{
  "best_global_step": 2000,
  "best_metric": 0.2992326021194458,
  "best_model_checkpoint": "./outputs/checkpoints/checkpoint-2000",
  "epoch": 3.0,
  "eval_steps": 250,
  "global_step": 3117,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.04814636494944632,
      "grad_norm": 0.49752116203308105,
      "learning_rate": 6.32258064516129e-05,
      "loss": 1.477303009033203,
      "step": 50
    },
    {
      "epoch": 0.09629272989889263,
      "grad_norm": 0.3652721643447876,
      "learning_rate": 0.00012774193548387096,
      "loss": 0.46554481506347656,
      "step": 100
    },
    {
      "epoch": 0.14443909484833894,
      "grad_norm": 0.32378166913986206,
      "learning_rate": 0.00019225806451612904,
      "loss": 0.4035662078857422,
      "step": 150
    },
    {
      "epoch": 0.19258545979778527,
      "grad_norm": 0.3265906870365143,
      "learning_rate": 0.0001998911255002373,
      "loss": 0.38129932403564454,
      "step": 200
    },
    {
      "epoch": 0.2407318247472316,
      "grad_norm": 0.2637921869754791,
      "learning_rate": 0.00019950341273368075,
      "loss": 0.36489856719970704,
      "step": 250
    },
    {
      "epoch": 0.2407318247472316,
      "eval_loss": 0.35290107131004333,
      "eval_runtime": 694.8697,
      "eval_samples_per_second": 2.658,
      "eval_steps_per_second": 0.665,
      "step": 250
    },
    {
      "epoch": 0.2888781896966779,
      "grad_norm": 0.23706986010074615,
      "learning_rate": 0.00019883592694304428,
      "loss": 0.3474379348754883,
      "step": 300
    },
    {
      "epoch": 0.33702455464612424,
      "grad_norm": 0.2362891435623169,
      "learning_rate": 0.00019789054489328546,
      "loss": 0.35004573822021484,
      "step": 350
    },
    {
      "epoch": 0.38517091959557054,
      "grad_norm": 0.2134203463792801,
      "learning_rate": 0.00019666992470825862,
      "loss": 0.33882694244384765,
      "step": 400
    },
    {
      "epoch": 0.43331728454501683,
      "grad_norm": 0.23494897782802582,
      "learning_rate": 0.0001951774983968872,
      "loss": 0.35397193908691404,
      "step": 450
    },
    {
      "epoch": 0.4814636494944632,
      "grad_norm": 0.20449194312095642,
      "learning_rate": 0.00019341746220340924,
      "loss": 0.3447082901000977,
      "step": 500
    },
    {
      "epoch": 0.4814636494944632,
      "eval_loss": 0.3319753110408783,
      "eval_runtime": 695.9468,
      "eval_samples_per_second": 2.654,
      "eval_steps_per_second": 0.664,
      "step": 500
    },
    {
      "epoch": 0.5296100144439095,
      "grad_norm": 0.19552361965179443,
      "learning_rate": 0.00019139476480882785,
      "loss": 0.34256481170654296,
      "step": 550
    },
    {
      "epoch": 0.5777563793933558,
      "grad_norm": 0.19304735958576202,
      "learning_rate": 0.00018911509341674071,
      "loss": 0.3311490631103516,
      "step": 600
    },
    {
      "epoch": 0.6259027443428021,
      "grad_norm": 0.22191040217876434,
      "learning_rate": 0.00018658485776267095,
      "loss": 0.32927177429199217,
      "step": 650
    },
    {
      "epoch": 0.6740491092922485,
      "grad_norm": 0.19663718342781067,
      "learning_rate": 0.00018381117209186033,
      "loss": 0.32792991638183594,
      "step": 700
    },
    {
      "epoch": 0.7221954742416947,
      "grad_norm": 0.20722714066505432,
      "learning_rate": 0.00018080183515619726,
      "loss": 0.3278979873657227,
      "step": 750
    },
    {
      "epoch": 0.7221954742416947,
      "eval_loss": 0.3201846480369568,
      "eval_runtime": 698.1861,
      "eval_samples_per_second": 2.645,
      "eval_steps_per_second": 0.662,
      "step": 750
    },
    {
      "epoch": 0.7703418391911411,
      "grad_norm": 0.18782874941825867,
      "learning_rate": 0.000177565308286523,
      "loss": 0.32439987182617186,
      "step": 800
    },
    {
      "epoch": 0.8184882041405874,
      "grad_norm": 0.19059380888938904,
      "learning_rate": 0.0001741106916019688,
      "loss": 0.32634170532226564,
      "step": 850
    },
    {
      "epoch": 0.8666345690900337,
      "grad_norm": 0.19336679577827454,
      "learning_rate": 0.00017044769842321705,
      "loss": 0.3187075614929199,
      "step": 900
    },
    {
      "epoch": 0.91478093403948,
      "grad_norm": 0.19474540650844574,
      "learning_rate": 0.00016658662796162794,
      "loss": 0.3124076271057129,
      "step": 950
    },
    {
      "epoch": 0.9629272989889264,
      "grad_norm": 0.19545841217041016,
      "learning_rate": 0.0001625383363610215,
      "loss": 0.32053829193115235,
      "step": 1000
    },
    {
      "epoch": 0.9629272989889264,
      "eval_loss": 0.3141631782054901,
      "eval_runtime": 696.8814,
      "eval_samples_per_second": 2.65,
      "eval_steps_per_second": 0.663,
      "step": 1000
    },
    {
      "epoch": 1.010592200288878,
      "grad_norm": 0.23214930295944214,
      "learning_rate": 0.00015831420617353685,
      "loss": 0.3157562255859375,
      "step": 1050
    },
    {
      "epoch": 1.0587385652383245,
      "grad_norm": 0.19467434287071228,
      "learning_rate": 0.0001539261143553928,
      "loss": 0.29116039276123046,
      "step": 1100
    },
    {
      "epoch": 1.1068849301877708,
      "grad_norm": 0.2111353874206543,
      "learning_rate": 0.0001493863988725362,
      "loss": 0.29466302871704103,
      "step": 1150
    },
    {
      "epoch": 1.155031295137217,
      "grad_norm": 0.21094423532485962,
      "learning_rate": 0.0001447078240100725,
      "loss": 0.2904193878173828,
      "step": 1200
    },
    {
      "epoch": 1.2031776600866635,
      "grad_norm": 0.20865824818611145,
      "learning_rate": 0.00013990354448301798,
      "loss": 0.29510746002197263,
      "step": 1250
    },
    {
      "epoch": 1.2031776600866635,
      "eval_loss": 0.3110591769218445,
      "eval_runtime": 694.8335,
      "eval_samples_per_second": 2.658,
      "eval_steps_per_second": 0.665,
      "step": 1250
    },
    {
      "epoch": 1.2513240250361097,
      "grad_norm": 0.20231032371520996,
      "learning_rate": 0.00013498706844928282,
      "loss": 0.29798513412475586,
      "step": 1300
    },
    {
      "epoch": 1.2994703899855562,
      "grad_norm": 0.23587271571159363,
      "learning_rate": 0.00012997221952888146,
      "loss": 0.29837047576904296,
      "step": 1350
    },
    {
      "epoch": 1.3476167549350024,
      "grad_norm": 0.19505874812602997,
      "learning_rate": 0.0001248730979361606,
      "loss": 0.2953862953186035,
      "step": 1400
    },
    {
      "epoch": 1.3957631198844487,
      "grad_norm": 0.18162201344966888,
      "learning_rate": 0.00011970404083432846,
      "loss": 0.28933364868164063,
      "step": 1450
    },
    {
      "epoch": 1.443909484833895,
      "grad_norm": 0.18973521888256073,
      "learning_rate": 0.0001144795820237573,
      "loss": 0.29408544540405274,
      "step": 1500
    },
    {
      "epoch": 1.443909484833895,
      "eval_loss": 0.30712059140205383,
      "eval_runtime": 694.8078,
      "eval_samples_per_second": 2.658,
      "eval_steps_per_second": 0.665,
      "step": 1500
    },
    {
      "epoch": 1.4920558497833414,
      "grad_norm": 0.18828196823596954,
      "learning_rate": 0.00010921441107740198,
      "loss": 0.28817583084106446,
      "step": 1550
    },
    {
      "epoch": 1.5402022147327876,
      "grad_norm": 0.19784030318260193,
      "learning_rate": 0.0001039233320382344,
      "loss": 0.292324333190918,
      "step": 1600
    },
    {
      "epoch": 1.588348579682234,
      "grad_norm": 0.21540944278240204,
      "learning_rate": 9.862122179482317e-05,
      "loss": 0.2901228141784668,
      "step": 1650
    },
    {
      "epoch": 1.6364949446316803,
      "grad_norm": 0.22679032385349274,
      "learning_rate": 9.332298825209385e-05,
      "loss": 0.28837751388549804,
      "step": 1700
    },
    {
      "epoch": 1.6846413095811266,
      "grad_norm": 0.19287723302841187,
      "learning_rate": 8.80435284148808e-05,
      "loss": 0.28883354187011717,
      "step": 1750
    },
    {
      "epoch": 1.6846413095811266,
      "eval_loss": 0.30308136343955994,
      "eval_runtime": 695.3246,
      "eval_samples_per_second": 2.656,
      "eval_steps_per_second": 0.664,
      "step": 1750
    },
    {
      "epoch": 1.7327876745305728,
      "grad_norm": 0.18742942810058594,
      "learning_rate": 8.279768650212679e-05,
      "loss": 0.2922953224182129,
      "step": 1800
    },
    {
      "epoch": 1.7809340394800193,
      "grad_norm": 0.17063775658607483,
      "learning_rate": 7.760021220950004e-05,
      "loss": 0.29046398162841797,
      "step": 1850
    },
    {
      "epoch": 1.8290804044294657,
      "grad_norm": 0.1963094025850296,
      "learning_rate": 7.246571923778208e-05,
      "loss": 0.28941730499267576,
      "step": 1900
    },
    {
      "epoch": 1.877226769378912,
      "grad_norm": 0.186972513794899,
      "learning_rate": 6.740864420363105e-05,
      "loss": 0.28976251602172853,
      "step": 1950
    },
    {
      "epoch": 1.9253731343283582,
      "grad_norm": 0.18225796520709991,
      "learning_rate": 6.244320604825131e-05,
      "loss": 0.2877297782897949,
      "step": 2000
    },
    {
      "epoch": 1.9253731343283582,
      "eval_loss": 0.2992326021194458,
      "eval_runtime": 695.681,
      "eval_samples_per_second": 2.655,
      "eval_steps_per_second": 0.664,
      "step": 2000
    },
    {
      "epoch": 1.9735194992778045,
      "grad_norm": 0.18147684633731842,
      "learning_rate": 5.7583366058099916e-05,
      "loss": 0.2825794219970703,
      "step": 2050
    },
    {
      "epoch": 2.021184400577756,
      "grad_norm": 0.21326377987861633,
      "learning_rate": 5.284278861003815e-05,
      "loss": 0.271549129486084,
      "step": 2100
    },
    {
      "epoch": 2.069330765527203,
      "grad_norm": 0.2095552235841751,
      "learning_rate": 4.823480275130263e-05,
      "loss": 0.24787240982055664,
      "step": 2150
    },
    {
      "epoch": 2.117477130476649,
      "grad_norm": 0.18589918315410614,
      "learning_rate": 4.3772364722319715e-05,
      "loss": 0.2502709770202637,
      "step": 2200
    },
    {
      "epoch": 2.1656234954260953,
      "grad_norm": 0.2025139182806015,
      "learning_rate": 3.946802152773811e-05,
      "loss": 0.25036895751953125,
      "step": 2250
    },
    {
      "epoch": 2.1656234954260953,
      "eval_loss": 0.3054317533969879,
      "eval_runtime": 695.5773,
      "eval_samples_per_second": 2.655,
      "eval_steps_per_second": 0.664,
      "step": 2250
    },
    {
      "epoch": 2.2137698603755416,
      "grad_norm": 0.23237130045890808,
      "learning_rate": 3.533387565810706e-05,
      "loss": 0.24967620849609376,
      "step": 2300
    },
    {
      "epoch": 2.261916225324988,
      "grad_norm": 0.21668598055839539,
      "learning_rate": 3.1381551061391366e-05,
      "loss": 0.2472244071960449,
      "step": 2350
    },
    {
      "epoch": 2.310062590274434,
      "grad_norm": 0.21904581785202026,
      "learning_rate": 2.7622160460001166e-05,
      "loss": 0.24260114669799804,
      "step": 2400
    },
    {
      "epoch": 2.3582089552238807,
      "grad_norm": 0.2477143406867981,
      "learning_rate": 2.406627410523087e-05,
      "loss": 0.24558507919311523,
      "step": 2450
    },
    {
      "epoch": 2.406355320173327,
      "grad_norm": 0.26366570591926575,
      "learning_rate": 2.0723890056960227e-05,
      "loss": 0.24778993606567382,
      "step": 2500
    },
    {
      "epoch": 2.406355320173327,
      "eval_loss": 0.30330467224121094,
      "eval_runtime": 695.5219,
      "eval_samples_per_second": 2.656,
      "eval_steps_per_second": 0.664,
      "step": 2500
    },
    {
      "epoch": 2.4545016851227732,
      "grad_norm": 0.25673332810401917,
      "learning_rate": 1.7604406072182122e-05,
      "loss": 0.24701383590698242,
      "step": 2550
    },
    {
      "epoch": 2.5026480500722195,
      "grad_norm": 0.2344675064086914,
      "learning_rate": 1.4716593181396964e-05,
      "loss": 0.23919960021972655,
      "step": 2600
    },
    {
      "epoch": 2.5507944150216657,
      "grad_norm": 0.24205727875232697,
      "learning_rate": 1.2068571027170161e-05,
      "loss": 0.2499522399902344,
      "step": 2650
    },
    {
      "epoch": 2.5989407799711124,
      "grad_norm": 0.18950000405311584,
      "learning_rate": 9.667785034191811e-06,
      "loss": 0.24822561264038087,
      "step": 2700
    },
    {
      "epoch": 2.6470871449205586,
      "grad_norm": 0.23864181339740753,
      "learning_rate": 7.5209854750301844e-06,
      "loss": 0.24670083999633788,
      "step": 2750
    },
    {
      "epoch": 2.6470871449205586,
      "eval_loss": 0.30303388833999634,
      "eval_runtime": 696.3045,
      "eval_samples_per_second": 2.653,
      "eval_steps_per_second": 0.664,
      "step": 2750
    },
    {
      "epoch": 2.695233509870005,
      "grad_norm": 0.2515101730823517,
      "learning_rate": 5.634208490439252e-06,
      "loss": 0.24807355880737306,
      "step": 2800
    },
    {
      "epoch": 2.743379874819451,
      "grad_norm": 0.22556646168231964,
      "learning_rate": 4.012759117585463e-06,
      "loss": 0.24218505859375,
      "step": 2850
    },
    {
      "epoch": 2.7915262397688974,
      "grad_norm": 0.21387307345867157,
      "learning_rate": 2.661196373913288e-06,
      "loss": 0.24130987167358398,
      "step": 2900
    },
    {
      "epoch": 2.839672604718344,
      "grad_norm": 0.20219890773296356,
      "learning_rate": 1.5833204385888867e-06,
      "loss": 0.24288219451904297,
      "step": 2950
    },
    {
      "epoch": 2.88781896966779,
      "grad_norm": 0.24099305272102356,
      "learning_rate": 7.821619675638658e-07,
      "loss": 0.25014289855957034,
      "step": 3000
    },
    {
      "epoch": 2.88781896966779,
      "eval_loss": 0.3028092682361603,
      "eval_runtime": 696.6975,
      "eval_samples_per_second": 2.651,
      "eval_steps_per_second": 0.663,
      "step": 3000
    },
    {
      "epoch": 2.9359653346172365,
      "grad_norm": 0.2587520182132721,
      "learning_rate": 2.5997357230198583e-07,
      "loss": 0.24698244094848631,
      "step": 3050
    },
    {
      "epoch": 2.9841116995666828,
      "grad_norm": 0.19957926869392395,
      "learning_rate": 1.8223486127799673e-08,
      "loss": 0.24292854309082032,
      "step": 3100
    }
  ],
  "logging_steps": 50,
  "max_steps": 3117,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 250,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 2.0828994331856486e+18,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}