    },
    {
      "epoch": 10.651824649988525,
      "grad_norm": 1.389333963394165,
      "learning_rate": 0.0001,
      "loss": 0.3962,
      "step": 11600
    },
    {
      "epoch": 10.743630938719303,
      "grad_norm": 1.4440951347351074,
      "learning_rate": 0.0001,
      "loss": 0.4049,
      "step": 11700
    },
    {
      "epoch": 10.83543722745008,
      "grad_norm": 1.4290118217468262,
      "learning_rate": 0.0001,
      "loss": 0.4093,
      "step": 11800
    },
    {
      "epoch": 10.927243516180859,
      "grad_norm": 1.46366548538208,
      "learning_rate": 0.0001,
      "loss": 0.413,
      "step": 11900
    },
    {
      "epoch": 10.999770484278173,
      "eval_accuracy": 0.7576768558951965,
      "eval_loss": 0.6291825175285339,
      "eval_runtime": 8.951,
      "eval_samples_per_second": 55.86,
      "eval_steps_per_second": 7.038,
      "step": 11979
    },
    {
      "epoch": 11.019049804911637,
      "grad_norm": 1.1713697910308838,
      "learning_rate": 0.0001,
      "loss": 0.395,
      "step": 12000
    },
    {
      "epoch": 11.110856093642415,
      "grad_norm": 1.277626395225525,
      "learning_rate": 0.0001,
      "loss": 0.3355,
      "step": 12100
    },
    {
      "epoch": 11.202662382373193,
      "grad_norm": 1.3597822189331055,
      "learning_rate": 0.0001,
      "loss": 0.3412,
      "step": 12200
    },
    {
      "epoch": 11.29446867110397,
      "grad_norm": 1.4017976522445679,
      "learning_rate": 0.0001,
      "loss": 0.3414,
      "step": 12300
    },
    {
      "epoch": 11.386274959834749,
      "grad_norm": 1.409915804862976,
      "learning_rate": 0.0001,
      "loss": 0.3558,
      "step": 12400
    },
    {
      "epoch": 11.478081248565527,
      "grad_norm": 1.400634765625,
      "learning_rate": 0.0001,
      "loss": 0.3577,
      "step": 12500
    },
    {
      "epoch": 11.569887537296305,
      "grad_norm": 1.5898892879486084,
      "learning_rate": 0.0001,
      "loss": 0.354,
      "step": 12600
    },
    {
      "epoch": 11.661693826027083,
      "grad_norm": 1.3252007961273193,
      "learning_rate": 0.0001,
      "loss": 0.3682,
      "step": 12700
    },
    {
      "epoch": 11.75350011475786,
      "grad_norm": 1.302128791809082,
      "learning_rate": 0.0001,
      "loss": 0.3715,
      "step": 12800
    },
    {
      "epoch": 11.845306403488639,
      "grad_norm": 1.3374468088150024,
      "learning_rate": 0.0001,
      "loss": 0.3707,
      "step": 12900
    },
    {
      "epoch": 11.937112692219417,
      "grad_norm": 1.1755791902542114,
      "learning_rate": 0.0001,
      "loss": 0.3731,
      "step": 13000
    },
    {
      "epoch": 11.999540968556346,
      "eval_accuracy": 0.76164192139738,
      "eval_loss": 0.5926400423049927,
      "eval_runtime": 9.0463,
      "eval_samples_per_second": 55.271,
      "eval_steps_per_second": 6.964,
      "step": 13068
    },
    {
      "epoch": 12.028918980950195,
      "grad_norm": 1.3085649013519287,
      "learning_rate": 0.0001,
      "loss": 0.3482,
      "step": 13100
    },
    {
      "epoch": 12.120725269680973,
      "grad_norm": 1.1860175132751465,
      "learning_rate": 0.0001,
      "loss": 0.2982,
      "step": 13200
    },
    {
      "epoch": 12.21253155841175,
      "grad_norm": 1.1902750730514526,
      "learning_rate": 0.0001,
      "loss": 0.3095,
      "step": 13300
    },
    {
      "epoch": 12.304337847142529,
      "grad_norm": 1.2473431825637817,
      "learning_rate": 0.0001,
      "loss": 0.3216,
      "step": 13400
    },
    {
      "epoch": 12.396144135873307,
      "grad_norm": 1.443493366241455,
      "learning_rate": 0.0001,
      "loss": 0.319,
      "step": 13500
    },
    {
      "epoch": 12.487950424604085,
      "grad_norm": 1.4389948844909668,
      "learning_rate": 0.0001,
      "loss": 0.328,
      "step": 13600
    },
    {
      "epoch": 12.579756713334863,
      "grad_norm": 1.1586631536483765,
      "learning_rate": 0.0001,
      "loss": 0.3285,
      "step": 13700
    },
    {
      "epoch": 12.671563002065641,
      "grad_norm": 1.180396318435669,
      "learning_rate": 0.0001,
      "loss": 0.3311,
      "step": 13800
    },
    {
      "epoch": 12.763369290796419,
      "grad_norm": 1.4230598211288452,
      "learning_rate": 0.0001,
      "loss": 0.3346,
      "step": 13900
    },
    {
      "epoch": 12.855175579527197,
      "grad_norm": 1.5782092809677124,
      "learning_rate": 0.0001,
      "loss": 0.3415,
      "step": 14000
    },
    {
      "epoch": 12.946981868257975,
      "grad_norm": 1.418642282485962,
      "learning_rate": 0.0001,
      "loss": 0.3423,
      "step": 14100
    },
    {
      "epoch": 12.999311452834519,
      "eval_accuracy": 0.7655866084425036,
      "eval_loss": 0.5619787573814392,
      "eval_runtime": 8.9635,
      "eval_samples_per_second": 55.782,
      "eval_steps_per_second": 7.029,
      "step": 14157
    },
    {
      "epoch": 13.038788156988753,
      "grad_norm": 1.1923723220825195,
      "learning_rate": 0.0001,
      "loss": 0.3119,
      "step": 14200
    },
    {
      "epoch": 13.130594445719531,
      "grad_norm": 1.2736058235168457,
      "learning_rate": 0.0001,
      "loss": 0.2762,
      "step": 14300
    },
    {
      "epoch": 13.22240073445031,
      "grad_norm": 0.9496171474456787,
      "learning_rate": 0.0001,
      "loss": 0.2844,
      "step": 14400
    },
    {
      "epoch": 13.314207023181089,
      "grad_norm": 1.22100031375885,
      "learning_rate": 0.0001,
      "loss": 0.2938,
      "step": 14500
    },
    {
      "epoch": 13.406013311911867,
      "grad_norm": 1.381606101989746,
      "learning_rate": 0.0001,
      "loss": 0.2978,
      "step": 14600
    },
    {
      "epoch": 13.497819600642645,
      "grad_norm": 1.43625807762146,
      "learning_rate": 0.0001,
      "loss": 0.3035,
      "step": 14700
    },
    {
      "epoch": 13.589625889373423,
      "grad_norm": 1.4393320083618164,
      "learning_rate": 0.0001,
      "loss": 0.3065,
      "step": 14800
    },
    {
      "epoch": 13.6814321781042,
      "grad_norm": 1.184833288192749,
      "learning_rate": 0.0001,
      "loss": 0.3091,
      "step": 14900
    },
    {
      "epoch": 13.773238466834979,
      "grad_norm": 1.4501614570617676,
      "learning_rate": 0.0001,
      "loss": 0.3103,
      "step": 15000
    },
    {
      "epoch": 13.865044755565757,
      "grad_norm": 1.368249535560608,
      "learning_rate": 0.0001,
      "loss": 0.3137,
      "step": 15100
    },
    {
      "epoch": 13.956851044296535,
      "grad_norm": 1.4249024391174316,
      "learning_rate": 0.0001,
      "loss": 0.3185,
      "step": 15200
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.7682037845705968,
      "eval_loss": 0.542601466178894,
      "eval_runtime": 8.9503,
      "eval_samples_per_second": 55.864,
      "eval_steps_per_second": 7.039,
      "step": 15247
    },
    {
      "epoch": 14.048657333027313,
      "grad_norm": 1.3438467979431152,
      "learning_rate": 0.0001,
      "loss": 0.2849,
      "step": 15300
    },
    {
      "epoch": 14.14046362175809,
      "grad_norm": 1.4617668390274048,
      "learning_rate": 0.0001,
      "loss": 0.2658,
      "step": 15400
    },
    {
      "epoch": 14.232269910488869,
      "grad_norm": 1.266655683517456,
      "learning_rate": 0.0001,
      "loss": 0.2679,
      "step": 15500
    },
    {
      "epoch": 14.324076199219647,
      "grad_norm": 1.2162944078445435,
      "learning_rate": 0.0001,
      "loss": 0.2711,
      "step": 15600
    },
    {
      "epoch": 14.415882487950425,
      "grad_norm": 1.10415518283844,
      "learning_rate": 0.0001,
      "loss": 0.2763,
      "step": 15700
    },
    {
      "epoch": 14.507688776681203,
      "grad_norm": 1.1962913274765015,
      "learning_rate": 0.0001,
      "loss": 0.2827,
      "step": 15800
    },
    {
      "epoch": 14.59949506541198,
      "grad_norm": 1.2264560461044312,
      "learning_rate": 0.0001,
      "loss": 0.2845,
      "step": 15900
    },
    {
      "epoch": 14.691301354142759,
      "grad_norm": 1.3857085704803467,
      "learning_rate": 0.0001,
      "loss": 0.2897,
      "step": 16000
    },
    {
      "epoch": 14.783107642873537,
      "grad_norm": 1.447581171989441,
      "learning_rate": 0.0001,
      "loss": 0.2894,
      "step": 16100
    },
    {
      "epoch": 14.874913931604315,
      "grad_norm": 1.3408719301223755,
      "learning_rate": 0.0001,
      "loss": 0.2899,
      "step": 16200
    },
    {
      "epoch": 14.966720220335093,
      "grad_norm": 1.695694088935852,
      "learning_rate": 0.0001,
      "loss": 0.2924,
      "step": 16300
    },
    {
      "epoch": 14.999770484278173,
      "eval_accuracy": 0.7708355167394468,
      "eval_loss": 0.5231938362121582,
      "eval_runtime": 9.1808,
      "eval_samples_per_second": 54.462,
      "eval_steps_per_second": 6.862,
      "step": 16336
    },
    {
      "epoch": 15.05852650906587,
      "grad_norm": 1.1147023439407349,
      "learning_rate": 0.0001,
      "loss": 0.2605,
      "step": 16400
    },
    {
      "epoch": 15.150332797796649,
      "grad_norm": 1.518908977508545,
      "learning_rate": 0.0001,
      "loss": 0.245,
      "step": 16500
    },
    {
      "epoch": 15.242139086527427,
      "grad_norm": 1.1342830657958984,
      "learning_rate": 0.0001,
      "loss": 0.2447,
      "step": 16600
    },
    {
      "epoch": 15.333945375258205,
      "grad_norm": 1.2657541036605835,
      "learning_rate": 0.0001,
      "loss": 0.2599,
      "step": 16700
    },
    {
      "epoch": 15.425751663988983,
      "grad_norm": 0.9707338809967041,
      "learning_rate": 0.0001,
      "loss": 0.2591,
      "step": 16800
    },
    {
      "epoch": 15.517557952719761,
      "grad_norm": 1.2904791831970215,
      "learning_rate": 0.0001,
      "loss": 0.264,
      "step": 16900
    },
    {
      "epoch": 15.609364241450539,
      "grad_norm": 1.4617804288864136,
      "learning_rate": 0.0001,
      "loss": 0.2665,
      "step": 17000
    },
    {
      "epoch": 15.701170530181317,
      "grad_norm": 1.1893932819366455,
      "learning_rate": 0.0001,
      "loss": 0.2689,
      "step": 17100
    },
    {
      "epoch": 15.792976818912095,
      "grad_norm": 1.3138148784637451,
      "learning_rate": 0.0001,
      "loss": 0.2731,
      "step": 17200
    },
    {
      "epoch": 15.884783107642873,
      "grad_norm": 1.2247110605239868,
      "learning_rate": 0.0001,
      "loss": 0.278,
      "step": 17300
    },
    {
      "epoch": 15.976589396373651,
      "grad_norm": 1.1995705366134644,
      "learning_rate": 0.0001,
      "loss": 0.2824,
      "step": 17400
    },
    {
      "epoch": 15.999540968556346,
      "eval_accuracy": 0.7727045123726346,
      "eval_loss": 0.5129293203353882,
      "eval_runtime": 8.9728,
      "eval_samples_per_second": 55.724,
      "eval_steps_per_second": 7.021,
      "step": 17425
    },
    {
      "epoch": 16.06839568510443,
      "grad_norm": 1.088183045387268,
      "learning_rate": 0.0001,
      "loss": 0.2408,
      "step": 17500
    },
    {
      "epoch": 16.160201973835207,
      "grad_norm": 1.27170991897583,
      "learning_rate": 0.0001,
      "loss": 0.2339,
      "step": 17600
    },
    {
      "epoch": 16.252008262565987,
      "grad_norm": 1.093220591545105,
      "learning_rate": 0.0001,
      "loss": 0.2381,
      "step": 17700
    },
    {
      "epoch": 16.343814551296763,
      "grad_norm": 1.3761118650436401,
      "learning_rate": 0.0001,
      "loss": 0.2361,
      "step": 17800
    },
    {
      "epoch": 16.435620840027543,
      "grad_norm": 1.3061089515686035,
      "learning_rate": 0.0001,
      "loss": 0.2437,
      "step": 17900
    },
    {
      "epoch": 16.52742712875832,
      "grad_norm": 1.318901538848877,
      "learning_rate": 0.0001,
      "loss": 0.2475,
      "step": 18000
    },
    {
      "epoch": 16.6192334174891,
      "grad_norm": 1.241626262664795,
      "learning_rate": 0.0001,
      "loss": 0.2542,
      "step": 18100
    },
    {
      "epoch": 16.711039706219875,
      "grad_norm": 1.1289949417114258,
      "learning_rate": 0.0001,
      "loss": 0.2566,
      "step": 18200
    },
    {
      "epoch": 16.802845994950655,
      "grad_norm": 1.4046275615692139,
      "learning_rate": 0.0001,
      "loss": 0.2594,
      "step": 18300
    },
    {
      "epoch": 16.89465228368143,
      "grad_norm": 1.1862374544143677,
      "learning_rate": 0.0001,
      "loss": 0.2611,
      "step": 18400
    },
    {
      "epoch": 16.98645857241221,
      "grad_norm": 1.3014901876449585,
      "learning_rate": 0.0001,
      "loss": 0.2669,
      "step": 18500
    },
    {
      "epoch": 16.99931145283452,
      "eval_accuracy": 0.774806404657933,
      "eval_loss": 0.49875929951667786,
      "eval_runtime": 9.0234,
      "eval_samples_per_second": 55.411,
      "eval_steps_per_second": 6.982,
      "step": 18514
    },
    {
      "epoch": 17.078264861142987,
      "grad_norm": 1.0681638717651367,
      "learning_rate": 0.0001,
      "loss": 0.2239,
      "step": 18600
    },
    {
      "epoch": 17.170071149873767,
      "grad_norm": 1.1279337406158447,
      "learning_rate": 0.0001,
      "loss": 0.2223,
      "step": 18700
    },
    {
      "epoch": 17.261877438604543,
      "grad_norm": 1.3798402547836304,
      "learning_rate": 0.0001,
      "loss": 0.2241,
      "step": 18800
    },
    {
      "epoch": 17.353683727335323,
      "grad_norm": 1.1741504669189453,
      "learning_rate": 0.0001,
      "loss": 0.2326,
      "step": 18900
    },
    {
      "epoch": 17.4454900160661,
      "grad_norm": 1.1289469003677368,
      "learning_rate": 0.0001,
      "loss": 0.2345,
      "step": 19000
    },
    {
      "epoch": 17.53729630479688,
      "grad_norm": 1.508701205253601,
      "learning_rate": 0.0001,
      "loss": 0.2421,
      "step": 19100
    },
    {
      "epoch": 17.629102593527655,
      "grad_norm": 1.449561357498169,
      "learning_rate": 0.0001,
      "loss": 0.2387,
      "step": 19200
    },
    {
      "epoch": 17.720908882258435,
      "grad_norm": 1.1868849992752075,
      "learning_rate": 0.0001,
      "loss": 0.2402,
      "step": 19300
    },
    {
      "epoch": 17.81271517098921,
      "grad_norm": 1.4335336685180664,
      "learning_rate": 0.0001,
      "loss": 0.249,
      "step": 19400
    },
    {
      "epoch": 17.90452145971999,
      "grad_norm": 1.3802162408828735,
      "learning_rate": 0.0001,
      "loss": 0.2491,
      "step": 19500
    },
    {
      "epoch": 17.996327748450767,
      "grad_norm": 1.3790746927261353,
      "learning_rate": 0.0001,
      "loss": 0.2517,
      "step": 19600
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.776174672489083,
      "eval_loss": 0.4891900420188904,
      "eval_runtime": 8.9325,
      "eval_samples_per_second": 55.975,
      "eval_steps_per_second": 7.053,
      "step": 19604
    },
    {
      "epoch": 18.088134037181547,
      "grad_norm": 1.1314564943313599,
      "learning_rate": 0.0001,
      "loss": 0.2088,
      "step": 19700
    },
    {
      "epoch": 18.179940325912327,
      "grad_norm": 1.2055948972702026,
      "learning_rate": 0.0001,
      "loss": 0.2128,
      "step": 19800
    },
    {
      "epoch": 18.271746614643103,
      "grad_norm": 1.1677360534667969,
      "learning_rate": 0.0001,
      "loss": 0.2178,
      "step": 19900
    },
    {
      "epoch": 18.363552903373883,
      "grad_norm": 1.2793176174163818,
      "learning_rate": 0.0001,
      "loss": 0.2216,
      "step": 20000
    },
    {
      "epoch": 18.45535919210466,
      "grad_norm": 1.187522292137146,
      "learning_rate": 0.0001,
      "loss": 0.2243,
      "step": 20100
    },
    {
      "epoch": 18.54716548083544,
      "grad_norm": 1.5564976930618286,
      "learning_rate": 0.0001,
      "loss": 0.2249,
      "step": 20200
    },
    {
      "epoch": 18.638971769566215,
      "grad_norm": 1.2912520170211792,
      "learning_rate": 0.0001,
      "loss": 0.2319,
      "step": 20300
    },
    {
      "epoch": 18.730778058296995,
      "grad_norm": 1.5046939849853516,
      "learning_rate": 0.0001,
      "loss": 0.2343,
      "step": 20400
    },
    {
      "epoch": 18.82258434702777,
      "grad_norm": 1.4738825559616089,
      "learning_rate": 0.0001,
      "loss": 0.2342,
      "step": 20500
    },
    {
      "epoch": 18.91439063575855,
      "grad_norm": 1.427435278892517,
      "learning_rate": 0.0001,
      "loss": 0.2376,
      "step": 20600
    },
    {
      "epoch": 18.999770484278173,
      "eval_accuracy": 0.7773391557496361,
      "eval_loss": 0.4808199405670166,
      "eval_runtime": 8.9815,
      "eval_samples_per_second": 55.67,
      "eval_steps_per_second": 7.014,
      "step": 20693
    },
    {
      "epoch": 19.006196924489327,
      "grad_norm": 1.294245719909668,
      "learning_rate": 0.0001,
      "loss": 0.2396,
      "step": 20700
    },
    {
      "epoch": 19.098003213220107,
      "grad_norm": 0.9566488862037659,
      "learning_rate": 0.0001,
      "loss": 0.2006,
      "step": 20800
    },
    {
      "epoch": 19.189809501950883,
      "grad_norm": 1.184180736541748,
      "learning_rate": 0.0001,
      "loss": 0.2049,
      "step": 20900
    },
    {
      "epoch": 19.281615790681663,
      "grad_norm": 1.1258317232131958,
      "learning_rate": 0.0001,
      "loss": 0.2081,
      "step": 21000
    },
    {
      "epoch": 19.37342207941244,
      "grad_norm": 1.2547038793563843,
      "learning_rate": 0.0001,
      "loss": 0.2133,
      "step": 21100
    },
    {
      "epoch": 19.46522836814322,
      "grad_norm": 1.3770051002502441,
      "learning_rate": 0.0001,
      "loss": 0.2175,
      "step": 21200
    },
    {
      "epoch": 19.557034656873995,
      "grad_norm": 1.3640483617782593,
      "learning_rate": 0.0001,
      "loss": 0.2178,
      "step": 21300
    },
    {
      "epoch": 19.648840945604775,
      "grad_norm": 1.2219371795654297,
      "learning_rate": 0.0001,
      "loss": 0.2233,
      "step": 21400
    },
    {
      "epoch": 19.74064723433555,
      "grad_norm": 1.3438184261322021,
      "learning_rate": 0.0001,
      "loss": 0.224,
      "step": 21500
    },
    {
      "epoch": 19.83245352306633,
      "grad_norm": 1.2909867763519287,
      "learning_rate": 0.0001,
      "loss": 0.2274,
      "step": 21600
    },
    {
      "epoch": 19.924259811797107,
      "grad_norm": 1.482640027999878,
      "learning_rate": 0.0001,
      "loss": 0.2316,
      "step": 21700
    },
    {
      "epoch": 19.99770484278173,
      "eval_accuracy": 0.7780232896652111,
      "eval_loss": 0.4803846478462219,
      "eval_runtime": 8.977,
      "eval_samples_per_second": 55.698,
      "eval_steps_per_second": 7.018,
      "step": 21780
    },
    {
      "epoch": 19.99770484278173,
      "step": 21780,
      "total_flos": 2.2953494160657613e+18,
      "train_loss": 0.0,
      "train_runtime": 0.0873,
      "train_samples_per_second": 7982672.15,
      "train_steps_per_second": 249415.561
    }
  ],
  "logging_steps": 100,
  "max_steps": 21780,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "save_steps": 500,
  "total_flos": 2.2953494160657613e+18,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}