{
  "best_metric": 3.290560483932495,
  "best_model_checkpoint": "./models/lora-finetuning/german-gpt2/checkpoint-292000",
  "epoch": 2.50689395838556,
  "eval_steps": 1000,
  "global_step": 300000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.004178156597309267,
      "grad_norm": 6.081511497497559,
      "learning_rate": 4.9800000000000004e-05,
      "loss": 5.8544,
      "step": 500
    },
    {
      "epoch": 0.008356313194618534,
      "grad_norm": 11.10040283203125,
      "learning_rate": 4.9791054795670054e-05,
      "loss": 5.1312,
      "step": 1000
    },
    {
      "epoch": 0.008356313194618534,
      "eval_loss": 4.83798360824585,
      "eval_runtime": 40.3269,
      "eval_samples_per_second": 110.646,
      "eval_steps_per_second": 13.837,
      "step": 1000
    },
    {
      "epoch": 0.012534469791927802,
      "grad_norm": 5.357274055480957,
      "learning_rate": 4.9581270453973316e-05,
      "loss": 4.8327,
      "step": 1500
    },
    {
      "epoch": 0.016712626389237067,
      "grad_norm": 4.852677345275879,
      "learning_rate": 4.9371486112276584e-05,
      "loss": 4.6285,
      "step": 2000
    },
    {
      "epoch": 0.016712626389237067,
      "eval_loss": 4.516997814178467,
      "eval_runtime": 40.2461,
      "eval_samples_per_second": 110.868,
      "eval_steps_per_second": 13.865,
      "step": 2000
    },
    {
      "epoch": 0.020890782986546334,
      "grad_norm": 5.627632141113281,
      "learning_rate": 4.916170177057985e-05,
      "loss": 4.6442,
      "step": 2500
    },
    {
      "epoch": 0.025068939583855605,
      "grad_norm": 4.464819431304932,
      "learning_rate": 4.8951917428883106e-05,
      "loss": 4.4546,
      "step": 3000
    },
    {
      "epoch": 0.025068939583855605,
      "eval_loss": 4.347652912139893,
      "eval_runtime": 40.226,
      "eval_samples_per_second": 110.923,
      "eval_steps_per_second": 13.872,
      "step": 3000
    },
    {
      "epoch": 0.02924709618116487,
      "grad_norm": 5.631376266479492,
      "learning_rate": 4.8742133087186374e-05,
      "loss": 4.4188,
      "step": 3500
    },
    {
      "epoch": 0.033425252778474135,
      "grad_norm": 5.489138603210449,
      "learning_rate": 4.8532348745489635e-05,
      "loss": 4.4816,
      "step": 4000
    },
    {
      "epoch": 0.033425252778474135,
      "eval_loss": 4.25124454498291,
      "eval_runtime": 40.2167,
      "eval_samples_per_second": 110.949,
      "eval_steps_per_second": 13.875,
      "step": 4000
    },
    {
      "epoch": 0.0376034093757834,
      "grad_norm": 6.327524185180664,
      "learning_rate": 4.83229839724763e-05,
      "loss": 4.3622,
      "step": 4500
    },
    {
      "epoch": 0.04178156597309267,
      "grad_norm": 5.45938777923584,
      "learning_rate": 4.811319963077956e-05,
      "loss": 4.2968,
      "step": 5000
    },
    {
      "epoch": 0.04178156597309267,
      "eval_loss": 4.177467346191406,
      "eval_runtime": 40.2059,
      "eval_samples_per_second": 110.979,
      "eval_steps_per_second": 13.879,
      "step": 5000
    },
    {
      "epoch": 0.045959722570401935,
      "grad_norm": 1.7540432214736938,
      "learning_rate": 4.7903415289082825e-05,
      "loss": 4.1926,
      "step": 5500
    },
    {
      "epoch": 0.05013787916771121,
      "grad_norm": 3.5564088821411133,
      "learning_rate": 4.769363094738609e-05,
      "loss": 4.1807,
      "step": 6000
    },
    {
      "epoch": 0.05013787916771121,
      "eval_loss": 4.1200947761535645,
      "eval_runtime": 40.2083,
      "eval_samples_per_second": 110.972,
      "eval_steps_per_second": 13.878,
      "step": 6000
    },
    {
      "epoch": 0.054316035765020476,
      "grad_norm": 4.009562969207764,
      "learning_rate": 4.7484266174372746e-05,
      "loss": 4.2543,
      "step": 6500
    },
    {
      "epoch": 0.05849419236232974,
      "grad_norm": 2.7500414848327637,
      "learning_rate": 4.727448183267601e-05,
      "loss": 4.2382,
      "step": 7000
    },
    {
      "epoch": 0.05849419236232974,
      "eval_loss": 4.056636810302734,
      "eval_runtime": 40.2301,
      "eval_samples_per_second": 110.912,
      "eval_steps_per_second": 13.87,
      "step": 7000
    },
    {
      "epoch": 0.06267234895963901,
      "grad_norm": 9.726302146911621,
      "learning_rate": 4.7064697490979275e-05,
      "loss": 4.1848,
      "step": 7500
    },
    {
      "epoch": 0.06685050555694827,
      "grad_norm": 8.977545738220215,
      "learning_rate": 4.685491314928254e-05,
      "loss": 4.1906,
      "step": 8000
    },
    {
      "epoch": 0.06685050555694827,
      "eval_loss": 4.021559715270996,
      "eval_runtime": 40.1844,
      "eval_samples_per_second": 111.038,
      "eval_steps_per_second": 13.886,
      "step": 8000
    },
    {
      "epoch": 0.07102866215425754,
      "grad_norm": 8.016194343566895,
      "learning_rate": 4.6645128807585805e-05,
      "loss": 4.2125,
      "step": 8500
    },
    {
      "epoch": 0.0752068187515668,
      "grad_norm": 7.392004489898682,
      "learning_rate": 4.643576403457246e-05,
      "loss": 4.1017,
      "step": 9000
    },
    {
      "epoch": 0.0752068187515668,
      "eval_loss": 3.993513822555542,
      "eval_runtime": 40.2054,
      "eval_samples_per_second": 110.98,
      "eval_steps_per_second": 13.879,
      "step": 9000
    },
    {
      "epoch": 0.07938497534887608,
      "grad_norm": 5.018495559692383,
      "learning_rate": 4.6225979692875726e-05,
      "loss": 4.1314,
      "step": 9500
    },
    {
      "epoch": 0.08356313194618534,
      "grad_norm": 4.584932804107666,
      "learning_rate": 4.6016195351178994e-05,
      "loss": 4.0724,
      "step": 10000
    },
    {
      "epoch": 0.08356313194618534,
      "eval_loss": 3.9573867321014404,
      "eval_runtime": 40.2188,
      "eval_samples_per_second": 110.943,
      "eval_steps_per_second": 13.874,
      "step": 10000
    },
    {
      "epoch": 0.08774128854349461,
      "grad_norm": 6.544865608215332,
      "learning_rate": 4.5806411009482255e-05,
      "loss": 4.0296,
      "step": 10500
    },
    {
      "epoch": 0.09191944514080387,
      "grad_norm": 8.033525466918945,
      "learning_rate": 4.5596626667785516e-05,
      "loss": 4.1032,
      "step": 11000
    },
    {
      "epoch": 0.09191944514080387,
      "eval_loss": 3.9294466972351074,
      "eval_runtime": 40.2168,
      "eval_samples_per_second": 110.949,
      "eval_steps_per_second": 13.875,
      "step": 11000
    },
    {
      "epoch": 0.09609760173811314,
      "grad_norm": 5.5422539710998535,
      "learning_rate": 4.5387261894772176e-05,
      "loss": 4.0642,
      "step": 11500
    },
    {
      "epoch": 0.10027575833542242,
      "grad_norm": 4.3626508712768555,
      "learning_rate": 4.5177477553075444e-05,
      "loss": 3.9922,
      "step": 12000
    },
    {
      "epoch": 0.10027575833542242,
      "eval_loss": 3.9016854763031006,
      "eval_runtime": 40.208,
      "eval_samples_per_second": 110.973,
      "eval_steps_per_second": 13.878,
      "step": 12000
    },
    {
      "epoch": 0.10445391493273168,
      "grad_norm": 4.636664867401123,
      "learning_rate": 4.4967693211378706e-05,
      "loss": 4.017,
      "step": 12500
    },
    {
      "epoch": 0.10863207153004095,
      "grad_norm": 5.429406642913818,
      "learning_rate": 4.475790886968197e-05,
      "loss": 3.9346,
      "step": 13000
    },
    {
      "epoch": 0.10863207153004095,
      "eval_loss": 3.881444215774536,
      "eval_runtime": 40.1856,
      "eval_samples_per_second": 111.035,
      "eval_steps_per_second": 13.886,
      "step": 13000
    },
    {
      "epoch": 0.11281022812735021,
      "grad_norm": 6.612671375274658,
      "learning_rate": 4.454854409666863e-05,
      "loss": 4.0277,
      "step": 13500
    },
    {
      "epoch": 0.11698838472465949,
      "grad_norm": 3.6754331588745117,
      "learning_rate": 4.4338759754971895e-05,
      "loss": 3.9975,
      "step": 14000
    },
    {
      "epoch": 0.11698838472465949,
      "eval_loss": 3.8531486988067627,
      "eval_runtime": 40.1972,
      "eval_samples_per_second": 111.003,
      "eval_steps_per_second": 13.882,
      "step": 14000
    },
    {
      "epoch": 0.12116654132196875,
      "grad_norm": 5.864243984222412,
      "learning_rate": 4.412981455064194e-05,
      "loss": 5.2354,
      "step": 14500
    },
    {
      "epoch": 0.12534469791927802,
      "grad_norm": 6.96849250793457,
      "learning_rate": 4.392003020894521e-05,
      "loss": 4.8952,
      "step": 15000
    },
    {
      "epoch": 0.12534469791927802,
      "eval_loss": 4.6278204917907715,
      "eval_runtime": 40.2925,
      "eval_samples_per_second": 110.74,
      "eval_steps_per_second": 13.849,
      "step": 15000
    },
    {
      "epoch": 0.1295228545165873,
      "grad_norm": 8.802956581115723,
      "learning_rate": 4.371024586724847e-05,
      "loss": 4.7056,
      "step": 15500
    },
    {
      "epoch": 0.13370101111389654,
      "grad_norm": 4.981395721435547,
      "learning_rate": 4.350046152555173e-05,
      "loss": 4.5538,
      "step": 16000
    },
    {
      "epoch": 0.13370101111389654,
      "eval_loss": 4.411046504974365,
      "eval_runtime": 40.2719,
      "eval_samples_per_second": 110.797,
      "eval_steps_per_second": 13.856,
      "step": 16000
    },
    {
      "epoch": 0.1378791677112058,
      "grad_norm": 5.577733993530273,
      "learning_rate": 4.3290677183855e-05,
      "loss": 4.507,
      "step": 16500
    },
    {
      "epoch": 0.1420573243085151,
      "grad_norm": 12.114670753479004,
      "learning_rate": 4.308089284215827e-05,
      "loss": 4.4785,
      "step": 17000
    },
    {
      "epoch": 0.1420573243085151,
      "eval_loss": 4.291799068450928,
      "eval_runtime": 40.2805,
      "eval_samples_per_second": 110.773,
      "eval_steps_per_second": 13.853,
      "step": 17000
    },
    {
      "epoch": 0.14623548090582436,
      "grad_norm": 8.596464157104492,
      "learning_rate": 4.287110850046153e-05,
      "loss": 4.4097,
      "step": 17500
    },
    {
      "epoch": 0.1504136375031336,
      "grad_norm": 11.195298194885254,
      "learning_rate": 4.2661324158764796e-05,
      "loss": 4.3955,
      "step": 18000
    },
    {
      "epoch": 0.1504136375031336,
      "eval_loss": 4.214114189147949,
      "eval_runtime": 40.2758,
      "eval_samples_per_second": 110.786,
      "eval_steps_per_second": 13.854,
      "step": 18000
    },
    {
      "epoch": 0.15459179410044288,
      "grad_norm": 7.286564826965332,
      "learning_rate": 4.245195938575145e-05,
      "loss": 4.2833,
      "step": 18500
    },
    {
      "epoch": 0.15876995069775215,
      "grad_norm": 4.253493309020996,
      "learning_rate": 4.224217504405472e-05,
      "loss": 4.3205,
      "step": 19000
    },
    {
      "epoch": 0.15876995069775215,
      "eval_loss": 4.141177177429199,
      "eval_runtime": 40.2675,
      "eval_samples_per_second": 110.809,
      "eval_steps_per_second": 13.857,
      "step": 19000
    },
    {
      "epoch": 0.16294810729506143,
      "grad_norm": 6.702279090881348,
      "learning_rate": 4.203239070235798e-05,
      "loss": 4.2901,
      "step": 19500
    },
    {
      "epoch": 0.16712626389237067,
      "grad_norm": 8.38072681427002,
      "learning_rate": 4.182260636066124e-05,
      "loss": 4.1862,
      "step": 20000
    },
    {
      "epoch": 0.16712626389237067,
      "eval_loss": 4.087652683258057,
      "eval_runtime": 40.2756,
      "eval_samples_per_second": 110.787,
      "eval_steps_per_second": 13.855,
      "step": 20000
    },
    {
      "epoch": 0.17130442048967995,
      "grad_norm": 7.476578712463379,
      "learning_rate": 4.16132415876479e-05,
      "loss": 4.1174,
      "step": 20500
    },
    {
      "epoch": 0.17548257708698922,
      "grad_norm": 6.918697357177734,
      "learning_rate": 4.140345724595117e-05,
      "loss": 4.2499,
      "step": 21000
    },
    {
      "epoch": 0.17548257708698922,
      "eval_loss": 4.053555011749268,
      "eval_runtime": 40.1812,
      "eval_samples_per_second": 111.047,
      "eval_steps_per_second": 13.887,
      "step": 21000
    },
    {
      "epoch": 0.1796607336842985,
      "grad_norm": 3.7625985145568848,
      "learning_rate": 4.119367290425443e-05,
      "loss": 4.1048,
      "step": 21500
    },
    {
      "epoch": 0.18383889028160774,
      "grad_norm": 14.019062042236328,
      "learning_rate": 4.098388856255769e-05,
      "loss": 4.0907,
      "step": 22000
    },
    {
      "epoch": 0.18383889028160774,
      "eval_loss": 4.000556468963623,
      "eval_runtime": 40.1881,
      "eval_samples_per_second": 111.028,
      "eval_steps_per_second": 13.885,
      "step": 22000
    },
    {
      "epoch": 0.18801704687891702,
      "grad_norm": 4.171750545501709,
      "learning_rate": 4.077410422086096e-05,
      "loss": 4.1064,
      "step": 22500
    },
    {
      "epoch": 0.1921952034762263,
      "grad_norm": 6.7127885818481445,
      "learning_rate": 4.056473944784762e-05,
      "loss": 4.1832,
      "step": 23000
    },
    {
      "epoch": 0.1921952034762263,
      "eval_loss": 3.97273588180542,
      "eval_runtime": 40.1852,
      "eval_samples_per_second": 111.036,
      "eval_steps_per_second": 13.886,
      "step": 23000
    },
    {
      "epoch": 0.19637336007353556,
      "grad_norm": 5.6008830070495605,
      "learning_rate": 4.035495510615088e-05,
      "loss": 4.0533,
      "step": 23500
    },
    {
      "epoch": 0.20055151667084484,
      "grad_norm": 6.787017345428467,
      "learning_rate": 4.014517076445414e-05,
      "loss": 4.1166,
      "step": 24000
    },
    {
      "epoch": 0.20055151667084484,
      "eval_loss": 3.9389452934265137,
      "eval_runtime": 40.1944,
      "eval_samples_per_second": 111.01,
      "eval_steps_per_second": 13.883,
      "step": 24000
    },
    {
      "epoch": 0.20472967326815408,
      "grad_norm": 4.794278144836426,
      "learning_rate": 3.993538642275741e-05,
      "loss": 4.101,
      "step": 24500
    },
    {
      "epoch": 0.20890782986546336,
      "grad_norm": 8.866153717041016,
      "learning_rate": 3.972602164974407e-05,
      "loss": 4.1032,
      "step": 25000
    },
    {
      "epoch": 0.20890782986546336,
      "eval_loss": 3.9197261333465576,
      "eval_runtime": 40.1697,
      "eval_samples_per_second": 111.079,
      "eval_steps_per_second": 13.891,
      "step": 25000
    },
    {
      "epoch": 0.21308598646277263,
      "grad_norm": 5.620716094970703,
      "learning_rate": 3.951623730804733e-05,
      "loss": 3.9787,
      "step": 25500
    },
    {
      "epoch": 0.2172641430600819,
      "grad_norm": 4.146615028381348,
      "learning_rate": 3.930645296635059e-05,
      "loss": 4.0408,
      "step": 26000
    },
    {
      "epoch": 0.2172641430600819,
      "eval_loss": 3.9010603427886963,
      "eval_runtime": 40.1779,
      "eval_samples_per_second": 111.056,
      "eval_steps_per_second": 13.888,
      "step": 26000
    },
    {
      "epoch": 0.22144229965739115,
      "grad_norm": 5.577437400817871,
      "learning_rate": 3.909666862465386e-05,
      "loss": 3.9532,
      "step": 26500
    },
    {
      "epoch": 0.22562045625470042,
      "grad_norm": 4.08618688583374,
      "learning_rate": 3.888730385164052e-05,
      "loss": 3.9802,
      "step": 27000
    },
    {
      "epoch": 0.22562045625470042,
      "eval_loss": 3.8729515075683594,
      "eval_runtime": 40.2084,
      "eval_samples_per_second": 110.972,
      "eval_steps_per_second": 13.878,
      "step": 27000
    },
    {
      "epoch": 0.2297986128520097,
      "grad_norm": 6.147714614868164,
      "learning_rate": 3.8677519509943774e-05,
      "loss": 4.0436,
      "step": 27500
    },
    {
      "epoch": 0.23397676944931897,
      "grad_norm": 8.099363327026367,
      "learning_rate": 3.846815473693044e-05,
      "loss": 4.0165,
      "step": 28000
    },
    {
      "epoch": 0.23397676944931897,
      "eval_loss": 3.847057342529297,
      "eval_runtime": 40.2222,
      "eval_samples_per_second": 110.934,
      "eval_steps_per_second": 13.873,
      "step": 28000
    },
    {
      "epoch": 0.23815492604662822,
      "grad_norm": 5.891151428222656,
      "learning_rate": 3.82583703952337e-05,
      "loss": 3.9964,
      "step": 28500
    },
    {
      "epoch": 0.2423330826439375,
      "grad_norm": 4.3900861740112305,
      "learning_rate": 3.804858605353696e-05,
      "loss": 3.9467,
      "step": 29000
    },
    {
      "epoch": 0.2423330826439375,
      "eval_loss": 3.8354432582855225,
      "eval_runtime": 40.3928,
      "eval_samples_per_second": 110.465,
      "eval_steps_per_second": 13.814,
      "step": 29000
    },
    {
      "epoch": 0.24651123924124677,
      "grad_norm": 11.848061561584473,
      "learning_rate": 3.783880171184023e-05,
      "loss": 3.894,
      "step": 29500
    },
    {
      "epoch": 0.25068939583855604,
      "grad_norm": 6.982874870300293,
      "learning_rate": 3.762901737014349e-05,
      "loss": 3.9895,
      "step": 30000
    },
    {
      "epoch": 0.25068939583855604,
      "eval_loss": 3.813512086868286,
      "eval_runtime": 40.3474,
      "eval_samples_per_second": 110.59,
      "eval_steps_per_second": 13.83,
      "step": 30000
    },
    {
      "epoch": 0.2548675524358653,
      "grad_norm": 3.815357208251953,
      "learning_rate": 3.741923302844676e-05,
      "loss": 3.9362,
      "step": 30500
    },
    {
      "epoch": 0.2590457090331746,
      "grad_norm": 4.882671356201172,
      "learning_rate": 3.720944868675002e-05,
      "loss": 3.9015,
      "step": 31000
    },
    {
      "epoch": 0.2590457090331746,
      "eval_loss": 3.801868438720703,
      "eval_runtime": 40.3453,
      "eval_samples_per_second": 110.595,
      "eval_steps_per_second": 13.831,
      "step": 31000
    },
    {
      "epoch": 0.26322386563048383,
      "grad_norm": 4.451164722442627,
      "learning_rate": 3.699966434505328e-05,
      "loss": 3.7901,
      "step": 31500
    },
    {
      "epoch": 0.2674020222277931,
      "grad_norm": 6.627746105194092,
      "learning_rate": 3.679029957203994e-05,
      "loss": 3.8765,
      "step": 32000
    },
    {
      "epoch": 0.2674020222277931,
      "eval_loss": 3.786133289337158,
      "eval_runtime": 40.3847,
      "eval_samples_per_second": 110.487,
      "eval_steps_per_second": 13.817,
      "step": 32000
    },
    {
      "epoch": 0.2715801788251024,
      "grad_norm": 6.209244251251221,
      "learning_rate": 3.658051523034321e-05,
      "loss": 3.89,
      "step": 32500
    },
    {
      "epoch": 0.2757583354224116,
      "grad_norm": 6.814667701721191,
      "learning_rate": 3.637073088864648e-05,
      "loss": 3.8415,
      "step": 33000
    },
    {
      "epoch": 0.2757583354224116,
      "eval_loss": 3.7628209590911865,
      "eval_runtime": 40.404,
      "eval_samples_per_second": 110.435,
      "eval_steps_per_second": 13.811,
      "step": 33000
    },
    {
      "epoch": 0.2799364920197209,
      "grad_norm": 4.806527614593506,
      "learning_rate": 3.6160946546949734e-05,
      "loss": 3.928,
      "step": 33500
    },
    {
      "epoch": 0.2841146486170302,
      "grad_norm": 6.326642036437988,
      "learning_rate": 3.5951581773936394e-05,
      "loss": 3.8653,
      "step": 34000
    },
    {
      "epoch": 0.2841146486170302,
      "eval_loss": 3.751917600631714,
      "eval_runtime": 40.3766,
      "eval_samples_per_second": 110.51,
      "eval_steps_per_second": 13.82,
      "step": 34000
    },
    {
      "epoch": 0.2882928052143394,
      "grad_norm": 4.734166145324707,
      "learning_rate": 3.574179743223966e-05,
      "loss": 3.8688,
      "step": 34500
    },
    {
      "epoch": 0.2924709618116487,
      "grad_norm": 5.2707343101501465,
      "learning_rate": 3.553201309054292e-05,
      "loss": 3.7493,
      "step": 35000
    },
    {
      "epoch": 0.2924709618116487,
      "eval_loss": 3.7429351806640625,
      "eval_runtime": 40.3777,
      "eval_samples_per_second": 110.506,
      "eval_steps_per_second": 13.819,
      "step": 35000
    },
    {
      "epoch": 0.29664911840895797,
      "grad_norm": 4.153214454650879,
      "learning_rate": 3.5322228748846184e-05,
      "loss": 3.8655,
      "step": 35500
    },
    {
      "epoch": 0.3008272750062672,
      "grad_norm": 8.407896995544434,
      "learning_rate": 3.511244440714945e-05,
      "loss": 3.7849,
      "step": 36000
    },
    {
      "epoch": 0.3008272750062672,
      "eval_loss": 3.7235593795776367,
      "eval_runtime": 40.3782,
      "eval_samples_per_second": 110.505,
      "eval_steps_per_second": 13.819,
      "step": 36000
    },
    {
      "epoch": 0.3050054316035765,
      "grad_norm": 5.558535099029541,
      "learning_rate": 3.4902660065452714e-05,
      "loss": 3.7974,
      "step": 36500
    },
    {
      "epoch": 0.30918358820088576,
      "grad_norm": 7.9293293952941895,
      "learning_rate": 3.4693295292439374e-05,
      "loss": 3.8025,
      "step": 37000
    },
    {
      "epoch": 0.30918358820088576,
      "eval_loss": 3.7119622230529785,
      "eval_runtime": 40.3046,
      "eval_samples_per_second": 110.707,
      "eval_steps_per_second": 13.845,
      "step": 37000
    },
    {
      "epoch": 0.31336174479819506,
      "grad_norm": 7.648885250091553,
      "learning_rate": 3.4483510950742635e-05,
      "loss": 3.7967,
      "step": 37500
    },
    {
      "epoch": 0.3175399013955043,
      "grad_norm": 5.128037929534912,
      "learning_rate": 3.42737266090459e-05,
      "loss": 3.7835,
      "step": 38000
    },
    {
      "epoch": 0.3175399013955043,
      "eval_loss": 3.6989333629608154,
      "eval_runtime": 40.4275,
      "eval_samples_per_second": 110.37,
      "eval_steps_per_second": 13.802,
      "step": 38000
    },
    {
      "epoch": 0.32171805799281356,
      "grad_norm": 5.813144207000732,
      "learning_rate": 3.406394226734917e-05,
      "loss": 3.8355,
      "step": 38500
    },
    {
      "epoch": 0.32589621459012286,
      "grad_norm": 5.470829010009766,
      "learning_rate": 3.3854577494335824e-05,
      "loss": 3.7739,
      "step": 39000
    },
    {
      "epoch": 0.32589621459012286,
      "eval_loss": 3.6981310844421387,
      "eval_runtime": 40.3386,
      "eval_samples_per_second": 110.614,
      "eval_steps_per_second": 13.833,
      "step": 39000
    },
    {
      "epoch": 0.3300743711874321,
      "grad_norm": 5.560943126678467,
      "learning_rate": 3.3644793152639085e-05,
      "loss": 3.8491,
      "step": 39500
    },
    {
      "epoch": 0.33425252778474135,
      "grad_norm": 6.327647686004639,
      "learning_rate": 3.3435008810942353e-05,
      "loss": 3.6971,
      "step": 40000
    },
    {
      "epoch": 0.33425252778474135,
      "eval_loss": 3.6876156330108643,
      "eval_runtime": 40.3373,
      "eval_samples_per_second": 110.617,
      "eval_steps_per_second": 13.833,
      "step": 40000
    },
    {
      "epoch": 0.33843068438205065,
      "grad_norm": 5.916841983795166,
      "learning_rate": 3.322522446924562e-05,
      "loss": 3.7719,
      "step": 40500
    },
    {
      "epoch": 0.3426088409793599,
      "grad_norm": 5.004364967346191,
      "learning_rate": 3.3015859696232275e-05,
      "loss": 3.7641,
      "step": 41000
    },
    {
      "epoch": 0.3426088409793599,
      "eval_loss": 3.672261953353882,
      "eval_runtime": 40.3068,
      "eval_samples_per_second": 110.701,
      "eval_steps_per_second": 13.844,
      "step": 41000
    },
    {
      "epoch": 0.3467869975766692,
      "grad_norm": 5.481464862823486,
      "learning_rate": 3.2806075354535536e-05,
      "loss": 3.75,
      "step": 41500
    },
    {
      "epoch": 0.35096515417397844,
      "grad_norm": 3.293029546737671,
      "learning_rate": 3.2596291012838804e-05,
      "loss": 3.7251,
      "step": 42000
    },
    {
      "epoch": 0.35096515417397844,
      "eval_loss": 3.660274028778076,
      "eval_runtime": 40.3581,
      "eval_samples_per_second": 110.56,
      "eval_steps_per_second": 13.826,
      "step": 42000
    },
    {
      "epoch": 0.3551433107712877,
      "grad_norm": 4.981443405151367,
      "learning_rate": 3.238650667114207e-05,
      "loss": 3.6741,
      "step": 42500
    },
    {
      "epoch": 0.359321467368597,
      "grad_norm": 4.1700263023376465,
      "learning_rate": 3.217672232944533e-05,
      "loss": 3.774,
      "step": 43000
    },
    {
      "epoch": 0.359321467368597,
      "eval_loss": 3.659872055053711,
      "eval_runtime": 40.335,
      "eval_samples_per_second": 110.624,
      "eval_steps_per_second": 13.834,
      "step": 43000
    },
    {
      "epoch": 0.36349962396590624,
      "grad_norm": 5.307525157928467,
      "learning_rate": 3.1966937987748595e-05,
      "loss": 3.7479,
      "step": 43500
    },
    {
      "epoch": 0.3676777805632155,
      "grad_norm": 4.981558322906494,
      "learning_rate": 3.1757573214735255e-05,
      "loss": 3.7878,
      "step": 44000
    },
    {
      "epoch": 0.3676777805632155,
      "eval_loss": 3.647749662399292,
      "eval_runtime": 40.341,
      "eval_samples_per_second": 110.607,
      "eval_steps_per_second": 13.832,
      "step": 44000
    },
    {
      "epoch": 0.3718559371605248,
      "grad_norm": 6.054855823516846,
      "learning_rate": 3.154778887303852e-05,
      "loss": 3.7851,
      "step": 44500
    },
    {
      "epoch": 0.37603409375783403,
      "grad_norm": 5.549240589141846,
      "learning_rate": 3.133800453134178e-05,
      "loss": 3.7971,
      "step": 45000
    },
    {
      "epoch": 0.37603409375783403,
      "eval_loss": 3.6410250663757324,
      "eval_runtime": 40.3919,
      "eval_samples_per_second": 110.468,
      "eval_steps_per_second": 13.815,
      "step": 45000
    },
    {
      "epoch": 0.38021225035514333,
      "grad_norm": 7.926109313964844,
      "learning_rate": 3.112863975832844e-05,
      "loss": 3.7444,
      "step": 45500
    },
    {
      "epoch": 0.3843904069524526,
      "grad_norm": 4.797922611236572,
      "learning_rate": 3.0918855416631705e-05,
      "loss": 3.7357,
      "step": 46000
    },
    {
      "epoch": 0.3843904069524526,
      "eval_loss": 3.6285250186920166,
      "eval_runtime": 40.2976,
      "eval_samples_per_second": 110.726,
      "eval_steps_per_second": 13.847,
      "step": 46000
    },
    {
      "epoch": 0.3885685635497618,
      "grad_norm": 3.2273850440979004,
      "learning_rate": 3.0709071074934966e-05,
      "loss": 3.6811,
      "step": 46500
    },
    {
      "epoch": 0.3927467201470711,
      "grad_norm": 3.6433870792388916,
      "learning_rate": 3.049928673323823e-05,
      "loss": 3.7285,
      "step": 47000
    },
    {
      "epoch": 0.3927467201470711,
      "eval_loss": 3.622321367263794,
      "eval_runtime": 40.3211,
      "eval_samples_per_second": 110.662,
      "eval_steps_per_second": 13.839,
      "step": 47000
    },
    {
      "epoch": 0.3969248767443804,
      "grad_norm": 3.6138274669647217,
      "learning_rate": 3.0289502391541496e-05,
      "loss": 3.7385,
      "step": 47500
    },
    {
      "epoch": 0.4011030333416897,
      "grad_norm": 3.9696879386901855,
      "learning_rate": 3.0079718049844764e-05,
      "loss": 3.7095,
      "step": 48000
    },
    {
      "epoch": 0.4011030333416897,
      "eval_loss": 3.6237733364105225,
      "eval_runtime": 40.3393,
      "eval_samples_per_second": 110.612,
      "eval_steps_per_second": 13.833,
      "step": 48000
    },
    {
      "epoch": 0.4052811899389989,
      "grad_norm": 4.0967912673950195,
      "learning_rate": 2.9869933708148025e-05,
      "loss": 3.7014,
      "step": 48500
    },
    {
      "epoch": 0.40945934653630817,
      "grad_norm": 6.807064533233643,
      "learning_rate": 2.966014936645129e-05,
      "loss": 3.7479,
      "step": 49000
    },
    {
      "epoch": 0.40945934653630817,
      "eval_loss": 3.614962100982666,
      "eval_runtime": 40.3225,
      "eval_samples_per_second": 110.658,
      "eval_steps_per_second": 13.838,
      "step": 49000
    },
    {
      "epoch": 0.41363750313361747,
      "grad_norm": 5.55667781829834,
      "learning_rate": 2.9450784593437946e-05,
      "loss": 3.6814,
      "step": 49500
    },
    {
      "epoch": 0.4178156597309267,
      "grad_norm": 3.904834032058716,
      "learning_rate": 2.9241000251741214e-05,
      "loss": 3.624,
      "step": 50000
    },
    {
      "epoch": 0.4178156597309267,
      "eval_loss": 3.604703187942505,
      "eval_runtime": 40.3492,
      "eval_samples_per_second": 110.585,
      "eval_steps_per_second": 13.829,
      "step": 50000
    },
    {
      "epoch": 0.42199381632823596,
      "grad_norm": 8.520374298095703,
      "learning_rate": 2.9031215910044472e-05,
      "loss": 3.6371,
      "step": 50500
    },
    {
      "epoch": 0.42617197292554526,
      "grad_norm": 8.05156135559082,
      "learning_rate": 2.8821851137031135e-05,
      "loss": 3.7153,
      "step": 51000
    },
    {
      "epoch": 0.42617197292554526,
      "eval_loss": 3.587707996368408,
      "eval_runtime": 40.3074,
      "eval_samples_per_second": 110.699,
      "eval_steps_per_second": 13.844,
      "step": 51000
    },
    {
      "epoch": 0.4303501295228545,
      "grad_norm": 5.482362270355225,
      "learning_rate": 2.8612066795334397e-05,
      "loss": 3.6853,
      "step": 51500
    },
    {
      "epoch": 0.4345282861201638,
      "grad_norm": 5.900150775909424,
      "learning_rate": 2.840228245363766e-05,
      "loss": 3.7009,
      "step": 52000
    },
    {
      "epoch": 0.4345282861201638,
      "eval_loss": 3.5961596965789795,
      "eval_runtime": 40.3077,
      "eval_samples_per_second": 110.698,
      "eval_steps_per_second": 13.843,
      "step": 52000
    },
    {
      "epoch": 0.43870644271747306,
      "grad_norm": 4.387772083282471,
      "learning_rate": 2.8192498111940923e-05,
      "loss": 3.6555,
      "step": 52500
    },
    {
      "epoch": 0.4428845993147823,
      "grad_norm": 4.479552745819092,
      "learning_rate": 2.798271377024419e-05,
      "loss": 3.6585,
      "step": 53000
    },
    {
      "epoch": 0.4428845993147823,
      "eval_loss": 3.587613582611084,
      "eval_runtime": 40.2994,
      "eval_samples_per_second": 110.721,
      "eval_steps_per_second": 13.846,
      "step": 53000
    },
    {
      "epoch": 0.4470627559120916,
      "grad_norm": 7.089645862579346,
      "learning_rate": 2.7772929428547455e-05,
      "loss": 3.6997,
      "step": 53500
    },
    {
      "epoch": 0.45124091250940085,
      "grad_norm": 4.489483833312988,
      "learning_rate": 2.7563145086850717e-05,
      "loss": 3.6804,
      "step": 54000
    },
    {
      "epoch": 0.45124091250940085,
      "eval_loss": 3.5782344341278076,
      "eval_runtime": 40.3396,
      "eval_samples_per_second": 110.611,
      "eval_steps_per_second": 13.833,
      "step": 54000
    },
    {
      "epoch": 0.4554190691067101,
      "grad_norm": 4.279026985168457,
      "learning_rate": 2.7353360745153985e-05,
      "loss": 3.6596,
      "step": 54500
    },
    {
      "epoch": 0.4595972257040194,
      "grad_norm": 5.833207607269287,
      "learning_rate": 2.714399597214064e-05,
      "loss": 3.6249,
      "step": 55000
    },
    {
      "epoch": 0.4595972257040194,
      "eval_loss": 3.5750606060028076,
      "eval_runtime": 40.2836,
      "eval_samples_per_second": 110.765,
      "eval_steps_per_second": 13.852,
      "step": 55000
    },
    {
      "epoch": 0.46377538230132864,
      "grad_norm": 1.6167926788330078,
      "learning_rate": 2.6934631199127298e-05,
      "loss": 3.6656,
      "step": 55500
    },
    {
      "epoch": 0.46795353889863794,
      "grad_norm": 8.225107192993164,
      "learning_rate": 2.6724846857430562e-05,
      "loss": 3.5755,
      "step": 56000
    },
    {
      "epoch": 0.46795353889863794,
      "eval_loss": 3.5670828819274902,
      "eval_runtime": 40.3346,
      "eval_samples_per_second": 110.625,
      "eval_steps_per_second": 13.834,
      "step": 56000
    },
    {
      "epoch": 0.4721316954959472,
      "grad_norm": 6.227290153503418,
      "learning_rate": 2.651506251573383e-05,
      "loss": 3.658,
      "step": 56500
    },
    {
      "epoch": 0.47630985209325644,
      "grad_norm": 2.954385995864868,
      "learning_rate": 2.6305278174037092e-05,
      "loss": 3.6594,
      "step": 57000
    },
    {
      "epoch": 0.47630985209325644,
      "eval_loss": 3.5596070289611816,
      "eval_runtime": 40.3256,
      "eval_samples_per_second": 110.649,
      "eval_steps_per_second": 13.837,
      "step": 57000
    },
    {
      "epoch": 0.48048800869056574,
      "grad_norm": 4.584821701049805,
      "learning_rate": 2.6095493832340356e-05,
      "loss": 3.6365,
      "step": 57500
    },
    {
      "epoch": 0.484666165287875,
      "grad_norm": 5.709506511688232,
      "learning_rate": 2.5885709490643618e-05,
      "loss": 3.7103,
      "step": 58000
    },
    {
      "epoch": 0.484666165287875,
      "eval_loss": 3.5584540367126465,
      "eval_runtime": 40.3216,
      "eval_samples_per_second": 110.66,
      "eval_steps_per_second": 13.839,
      "step": 58000
    },
    {
      "epoch": 0.48884432188518423,
      "grad_norm": 2.718370199203491,
      "learning_rate": 2.5675925148946882e-05,
      "loss": 3.6135,
      "step": 58500
    },
    {
      "epoch": 0.49302247848249353,
      "grad_norm": 5.212717056274414,
      "learning_rate": 2.546614080725015e-05,
      "loss": 3.6775,
      "step": 59000
    },
    {
      "epoch": 0.49302247848249353,
      "eval_loss": 3.54938006401062,
      "eval_runtime": 40.2984,
      "eval_samples_per_second": 110.724,
      "eval_steps_per_second": 13.847,
      "step": 59000
    },
    {
      "epoch": 0.4972006350798028,
      "grad_norm": 5.458935737609863,
      "learning_rate": 2.5256776034236807e-05,
      "loss": 3.6635,
      "step": 59500
    },
    {
      "epoch": 0.5013787916771121,
      "grad_norm": 6.3775248527526855,
      "learning_rate": 2.504699169254007e-05,
      "loss": 3.6256,
      "step": 60000
    },
    {
      "epoch": 0.5013787916771121,
      "eval_loss": 3.5428049564361572,
      "eval_runtime": 40.2928,
      "eval_samples_per_second": 110.739,
      "eval_steps_per_second": 13.849,
      "step": 60000
    },
    {
      "epoch": 0.5055569482744213,
      "grad_norm": 3.5704195499420166,
      "learning_rate": 2.4837207350843333e-05,
      "loss": 3.591,
      "step": 60500
    },
    {
      "epoch": 0.5097351048717306,
      "grad_norm": 4.804786682128906,
      "learning_rate": 2.46274230091466e-05,
      "loss": 3.6524,
      "step": 61000
    },
    {
      "epoch": 0.5097351048717306,
      "eval_loss": 3.5428876876831055,
      "eval_runtime": 40.303,
      "eval_samples_per_second": 110.711,
      "eval_steps_per_second": 13.845,
      "step": 61000
    },
    {
      "epoch": 0.5139132614690398,
      "grad_norm": 5.669532775878906,
      "learning_rate": 2.4418058236133258e-05,
      "loss": 3.613,
      "step": 61500
    },
    {
      "epoch": 0.5180914180663492,
      "grad_norm": 6.797584056854248,
      "learning_rate": 2.420827389443652e-05,
      "loss": 3.6231,
      "step": 62000
    },
    {
      "epoch": 0.5180914180663492,
      "eval_loss": 3.537569046020508,
      "eval_runtime": 40.3095,
      "eval_samples_per_second": 110.694,
      "eval_steps_per_second": 13.843,
      "step": 62000
    },
    {
      "epoch": 0.5222695746636584,
      "grad_norm": 3.8286426067352295,
      "learning_rate": 2.3998489552739783e-05,
      "loss": 3.6407,
      "step": 62500
    },
    {
      "epoch": 0.5264477312609677,
      "grad_norm": 6.5328850746154785,
      "learning_rate": 2.3788705211043048e-05,
      "loss": 3.6704,
      "step": 63000
    },
    {
      "epoch": 0.5264477312609677,
      "eval_loss": 3.5320422649383545,
      "eval_runtime": 40.3117,
      "eval_samples_per_second": 110.687,
      "eval_steps_per_second": 13.842,
      "step": 63000
    },
    {
      "epoch": 0.5306258878582769,
      "grad_norm": 5.264260768890381,
      "learning_rate": 2.3579340438029708e-05,
      "loss": 3.6362,
      "step": 63500
    },
    {
      "epoch": 0.5348040444555862,
      "grad_norm": 4.104739189147949,
      "learning_rate": 2.336955609633297e-05,
      "loss": 3.5952,
      "step": 64000
    },
    {
      "epoch": 0.5348040444555862,
      "eval_loss": 3.5205154418945312,
      "eval_runtime": 40.2973,
      "eval_samples_per_second": 110.727,
      "eval_steps_per_second": 13.847,
      "step": 64000
    },
    {
      "epoch": 0.5389822010528955,
      "grad_norm": 3.8011248111724854,
      "learning_rate": 2.3159771754636237e-05,
      "loss": 3.6439,
      "step": 64500
    },
    {
      "epoch": 0.5431603576502048,
      "grad_norm": 3.5783491134643555,
      "learning_rate": 2.29499874129395e-05,
      "loss": 3.6597,
      "step": 65000
    },
    {
      "epoch": 0.5431603576502048,
      "eval_loss": 3.5196046829223633,
      "eval_runtime": 40.4106,
      "eval_samples_per_second": 110.416,
      "eval_steps_per_second": 13.808,
      "step": 65000
    },
    {
      "epoch": 0.547338514247514,
      "grad_norm": 6.129964351654053,
      "learning_rate": 2.2740203071242763e-05,
      "loss": 3.6149,
      "step": 65500
    },
    {
      "epoch": 0.5515166708448233,
      "grad_norm": 5.428370952606201,
      "learning_rate": 2.253083829822942e-05,
      "loss": 3.543,
      "step": 66000
    },
    {
      "epoch": 0.5515166708448233,
      "eval_loss": 3.5081422328948975,
      "eval_runtime": 40.3325,
      "eval_samples_per_second": 110.63,
      "eval_steps_per_second": 13.835,
      "step": 66000
    },
    {
      "epoch": 0.5556948274421325,
      "grad_norm": 4.526833534240723,
      "learning_rate": 2.232147352521608e-05,
      "loss": 5.1116,
      "step": 66500
    },
    {
      "epoch": 0.5598729840394417,
      "grad_norm": 11.199429512023926,
      "learning_rate": 2.2111689183519345e-05,
      "loss": 4.8238,
      "step": 67000
    },
    {
      "epoch": 0.5598729840394417,
      "eval_loss": 4.616936683654785,
      "eval_runtime": 40.3658,
      "eval_samples_per_second": 110.539,
      "eval_steps_per_second": 13.824,
      "step": 67000
    },
    {
      "epoch": 0.5640511406367511,
      "grad_norm": 8.317018508911133,
      "learning_rate": 2.1901904841822606e-05,
      "loss": 4.683,
      "step": 67500
    },
    {
      "epoch": 0.5682292972340603,
      "grad_norm": 4.753358364105225,
      "learning_rate": 2.169212050012587e-05,
      "loss": 4.5343,
      "step": 68000
    },
    {
      "epoch": 0.5682292972340603,
      "eval_loss": 4.415843963623047,
      "eval_runtime": 40.3527,
      "eval_samples_per_second": 110.575,
      "eval_steps_per_second": 13.828,
      "step": 68000
    },
    {
      "epoch": 0.5724074538313696,
      "grad_norm": 13.623976707458496,
      "learning_rate": 2.148275572711253e-05,
      "loss": 4.4902,
      "step": 68500
    },
    {
      "epoch": 0.5765856104286788,
      "grad_norm": 6.011263847351074,
      "learning_rate": 2.1272971385415795e-05,
      "loss": 4.4547,
      "step": 69000
    },
    {
      "epoch": 0.5765856104286788,
      "eval_loss": 4.297609806060791,
      "eval_runtime": 40.3782,
      "eval_samples_per_second": 110.505,
      "eval_steps_per_second": 13.819,
      "step": 69000
    },
    {
      "epoch": 0.5807637670259881,
      "grad_norm": 5.958744049072266,
      "learning_rate": 2.1063187043719056e-05,
      "loss": 4.413,
      "step": 69500
    },
    {
      "epoch": 0.5849419236232974,
      "grad_norm": 3.0502681732177734,
      "learning_rate": 2.0853402702022324e-05,
      "loss": 4.3609,
      "step": 70000
    },
    {
      "epoch": 0.5849419236232974,
      "eval_loss": 4.230189323425293,
      "eval_runtime": 40.3883,
      "eval_samples_per_second": 110.478,
      "eval_steps_per_second": 13.816,
      "step": 70000
    },
    {
      "epoch": 0.5891200802206067,
      "grad_norm": 4.797601699829102,
      "learning_rate": 2.064403792900898e-05,
      "loss": 4.3879,
      "step": 70500
    },
    {
      "epoch": 0.5932982368179159,
      "grad_norm": 7.879065990447998,
      "learning_rate": 2.0434253587312242e-05,
      "loss": 4.3503,
      "step": 71000
    },
    {
      "epoch": 0.5932982368179159,
      "eval_loss": 4.177266597747803,
      "eval_runtime": 40.395,
      "eval_samples_per_second": 110.459,
      "eval_steps_per_second": 13.814,
      "step": 71000
    },
    {
      "epoch": 0.5974763934152252,
      "grad_norm": 5.502597332000732,
      "learning_rate": 2.0224469245615507e-05,
      "loss": 4.3594,
      "step": 71500
    },
    {
      "epoch": 0.6016545500125344,
      "grad_norm": 14.831869125366211,
      "learning_rate": 2.0014684903918775e-05,
      "loss": 4.2399,
      "step": 72000
    },
    {
      "epoch": 0.6016545500125344,
      "eval_loss": 4.133894443511963,
      "eval_runtime": 40.3876,
      "eval_samples_per_second": 110.479,
      "eval_steps_per_second": 13.816,
      "step": 72000
    },
    {
      "epoch": 0.6058327066098438,
      "grad_norm": 7.395357608795166,
      "learning_rate": 1.980532013090543e-05,
      "loss": 4.2413,
      "step": 72500
    },
    {
      "epoch": 0.610010863207153,
      "grad_norm": 8.504011154174805,
      "learning_rate": 1.9595535789208693e-05,
      "loss": 4.3134,
      "step": 73000
    },
    {
      "epoch": 0.610010863207153,
      "eval_loss": 4.093682765960693,
      "eval_runtime": 40.4955,
      "eval_samples_per_second": 110.185,
      "eval_steps_per_second": 13.779,
      "step": 73000
    },
    {
      "epoch": 0.6141890198044623,
      "grad_norm": 4.6213698387146,
      "learning_rate": 1.9385751447511957e-05,
      "loss": 4.2654,
      "step": 73500
    },
    {
      "epoch": 0.6183671764017715,
      "grad_norm": 7.031868934631348,
      "learning_rate": 1.9176386674498617e-05,
      "loss": 4.1979,
      "step": 74000
    },
    {
      "epoch": 0.6183671764017715,
      "eval_loss": 4.06504487991333,
      "eval_runtime": 40.3928,
      "eval_samples_per_second": 110.465,
      "eval_steps_per_second": 13.814,
      "step": 74000
    },
    {
      "epoch": 0.6225453329990808,
      "grad_norm": 6.018976211547852,
      "learning_rate": 1.8966602332801882e-05,
      "loss": 4.1942,
      "step": 74500
    },
    {
      "epoch": 0.6267234895963901,
      "grad_norm": 5.645545959472656,
      "learning_rate": 1.8756817991105143e-05,
      "loss": 4.2067,
      "step": 75000
    },
    {
      "epoch": 0.6267234895963901,
      "eval_loss": 4.035527229309082,
      "eval_runtime": 40.4831,
      "eval_samples_per_second": 110.219,
      "eval_steps_per_second": 13.784,
      "step": 75000
    },
    {
      "epoch": 0.6309016461936994,
      "grad_norm": 9.228676795959473,
      "learning_rate": 1.854703364940841e-05,
      "loss": 4.1713,
      "step": 75500
    },
    {
      "epoch": 0.6350798027910086,
      "grad_norm": 5.62150239944458,
      "learning_rate": 1.8337668876395068e-05,
      "loss": 4.2362,
      "step": 76000
    },
    {
      "epoch": 0.6350798027910086,
      "eval_loss": 4.016374588012695,
      "eval_runtime": 40.3961,
      "eval_samples_per_second": 110.456,
      "eval_steps_per_second": 13.813,
      "step": 76000
    },
    {
      "epoch": 0.6392579593883179,
      "grad_norm": 6.281893730163574,
      "learning_rate": 1.812788453469833e-05,
      "loss": 4.1349,
      "step": 76500
    },
    {
      "epoch": 0.6434361159856271,
      "grad_norm": 5.5328497886657715,
      "learning_rate": 1.7918100193001594e-05,
      "loss": 4.0535,
      "step": 77000
    },
    {
      "epoch": 0.6434361159856271,
      "eval_loss": 3.986826181411743,
      "eval_runtime": 40.408,
      "eval_samples_per_second": 110.424,
      "eval_steps_per_second": 13.809,
      "step": 77000
    },
    {
      "epoch": 0.6476142725829364,
      "grad_norm": 11.436102867126465,
      "learning_rate": 1.7708315851304862e-05,
      "loss": 4.1203,
      "step": 77500
    },
    {
      "epoch": 0.6517924291802457,
      "grad_norm": 6.688998222351074,
      "learning_rate": 1.7498531509608123e-05,
      "loss": 4.0486,
      "step": 78000
    },
    {
      "epoch": 0.6517924291802457,
      "eval_loss": 3.9673120975494385,
      "eval_runtime": 40.4149,
      "eval_samples_per_second": 110.405,
      "eval_steps_per_second": 13.807,
      "step": 78000
    },
    {
      "epoch": 0.655970585777555,
      "grad_norm": 8.663886070251465,
      "learning_rate": 1.7288747167911388e-05,
      "loss": 4.0426,
      "step": 78500
    },
    {
      "epoch": 0.6601487423748642,
      "grad_norm": 5.486411094665527,
      "learning_rate": 1.7078962826214652e-05,
      "loss": 4.0643,
      "step": 79000
    },
    {
      "epoch": 0.6601487423748642,
      "eval_loss": 3.941148042678833,
      "eval_runtime": 40.4048,
      "eval_samples_per_second": 110.432,
      "eval_steps_per_second": 13.81,
      "step": 79000
    },
    {
      "epoch": 0.6643268989721735,
      "grad_norm": 6.589982509613037,
      "learning_rate": 1.6869178484517917e-05,
      "loss": 4.1028,
      "step": 79500
    },
    {
      "epoch": 0.6685050555694827,
      "grad_norm": 5.7215800285339355,
      "learning_rate": 1.6659813711504574e-05,
      "loss": 4.0685,
      "step": 80000
    },
    {
      "epoch": 0.6685050555694827,
      "eval_loss": 3.935919761657715,
      "eval_runtime": 40.3891,
      "eval_samples_per_second": 110.475,
      "eval_steps_per_second": 13.816,
      "step": 80000
    },
    {
      "epoch": 0.672683212166792,
      "grad_norm": 4.948646068572998,
      "learning_rate": 1.645002936980784e-05,
      "loss": 4.0631,
      "step": 80500
    },
    {
      "epoch": 0.6768613687641013,
      "grad_norm": 10.537930488586426,
      "learning_rate": 1.6240245028111103e-05,
      "loss": 4.0076,
      "step": 81000
    },
    {
      "epoch": 0.6768613687641013,
      "eval_loss": 3.921271800994873,
      "eval_runtime": 40.394,
      "eval_samples_per_second": 110.462,
      "eval_steps_per_second": 13.814,
      "step": 81000
    },
    {
      "epoch": 0.6810395253614105,
      "grad_norm": 8.366657257080078,
      "learning_rate": 1.603088025509776e-05,
      "loss": 4.0462,
      "step": 81500
    },
    {
      "epoch": 0.6852176819587198,
      "grad_norm": 8.010354042053223,
      "learning_rate": 1.5821095913401024e-05,
      "loss": 4.0686,
      "step": 82000
    },
    {
      "epoch": 0.6852176819587198,
      "eval_loss": 3.901902914047241,
      "eval_runtime": 40.3855,
      "eval_samples_per_second": 110.485,
      "eval_steps_per_second": 13.817,
      "step": 82000
    },
    {
      "epoch": 0.689395838556029,
      "grad_norm": 7.819444179534912,
      "learning_rate": 1.561131157170429e-05,
      "loss": 3.9706,
      "step": 82500
    },
    {
      "epoch": 0.6935739951533384,
      "grad_norm": 5.248921871185303,
      "learning_rate": 1.5401527230007554e-05,
      "loss": 3.9257,
      "step": 83000
    },
    {
      "epoch": 0.6935739951533384,
      "eval_loss": 3.8821218013763428,
      "eval_runtime": 40.3861,
      "eval_samples_per_second": 110.484,
      "eval_steps_per_second": 13.817,
      "step": 83000
    },
    {
      "epoch": 0.6977521517506476,
      "grad_norm": 6.49260139465332,
      "learning_rate": 1.5191742888310817e-05,
      "loss": 4.0404,
      "step": 83500
    },
    {
      "epoch": 0.7019303083479569,
      "grad_norm": 5.810378074645996,
      "learning_rate": 1.4981958546614081e-05,
      "loss": 3.9735,
      "step": 84000
    },
    {
      "epoch": 0.7019303083479569,
      "eval_loss": 3.877354145050049,
      "eval_runtime": 40.4901,
      "eval_samples_per_second": 110.2,
      "eval_steps_per_second": 13.781,
      "step": 84000
    },
    {
      "epoch": 0.7061084649452661,
      "grad_norm": 5.234148025512695,
      "learning_rate": 1.4772174204917347e-05,
      "loss": 3.8852,
      "step": 84500
    },
    {
      "epoch": 0.7102866215425754,
      "grad_norm": 10.57536506652832,
      "learning_rate": 1.456238986322061e-05,
      "loss": 3.9721,
      "step": 85000
    },
    {
      "epoch": 0.7102866215425754,
      "eval_loss": 3.865615129470825,
      "eval_runtime": 40.3575,
      "eval_samples_per_second": 110.562,
      "eval_steps_per_second": 13.826,
      "step": 85000
    },
    {
      "epoch": 0.7144647781398847,
      "grad_norm": 4.43326473236084,
      "learning_rate": 1.4353025090207267e-05,
      "loss": 4.0048,
      "step": 85500
    },
    {
      "epoch": 0.718642934737194,
      "grad_norm": 12.378066062927246,
      "learning_rate": 1.414324074851053e-05,
      "loss": 3.9356,
      "step": 86000
    },
    {
      "epoch": 0.718642934737194,
      "eval_loss": 3.8526625633239746,
      "eval_runtime": 40.4019,
      "eval_samples_per_second": 110.44,
      "eval_steps_per_second": 13.811,
      "step": 86000
    },
    {
      "epoch": 0.7228210913345032,
      "grad_norm": 6.761714458465576,
      "learning_rate": 1.3933456406813796e-05,
      "loss": 3.9714,
      "step": 86500
    },
    {
      "epoch": 0.7269992479318125,
      "grad_norm": 7.718149662017822,
      "learning_rate": 1.3723672065117061e-05,
      "loss": 3.9137,
      "step": 87000
    },
    {
      "epoch": 0.7269992479318125,
      "eval_loss": 3.8465211391448975,
      "eval_runtime": 40.4199,
      "eval_samples_per_second": 110.391,
      "eval_steps_per_second": 13.805,
      "step": 87000
    },
    {
      "epoch": 0.7311774045291217,
      "grad_norm": 6.301932334899902,
      "learning_rate": 1.3514307292103718e-05,
      "loss": 3.9436,
      "step": 87500
    },
    {
      "epoch": 0.735355561126431,
      "grad_norm": 5.401347637176514,
      "learning_rate": 1.3304522950406984e-05,
      "loss": 3.908,
      "step": 88000
    },
    {
      "epoch": 0.735355561126431,
      "eval_loss": 3.828646659851074,
      "eval_runtime": 40.4627,
      "eval_samples_per_second": 110.275,
      "eval_steps_per_second": 13.79,
      "step": 88000
    },
    {
      "epoch": 0.7395337177237403,
      "grad_norm": 5.284492492675781,
      "learning_rate": 1.3094738608710247e-05,
      "loss": 3.9225,
      "step": 88500
    },
    {
      "epoch": 0.7437118743210496,
      "grad_norm": 7.627567768096924,
      "learning_rate": 1.288495426701351e-05,
      "loss": 3.9705,
      "step": 89000
    },
    {
      "epoch": 0.7437118743210496,
      "eval_loss": 3.8248815536499023,
      "eval_runtime": 40.4521,
      "eval_samples_per_second": 110.303,
      "eval_steps_per_second": 13.794,
      "step": 89000
    },
    {
      "epoch": 0.7478900309183588,
      "grad_norm": 10.147069931030273,
      "learning_rate": 1.2675169925316775e-05,
      "loss": 3.9042,
      "step": 89500
    },
    {
      "epoch": 0.7520681875156681,
      "grad_norm": 5.002488136291504,
      "learning_rate": 1.2465805152303433e-05,
      "loss": 3.9459,
      "step": 90000
    },
    {
      "epoch": 0.7520681875156681,
      "eval_loss": 3.8160951137542725,
      "eval_runtime": 40.3907,
      "eval_samples_per_second": 110.471,
      "eval_steps_per_second": 13.815,
      "step": 90000
    },
    {
      "epoch": 0.7562463441129773,
      "grad_norm": 17.16789436340332,
      "learning_rate": 1.2256440379290091e-05,
      "loss": 4.0108,
      "step": 90500
    },
    {
      "epoch": 0.7604245007102867,
      "grad_norm": 7.082775115966797,
      "learning_rate": 1.2046656037593354e-05,
      "loss": 3.9803,
      "step": 91000
    },
    {
      "epoch": 0.7604245007102867,
      "eval_loss": 3.807090997695923,
      "eval_runtime": 40.3824,
      "eval_samples_per_second": 110.494,
      "eval_steps_per_second": 13.818,
      "step": 91000
    },
    {
      "epoch": 0.7646026573075959,
      "grad_norm": 5.471470832824707,
      "learning_rate": 1.1836871695896619e-05,
      "loss": 3.9751,
      "step": 91500
    },
    {
      "epoch": 0.7687808139049052,
      "grad_norm": 10.89763355255127,
      "learning_rate": 1.1627087354199883e-05,
      "loss": 3.9095,
      "step": 92000
    },
    {
      "epoch": 0.7687808139049052,
      "eval_loss": 3.7957355976104736,
      "eval_runtime": 40.3893,
      "eval_samples_per_second": 110.475,
      "eval_steps_per_second": 13.816,
      "step": 92000
    },
    {
      "epoch": 0.7729589705022144,
      "grad_norm": 9.122933387756348,
      "learning_rate": 1.1417303012503146e-05,
      "loss": 3.9344,
      "step": 92500
    },
    {
      "epoch": 0.7771371270995237,
      "grad_norm": 12.458266258239746,
      "learning_rate": 1.1207518670806413e-05,
      "loss": 3.7938,
      "step": 93000
    },
    {
      "epoch": 0.7771371270995237,
      "eval_loss": 3.790499210357666,
      "eval_runtime": 40.3688,
      "eval_samples_per_second": 110.531,
      "eval_steps_per_second": 13.823,
      "step": 93000
    },
    {
      "epoch": 0.781315283696833,
      "grad_norm": 7.447702407836914,
      "learning_rate": 1.0997734329109676e-05,
      "loss": 3.9318,
      "step": 93500
    },
    {
      "epoch": 0.7854934402941423,
      "grad_norm": 5.904074668884277,
      "learning_rate": 1.078794998741294e-05,
      "loss": 3.851,
      "step": 94000
    },
    {
      "epoch": 0.7854934402941423,
      "eval_loss": 3.7797744274139404,
      "eval_runtime": 40.3311,
      "eval_samples_per_second": 110.634,
      "eval_steps_per_second": 13.835,
      "step": 94000
    },
    {
      "epoch": 0.7896715968914515,
      "grad_norm": 8.686776161193848,
      "learning_rate": 1.0578585214399597e-05,
      "loss": 3.886,
      "step": 94500
    },
    {
      "epoch": 0.7938497534887607,
      "grad_norm": 9.1967134475708,
      "learning_rate": 1.0368800872702862e-05,
      "loss": 3.8963,
      "step": 95000
    },
    {
      "epoch": 0.7938497534887607,
      "eval_loss": 3.777365207672119,
      "eval_runtime": 40.3669,
      "eval_samples_per_second": 110.536,
      "eval_steps_per_second": 13.823,
      "step": 95000
    },
    {
      "epoch": 0.79802791008607,
      "grad_norm": 8.713147163391113,
      "learning_rate": 1.0159016531006126e-05,
      "loss": 3.9161,
      "step": 95500
    },
    {
      "epoch": 0.8022060666833793,
      "grad_norm": 4.081088542938232,
      "learning_rate": 9.94923218930939e-06,
      "loss": 3.8978,
      "step": 96000
    },
    {
      "epoch": 0.8022060666833793,
      "eval_loss": 3.769115447998047,
      "eval_runtime": 40.3884,
      "eval_samples_per_second": 110.477,
      "eval_steps_per_second": 13.816,
      "step": 96000
    },
    {
      "epoch": 0.8063842232806886,
      "grad_norm": 6.48073673248291,
      "learning_rate": 9.739447847612655e-06,
      "loss": 3.8779,
      "step": 96500
    },
    {
      "epoch": 0.8105623798779978,
      "grad_norm": 10.993706703186035,
      "learning_rate": 9.530083074599312e-06,
      "loss": 3.8931,
      "step": 97000
    },
    {
      "epoch": 0.8105623798779978,
      "eval_loss": 3.7640254497528076,
      "eval_runtime": 40.412,
      "eval_samples_per_second": 110.413,
      "eval_steps_per_second": 13.808,
      "step": 97000
    },
    {
      "epoch": 0.8147405364753071,
      "grad_norm": 9.473855018615723,
      "learning_rate": 9.320298732902577e-06,
      "loss": 3.8533,
      "step": 97500
    },
    {
      "epoch": 0.8189186930726163,
      "grad_norm": 6.103699684143066,
      "learning_rate": 9.110933959889233e-06,
      "loss": 3.7507,
      "step": 98000
    },
    {
      "epoch": 0.8189186930726163,
      "eval_loss": 3.759049415588379,
      "eval_runtime": 40.3873,
      "eval_samples_per_second": 110.48,
      "eval_steps_per_second": 13.816,
      "step": 98000
    },
    {
      "epoch": 0.8230968496699256,
      "grad_norm": 7.411025524139404,
      "learning_rate": 8.9011496181925e-06,
      "loss": 3.8444,
      "step": 98500
    },
    {
      "epoch": 0.8272750062672349,
      "grad_norm": 10.114791870117188,
      "learning_rate": 8.691365276495763e-06,
      "loss": 3.8984,
      "step": 99000
    },
    {
      "epoch": 0.8272750062672349,
      "eval_loss": 3.753899574279785,
      "eval_runtime": 40.3896,
      "eval_samples_per_second": 110.474,
      "eval_steps_per_second": 13.815,
      "step": 99000
    },
    {
      "epoch": 0.8314531628645442,
      "grad_norm": 5.56119441986084,
      "learning_rate": 8.481580934799027e-06,
      "loss": 3.8441,
      "step": 99500
    },
    {
      "epoch": 0.8356313194618534,
      "grad_norm": 7.270823001861572,
      "learning_rate": 8.27179659310229e-06,
      "loss": 3.8968,
      "step": 100000
    },
    {
      "epoch": 0.8356313194618534,
      "eval_loss": 3.7469892501831055,
      "eval_runtime": 40.37,
      "eval_samples_per_second": 110.528,
      "eval_steps_per_second": 13.822,
      "step": 100000
    },
    {
      "epoch": 0.8398094760591627,
      "grad_norm": 5.920239448547363,
      "learning_rate": 8.062012251405555e-06,
      "loss": 3.8607,
      "step": 100500
    },
    {
      "epoch": 0.8439876326564719,
      "grad_norm": 6.5870442390441895,
      "learning_rate": 7.85222790970882e-06,
      "loss": 3.8543,
      "step": 101000
    },
    {
      "epoch": 0.8439876326564719,
      "eval_loss": 3.7423293590545654,
      "eval_runtime": 40.3917,
      "eval_samples_per_second": 110.468,
      "eval_steps_per_second": 13.815,
      "step": 101000
    },
    {
      "epoch": 0.8481657892537813,
      "grad_norm": 11.815503120422363,
      "learning_rate": 7.642443568012084e-06,
      "loss": 3.7859,
      "step": 101500
    },
    {
      "epoch": 0.8523439458510905,
      "grad_norm": 9.67881965637207,
      "learning_rate": 7.4330787949987425e-06,
      "loss": 3.8422,
      "step": 102000
    },
    {
      "epoch": 0.8523439458510905,
      "eval_loss": 3.739718198776245,
      "eval_runtime": 40.4029,
      "eval_samples_per_second": 110.438,
      "eval_steps_per_second": 13.811,
      "step": 102000
    },
    {
      "epoch": 0.8565221024483998,
      "grad_norm": 9.272939682006836,
      "learning_rate": 7.223294453302005e-06,
      "loss": 3.7835,
      "step": 102500
    },
    {
      "epoch": 0.860700259045709,
      "grad_norm": 5.5945000648498535,
      "learning_rate": 7.01351011160527e-06,
      "loss": 3.8318,
      "step": 103000
    },
    {
      "epoch": 0.860700259045709,
      "eval_loss": 3.7335968017578125,
      "eval_runtime": 40.3673,
      "eval_samples_per_second": 110.535,
      "eval_steps_per_second": 13.823,
      "step": 103000
    },
    {
      "epoch": 0.8648784156430183,
      "grad_norm": 9.059176445007324,
      "learning_rate": 6.803725769908534e-06,
      "loss": 3.822,
      "step": 103500
    },
    {
      "epoch": 0.8690565722403276,
      "grad_norm": 5.025524139404297,
      "learning_rate": 6.5939414282117985e-06,
      "loss": 3.7921,
      "step": 104000
    },
    {
      "epoch": 0.8690565722403276,
      "eval_loss": 3.726698637008667,
      "eval_runtime": 40.365,
      "eval_samples_per_second": 110.541,
      "eval_steps_per_second": 13.824,
      "step": 104000
    },
    {
      "epoch": 0.8732347288376369,
      "grad_norm": 6.096043586730957,
      "learning_rate": 6.384157086515063e-06,
      "loss": 3.7771,
      "step": 104500
    },
    {
      "epoch": 0.8774128854349461,
      "grad_norm": 10.153267860412598,
      "learning_rate": 6.174792313501721e-06,
      "loss": 3.7626,
      "step": 105000
    },
    {
      "epoch": 0.8774128854349461,
      "eval_loss": 3.728839635848999,
      "eval_runtime": 40.2024,
      "eval_samples_per_second": 110.988,
      "eval_steps_per_second": 13.88,
      "step": 105000
    },
    {
      "epoch": 0.8815910420322554,
      "grad_norm": 11.755720138549805,
      "learning_rate": 5.965007971804984e-06,
      "loss": 3.8455,
      "step": 105500
    },
    {
      "epoch": 0.8857691986295646,
      "grad_norm": 7.943704605102539,
      "learning_rate": 5.755223630108249e-06,
      "loss": 3.8765,
      "step": 106000
    },
    {
      "epoch": 0.8857691986295646,
      "eval_loss": 3.7219812870025635,
      "eval_runtime": 40.2173,
      "eval_samples_per_second": 110.947,
      "eval_steps_per_second": 13.875,
      "step": 106000
    },
    {
      "epoch": 0.8899473552268738,
      "grad_norm": 3.804535150527954,
      "learning_rate": 5.545439288411513e-06,
      "loss": 3.8256,
      "step": 106500
    },
    {
      "epoch": 0.8941255118241832,
      "grad_norm": 11.546579360961914,
      "learning_rate": 5.3356549467147775e-06,
      "loss": 3.7837,
      "step": 107000
    },
    {
      "epoch": 0.8941255118241832,
      "eval_loss": 3.7193517684936523,
      "eval_runtime": 40.3063,
      "eval_samples_per_second": 110.702,
      "eval_steps_per_second": 13.844,
      "step": 107000
    },
    {
      "epoch": 0.8983036684214925,
      "grad_norm": 5.996183395385742,
      "learning_rate": 5.125870605018042e-06,
      "loss": 3.8781,
      "step": 107500
    },
    {
      "epoch": 0.9024818250188017,
      "grad_norm": 7.48162841796875,
      "learning_rate": 4.916086263321306e-06,
      "loss": 3.8106,
      "step": 108000
    },
    {
      "epoch": 0.9024818250188017,
      "eval_loss": 3.713139295578003,
      "eval_runtime": 40.3495,
      "eval_samples_per_second": 110.584,
      "eval_steps_per_second": 13.829,
      "step": 108000
    },
    {
      "epoch": 0.9066599816161109,
      "grad_norm": 6.757467746734619,
      "learning_rate": 4.706721490307963e-06,
      "loss": 5.0696,
      "step": 108500
    },
    {
      "epoch": 0.9108381382134202,
      "grad_norm": 7.004515647888184,
      "learning_rate": 4.496937148611228e-06,
      "loss": 4.8587,
      "step": 109000
    },
    {
      "epoch": 0.9108381382134202,
      "eval_loss": 4.723120212554932,
      "eval_runtime": 40.383,
      "eval_samples_per_second": 110.492,
      "eval_steps_per_second": 13.818,
      "step": 109000
    },
    {
      "epoch": 0.9150162948107295,
      "grad_norm": 9.177141189575195,
      "learning_rate": 4.287152806914492e-06,
      "loss": 4.8224,
      "step": 109500
    },
    {
      "epoch": 0.9191944514080388,
      "grad_norm": 5.546660900115967,
      "learning_rate": 4.0773684652177565e-06,
      "loss": 4.7712,
      "step": 110000
    },
    {
      "epoch": 0.9191944514080388,
      "eval_loss": 4.593838214874268,
      "eval_runtime": 40.3004,
      "eval_samples_per_second": 110.719,
      "eval_steps_per_second": 13.846,
      "step": 110000
    },
    {
      "epoch": 0.923372608005348,
      "grad_norm": 7.7168169021606445,
      "learning_rate": 3.868003692204414e-06,
      "loss": 4.7148,
      "step": 110500
    },
    {
      "epoch": 0.9275507646026573,
      "grad_norm": 6.541009426116943,
      "learning_rate": 3.658219350507678e-06,
      "loss": 4.6866,
      "step": 111000
    },
    {
      "epoch": 0.9275507646026573,
      "eval_loss": 4.512722969055176,
      "eval_runtime": 40.2842,
      "eval_samples_per_second": 110.763,
      "eval_steps_per_second": 13.852,
      "step": 111000
    },
    {
      "epoch": 0.9317289211999665,
      "grad_norm": 8.877721786499023,
      "learning_rate": 3.448435008810943e-06,
      "loss": 4.6519,
      "step": 111500
    },
    {
      "epoch": 0.9359070777972759,
      "grad_norm": 11.03339958190918,
      "learning_rate": 3.238650667114207e-06,
      "loss": 4.5986,
      "step": 112000
    },
    {
      "epoch": 0.9359070777972759,
      "eval_loss": 4.453566074371338,
      "eval_runtime": 40.3468,
      "eval_samples_per_second": 110.591,
      "eval_steps_per_second": 13.83,
      "step": 112000
    },
    {
      "epoch": 0.9400852343945851,
      "grad_norm": 14.915512084960938,
      "learning_rate": 3.0292858941008645e-06,
      "loss": 4.6582,
      "step": 112500
    },
    {
      "epoch": 0.9442633909918944,
      "grad_norm": 7.842238903045654,
      "learning_rate": 2.8199211210875224e-06,
      "loss": 4.6007,
      "step": 113000
    },
    {
      "epoch": 0.9442633909918944,
      "eval_loss": 4.413528919219971,
      "eval_runtime": 40.3191,
      "eval_samples_per_second": 110.667,
      "eval_steps_per_second": 13.84,
      "step": 113000
    },
    {
      "epoch": 0.9484415475892036,
      "grad_norm": 5.456047534942627,
      "learning_rate": 2.6101367793907866e-06,
      "loss": 4.5077,
      "step": 113500
    },
    {
      "epoch": 0.9526197041865129,
      "grad_norm": 8.683294296264648,
      "learning_rate": 2.4003524376940504e-06,
      "loss": 4.5176,
      "step": 114000
    },
    {
      "epoch": 0.9526197041865129,
      "eval_loss": 4.385701656341553,
      "eval_runtime": 40.2816,
      "eval_samples_per_second": 110.77,
      "eval_steps_per_second": 13.852,
      "step": 114000
    },
    {
      "epoch": 0.9567978607838222,
      "grad_norm": 6.126138210296631,
      "learning_rate": 2.1905680959973146e-06,
      "loss": 4.4375,
      "step": 114500
    },
    {
      "epoch": 0.9609760173811315,
      "grad_norm": 8.375338554382324,
      "learning_rate": 1.9807837543005793e-06,
      "loss": 4.505,
      "step": 115000
    },
    {
      "epoch": 0.9609760173811315,
      "eval_loss": 4.364550590515137,
      "eval_runtime": 40.3282,
      "eval_samples_per_second": 110.642,
      "eval_steps_per_second": 13.836,
      "step": 115000
    },
    {
      "epoch": 0.9651541739784407,
      "grad_norm": 10.08293628692627,
      "learning_rate": 1.7709994126038433e-06,
      "loss": 4.5584,
      "step": 115500
    },
    {
      "epoch": 0.96933233057575,
      "grad_norm": 6.49351692199707,
      "learning_rate": 1.5616346395905012e-06,
      "loss": 4.411,
      "step": 116000
    },
    {
      "epoch": 0.96933233057575,
      "eval_loss": 4.347931385040283,
      "eval_runtime": 40.3323,
      "eval_samples_per_second": 110.631,
      "eval_steps_per_second": 13.835,
      "step": 116000
    },
    {
      "epoch": 0.9735104871730592,
      "grad_norm": 8.52354621887207,
      "learning_rate": 1.3518502978937652e-06,
      "loss": 4.4968,
      "step": 116500
    },
    {
      "epoch": 0.9776886437703685,
      "grad_norm": 4.475661754608154,
      "learning_rate": 1.1420659561970296e-06,
      "loss": 4.4801,
      "step": 117000
    },
    {
      "epoch": 0.9776886437703685,
      "eval_loss": 4.335597991943359,
      "eval_runtime": 40.253,
      "eval_samples_per_second": 110.849,
      "eval_steps_per_second": 13.862,
      "step": 117000
    },
    {
      "epoch": 0.9818668003676778,
      "grad_norm": 7.370941638946533,
      "learning_rate": 9.322816145002936e-07,
      "loss": 4.4369,
      "step": 117500
    },
    {
      "epoch": 0.9860449569649871,
      "grad_norm": 7.109913349151611,
      "learning_rate": 7.224972728035579e-07,
      "loss": 4.4866,
      "step": 118000
    },
    {
      "epoch": 0.9860449569649871,
      "eval_loss": 4.327831745147705,
      "eval_runtime": 40.3607,
      "eval_samples_per_second": 110.553,
      "eval_steps_per_second": 13.825,
      "step": 118000
    },
    {
      "epoch": 0.9902231135622963,
      "grad_norm": 6.5864434242248535,
      "learning_rate": 5.131324997902156e-07,
      "loss": 4.3809,
      "step": 118500
    },
    {
      "epoch": 0.9944012701596056,
      "grad_norm": 12.8942232131958,
      "learning_rate": 3.0334815809347995e-07,
      "loss": 4.5074,
      "step": 119000
    },
    {
      "epoch": 0.9944012701596056,
      "eval_loss": 4.323913097381592,
      "eval_runtime": 40.3487,
      "eval_samples_per_second": 110.586,
      "eval_steps_per_second": 13.829,
      "step": 119000
    },
    {
      "epoch": 0.9985794267569148,
      "grad_norm": 5.209051132202148,
      "learning_rate": 9.356381639674414e-08,
      "loss": 4.528,
      "step": 119500
    },
    {
      "epoch": 1.0027575833542242,
      "grad_norm": 5.586207866668701,
      "learning_rate": 3.7510351750386886e-05,
      "loss": 4.8288,
      "step": 120000
    },
    {
      "epoch": 1.0027575833542242,
      "eval_loss": 4.565053462982178,
      "eval_runtime": 40.4476,
      "eval_samples_per_second": 110.316,
      "eval_steps_per_second": 13.796,
      "step": 120000
    },
    {
      "epoch": 1.0069357399515333,
      "grad_norm": 7.037691593170166,
      "learning_rate": 3.745807018277636e-05,
      "loss": 4.5428,
      "step": 120500
    },
    {
      "epoch": 1.0111138965488426,
      "grad_norm": 15.185105323791504,
      "learning_rate": 3.740589317830106e-05,
      "loss": 4.4273,
      "step": 121000
    },
    {
      "epoch": 1.0111138965488426,
      "eval_loss": 4.302701950073242,
      "eval_runtime": 40.408,
      "eval_samples_per_second": 110.424,
      "eval_steps_per_second": 13.809,
      "step": 121000
    },
    {
      "epoch": 1.015292053146152,
      "grad_norm": 6.465140342712402,
      "learning_rate": 3.735361161069054e-05,
      "loss": 4.4009,
      "step": 121500
    },
    {
      "epoch": 1.0194702097434611,
      "grad_norm": 5.409151077270508,
      "learning_rate": 3.730133004308001e-05,
      "loss": 4.3574,
      "step": 122000
    },
    {
      "epoch": 1.0194702097434611,
      "eval_loss": 4.197505474090576,
      "eval_runtime": 40.4321,
      "eval_samples_per_second": 110.358,
      "eval_steps_per_second": 13.801,
      "step": 122000
    },
    {
      "epoch": 1.0236483663407705,
      "grad_norm": 3.3518075942993164,
      "learning_rate": 3.724904847546949e-05,
      "loss": 4.287,
      "step": 122500
    },
    {
      "epoch": 1.0278265229380796,
      "grad_norm": 7.260291576385498,
      "learning_rate": 3.719676690785896e-05,
      "loss": 4.2858,
      "step": 123000
    },
    {
      "epoch": 1.0278265229380796,
      "eval_loss": 4.131897449493408,
      "eval_runtime": 40.5776,
      "eval_samples_per_second": 109.962,
      "eval_steps_per_second": 13.751,
      "step": 123000
    },
    {
      "epoch": 1.032004679535389,
      "grad_norm": 5.965569972991943,
      "learning_rate": 3.714448534024844e-05,
      "loss": 4.2024,
      "step": 123500
    },
    {
      "epoch": 1.0361828361326983,
      "grad_norm": 6.48886251449585,
      "learning_rate": 3.709220377263792e-05,
      "loss": 4.2055,
      "step": 124000
    },
    {
      "epoch": 1.0361828361326983,
      "eval_loss": 4.078915596008301,
      "eval_runtime": 40.4564,
      "eval_samples_per_second": 110.291,
      "eval_steps_per_second": 13.793,
      "step": 124000
    },
    {
      "epoch": 1.0403609927300075,
      "grad_norm": 6.582276344299316,
      "learning_rate": 3.7040026768162615e-05,
      "loss": 4.1067,
      "step": 124500
    },
    {
      "epoch": 1.0445391493273168,
      "grad_norm": 5.228335857391357,
      "learning_rate": 3.6987745200552094e-05,
      "loss": 4.1359,
      "step": 125000
    },
    {
      "epoch": 1.0445391493273168,
      "eval_loss": 4.038590908050537,
      "eval_runtime": 40.4792,
      "eval_samples_per_second": 110.229,
      "eval_steps_per_second": 13.785,
      "step": 125000
    },
    {
      "epoch": 1.048717305924626,
      "grad_norm": 8.54234790802002,
      "learning_rate": 3.6935463632941566e-05,
      "loss": 4.1414,
      "step": 125500
    },
    {
      "epoch": 1.0528954625219353,
      "grad_norm": 6.923229694366455,
      "learning_rate": 3.688318206533105e-05,
      "loss": 4.1308,
      "step": 126000
    },
    {
      "epoch": 1.0528954625219353,
      "eval_loss": 3.986788272857666,
      "eval_runtime": 40.465,
      "eval_samples_per_second": 110.268,
      "eval_steps_per_second": 13.79,
      "step": 126000
    },
    {
      "epoch": 1.0570736191192447,
      "grad_norm": 8.614554405212402,
      "learning_rate": 3.683090049772053e-05,
      "loss": 4.1132,
      "step": 126500
    },
    {
      "epoch": 1.0612517757165538,
      "grad_norm": 11.300436019897461,
      "learning_rate": 3.677861893011e-05,
      "loss": 4.0469,
      "step": 127000
    },
    {
      "epoch": 1.0612517757165538,
      "eval_loss": 3.9560070037841797,
      "eval_runtime": 40.5254,
      "eval_samples_per_second": 110.104,
      "eval_steps_per_second": 13.769,
      "step": 127000
    },
    {
      "epoch": 1.0654299323138632,
      "grad_norm": 6.241533279418945,
      "learning_rate": 3.672633736249948e-05,
      "loss": 4.0292,
      "step": 127500
    },
    {
      "epoch": 1.0696080889111723,
      "grad_norm": 4.231565952301025,
      "learning_rate": 3.6674055794888954e-05,
      "loss": 4.0256,
      "step": 128000
    },
    {
      "epoch": 1.0696080889111723,
      "eval_loss": 3.9226207733154297,
      "eval_runtime": 40.4419,
      "eval_samples_per_second": 110.331,
      "eval_steps_per_second": 13.798,
      "step": 128000
    },
    {
      "epoch": 1.0737862455084817,
      "grad_norm": 6.553409099578857,
      "learning_rate": 3.662177422727843e-05,
      "loss": 3.9736,
      "step": 128500
    },
    {
      "epoch": 1.077964402105791,
      "grad_norm": 5.839107990264893,
      "learning_rate": 3.6569597222803135e-05,
      "loss": 4.0075,
      "step": 129000
    },
    {
      "epoch": 1.077964402105791,
      "eval_loss": 3.9092047214508057,
      "eval_runtime": 40.4523,
      "eval_samples_per_second": 110.303,
      "eval_steps_per_second": 13.794,
      "step": 129000
    },
    {
      "epoch": 1.0821425587031002,
      "grad_norm": 9.884592056274414,
      "learning_rate": 3.651731565519261e-05,
      "loss": 4.0616,
      "step": 129500
    },
    {
      "epoch": 1.0863207153004095,
      "grad_norm": 9.318048477172852,
      "learning_rate": 3.6465034087582086e-05,
      "loss": 3.9327,
      "step": 130000
    },
    {
      "epoch": 1.0863207153004095,
      "eval_loss": 3.877835988998413,
      "eval_runtime": 40.4173,
      "eval_samples_per_second": 110.398,
      "eval_steps_per_second": 13.806,
      "step": 130000
    },
    {
      "epoch": 1.0904988718977187,
      "grad_norm": 5.386814117431641,
      "learning_rate": 3.641275251997156e-05,
      "loss": 3.9593,
      "step": 130500
    },
    {
      "epoch": 1.094677028495028,
      "grad_norm": 6.430805683135986,
      "learning_rate": 3.636057551549626e-05,
      "loss": 3.8757,
      "step": 131000
    },
    {
      "epoch": 1.094677028495028,
      "eval_loss": 3.8680901527404785,
      "eval_runtime": 40.4859,
      "eval_samples_per_second": 110.211,
      "eval_steps_per_second": 13.783,
      "step": 131000
    },
    {
      "epoch": 1.0988551850923374,
      "grad_norm": 4.691482067108154,
      "learning_rate": 3.630829394788574e-05,
      "loss": 3.9689,
      "step": 131500
    },
    {
      "epoch": 1.1030333416896465,
      "grad_norm": 5.442196369171143,
      "learning_rate": 3.625601238027521e-05,
      "loss": 3.9561,
      "step": 132000
    },
    {
      "epoch": 1.1030333416896465,
      "eval_loss": 3.8496084213256836,
      "eval_runtime": 40.4836,
      "eval_samples_per_second": 110.217,
      "eval_steps_per_second": 13.783,
      "step": 132000
    },
    {
      "epoch": 1.1072114982869559,
      "grad_norm": 5.613692760467529,
      "learning_rate": 3.620373081266469e-05,
      "loss": 3.9522,
      "step": 132500
    },
    {
      "epoch": 1.111389654884265,
      "grad_norm": 6.412971496582031,
      "learning_rate": 3.6151553808189384e-05,
      "loss": 3.9606,
      "step": 133000
    },
    {
      "epoch": 1.111389654884265,
      "eval_loss": 3.827300786972046,
      "eval_runtime": 40.4412,
      "eval_samples_per_second": 110.333,
      "eval_steps_per_second": 13.798,
      "step": 133000
    },
    {
      "epoch": 1.1155678114815744,
      "grad_norm": 5.4674072265625,
      "learning_rate": 3.609927224057886e-05,
      "loss": 4.018,
      "step": 133500
    },
    {
      "epoch": 1.1197459680788837,
      "grad_norm": 8.124231338500977,
      "learning_rate": 3.604699067296834e-05,
      "loss": 3.9462,
      "step": 134000
    },
    {
      "epoch": 1.1197459680788837,
      "eval_loss": 3.802276611328125,
      "eval_runtime": 40.4531,
      "eval_samples_per_second": 110.301,
      "eval_steps_per_second": 13.794,
      "step": 134000
    },
    {
      "epoch": 1.1239241246761928,
      "grad_norm": 6.545614719390869,
      "learning_rate": 3.5994709105357814e-05,
      "loss": 3.9833,
      "step": 134500
    },
    {
      "epoch": 1.1281022812735022,
      "grad_norm": 4.819684028625488,
      "learning_rate": 3.5942532100882516e-05,
      "loss": 3.8276,
      "step": 135000
    },
    {
      "epoch": 1.1281022812735022,
      "eval_loss": 3.7954277992248535,
      "eval_runtime": 40.4488,
      "eval_samples_per_second": 110.312,
      "eval_steps_per_second": 13.795,
      "step": 135000
    },
    {
      "epoch": 1.1322804378708113,
      "grad_norm": 5.222882270812988,
      "learning_rate": 3.589035509640721e-05,
      "loss": 3.8508,
      "step": 135500
    },
    {
      "epoch": 1.1364585944681207,
      "grad_norm": 4.044717788696289,
      "learning_rate": 3.583807352879669e-05,
      "loss": 3.9272,
      "step": 136000
    },
    {
      "epoch": 1.1364585944681207,
      "eval_loss": 3.774327039718628,
      "eval_runtime": 40.444,
      "eval_samples_per_second": 110.325,
      "eval_steps_per_second": 13.797,
      "step": 136000
    },
    {
      "epoch": 1.14063675106543,
      "grad_norm": 7.321952819824219,
      "learning_rate": 3.578579196118617e-05,
      "loss": 3.9062,
      "step": 136500
    },
    {
      "epoch": 1.1448149076627392,
      "grad_norm": 7.0700297355651855,
      "learning_rate": 3.573351039357564e-05,
      "loss": 3.9186,
      "step": 137000
    },
    {
      "epoch": 1.1448149076627392,
      "eval_loss": 3.7648792266845703,
      "eval_runtime": 40.4894,
      "eval_samples_per_second": 110.202,
      "eval_steps_per_second": 13.781,
      "step": 137000
    },
    {
      "epoch": 1.1489930642600485,
      "grad_norm": 10.587075233459473,
      "learning_rate": 3.568122882596512e-05,
      "loss": 3.8594,
      "step": 137500
    },
    {
      "epoch": 1.1531712208573577,
      "grad_norm": 9.123832702636719,
      "learning_rate": 3.562894725835459e-05,
      "loss": 3.7987,
      "step": 138000
    },
    {
      "epoch": 1.1531712208573577,
      "eval_loss": 3.750222682952881,
      "eval_runtime": 40.4606,
      "eval_samples_per_second": 110.28,
      "eval_steps_per_second": 13.791,
      "step": 138000
    },
    {
      "epoch": 1.157349377454667,
      "grad_norm": 9.549878120422363,
      "learning_rate": 3.557666569074407e-05,
      "loss": 3.8156,
      "step": 138500
    },
    {
      "epoch": 1.1615275340519762,
      "grad_norm": 7.350008964538574,
      "learning_rate": 3.552438412313355e-05,
      "loss": 3.9018,
      "step": 139000
    },
    {
      "epoch": 1.1615275340519762,
      "eval_loss": 3.7414135932922363,
      "eval_runtime": 40.4584,
      "eval_samples_per_second": 110.286,
      "eval_steps_per_second": 13.792,
      "step": 139000
    },
    {
      "epoch": 1.1657056906492855,
      "grad_norm": 8.974669456481934,
      "learning_rate": 3.5472207118658245e-05,
      "loss": 3.833,
      "step": 139500
    },
    {
      "epoch": 1.1698838472465949,
      "grad_norm": 9.602900505065918,
      "learning_rate": 3.5419925551047724e-05,
      "loss": 3.829,
      "step": 140000
    },
    {
      "epoch": 1.1698838472465949,
      "eval_loss": 3.735084056854248,
      "eval_runtime": 40.431,
      "eval_samples_per_second": 110.361,
      "eval_steps_per_second": 13.801,
      "step": 140000
    },
    {
      "epoch": 1.174062003843904,
      "grad_norm": 4.472715854644775,
      "learning_rate": 3.5367643983437196e-05,
      "loss": 3.8553,
      "step": 140500
    },
    {
      "epoch": 1.1782401604412134,
      "grad_norm": 4.927888870239258,
      "learning_rate": 3.531536241582668e-05,
      "loss": 3.8512,
      "step": 141000
    },
    {
      "epoch": 1.1782401604412134,
      "eval_loss": 3.724465847015381,
      "eval_runtime": 40.4785,
      "eval_samples_per_second": 110.231,
      "eval_steps_per_second": 13.785,
      "step": 141000
    },
    {
      "epoch": 1.1824183170385225,
      "grad_norm": 7.687999725341797,
      "learning_rate": 3.526318541135138e-05,
      "loss": 3.8196,
      "step": 141500
    },
    {
      "epoch": 1.1865964736358319,
      "grad_norm": 5.985261917114258,
      "learning_rate": 3.521090384374085e-05,
      "loss": 3.842,
      "step": 142000
    },
    {
      "epoch": 1.1865964736358319,
      "eval_loss": 3.7113516330718994,
      "eval_runtime": 40.4693,
      "eval_samples_per_second": 110.256,
      "eval_steps_per_second": 13.788,
      "step": 142000
    },
    {
      "epoch": 1.1907746302331412,
      "grad_norm": 6.08819055557251,
      "learning_rate": 3.5158622276130335e-05,
      "loss": 3.852,
      "step": 142500
    },
    {
      "epoch": 1.1949527868304504,
      "grad_norm": 8.868494033813477,
      "learning_rate": 3.510634070851981e-05,
      "loss": 3.7532,
      "step": 143000
    },
    {
      "epoch": 1.1949527868304504,
      "eval_loss": 3.7093913555145264,
      "eval_runtime": 40.4739,
      "eval_samples_per_second": 110.244,
      "eval_steps_per_second": 13.787,
      "step": 143000
    },
    {
      "epoch": 1.1991309434277597,
      "grad_norm": 7.246727466583252,
      "learning_rate": 3.50541637040445e-05,
      "loss": 3.7916,
      "step": 143500
    },
    {
      "epoch": 1.2033091000250689,
      "grad_norm": 5.877631187438965,
      "learning_rate": 3.500188213643398e-05,
      "loss": 3.7639,
      "step": 144000
    },
    {
      "epoch": 1.2033091000250689,
      "eval_loss": 3.6901369094848633,
      "eval_runtime": 40.4322,
      "eval_samples_per_second": 110.358,
      "eval_steps_per_second": 13.801,
      "step": 144000
    },
    {
      "epoch": 1.2074872566223782,
      "grad_norm": 9.112384796142578,
      "learning_rate": 3.494960056882346e-05,
      "loss": 3.7623,
      "step": 144500
    },
    {
      "epoch": 1.2116654132196876,
      "grad_norm": 5.740987777709961,
      "learning_rate": 3.489731900121294e-05,
      "loss": 3.8049,
      "step": 145000
    },
    {
      "epoch": 1.2116654132196876,
      "eval_loss": 3.6749048233032227,
      "eval_runtime": 40.4704,
      "eval_samples_per_second": 110.254,
      "eval_steps_per_second": 13.788,
      "step": 145000
    },
    {
      "epoch": 1.2158435698169967,
      "grad_norm": 3.883148431777954,
      "learning_rate": 3.4845141996737627e-05,
      "loss": 3.7315,
      "step": 145500
    },
    {
      "epoch": 1.220021726414306,
      "grad_norm": 5.081009387969971,
      "learning_rate": 3.479286042912711e-05,
      "loss": 3.7657,
      "step": 146000
    },
    {
      "epoch": 1.220021726414306,
      "eval_loss": 3.678842306137085,
      "eval_runtime": 40.4772,
      "eval_samples_per_second": 110.235,
      "eval_steps_per_second": 13.786,
      "step": 146000
    },
    {
      "epoch": 1.2241998830116152,
      "grad_norm": 4.904498100280762,
      "learning_rate": 3.4740578861516584e-05,
      "loss": 3.7378,
      "step": 146500
    },
    {
      "epoch": 1.2283780396089246,
      "grad_norm": 6.102782249450684,
      "learning_rate": 3.468829729390606e-05,
      "loss": 3.728,
      "step": 147000
    },
    {
      "epoch": 1.2283780396089246,
      "eval_loss": 3.6656246185302734,
      "eval_runtime": 40.4863,
      "eval_samples_per_second": 110.21,
      "eval_steps_per_second": 13.782,
      "step": 147000
    },
    {
      "epoch": 1.2325561962062337,
      "grad_norm": 6.608770847320557,
      "learning_rate": 3.463601572629554e-05,
      "loss": 3.7734,
      "step": 147500
    },
    {
      "epoch": 1.236734352803543,
      "grad_norm": 10.70559310913086,
      "learning_rate": 3.458383872182024e-05,
      "loss": 3.7575,
      "step": 148000
    },
    {
      "epoch": 1.236734352803543,
      "eval_loss": 3.65804123878479,
      "eval_runtime": 40.4413,
      "eval_samples_per_second": 110.333,
      "eval_steps_per_second": 13.798,
      "step": 148000
    },
    {
      "epoch": 1.2409125094008524,
      "grad_norm": 4.311787128448486,
      "learning_rate": 3.4531557154209716e-05,
      "loss": 3.7713,
      "step": 148500
    },
    {
      "epoch": 1.2450906659981615,
      "grad_norm": 5.005003929138184,
      "learning_rate": 3.447927558659919e-05,
      "loss": 3.7568,
      "step": 149000
    },
    {
      "epoch": 1.2450906659981615,
      "eval_loss": 3.6577367782592773,
      "eval_runtime": 40.4197,
      "eval_samples_per_second": 110.392,
      "eval_steps_per_second": 13.805,
      "step": 149000
    },
    {
      "epoch": 1.249268822595471,
      "grad_norm": 5.860953330993652,
      "learning_rate": 3.442709858212389e-05,
      "loss": 3.8178,
      "step": 149500
    },
    {
      "epoch": 1.25344697919278,
      "grad_norm": 7.703333377838135,
      "learning_rate": 3.437481701451337e-05,
      "loss": 3.6736,
      "step": 150000
    },
    {
      "epoch": 1.25344697919278,
      "eval_loss": 3.6417226791381836,
      "eval_runtime": 40.4622,
      "eval_samples_per_second": 110.276,
      "eval_steps_per_second": 13.791,
      "step": 150000
    },
    {
      "epoch": 1.2576251357900894,
      "grad_norm": 4.65405797958374,
      "learning_rate": 3.432253544690284e-05,
      "loss": 3.5936,
      "step": 150500
    },
    {
      "epoch": 1.2618032923873987,
      "grad_norm": 9.61950969696045,
      "learning_rate": 3.427025387929232e-05,
      "loss": 3.7404,
      "step": 151000
    },
    {
      "epoch": 1.2618032923873987,
      "eval_loss": 3.631601333618164,
      "eval_runtime": 40.4481,
      "eval_samples_per_second": 110.314,
      "eval_steps_per_second": 13.795,
      "step": 151000
    },
    {
      "epoch": 1.2659814489847079,
      "grad_norm": 4.689013957977295,
      "learning_rate": 3.421797231168179e-05,
      "loss": 3.7412,
      "step": 151500
    },
    {
      "epoch": 1.2701596055820172,
      "grad_norm": 5.773473739624023,
      "learning_rate": 3.4165795307206494e-05,
      "loss": 3.749,
      "step": 152000
    },
    {
      "epoch": 1.2701596055820172,
      "eval_loss": 3.6227645874023438,
      "eval_runtime": 40.4716,
      "eval_samples_per_second": 110.25,
      "eval_steps_per_second": 13.787,
      "step": 152000
    },
    {
      "epoch": 1.2743377621793264,
      "grad_norm": 4.025112152099609,
      "learning_rate": 3.411351373959597e-05,
      "loss": 3.688,
      "step": 152500
    },
    {
      "epoch": 1.2785159187766357,
      "grad_norm": 4.763550281524658,
      "learning_rate": 3.4061232171985445e-05,
      "loss": 3.7055,
      "step": 153000
    },
    {
      "epoch": 1.2785159187766357,
      "eval_loss": 3.6147379875183105,
      "eval_runtime": 40.4216,
      "eval_samples_per_second": 110.387,
      "eval_steps_per_second": 13.804,
      "step": 153000
    },
    {
      "epoch": 1.282694075373945,
      "grad_norm": 6.419916152954102,
      "learning_rate": 3.4008950604374924e-05,
      "loss": 3.6067,
      "step": 153500
    },
    {
      "epoch": 1.2868722319712542,
      "grad_norm": 4.361742973327637,
      "learning_rate": 3.3956669036764396e-05,
      "loss": 3.7258,
      "step": 154000
    },
    {
      "epoch": 1.2868722319712542,
      "eval_loss": 3.607306957244873,
      "eval_runtime": 40.4593,
      "eval_samples_per_second": 110.284,
      "eval_steps_per_second": 13.792,
      "step": 154000
    },
    {
      "epoch": 1.2910503885685636,
      "grad_norm": 6.296881675720215,
      "learning_rate": 3.3904387469153875e-05,
      "loss": 3.6885,
      "step": 154500
    },
    {
      "epoch": 1.2952285451658727,
      "grad_norm": 4.604855537414551,
      "learning_rate": 3.3852105901543354e-05,
      "loss": 3.6569,
      "step": 155000
    },
    {
      "epoch": 1.2952285451658727,
      "eval_loss": 3.604320764541626,
      "eval_runtime": 40.4612,
      "eval_samples_per_second": 110.279,
      "eval_steps_per_second": 13.791,
      "step": 155000
    },
    {
      "epoch": 1.299406701763182,
      "grad_norm": 4.907881736755371,
      "learning_rate": 3.3799824333932826e-05,
      "loss": 3.6488,
      "step": 155500
    },
    {
      "epoch": 1.3035848583604914,
      "grad_norm": 5.200791358947754,
      "learning_rate": 3.374764732945753e-05,
      "loss": 3.7493,
      "step": 156000
    },
    {
      "epoch": 1.3035848583604914,
      "eval_loss": 3.596531629562378,
      "eval_runtime": 40.4463,
      "eval_samples_per_second": 110.319,
      "eval_steps_per_second": 13.796,
      "step": 156000
    },
    {
      "epoch": 1.3077630149578006,
      "grad_norm": 6.453032970428467,
      "learning_rate": 3.3695365761847e-05,
      "loss": 3.6468,
      "step": 156500
    },
    {
      "epoch": 1.31194117155511,
      "grad_norm": 5.0477142333984375,
      "learning_rate": 3.364308419423648e-05,
      "loss": 3.6687,
      "step": 157000
    },
    {
      "epoch": 1.31194117155511,
      "eval_loss": 3.5927352905273438,
      "eval_runtime": 40.4591,
      "eval_samples_per_second": 110.284,
      "eval_steps_per_second": 13.792,
      "step": 157000
    },
    {
      "epoch": 1.316119328152419,
      "grad_norm": 8.696310997009277,
      "learning_rate": 3.359080262662596e-05,
      "loss": 3.6277,
      "step": 157500
    },
    {
      "epoch": 1.3202974847497284,
      "grad_norm": 4.83682107925415,
      "learning_rate": 3.353862562215065e-05,
      "loss": 3.6981,
      "step": 158000
    },
    {
      "epoch": 1.3202974847497284,
      "eval_loss": 3.5905144214630127,
      "eval_runtime": 40.438,
      "eval_samples_per_second": 110.342,
      "eval_steps_per_second": 13.799,
      "step": 158000
    },
    {
      "epoch": 1.3244756413470378,
      "grad_norm": 5.432042121887207,
      "learning_rate": 3.348634405454013e-05,
      "loss": 3.598,
      "step": 158500
    },
    {
      "epoch": 1.328653797944347,
      "grad_norm": 6.204931735992432,
      "learning_rate": 3.3434062486929604e-05,
      "loss": 3.6486,
      "step": 159000
    },
    {
      "epoch": 1.328653797944347,
      "eval_loss": 3.581803321838379,
      "eval_runtime": 40.4743,
      "eval_samples_per_second": 110.243,
      "eval_steps_per_second": 13.787,
      "step": 159000
    },
    {
      "epoch": 1.3328319545416563,
      "grad_norm": 6.183650016784668,
      "learning_rate": 3.338178091931909e-05,
      "loss": 3.7491,
      "step": 159500
    },
    {
      "epoch": 1.3370101111389654,
      "grad_norm": 7.474965572357178,
      "learning_rate": 3.3329603914843784e-05,
      "loss": 3.6358,
      "step": 160000
    },
    {
      "epoch": 1.3370101111389654,
      "eval_loss": 3.5771028995513916,
      "eval_runtime": 40.4482,
      "eval_samples_per_second": 110.314,
      "eval_steps_per_second": 13.795,
      "step": 160000
    },
    {
      "epoch": 1.3411882677362748,
      "grad_norm": 5.560182094573975,
      "learning_rate": 3.3277322347233256e-05,
      "loss": 3.6578,
      "step": 160500
    },
    {
      "epoch": 1.345366424333584,
      "grad_norm": 7.701005458831787,
      "learning_rate": 3.322504077962274e-05,
      "loss": 3.5839,
      "step": 161000
    },
    {
      "epoch": 1.345366424333584,
      "eval_loss": 3.5681257247924805,
      "eval_runtime": 40.4648,
      "eval_samples_per_second": 110.269,
      "eval_steps_per_second": 13.79,
      "step": 161000
    },
    {
      "epoch": 1.3495445809308932,
      "grad_norm": 2.855236053466797,
      "learning_rate": 3.3172759212012214e-05,
      "loss": 3.6106,
      "step": 161500
    },
    {
      "epoch": 1.3537227375282026,
      "grad_norm": 7.088588714599609,
      "learning_rate": 3.312058220753691e-05,
      "loss": 3.5712,
      "step": 162000
    },
    {
      "epoch": 1.3537227375282026,
      "eval_loss": 3.5612449645996094,
      "eval_runtime": 40.457,
      "eval_samples_per_second": 110.29,
      "eval_steps_per_second": 13.792,
      "step": 162000
    },
    {
      "epoch": 1.3579008941255117,
      "grad_norm": 5.757685661315918,
      "learning_rate": 3.3068300639926395e-05,
      "loss": 3.6394,
      "step": 162500
    },
    {
      "epoch": 1.362079050722821,
      "grad_norm": 3.8591792583465576,
      "learning_rate": 3.301601907231587e-05,
      "loss": 3.6229,
      "step": 163000
    },
    {
      "epoch": 1.362079050722821,
      "eval_loss": 3.562960624694824,
      "eval_runtime": 40.456,
      "eval_samples_per_second": 110.293,
      "eval_steps_per_second": 13.793,
      "step": 163000
    },
    {
      "epoch": 1.3662572073201305,
      "grad_norm": 5.039220809936523,
      "learning_rate": 3.2963737504705346e-05,
      "loss": 3.5562,
      "step": 163500
    },
    {
      "epoch": 1.3704353639174396,
      "grad_norm": 3.936936616897583,
      "learning_rate": 3.291145593709482e-05,
      "loss": 3.5953,
      "step": 164000
    },
    {
      "epoch": 1.3704353639174396,
      "eval_loss": 3.555675983428955,
      "eval_runtime": 40.4593,
      "eval_samples_per_second": 110.284,
      "eval_steps_per_second": 13.792,
      "step": 164000
    },
    {
      "epoch": 1.374613520514749,
      "grad_norm": 5.979030609130859,
      "learning_rate": 3.28591743694843e-05,
      "loss": 3.7403,
      "step": 164500
    },
    {
      "epoch": 1.378791677112058,
      "grad_norm": 3.9190635681152344,
      "learning_rate": 3.2806892801873776e-05,
      "loss": 3.6413,
      "step": 165000
    },
    {
      "epoch": 1.378791677112058,
      "eval_loss": 3.5578091144561768,
      "eval_runtime": 40.4291,
      "eval_samples_per_second": 110.366,
      "eval_steps_per_second": 13.802,
      "step": 165000
    },
    {
      "epoch": 1.3829698337093674,
      "grad_norm": 3.9135348796844482,
      "learning_rate": 3.275461123426325e-05,
      "loss": 3.5954,
      "step": 165500
    },
    {
      "epoch": 1.3871479903066768,
      "grad_norm": 4.480642795562744,
      "learning_rate": 3.270243422978795e-05,
      "loss": 3.6482,
      "step": 166000
    },
    {
      "epoch": 1.3871479903066768,
      "eval_loss": 3.5539774894714355,
      "eval_runtime": 40.4703,
      "eval_samples_per_second": 110.254,
      "eval_steps_per_second": 13.788,
      "step": 166000
    },
    {
      "epoch": 1.391326146903986,
      "grad_norm": 4.073483467102051,
      "learning_rate": 3.2650257225312645e-05,
      "loss": 3.6099,
      "step": 166500
    },
    {
      "epoch": 1.3955043035012953,
      "grad_norm": 6.1564860343933105,
      "learning_rate": 3.2597975657702124e-05,
      "loss": 3.5885,
      "step": 167000
    },
    {
      "epoch": 1.3955043035012953,
      "eval_loss": 3.55012583732605,
      "eval_runtime": 40.43,
      "eval_samples_per_second": 110.364,
      "eval_steps_per_second": 13.802,
      "step": 167000
    },
    {
      "epoch": 1.3996824600986044,
      "grad_norm": 5.0414605140686035,
      "learning_rate": 3.25456940900916e-05,
      "loss": 3.6201,
      "step": 167500
    },
    {
      "epoch": 1.4038606166959138,
      "grad_norm": 4.474603652954102,
      "learning_rate": 3.2493412522481075e-05,
      "loss": 3.6112,
      "step": 168000
    },
    {
      "epoch": 1.4038606166959138,
      "eval_loss": 3.5365307331085205,
      "eval_runtime": 40.4685,
      "eval_samples_per_second": 110.259,
      "eval_steps_per_second": 13.789,
      "step": 168000
    },
    {
      "epoch": 1.4080387732932231,
      "grad_norm": 5.0425801277160645,
      "learning_rate": 3.2441130954870554e-05,
      "loss": 3.5891,
      "step": 168500
    },
    {
      "epoch": 1.4122169298905323,
      "grad_norm": 4.1331915855407715,
      "learning_rate": 3.2388849387260026e-05,
      "loss": 3.5289,
      "step": 169000
    },
    {
      "epoch": 1.4122169298905323,
      "eval_loss": 3.5384674072265625,
      "eval_runtime": 40.5065,
      "eval_samples_per_second": 110.155,
      "eval_steps_per_second": 13.776,
      "step": 169000
    },
    {
      "epoch": 1.4163950864878416,
      "grad_norm": 5.106828212738037,
      "learning_rate": 3.2336567819649505e-05,
      "loss": 3.6218,
      "step": 169500
    },
    {
      "epoch": 1.4205732430851508,
      "grad_norm": 7.320224761962891,
      "learning_rate": 3.2284390815174207e-05,
      "loss": 3.6387,
      "step": 170000
    },
    {
      "epoch": 1.4205732430851508,
      "eval_loss": 3.5212485790252686,
      "eval_runtime": 40.4469,
      "eval_samples_per_second": 110.317,
      "eval_steps_per_second": 13.796,
      "step": 170000
    },
    {
      "epoch": 1.4247513996824601,
      "grad_norm": 5.692435264587402,
      "learning_rate": 3.223210924756368e-05,
      "loss": 3.6134,
      "step": 170500
    },
    {
      "epoch": 1.4289295562797695,
      "grad_norm": 6.391637325286865,
      "learning_rate": 3.217982767995316e-05,
      "loss": 3.6322,
      "step": 171000
    },
    {
      "epoch": 1.4289295562797695,
      "eval_loss": 3.5183897018432617,
      "eval_runtime": 41.7439,
      "eval_samples_per_second": 106.89,
      "eval_steps_per_second": 13.367,
      "step": 171000
    },
    {
      "epoch": 1.4331077128770786,
      "grad_norm": 9.630993843078613,
      "learning_rate": 3.212754611234263e-05,
      "loss": 3.5568,
      "step": 171500
    },
    {
      "epoch": 1.437285869474388,
      "grad_norm": 4.306568145751953,
      "learning_rate": 3.207526454473211e-05,
      "loss": 3.6337,
      "step": 172000
    },
    {
      "epoch": 1.437285869474388,
      "eval_loss": 3.511368751525879,
      "eval_runtime": 41.0631,
      "eval_samples_per_second": 108.662,
      "eval_steps_per_second": 13.589,
      "step": 172000
    },
    {
      "epoch": 1.441464026071697,
      "grad_norm": 4.794216156005859,
      "learning_rate": 3.202298297712159e-05,
      "loss": 3.6292,
      "step": 172500
    },
    {
      "epoch": 1.4456421826690065,
      "grad_norm": 8.913003921508789,
      "learning_rate": 3.1970701409511067e-05,
      "loss": 3.5451,
      "step": 173000
    },
    {
      "epoch": 1.4456421826690065,
      "eval_loss": 3.5066542625427246,
      "eval_runtime": 40.4633,
      "eval_samples_per_second": 110.273,
      "eval_steps_per_second": 13.79,
      "step": 173000
    },
    {
      "epoch": 1.4498203392663158,
      "grad_norm": 4.385974884033203,
      "learning_rate": 3.1918419841900545e-05,
      "loss": 3.5534,
      "step": 173500
    },
    {
      "epoch": 1.453998495863625,
      "grad_norm": 3.306346893310547,
      "learning_rate": 3.186613827429002e-05,
      "loss": 3.5008,
      "step": 174000
    },
    {
      "epoch": 1.453998495863625,
      "eval_loss": 3.496458053588867,
      "eval_runtime": 40.461,
      "eval_samples_per_second": 110.279,
      "eval_steps_per_second": 13.791,
      "step": 174000
    },
    {
      "epoch": 1.4581766524609343,
      "grad_norm": 6.9272050857543945,
      "learning_rate": 3.181396126981472e-05,
      "loss": 3.6206,
      "step": 174500
    },
    {
      "epoch": 1.4623548090582434,
      "grad_norm": 6.005661964416504,
      "learning_rate": 3.176167970220419e-05,
      "loss": 3.554,
      "step": 175000
    },
    {
      "epoch": 1.4623548090582434,
      "eval_loss": 3.4955668449401855,
      "eval_runtime": 40.47,
      "eval_samples_per_second": 110.255,
      "eval_steps_per_second": 13.788,
      "step": 175000
    },
    {
      "epoch": 1.4665329656555528,
      "grad_norm": 5.330180644989014,
      "learning_rate": 3.170939813459367e-05,
      "loss": 3.5613,
      "step": 175500
    },
    {
      "epoch": 1.4707111222528622,
      "grad_norm": 4.7661027908325195,
      "learning_rate": 3.165711656698315e-05,
      "loss": 3.6066,
      "step": 176000
    },
    {
      "epoch": 1.4707111222528622,
      "eval_loss": 3.491217851638794,
      "eval_runtime": 40.7953,
      "eval_samples_per_second": 109.375,
      "eval_steps_per_second": 13.678,
      "step": 176000
    },
    {
      "epoch": 1.4748892788501713,
      "grad_norm": 5.372822284698486,
      "learning_rate": 3.160483499937262e-05,
      "loss": 3.5948,
      "step": 176500
    },
    {
      "epoch": 1.4790674354474806,
      "grad_norm": 4.185611248016357,
      "learning_rate": 3.155265799489732e-05,
      "loss": 3.5511,
      "step": 177000
    },
    {
      "epoch": 1.4790674354474806,
      "eval_loss": 3.4882309436798096,
      "eval_runtime": 40.4344,
      "eval_samples_per_second": 110.352,
      "eval_steps_per_second": 13.8,
      "step": 177000
    },
    {
      "epoch": 1.4832455920447898,
      "grad_norm": 4.737360000610352,
      "learning_rate": 3.1500376427286795e-05,
      "loss": 3.5946,
      "step": 177500
    },
    {
      "epoch": 1.4874237486420991,
      "grad_norm": 6.088687419891357,
      "learning_rate": 3.1448094859676274e-05,
      "loss": 3.6226,
      "step": 178000
    },
    {
      "epoch": 1.4874237486420991,
      "eval_loss": 3.4909913539886475,
      "eval_runtime": 40.4208,
      "eval_samples_per_second": 110.389,
      "eval_steps_per_second": 13.805,
      "step": 178000
    },
    {
      "epoch": 1.4916019052394085,
      "grad_norm": 4.8783369064331055,
      "learning_rate": 3.139581329206575e-05,
      "loss": 3.5074,
      "step": 178500
    },
    {
      "epoch": 1.4957800618367176,
      "grad_norm": 4.050315856933594,
      "learning_rate": 3.134363628759045e-05,
      "loss": 3.5209,
      "step": 179000
    },
    {
      "epoch": 1.4957800618367176,
      "eval_loss": 3.4775376319885254,
      "eval_runtime": 40.4543,
      "eval_samples_per_second": 110.297,
      "eval_steps_per_second": 13.793,
      "step": 179000
    },
    {
      "epoch": 1.4999582184340268,
      "grad_norm": 7.867382526397705,
      "learning_rate": 3.129135471997993e-05,
      "loss": 3.5293,
      "step": 179500
    },
    {
      "epoch": 1.5041363750313361,
      "grad_norm": 4.2094597816467285,
      "learning_rate": 3.12390731523694e-05,
      "loss": 3.5055,
      "step": 180000
    },
    {
      "epoch": 1.5041363750313361,
      "eval_loss": 3.4833099842071533,
      "eval_runtime": 40.4847,
      "eval_samples_per_second": 110.214,
      "eval_steps_per_second": 13.783,
      "step": 180000
    },
    {
      "epoch": 1.5083145316286455,
      "grad_norm": 5.466799259185791,
      "learning_rate": 3.118679158475888e-05,
      "loss": 3.4833,
      "step": 180500
    },
    {
      "epoch": 1.5124926882259548,
      "grad_norm": 8.360751152038574,
      "learning_rate": 3.113461458028358e-05,
      "loss": 3.4472,
      "step": 181000
    },
    {
      "epoch": 1.5124926882259548,
      "eval_loss": 3.4807190895080566,
      "eval_runtime": 40.437,
      "eval_samples_per_second": 110.345,
      "eval_steps_per_second": 13.799,
      "step": 181000
    },
    {
      "epoch": 1.516670844823264,
      "grad_norm": 4.71815824508667,
      "learning_rate": 3.108233301267305e-05,
      "loss": 3.5698,
      "step": 181500
    },
    {
      "epoch": 1.520849001420573,
      "grad_norm": 5.6605658531188965,
      "learning_rate": 3.103005144506253e-05,
      "loss": 3.5575,
      "step": 182000
    },
    {
      "epoch": 1.520849001420573,
      "eval_loss": 3.4685845375061035,
      "eval_runtime": 40.4646,
      "eval_samples_per_second": 110.269,
      "eval_steps_per_second": 13.79,
      "step": 182000
    },
    {
      "epoch": 1.5250271580178825,
      "grad_norm": 5.890238285064697,
      "learning_rate": 3.097776987745201e-05,
      "loss": 3.5999,
      "step": 182500
    },
    {
      "epoch": 1.5292053146151918,
      "grad_norm": 6.8811140060424805,
      "learning_rate": 3.0925592872976705e-05,
      "loss": 3.4265,
      "step": 183000
    },
    {
      "epoch": 1.5292053146151918,
      "eval_loss": 3.465893030166626,
      "eval_runtime": 40.4398,
      "eval_samples_per_second": 110.337,
      "eval_steps_per_second": 13.798,
      "step": 183000
    },
    {
      "epoch": 1.5333834712125012,
      "grad_norm": 6.643799304962158,
      "learning_rate": 3.0873311305366184e-05,
      "loss": 3.6255,
      "step": 183500
    },
    {
      "epoch": 1.5375616278098103,
      "grad_norm": 6.201560020446777,
      "learning_rate": 3.0821029737755656e-05,
      "loss": 3.622,
      "step": 184000
    },
    {
      "epoch": 1.5375616278098103,
      "eval_loss": 3.464231491088867,
      "eval_runtime": 40.4428,
      "eval_samples_per_second": 110.329,
      "eval_steps_per_second": 13.797,
      "step": 184000
    },
    {
      "epoch": 1.5417397844071195,
      "grad_norm": 4.211904048919678,
      "learning_rate": 3.0768748170145135e-05,
      "loss": 3.5642,
      "step": 184500
    },
    {
      "epoch": 1.5459179410044288,
      "grad_norm": 5.1118011474609375,
      "learning_rate": 3.0716466602534614e-05,
      "loss": 3.5833,
      "step": 185000
    },
    {
      "epoch": 1.5459179410044288,
      "eval_loss": 3.461984157562256,
      "eval_runtime": 40.4142,
      "eval_samples_per_second": 110.407,
      "eval_steps_per_second": 13.807,
      "step": 185000
    },
    {
      "epoch": 1.5500960976017382,
      "grad_norm": 5.767035484313965,
      "learning_rate": 3.066428959805931e-05,
      "loss": 3.6235,
      "step": 185500
    },
    {
      "epoch": 1.5542742541990475,
      "grad_norm": 3.612896680831909,
      "learning_rate": 3.061200803044879e-05,
      "loss": 3.5037,
      "step": 186000
    },
    {
      "epoch": 1.5542742541990475,
      "eval_loss": 3.4645168781280518,
      "eval_runtime": 40.4303,
      "eval_samples_per_second": 110.363,
      "eval_steps_per_second": 13.802,
      "step": 186000
    },
    {
      "epoch": 1.5584524107963567,
      "grad_norm": 7.333379745483398,
      "learning_rate": 3.055972646283826e-05,
      "loss": 3.5563,
      "step": 186500
    },
    {
      "epoch": 1.5626305673936658,
      "grad_norm": 4.652407646179199,
      "learning_rate": 3.050744489522774e-05,
      "loss": 3.5311,
      "step": 187000
    },
    {
      "epoch": 1.5626305673936658,
      "eval_loss": 3.456352949142456,
      "eval_runtime": 40.4378,
      "eval_samples_per_second": 110.342,
      "eval_steps_per_second": 13.799,
      "step": 187000
    },
    {
      "epoch": 1.5668087239909751,
      "grad_norm": 6.25783634185791,
      "learning_rate": 3.0455267890752437e-05,
      "loss": 3.5793,
      "step": 187500
    },
    {
      "epoch": 1.5709868805882845,
      "grad_norm": 10.195725440979004,
      "learning_rate": 3.0402986323141912e-05,
      "loss": 3.4834,
      "step": 188000
    },
    {
      "epoch": 1.5709868805882845,
      "eval_loss": 3.4547135829925537,
      "eval_runtime": 40.522,
      "eval_samples_per_second": 110.113,
      "eval_steps_per_second": 13.77,
      "step": 188000
    },
    {
      "epoch": 1.5751650371855939,
      "grad_norm": 3.9726510047912598,
      "learning_rate": 3.0350704755531388e-05,
      "loss": 3.4863,
      "step": 188500
    },
    {
      "epoch": 1.579343193782903,
      "grad_norm": 5.096458911895752,
      "learning_rate": 3.0298423187920867e-05,
      "loss": 3.4679,
      "step": 189000
    },
    {
      "epoch": 1.579343193782903,
      "eval_loss": 3.448751926422119,
      "eval_runtime": 40.4496,
      "eval_samples_per_second": 110.31,
      "eval_steps_per_second": 13.795,
      "step": 189000
    },
    {
      "epoch": 1.5835213503802121,
      "grad_norm": 5.1278157234191895,
      "learning_rate": 3.024614162031035e-05,
      "loss": 3.5127,
      "step": 189500
    },
    {
      "epoch": 1.5876995069775215,
      "grad_norm": 3.9133641719818115,
      "learning_rate": 3.0193860052699825e-05,
      "loss": 3.4699,
      "step": 190000
    },
    {
      "epoch": 1.5876995069775215,
      "eval_loss": 3.453089952468872,
      "eval_runtime": 40.4458,
      "eval_samples_per_second": 110.32,
      "eval_steps_per_second": 13.796,
      "step": 190000
    },
    {
      "epoch": 1.5918776635748308,
      "grad_norm": 3.5668985843658447,
      "learning_rate": 3.0141683048224516e-05,
      "loss": 3.5933,
      "step": 190500
    },
    {
      "epoch": 1.5960558201721402,
      "grad_norm": 4.461442947387695,
      "learning_rate": 3.0089506043749215e-05,
      "loss": 3.4547,
      "step": 191000
    },
    {
      "epoch": 1.5960558201721402,
      "eval_loss": 3.4460525512695312,
      "eval_runtime": 40.4625,
      "eval_samples_per_second": 110.275,
      "eval_steps_per_second": 13.791,
      "step": 191000
    },
    {
      "epoch": 1.6002339767694493,
      "grad_norm": 9.258716583251953,
      "learning_rate": 3.0037224476138694e-05,
      "loss": 3.4844,
      "step": 191500
    },
    {
      "epoch": 1.6044121333667585,
      "grad_norm": 5.085646629333496,
      "learning_rate": 2.998494290852817e-05,
      "loss": 3.4608,
      "step": 192000
    },
    {
      "epoch": 1.6044121333667585,
      "eval_loss": 3.4406673908233643,
      "eval_runtime": 40.4745,
      "eval_samples_per_second": 110.242,
      "eval_steps_per_second": 13.786,
      "step": 192000
    },
    {
      "epoch": 1.6085902899640678,
      "grad_norm": 4.961802005767822,
      "learning_rate": 2.9932661340917645e-05,
      "loss": 3.4265,
      "step": 192500
    },
    {
      "epoch": 1.6127684465613772,
      "grad_norm": 5.648656845092773,
      "learning_rate": 2.9880379773307127e-05,
      "loss": 3.5027,
      "step": 193000
    },
    {
      "epoch": 1.6127684465613772,
      "eval_loss": 3.437014579772949,
      "eval_runtime": 40.4642,
      "eval_samples_per_second": 110.27,
      "eval_steps_per_second": 13.79,
      "step": 193000
    },
    {
      "epoch": 1.6169466031586865,
      "grad_norm": 5.819306373596191,
      "learning_rate": 2.9828098205696603e-05,
      "loss": 3.5167,
      "step": 193500
    },
    {
      "epoch": 1.6211247597559957,
      "grad_norm": 5.806619167327881,
      "learning_rate": 2.9775816638086078e-05,
      "loss": 3.4641,
      "step": 194000
    },
    {
      "epoch": 1.6211247597559957,
      "eval_loss": 3.435065746307373,
      "eval_runtime": 40.453,
      "eval_samples_per_second": 110.301,
      "eval_steps_per_second": 13.794,
      "step": 194000
    },
    {
      "epoch": 1.6253029163533048,
      "grad_norm": 7.277612209320068,
      "learning_rate": 2.9723535070475557e-05,
      "loss": 3.5148,
      "step": 194500
    },
    {
      "epoch": 1.6294810729506142,
      "grad_norm": 5.517014503479004,
      "learning_rate": 2.9671358066000255e-05,
      "loss": 3.4822,
      "step": 195000
    },
    {
      "epoch": 1.6294810729506142,
      "eval_loss": 3.4265546798706055,
      "eval_runtime": 40.492,
      "eval_samples_per_second": 110.194,
      "eval_steps_per_second": 13.78,
      "step": 195000
    },
    {
      "epoch": 1.6336592295479235,
      "grad_norm": 5.561305046081543,
      "learning_rate": 2.961907649838973e-05,
      "loss": 3.4422,
      "step": 195500
    },
    {
      "epoch": 1.6378373861452329,
      "grad_norm": 4.937339782714844,
      "learning_rate": 2.9566794930779206e-05,
      "loss": 3.4682,
      "step": 196000
    },
    {
      "epoch": 1.6378373861452329,
      "eval_loss": 3.425950765609741,
      "eval_runtime": 40.469,
      "eval_samples_per_second": 110.257,
      "eval_steps_per_second": 13.788,
      "step": 196000
    },
    {
      "epoch": 1.642015542742542,
      "grad_norm": 4.299443244934082,
      "learning_rate": 2.9514513363168682e-05,
      "loss": 3.4398,
      "step": 196500
    },
    {
      "epoch": 1.6461936993398512,
      "grad_norm": 3.591259479522705,
      "learning_rate": 2.9462336358693384e-05,
      "loss": 3.4695,
      "step": 197000
    },
    {
      "epoch": 1.6461936993398512,
      "eval_loss": 3.4207708835601807,
      "eval_runtime": 40.4515,
      "eval_samples_per_second": 110.305,
      "eval_steps_per_second": 13.794,
      "step": 197000
    },
    {
      "epoch": 1.6503718559371605,
      "grad_norm": 4.869890213012695,
      "learning_rate": 2.941005479108286e-05,
      "loss": 3.5136,
      "step": 197500
    },
    {
      "epoch": 1.6545500125344699,
      "grad_norm": 4.020719051361084,
      "learning_rate": 2.9357773223472335e-05,
      "loss": 3.4545,
      "step": 198000
    },
    {
      "epoch": 1.6545500125344699,
      "eval_loss": 3.4241840839385986,
      "eval_runtime": 40.4973,
      "eval_samples_per_second": 110.18,
      "eval_steps_per_second": 13.779,
      "step": 198000
    },
    {
      "epoch": 1.658728169131779,
      "grad_norm": 5.599808692932129,
      "learning_rate": 2.930549165586181e-05,
      "loss": 3.4355,
      "step": 198500
    },
    {
      "epoch": 1.6629063257290884,
      "grad_norm": 5.005913734436035,
      "learning_rate": 2.9253210088251286e-05,
      "loss": 3.4585,
      "step": 199000
    },
    {
      "epoch": 1.6629063257290884,
      "eval_loss": 3.4280543327331543,
      "eval_runtime": 40.4983,
      "eval_samples_per_second": 110.177,
      "eval_steps_per_second": 13.778,
      "step": 199000
    },
    {
      "epoch": 1.6670844823263975,
      "grad_norm": 3.628136396408081,
      "learning_rate": 2.9200928520640765e-05,
      "loss": 3.5212,
      "step": 199500
    },
    {
      "epoch": 1.6712626389237069,
      "grad_norm": 3.2922470569610596,
      "learning_rate": 2.914864695303024e-05,
      "loss": 3.5205,
      "step": 200000
    },
    {
      "epoch": 1.6712626389237069,
      "eval_loss": 3.416393280029297,
      "eval_runtime": 40.4533,
      "eval_samples_per_second": 110.3,
      "eval_steps_per_second": 13.794,
      "step": 200000
    },
    {
      "epoch": 1.6754407955210162,
      "grad_norm": 4.794888496398926,
      "learning_rate": 2.9096365385419716e-05,
      "loss": 3.4798,
      "step": 200500
    },
    {
      "epoch": 1.6796189521183253,
      "grad_norm": 4.415316581726074,
      "learning_rate": 2.9044188380944414e-05,
      "loss": 3.4709,
      "step": 201000
    },
    {
      "epoch": 1.6796189521183253,
      "eval_loss": 3.4112062454223633,
      "eval_runtime": 40.4752,
      "eval_samples_per_second": 110.24,
      "eval_steps_per_second": 13.786,
      "step": 201000
    },
    {
      "epoch": 1.6837971087156347,
      "grad_norm": 3.854452610015869,
      "learning_rate": 2.899190681333389e-05,
      "loss": 3.427,
      "step": 201500
    },
    {
      "epoch": 1.6879752653129438,
      "grad_norm": 4.938957214355469,
      "learning_rate": 2.893962524572337e-05,
      "loss": 3.407,
      "step": 202000
    },
    {
      "epoch": 1.6879752653129438,
      "eval_loss": 3.410858154296875,
      "eval_runtime": 40.4568,
      "eval_samples_per_second": 110.29,
      "eval_steps_per_second": 13.792,
      "step": 202000
    },
    {
      "epoch": 1.6921534219102532,
      "grad_norm": 5.272583961486816,
      "learning_rate": 2.8887343678112844e-05,
      "loss": 3.4582,
      "step": 202500
    },
    {
      "epoch": 1.6963315785075626,
      "grad_norm": 5.953027725219727,
      "learning_rate": 2.8835166673637542e-05,
      "loss": 3.3917,
      "step": 203000
    },
    {
      "epoch": 1.6963315785075626,
      "eval_loss": 3.4051761627197266,
      "eval_runtime": 40.4397,
      "eval_samples_per_second": 110.337,
      "eval_steps_per_second": 13.798,
      "step": 203000
    },
    {
      "epoch": 1.7005097351048717,
      "grad_norm": 9.10191535949707,
      "learning_rate": 2.8782885106027018e-05,
      "loss": 3.4971,
      "step": 203500
    },
    {
      "epoch": 1.704687891702181,
      "grad_norm": 7.353392124176025,
      "learning_rate": 2.8730603538416493e-05,
      "loss": 3.4285,
      "step": 204000
    },
    {
      "epoch": 1.704687891702181,
      "eval_loss": 3.3989508152008057,
      "eval_runtime": 40.4649,
      "eval_samples_per_second": 110.268,
      "eval_steps_per_second": 13.79,
      "step": 204000
    },
    {
      "epoch": 1.7088660482994902,
      "grad_norm": 5.292013168334961,
      "learning_rate": 2.8678321970805972e-05,
      "loss": 3.4771,
      "step": 204500
    },
    {
      "epoch": 1.7130442048967995,
      "grad_norm": 4.941309452056885,
      "learning_rate": 2.862614496633067e-05,
      "loss": 3.4328,
      "step": 205000
    },
    {
      "epoch": 1.7130442048967995,
      "eval_loss": 3.404594898223877,
      "eval_runtime": 40.469,
      "eval_samples_per_second": 110.257,
      "eval_steps_per_second": 13.788,
      "step": 205000
    },
    {
      "epoch": 1.717222361494109,
      "grad_norm": 5.848660945892334,
      "learning_rate": 2.8573863398720146e-05,
      "loss": 3.5314,
      "step": 205500
    },
    {
      "epoch": 1.721400518091418,
      "grad_norm": 3.0623245239257812,
      "learning_rate": 2.8521581831109622e-05,
      "loss": 3.4575,
      "step": 206000
    },
    {
      "epoch": 1.721400518091418,
      "eval_loss": 3.4002387523651123,
      "eval_runtime": 40.4802,
      "eval_samples_per_second": 110.227,
      "eval_steps_per_second": 13.785,
      "step": 206000
    },
    {
      "epoch": 1.7255786746887274,
      "grad_norm": 2.902146816253662,
      "learning_rate": 2.8469300263499104e-05,
      "loss": 3.4533,
      "step": 206500
    },
    {
      "epoch": 1.7297568312860365,
      "grad_norm": 4.314316749572754,
      "learning_rate": 2.841701869588858e-05,
      "loss": 3.4924,
      "step": 207000
    },
    {
      "epoch": 1.7297568312860365,
      "eval_loss": 3.398782253265381,
      "eval_runtime": 40.4718,
      "eval_samples_per_second": 110.25,
      "eval_steps_per_second": 13.787,
      "step": 207000
    },
    {
      "epoch": 1.7339349878833459,
      "grad_norm": 4.598596096038818,
      "learning_rate": 2.8364841691413275e-05,
      "loss": 3.4796,
      "step": 207500
    },
    {
      "epoch": 1.7381131444806552,
      "grad_norm": 7.308864593505859,
      "learning_rate": 2.8312560123802757e-05,
      "loss": 3.4804,
      "step": 208000
    },
    {
      "epoch": 1.7381131444806552,
      "eval_loss": 3.393099308013916,
      "eval_runtime": 40.4423,
      "eval_samples_per_second": 110.33,
      "eval_steps_per_second": 13.797,
      "step": 208000
    },
    {
      "epoch": 1.7422913010779644,
      "grad_norm": 2.0738399028778076,
      "learning_rate": 2.8260278556192232e-05,
      "loss": 3.5247,
      "step": 208500
    },
    {
      "epoch": 1.7464694576752735,
      "grad_norm": 4.9979753494262695,
      "learning_rate": 2.8207996988581708e-05,
      "loss": 3.445,
      "step": 209000
    },
    {
      "epoch": 1.7464694576752735,
      "eval_loss": 3.389554977416992,
      "eval_runtime": 40.4907,
      "eval_samples_per_second": 110.198,
      "eval_steps_per_second": 13.781,
      "step": 209000
    },
    {
      "epoch": 1.7506476142725829,
      "grad_norm": 6.01882791519165,
      "learning_rate": 2.8155819984106406e-05,
      "loss": 3.461,
      "step": 209500
    },
    {
      "epoch": 1.7548257708698922,
      "grad_norm": 4.392005920410156,
      "learning_rate": 2.8103538416495885e-05,
      "loss": 3.5033,
      "step": 210000
    },
    {
      "epoch": 1.7548257708698922,
      "eval_loss": 3.391627073287964,
      "eval_runtime": 40.4818,
      "eval_samples_per_second": 110.222,
      "eval_steps_per_second": 13.784,
      "step": 210000
    },
    {
      "epoch": 1.7590039274672016,
      "grad_norm": 3.1449947357177734,
      "learning_rate": 2.805125684888536e-05,
      "loss": 3.4507,
      "step": 210500
    },
    {
      "epoch": 1.7631820840645107,
      "grad_norm": 5.737192153930664,
      "learning_rate": 2.7998975281274836e-05,
      "loss": 3.4617,
      "step": 211000
    },
    {
      "epoch": 1.7631820840645107,
      "eval_loss": 3.3873517513275146,
      "eval_runtime": 40.4668,
      "eval_samples_per_second": 110.263,
      "eval_steps_per_second": 13.789,
      "step": 211000
    },
    {
      "epoch": 1.7673602406618198,
      "grad_norm": 5.513020038604736,
      "learning_rate": 2.7946693713664312e-05,
      "loss": 3.4475,
      "step": 211500
    },
    {
      "epoch": 1.7715383972591292,
      "grad_norm": 3.074239492416382,
      "learning_rate": 2.789451670918901e-05,
      "loss": 3.4343,
      "step": 212000
    },
    {
      "epoch": 1.7715383972591292,
      "eval_loss": 3.38964581489563,
      "eval_runtime": 41.7425,
      "eval_samples_per_second": 106.893,
      "eval_steps_per_second": 13.368,
      "step": 212000
    },
    {
      "epoch": 1.7757165538564386,
      "grad_norm": 3.9970204830169678,
      "learning_rate": 2.784223514157849e-05,
      "loss": 3.3716,
      "step": 212500
    },
    {
      "epoch": 1.779894710453748,
      "grad_norm": 4.594680309295654,
      "learning_rate": 2.7789953573967965e-05,
      "loss": 3.4534,
      "step": 213000
    },
    {
      "epoch": 1.779894710453748,
      "eval_loss": 3.382511854171753,
      "eval_runtime": 40.4565,
      "eval_samples_per_second": 110.291,
      "eval_steps_per_second": 13.793,
      "step": 213000
    },
    {
      "epoch": 1.784072867051057,
      "grad_norm": 6.570225238800049,
      "learning_rate": 2.773767200635744e-05,
      "loss": 3.4291,
      "step": 213500
    },
    {
      "epoch": 1.7882510236483662,
      "grad_norm": 5.155056953430176,
      "learning_rate": 2.768549500188214e-05,
      "loss": 3.4402,
      "step": 214000
    },
    {
      "epoch": 1.7882510236483662,
      "eval_loss": 3.381510019302368,
      "eval_runtime": 40.5141,
      "eval_samples_per_second": 110.134,
      "eval_steps_per_second": 13.773,
      "step": 214000
    },
    {
      "epoch": 1.7924291802456755,
      "grad_norm": 4.226760387420654,
      "learning_rate": 2.7633213434271614e-05,
      "loss": 3.4705,
      "step": 214500
    },
    {
      "epoch": 1.796607336842985,
      "grad_norm": 6.171258926391602,
      "learning_rate": 2.7581036429796316e-05,
      "loss": 3.4773,
      "step": 215000
    },
    {
      "epoch": 1.796607336842985,
      "eval_loss": 3.3844947814941406,
      "eval_runtime": 40.4689,
      "eval_samples_per_second": 110.258,
      "eval_steps_per_second": 13.788,
      "step": 215000
    },
    {
      "epoch": 1.8007854934402943,
      "grad_norm": 5.332149982452393,
      "learning_rate": 2.752875486218579e-05,
      "loss": 3.4529,
      "step": 215500
    },
    {
      "epoch": 1.8049636500376034,
      "grad_norm": 6.050594806671143,
      "learning_rate": 2.7476473294575267e-05,
      "loss": 3.4419,
      "step": 216000
    },
    {
      "epoch": 1.8049636500376034,
      "eval_loss": 3.3761329650878906,
      "eval_runtime": 40.4659,
      "eval_samples_per_second": 110.266,
      "eval_steps_per_second": 13.789,
      "step": 216000
    },
    {
      "epoch": 1.8091418066349125,
      "grad_norm": 4.910857200622559,
      "learning_rate": 2.7424191726964742e-05,
      "loss": 3.424,
      "step": 216500
    },
    {
      "epoch": 1.8133199632322219,
      "grad_norm": 8.457829475402832,
      "learning_rate": 2.7371910159354218e-05,
      "loss": 3.4431,
      "step": 217000
    },
    {
      "epoch": 1.8133199632322219,
      "eval_loss": 3.38053297996521,
      "eval_runtime": 40.4472,
      "eval_samples_per_second": 110.317,
      "eval_steps_per_second": 13.796,
      "step": 217000
    },
    {
      "epoch": 1.8174981198295312,
      "grad_norm": 4.863792896270752,
      "learning_rate": 2.7319628591743697e-05,
      "loss": 3.4556,
      "step": 217500
    },
    {
      "epoch": 1.8216762764268406,
      "grad_norm": 6.228066921234131,
      "learning_rate": 2.7267347024133172e-05,
      "loss": 3.4047,
      "step": 218000
    },
    {
      "epoch": 1.8216762764268406,
      "eval_loss": 3.3676483631134033,
      "eval_runtime": 40.4895,
      "eval_samples_per_second": 110.201,
      "eval_steps_per_second": 13.781,
      "step": 218000
    },
    {
      "epoch": 1.8258544330241497,
      "grad_norm": 6.320400714874268,
      "learning_rate": 2.7215065456522648e-05,
      "loss": 3.4152,
      "step": 218500
    },
    {
      "epoch": 1.8300325896214589,
      "grad_norm": 3.3179121017456055,
      "learning_rate": 2.7162888452047346e-05,
      "loss": 3.3419,
      "step": 219000
    },
    {
      "epoch": 1.8300325896214589,
      "eval_loss": 3.3715813159942627,
      "eval_runtime": 40.4828,
      "eval_samples_per_second": 110.22,
      "eval_steps_per_second": 13.784,
      "step": 219000
    },
    {
      "epoch": 1.8342107462187682,
      "grad_norm": 4.821847915649414,
      "learning_rate": 2.7110606884436822e-05,
      "loss": 3.4722,
      "step": 219500
    },
    {
      "epoch": 1.8383889028160776,
      "grad_norm": 5.0839972496032715,
      "learning_rate": 2.70583253168263e-05,
      "loss": 3.3995,
      "step": 220000
    },
    {
      "epoch": 1.8383889028160776,
      "eval_loss": 3.3703973293304443,
      "eval_runtime": 40.4815,
      "eval_samples_per_second": 110.223,
      "eval_steps_per_second": 13.784,
      "step": 220000
    },
    {
      "epoch": 1.842567059413387,
      "grad_norm": 6.820371627807617,
      "learning_rate": 2.7006043749215776e-05,
      "loss": 3.4211,
      "step": 220500
    },
    {
      "epoch": 1.846745216010696,
      "grad_norm": 6.776442527770996,
      "learning_rate": 2.6953866744740475e-05,
      "loss": 3.4227,
      "step": 221000
    },
    {
      "epoch": 1.846745216010696,
      "eval_loss": 3.372316360473633,
      "eval_runtime": 40.4479,
      "eval_samples_per_second": 110.315,
      "eval_steps_per_second": 13.796,
      "step": 221000
    },
    {
      "epoch": 1.8509233726080052,
      "grad_norm": 5.450851917266846,
      "learning_rate": 2.690158517712995e-05,
      "loss": 3.5122,
      "step": 221500
    },
    {
      "epoch": 1.8551015292053146,
      "grad_norm": 7.390010833740234,
      "learning_rate": 2.6849303609519426e-05,
      "loss": 3.4446,
      "step": 222000
    },
    {
      "epoch": 1.8551015292053146,
      "eval_loss": 3.3640332221984863,
      "eval_runtime": 40.4941,
      "eval_samples_per_second": 110.189,
      "eval_steps_per_second": 13.78,
      "step": 222000
    },
    {
      "epoch": 1.859279685802624,
      "grad_norm": 5.441040992736816,
      "learning_rate": 2.6797022041908905e-05,
      "loss": 3.4843,
      "step": 222500
    },
    {
      "epoch": 1.8634578423999333,
      "grad_norm": 6.238590717315674,
      "learning_rate": 2.6744845037433603e-05,
      "loss": 3.3788,
      "step": 223000
    },
    {
      "epoch": 1.8634578423999333,
      "eval_loss": 3.360541582107544,
      "eval_runtime": 40.4855,
      "eval_samples_per_second": 110.212,
      "eval_steps_per_second": 13.783,
      "step": 223000
    },
    {
      "epoch": 1.8676359989972424,
      "grad_norm": 1.9441295862197876,
      "learning_rate": 2.669256346982308e-05,
      "loss": 3.4215,
      "step": 223500
    },
    {
      "epoch": 1.8718141555945516,
      "grad_norm": 4.011134624481201,
      "learning_rate": 2.6640281902212554e-05,
      "loss": 3.4202,
      "step": 224000
    },
    {
      "epoch": 1.8718141555945516,
      "eval_loss": 3.3591411113739014,
      "eval_runtime": 40.4741,
      "eval_samples_per_second": 110.243,
      "eval_steps_per_second": 13.787,
      "step": 224000
    },
    {
      "epoch": 1.875992312191861,
      "grad_norm": 4.267407417297363,
      "learning_rate": 2.6588000334602033e-05,
      "loss": 3.3769,
      "step": 224500
    },
    {
      "epoch": 1.8801704687891703,
      "grad_norm": 4.621767520904541,
      "learning_rate": 2.6535718766991512e-05,
      "loss": 3.408,
      "step": 225000
    },
    {
      "epoch": 1.8801704687891703,
      "eval_loss": 3.3587403297424316,
      "eval_runtime": 40.4994,
      "eval_samples_per_second": 110.174,
      "eval_steps_per_second": 13.778,
      "step": 225000
    },
    {
      "epoch": 1.8843486253864796,
      "grad_norm": 4.63948392868042,
      "learning_rate": 2.6483541762516207e-05,
      "loss": 3.3376,
      "step": 225500
    },
    {
      "epoch": 1.8885267819837888,
      "grad_norm": 3.7362630367279053,
      "learning_rate": 2.6431260194905682e-05,
      "loss": 3.3261,
      "step": 226000
    },
    {
      "epoch": 1.8885267819837888,
      "eval_loss": 3.357508420944214,
      "eval_runtime": 40.4249,
      "eval_samples_per_second": 110.378,
      "eval_steps_per_second": 13.803,
      "step": 226000
    },
    {
      "epoch": 1.892704938581098,
      "grad_norm": 3.484428644180298,
      "learning_rate": 2.6378978627295165e-05,
      "loss": 3.3812,
      "step": 226500
    },
    {
      "epoch": 1.8968830951784073,
      "grad_norm": 5.982702255249023,
      "learning_rate": 2.632669705968464e-05,
      "loss": 3.4233,
      "step": 227000
    },
    {
      "epoch": 1.8968830951784073,
      "eval_loss": 3.355980396270752,
      "eval_runtime": 40.4822,
      "eval_samples_per_second": 110.221,
      "eval_steps_per_second": 13.784,
      "step": 227000
    },
    {
      "epoch": 1.9010612517757166,
      "grad_norm": 3.434718132019043,
      "learning_rate": 2.6274415492074116e-05,
      "loss": 3.3951,
      "step": 227500
    },
    {
      "epoch": 1.905239408373026,
      "grad_norm": 4.6421709060668945,
      "learning_rate": 2.6222238487598817e-05,
      "loss": 3.3664,
      "step": 228000
    },
    {
      "epoch": 1.905239408373026,
      "eval_loss": 3.351598024368286,
      "eval_runtime": 40.4676,
      "eval_samples_per_second": 110.261,
      "eval_steps_per_second": 13.789,
      "step": 228000
    },
    {
      "epoch": 1.909417564970335,
      "grad_norm": 10.201983451843262,
      "learning_rate": 2.6169956919988293e-05,
      "loss": 3.4341,
      "step": 228500
    },
    {
      "epoch": 1.9135957215676442,
      "grad_norm": 4.66003942489624,
      "learning_rate": 2.611767535237777e-05,
      "loss": 3.3314,
      "step": 229000
    },
    {
      "epoch": 1.9135957215676442,
      "eval_loss": 3.3540987968444824,
      "eval_runtime": 40.4587,
      "eval_samples_per_second": 110.285,
      "eval_steps_per_second": 13.792,
      "step": 229000
    },
    {
      "epoch": 1.9177738781649536,
      "grad_norm": 1.5573841333389282,
      "learning_rate": 2.6065393784767244e-05,
      "loss": 3.3577,
      "step": 229500
    },
    {
      "epoch": 1.921952034762263,
      "grad_norm": 4.663156509399414,
      "learning_rate": 2.6013216780291942e-05,
      "loss": 3.3725,
      "step": 230000
    },
    {
      "epoch": 1.921952034762263,
      "eval_loss": 3.3473751544952393,
      "eval_runtime": 40.4615,
      "eval_samples_per_second": 110.278,
      "eval_steps_per_second": 13.791,
      "step": 230000
    },
    {
      "epoch": 1.9261301913595723,
      "grad_norm": 4.893932342529297,
      "learning_rate": 2.596093521268142e-05,
      "loss": 3.3943,
      "step": 230500
    },
    {
      "epoch": 1.9303083479568814,
      "grad_norm": 1.92241370677948,
      "learning_rate": 2.5908653645070897e-05,
      "loss": 3.3888,
      "step": 231000
    },
    {
      "epoch": 1.9303083479568814,
      "eval_loss": 3.3468871116638184,
      "eval_runtime": 40.474,
      "eval_samples_per_second": 110.244,
      "eval_steps_per_second": 13.787,
      "step": 231000
    },
    {
      "epoch": 1.9344865045541906,
      "grad_norm": 4.520233631134033,
      "learning_rate": 2.5856372077460372e-05,
      "loss": 3.3683,
      "step": 231500
    },
    {
      "epoch": 1.9386646611515,
      "grad_norm": 4.232703685760498,
      "learning_rate": 2.5804090509849848e-05,
      "loss": 3.3707,
      "step": 232000
    },
    {
      "epoch": 1.9386646611515,
      "eval_loss": 3.345097780227661,
      "eval_runtime": 40.4911,
      "eval_samples_per_second": 110.197,
      "eval_steps_per_second": 13.781,
      "step": 232000
    },
    {
      "epoch": 1.9428428177488093,
      "grad_norm": 3.1016223430633545,
      "learning_rate": 2.5751913505374546e-05,
      "loss": 3.3736,
      "step": 232500
    },
    {
      "epoch": 1.9470209743461186,
      "grad_norm": 4.9680047035217285,
      "learning_rate": 2.5699631937764025e-05,
      "loss": 3.2975,
      "step": 233000
    },
    {
      "epoch": 1.9470209743461186,
      "eval_loss": 3.3408877849578857,
      "eval_runtime": 40.4839,
      "eval_samples_per_second": 110.217,
      "eval_steps_per_second": 13.783,
      "step": 233000
    },
    {
      "epoch": 1.9511991309434278,
      "grad_norm": 3.1786868572235107,
      "learning_rate": 2.56473503701535e-05,
      "loss": 3.4885,
      "step": 233500
    },
    {
      "epoch": 1.955377287540737,
      "grad_norm": 4.074703693389893,
      "learning_rate": 2.5595068802542976e-05,
      "loss": 3.3695,
      "step": 234000
    },
    {
      "epoch": 1.955377287540737,
      "eval_loss": 3.3347997665405273,
      "eval_runtime": 40.4887,
      "eval_samples_per_second": 110.204,
      "eval_steps_per_second": 13.782,
      "step": 234000
    },
    {
      "epoch": 1.9595554441380463,
      "grad_norm": 3.2261080741882324,
      "learning_rate": 2.554278723493245e-05,
      "loss": 3.3795,
      "step": 234500
    },
    {
      "epoch": 1.9637336007353556,
      "grad_norm": 4.230503559112549,
      "learning_rate": 2.549061023045715e-05,
      "loss": 3.4472,
      "step": 235000
    },
    {
      "epoch": 1.9637336007353556,
      "eval_loss": 3.3395724296569824,
      "eval_runtime": 40.4672,
      "eval_samples_per_second": 110.262,
      "eval_steps_per_second": 13.789,
      "step": 235000
    },
    {
      "epoch": 1.967911757332665,
      "grad_norm": 7.1046833992004395,
      "learning_rate": 2.543832866284663e-05,
      "loss": 3.4055,
      "step": 235500
    },
    {
      "epoch": 1.9720899139299741,
      "grad_norm": 5.35427713394165,
      "learning_rate": 2.5386047095236104e-05,
      "loss": 3.3572,
      "step": 236000
    },
    {
      "epoch": 1.9720899139299741,
      "eval_loss": 3.3387038707733154,
      "eval_runtime": 40.4842,
      "eval_samples_per_second": 110.216,
      "eval_steps_per_second": 13.783,
      "step": 236000
    },
    {
      "epoch": 1.9762680705272833,
      "grad_norm": 3.269439697265625,
      "learning_rate": 2.533376552762558e-05,
      "loss": 3.3778,
      "step": 236500
    },
    {
      "epoch": 1.9804462271245926,
      "grad_norm": 6.2255730628967285,
      "learning_rate": 2.528158852315028e-05,
      "loss": 3.3925,
      "step": 237000
    },
    {
      "epoch": 1.9804462271245926,
      "eval_loss": 3.3347251415252686,
      "eval_runtime": 40.4931,
      "eval_samples_per_second": 110.191,
      "eval_steps_per_second": 13.78,
      "step": 237000
    },
    {
      "epoch": 1.984624383721902,
      "grad_norm": 8.227920532226562,
      "learning_rate": 2.5229306955539754e-05,
      "loss": 3.2903,
      "step": 237500
    },
    {
      "epoch": 1.988802540319211,
      "grad_norm": 3.8814873695373535,
      "learning_rate": 2.5177025387929233e-05,
      "loss": 3.3328,
      "step": 238000
    },
    {
      "epoch": 1.988802540319211,
      "eval_loss": 3.3345508575439453,
      "eval_runtime": 40.4733,
      "eval_samples_per_second": 110.246,
      "eval_steps_per_second": 13.787,
      "step": 238000
    },
    {
      "epoch": 1.9929806969165205,
      "grad_norm": 4.1887006759643555,
      "learning_rate": 2.512474382031871e-05,
      "loss": 3.3947,
      "step": 238500
    },
    {
      "epoch": 1.9971588535138296,
      "grad_norm": 5.336971759796143,
      "learning_rate": 2.5072462252708184e-05,
      "loss": 3.3977,
      "step": 239000
    },
    {
      "epoch": 1.9971588535138296,
      "eval_loss": 3.333317756652832,
      "eval_runtime": 40.4282,
      "eval_samples_per_second": 110.369,
      "eval_steps_per_second": 13.802,
      "step": 239000
    },
    {
      "epoch": 2.001337010111139,
      "grad_norm": 5.35257625579834,
      "learning_rate": 2.502018068509766e-05,
      "loss": 3.2881,
      "step": 239500
    },
    {
      "epoch": 2.0055151667084483,
      "grad_norm": 4.52628231048584,
      "learning_rate": 2.496789911748714e-05,
      "loss": 3.1522,
      "step": 240000
    },
    {
      "epoch": 2.0055151667084483,
      "eval_loss": 3.3466105461120605,
      "eval_runtime": 40.4274,
      "eval_samples_per_second": 110.371,
      "eval_steps_per_second": 13.803,
      "step": 240000
    },
    {
      "epoch": 2.0096933233057577,
      "grad_norm": 5.243988037109375,
      "learning_rate": 2.491572211301184e-05,
      "loss": 3.1144,
      "step": 240500
    },
    {
      "epoch": 2.0138714799030666,
      "grad_norm": 6.480501174926758,
      "learning_rate": 2.4863440545401316e-05,
      "loss": 3.1233,
      "step": 241000
    },
    {
      "epoch": 2.0138714799030666,
      "eval_loss": 3.3396503925323486,
      "eval_runtime": 40.4455,
      "eval_samples_per_second": 110.321,
      "eval_steps_per_second": 13.796,
      "step": 241000
    },
    {
      "epoch": 2.018049636500376,
      "grad_norm": 4.595026016235352,
      "learning_rate": 2.481115897779079e-05,
      "loss": 3.0936,
      "step": 241500
    },
    {
      "epoch": 2.0222277930976853,
      "grad_norm": 7.027838706970215,
      "learning_rate": 2.4758877410180267e-05,
      "loss": 3.1178,
      "step": 242000
    },
    {
      "epoch": 2.0222277930976853,
      "eval_loss": 3.3471105098724365,
      "eval_runtime": 40.4778,
      "eval_samples_per_second": 110.233,
      "eval_steps_per_second": 13.785,
      "step": 242000
    },
    {
      "epoch": 2.0264059496949947,
      "grad_norm": 3.6396079063415527,
      "learning_rate": 2.4706700405704965e-05,
      "loss": 3.1851,
      "step": 242500
    },
    {
      "epoch": 2.030584106292304,
      "grad_norm": 6.937821865081787,
      "learning_rate": 2.4654418838094444e-05,
      "loss": 3.0675,
      "step": 243000
    },
    {
      "epoch": 2.030584106292304,
      "eval_loss": 3.3515665531158447,
      "eval_runtime": 40.4644,
      "eval_samples_per_second": 110.27,
      "eval_steps_per_second": 13.79,
      "step": 243000
    },
    {
      "epoch": 2.034762262889613,
      "grad_norm": 6.9290008544921875,
      "learning_rate": 2.460213727048392e-05,
      "loss": 3.1355,
      "step": 243500
    },
    {
      "epoch": 2.0389404194869223,
      "grad_norm": 4.02460241317749,
      "learning_rate": 2.4549960266008618e-05,
      "loss": 3.1461,
      "step": 244000
    },
    {
      "epoch": 2.0389404194869223,
      "eval_loss": 3.341926336288452,
      "eval_runtime": 40.4246,
      "eval_samples_per_second": 110.378,
      "eval_steps_per_second": 13.803,
      "step": 244000
    },
    {
      "epoch": 2.0431185760842316,
      "grad_norm": 8.292485237121582,
      "learning_rate": 2.4497678698398093e-05,
      "loss": 3.2199,
      "step": 244500
    },
    {
      "epoch": 2.047296732681541,
      "grad_norm": 10.30803108215332,
      "learning_rate": 2.444539713078757e-05,
      "loss": 3.1664,
      "step": 245000
    },
    {
      "epoch": 2.047296732681541,
      "eval_loss": 3.3483948707580566,
      "eval_runtime": 40.4677,
      "eval_samples_per_second": 110.261,
      "eval_steps_per_second": 13.789,
      "step": 245000
    },
    {
      "epoch": 2.0514748892788504,
      "grad_norm": 4.300713539123535,
      "learning_rate": 2.4393115563177048e-05,
      "loss": 3.1249,
      "step": 245500
    },
    {
      "epoch": 2.0556530458761593,
      "grad_norm": 4.996560096740723,
      "learning_rate": 2.4340833995566527e-05,
      "loss": 3.1182,
      "step": 246000
    },
    {
      "epoch": 2.0556530458761593,
      "eval_loss": 3.342724323272705,
      "eval_runtime": 40.4703,
      "eval_samples_per_second": 110.254,
      "eval_steps_per_second": 13.788,
      "step": 246000
    },
    {
      "epoch": 2.0598312024734686,
      "grad_norm": 4.595823287963867,
      "learning_rate": 2.4288552427956002e-05,
      "loss": 3.1257,
      "step": 246500
    },
    {
      "epoch": 2.064009359070778,
      "grad_norm": 7.316824913024902,
      "learning_rate": 2.4236270860345478e-05,
      "loss": 3.2508,
      "step": 247000
    },
    {
      "epoch": 2.064009359070778,
      "eval_loss": 3.341085910797119,
      "eval_runtime": 40.4526,
      "eval_samples_per_second": 110.302,
      "eval_steps_per_second": 13.794,
      "step": 247000
    },
    {
      "epoch": 2.0681875156680873,
      "grad_norm": 4.907251358032227,
      "learning_rate": 2.4183989292734953e-05,
      "loss": 3.0809,
      "step": 247500
    },
    {
      "epoch": 2.0723656722653967,
      "grad_norm": 5.387881755828857,
      "learning_rate": 2.413181228825965e-05,
      "loss": 3.1521,
      "step": 248000
    },
    {
      "epoch": 2.0723656722653967,
      "eval_loss": 3.3404438495635986,
      "eval_runtime": 40.4564,
      "eval_samples_per_second": 110.292,
      "eval_steps_per_second": 13.793,
      "step": 248000
    },
    {
      "epoch": 2.0765438288627056,
      "grad_norm": 5.49289608001709,
      "learning_rate": 2.407953072064913e-05,
      "loss": 3.2411,
      "step": 248500
    },
    {
      "epoch": 2.080721985460015,
      "grad_norm": 4.194238662719727,
      "learning_rate": 2.4027249153038606e-05,
      "loss": 3.1179,
      "step": 249000
    },
    {
      "epoch": 2.080721985460015,
      "eval_loss": 3.342294931411743,
      "eval_runtime": 40.4613,
      "eval_samples_per_second": 110.278,
      "eval_steps_per_second": 13.791,
      "step": 249000
    },
    {
      "epoch": 2.0849001420573243,
      "grad_norm": 6.045317649841309,
      "learning_rate": 2.397496758542808e-05,
      "loss": 3.186,
      "step": 249500
    },
    {
      "epoch": 2.0890782986546337,
      "grad_norm": 6.206670761108398,
      "learning_rate": 2.3922686017817557e-05,
      "loss": 3.0654,
      "step": 250000
    },
    {
      "epoch": 2.0890782986546337,
      "eval_loss": 3.345398426055908,
      "eval_runtime": 40.4592,
      "eval_samples_per_second": 110.284,
      "eval_steps_per_second": 13.792,
      "step": 250000
    },
    {
      "epoch": 2.093256455251943,
      "grad_norm": 4.87103796005249,
      "learning_rate": 2.3870404450207036e-05,
      "loss": 3.2034,
      "step": 250500
    },
    {
      "epoch": 2.097434611849252,
      "grad_norm": 4.335954189300537,
      "learning_rate": 2.3818227445731734e-05,
      "loss": 3.2126,
      "step": 251000
    },
    {
      "epoch": 2.097434611849252,
      "eval_loss": 3.340569019317627,
      "eval_runtime": 40.4548,
      "eval_samples_per_second": 110.296,
      "eval_steps_per_second": 13.793,
      "step": 251000
    },
    {
      "epoch": 2.1016127684465613,
      "grad_norm": 6.09652042388916,
      "learning_rate": 2.376594587812121e-05,
      "loss": 3.1864,
      "step": 251500
    },
    {
      "epoch": 2.1057909250438707,
      "grad_norm": 9.233617782592773,
      "learning_rate": 2.371366431051069e-05,
      "loss": 3.1725,
      "step": 252000
    },
    {
      "epoch": 2.1057909250438707,
      "eval_loss": 3.342285394668579,
      "eval_runtime": 40.5048,
      "eval_samples_per_second": 110.16,
      "eval_steps_per_second": 13.776,
      "step": 252000
    },
    {
      "epoch": 2.10996908164118,
      "grad_norm": 6.006067276000977,
      "learning_rate": 2.3661382742900164e-05,
      "loss": 3.1333,
      "step": 252500
    },
    {
      "epoch": 2.1141472382384894,
      "grad_norm": 6.246615886688232,
      "learning_rate": 2.360910117528964e-05,
      "loss": 3.0531,
      "step": 253000
    },
    {
      "epoch": 2.1141472382384894,
      "eval_loss": 3.3393425941467285,
      "eval_runtime": 40.4551,
      "eval_samples_per_second": 110.295,
      "eval_steps_per_second": 13.793,
      "step": 253000
    },
    {
      "epoch": 2.1183253948357983,
      "grad_norm": 6.1417765617370605,
      "learning_rate": 2.355692417081434e-05,
      "loss": 3.1325,
      "step": 253500
    },
    {
      "epoch": 2.1225035514331076,
      "grad_norm": 4.473926067352295,
      "learning_rate": 2.3504642603203817e-05,
      "loss": 3.1911,
      "step": 254000
    },
    {
      "epoch": 2.1225035514331076,
      "eval_loss": 3.3351833820343018,
      "eval_runtime": 40.4613,
      "eval_samples_per_second": 110.278,
      "eval_steps_per_second": 13.791,
      "step": 254000
    },
    {
      "epoch": 2.126681708030417,
      "grad_norm": 8.741314888000488,
      "learning_rate": 2.3452361035593293e-05,
      "loss": 3.1,
      "step": 254500
    },
    {
      "epoch": 2.1308598646277264,
      "grad_norm": 7.776458740234375,
      "learning_rate": 2.3400079467982768e-05,
      "loss": 3.1181,
      "step": 255000
    },
    {
      "epoch": 2.1308598646277264,
      "eval_loss": 3.3358802795410156,
      "eval_runtime": 40.4824,
      "eval_samples_per_second": 110.221,
      "eval_steps_per_second": 13.784,
      "step": 255000
    },
    {
      "epoch": 2.1350380212250357,
      "grad_norm": 4.580501556396484,
      "learning_rate": 2.3347902463507467e-05,
      "loss": 3.1562,
      "step": 255500
    },
    {
      "epoch": 2.1392161778223446,
      "grad_norm": 4.885230541229248,
      "learning_rate": 2.3295620895896946e-05,
      "loss": 3.1109,
      "step": 256000
    },
    {
      "epoch": 2.1392161778223446,
      "eval_loss": 3.3395299911499023,
      "eval_runtime": 40.4629,
      "eval_samples_per_second": 110.274,
      "eval_steps_per_second": 13.79,
      "step": 256000
    },
    {
      "epoch": 2.143394334419654,
      "grad_norm": 5.507672309875488,
      "learning_rate": 2.324333932828642e-05,
      "loss": 3.2034,
      "step": 256500
    },
    {
      "epoch": 2.1475724910169633,
      "grad_norm": 4.263452529907227,
      "learning_rate": 2.3191057760675897e-05,
      "loss": 3.1226,
      "step": 257000
    },
    {
      "epoch": 2.1475724910169633,
      "eval_loss": 3.333815336227417,
      "eval_runtime": 40.4332,
      "eval_samples_per_second": 110.355,
      "eval_steps_per_second": 13.801,
      "step": 257000
    },
    {
      "epoch": 2.1517506476142727,
      "grad_norm": 4.941694736480713,
      "learning_rate": 2.3138776193065372e-05,
      "loss": 3.1588,
      "step": 257500
    },
    {
      "epoch": 2.155928804211582,
      "grad_norm": 3.2227394580841064,
      "learning_rate": 2.308659918859007e-05,
      "loss": 3.0437,
      "step": 258000
    },
    {
      "epoch": 2.155928804211582,
      "eval_loss": 3.3339126110076904,
      "eval_runtime": 40.4431,
      "eval_samples_per_second": 110.328,
      "eval_steps_per_second": 13.797,
      "step": 258000
    },
    {
      "epoch": 2.160106960808891,
      "grad_norm": 6.820015907287598,
      "learning_rate": 2.303431762097955e-05,
      "loss": 3.1968,
      "step": 258500
    },
    {
      "epoch": 2.1642851174062003,
      "grad_norm": 7.1376800537109375,
      "learning_rate": 2.2982036053369025e-05,
      "loss": 3.1401,
      "step": 259000
    },
    {
      "epoch": 2.1642851174062003,
      "eval_loss": 3.3317220211029053,
      "eval_runtime": 40.454,
      "eval_samples_per_second": 110.298,
      "eval_steps_per_second": 13.793,
      "step": 259000
    },
    {
      "epoch": 2.1684632740035097,
      "grad_norm": 10.74910831451416,
      "learning_rate": 2.2929754485758504e-05,
      "loss": 3.1429,
      "step": 259500
    },
    {
      "epoch": 2.172641430600819,
      "grad_norm": 5.5305585861206055,
      "learning_rate": 2.28775774812832e-05,
      "loss": 3.1238,
      "step": 260000
    },
    {
      "epoch": 2.172641430600819,
      "eval_loss": 3.33156156539917,
      "eval_runtime": 40.4787,
      "eval_samples_per_second": 110.231,
      "eval_steps_per_second": 13.785,
      "step": 260000
    },
    {
      "epoch": 2.1768195871981284,
      "grad_norm": 4.4482927322387695,
      "learning_rate": 2.2825295913672674e-05,
      "loss": 3.1677,
      "step": 260500
    },
    {
      "epoch": 2.1809977437954373,
      "grad_norm": 7.104944229125977,
      "learning_rate": 2.2773014346062153e-05,
      "loss": 3.1312,
      "step": 261000
    },
    {
      "epoch": 2.1809977437954373,
      "eval_loss": 3.3333513736724854,
      "eval_runtime": 40.4765,
      "eval_samples_per_second": 110.237,
      "eval_steps_per_second": 13.786,
      "step": 261000
    },
    {
      "epoch": 2.1851759003927467,
      "grad_norm": 4.176535129547119,
      "learning_rate": 2.2720732778451632e-05,
      "loss": 3.1194,
      "step": 261500
    },
    {
      "epoch": 2.189354056990056,
      "grad_norm": 10.3728609085083,
      "learning_rate": 2.2668451210841108e-05,
      "loss": 3.1206,
      "step": 262000
    },
    {
      "epoch": 2.189354056990056,
      "eval_loss": 3.3327999114990234,
      "eval_runtime": 40.4475,
      "eval_samples_per_second": 110.316,
      "eval_steps_per_second": 13.796,
      "step": 262000
    },
    {
      "epoch": 2.1935322135873654,
      "grad_norm": 5.367616653442383,
      "learning_rate": 2.2616274206365803e-05,
      "loss": 3.1405,
      "step": 262500
    },
    {
      "epoch": 2.1977103701846747,
      "grad_norm": 4.320235729217529,
      "learning_rate": 2.256399263875528e-05,
      "loss": 3.0658,
      "step": 263000
    },
    {
      "epoch": 2.1977103701846747,
      "eval_loss": 3.3294503688812256,
      "eval_runtime": 40.4559,
      "eval_samples_per_second": 110.293,
      "eval_steps_per_second": 13.793,
      "step": 263000
    },
    {
      "epoch": 2.2018885267819837,
      "grad_norm": 6.458812236785889,
      "learning_rate": 2.2511711071144757e-05,
      "loss": 3.1988,
      "step": 263500
    },
    {
      "epoch": 2.206066683379293,
      "grad_norm": 6.220398426055908,
      "learning_rate": 2.2459429503534236e-05,
      "loss": 3.1453,
      "step": 264000
    },
    {
      "epoch": 2.206066683379293,
      "eval_loss": 3.3279049396514893,
      "eval_runtime": 40.4446,
      "eval_samples_per_second": 110.324,
      "eval_steps_per_second": 13.797,
      "step": 264000
    },
    {
      "epoch": 2.2102448399766024,
      "grad_norm": 4.93303108215332,
      "learning_rate": 2.2407252499058934e-05,
      "loss": 3.1548,
      "step": 264500
    },
    {
      "epoch": 2.2144229965739117,
      "grad_norm": 4.953177452087402,
      "learning_rate": 2.235497093144841e-05,
      "loss": 3.1606,
      "step": 265000
    },
    {
      "epoch": 2.2144229965739117,
      "eval_loss": 3.3168184757232666,
      "eval_runtime": 40.4624,
      "eval_samples_per_second": 110.275,
      "eval_steps_per_second": 13.791,
      "step": 265000
    },
    {
      "epoch": 2.218601153171221,
      "grad_norm": 7.259415626525879,
      "learning_rate": 2.2302689363837885e-05,
      "loss": 3.1821,
      "step": 265500
    },
    {
      "epoch": 2.22277930976853,
      "grad_norm": 3.314384937286377,
      "learning_rate": 2.2250407796227364e-05,
      "loss": 3.1333,
      "step": 266000
    },
    {
      "epoch": 2.22277930976853,
      "eval_loss": 3.323383092880249,
      "eval_runtime": 40.4626,
      "eval_samples_per_second": 110.275,
      "eval_steps_per_second": 13.791,
      "step": 266000
    },
    {
      "epoch": 2.2269574663658394,
      "grad_norm": 6.855319499969482,
      "learning_rate": 2.219812622861684e-05,
      "loss": 3.1626,
      "step": 266500
    },
    {
      "epoch": 2.2311356229631487,
      "grad_norm": 8.510984420776367,
      "learning_rate": 2.2145949224141538e-05,
      "loss": 3.1123,
      "step": 267000
    },
    {
      "epoch": 2.2311356229631487,
      "eval_loss": 3.3220582008361816,
      "eval_runtime": 40.4492,
      "eval_samples_per_second": 110.311,
      "eval_steps_per_second": 13.795,
      "step": 267000
    },
    {
      "epoch": 2.235313779560458,
      "grad_norm": 5.909183502197266,
      "learning_rate": 2.2093667656531014e-05,
      "loss": 3.1804,
      "step": 267500
    },
    {
      "epoch": 2.2394919361577674,
      "grad_norm": 4.291212558746338,
      "learning_rate": 2.204138608892049e-05,
      "loss": 3.1723,
      "step": 268000
    },
    {
      "epoch": 2.2394919361577674,
      "eval_loss": 3.3167879581451416,
      "eval_runtime": 40.4755,
      "eval_samples_per_second": 110.24,
      "eval_steps_per_second": 13.786,
      "step": 268000
    },
    {
      "epoch": 2.2436700927550763,
      "grad_norm": 6.5470404624938965,
      "learning_rate": 2.1989104521309968e-05,
      "loss": 3.0986,
      "step": 268500
    },
    {
      "epoch": 2.2478482493523857,
      "grad_norm": 3.792804718017578,
      "learning_rate": 2.1936822953699447e-05,
      "loss": 3.1596,
      "step": 269000
    },
    {
      "epoch": 2.2478482493523857,
      "eval_loss": 3.3173413276672363,
      "eval_runtime": 40.4588,
      "eval_samples_per_second": 110.285,
      "eval_steps_per_second": 13.792,
      "step": 269000
    },
    {
      "epoch": 2.252026405949695,
      "grad_norm": 4.821345806121826,
      "learning_rate": 2.1884645949224142e-05,
      "loss": 3.1677,
      "step": 269500
    },
    {
      "epoch": 2.2562045625470044,
      "grad_norm": 11.906582832336426,
      "learning_rate": 2.1832364381613618e-05,
      "loss": 3.1157,
      "step": 270000
    },
    {
      "epoch": 2.2562045625470044,
      "eval_loss": 3.321375846862793,
      "eval_runtime": 40.4674,
      "eval_samples_per_second": 110.262,
      "eval_steps_per_second": 13.789,
      "step": 270000
    },
    {
      "epoch": 2.2603827191443138,
      "grad_norm": 5.045265197753906,
      "learning_rate": 2.1780082814003097e-05,
      "loss": 3.0692,
      "step": 270500
    },
    {
      "epoch": 2.2645608757416227,
      "grad_norm": 8.908251762390137,
      "learning_rate": 2.1727801246392572e-05,
      "loss": 3.0644,
      "step": 271000
    },
    {
      "epoch": 2.2645608757416227,
      "eval_loss": 3.3190419673919678,
      "eval_runtime": 40.4315,
      "eval_samples_per_second": 110.359,
      "eval_steps_per_second": 13.801,
      "step": 271000
    },
    {
      "epoch": 2.268739032338932,
      "grad_norm": 7.89150857925415,
      "learning_rate": 2.167551967878205e-05,
      "loss": 3.0926,
      "step": 271500
    },
    {
      "epoch": 2.2729171889362414,
      "grad_norm": 4.523589134216309,
      "learning_rate": 2.1623238111171526e-05,
      "loss": 3.1457,
      "step": 272000
    },
    {
      "epoch": 2.2729171889362414,
      "eval_loss": 3.316985845565796,
      "eval_runtime": 40.4542,
      "eval_samples_per_second": 110.298,
      "eval_steps_per_second": 13.793,
      "step": 272000
    },
    {
      "epoch": 2.2770953455335508,
      "grad_norm": 3.6266989707946777,
      "learning_rate": 2.1571061106696225e-05,
      "loss": 3.1394,
      "step": 272500
    },
    {
      "epoch": 2.28127350213086,
      "grad_norm": 3.1620638370513916,
      "learning_rate": 2.15187795390857e-05,
      "loss": 3.1771,
      "step": 273000
    },
    {
      "epoch": 2.28127350213086,
      "eval_loss": 3.316102981567383,
      "eval_runtime": 40.5105,
      "eval_samples_per_second": 110.144,
      "eval_steps_per_second": 13.774,
      "step": 273000
    },
    {
      "epoch": 2.285451658728169,
      "grad_norm": 12.058000564575195,
      "learning_rate": 2.1466497971475176e-05,
      "loss": 3.148,
      "step": 273500
    },
    {
      "epoch": 2.2896298153254784,
      "grad_norm": 5.249194145202637,
      "learning_rate": 2.1414216403864655e-05,
      "loss": 3.1185,
      "step": 274000
    },
    {
      "epoch": 2.2896298153254784,
      "eval_loss": 3.31988787651062,
      "eval_runtime": 40.4699,
      "eval_samples_per_second": 110.255,
      "eval_steps_per_second": 13.788,
      "step": 274000
    },
    {
      "epoch": 2.2938079719227877,
      "grad_norm": 7.467369079589844,
      "learning_rate": 2.1362039399389353e-05,
      "loss": 3.1526,
      "step": 274500
    },
    {
      "epoch": 2.297986128520097,
      "grad_norm": 6.099387168884277,
      "learning_rate": 2.130975783177883e-05,
      "loss": 3.1502,
      "step": 275000
    },
    {
      "epoch": 2.297986128520097,
      "eval_loss": 3.316650390625,
      "eval_runtime": 40.4528,
      "eval_samples_per_second": 110.301,
      "eval_steps_per_second": 13.794,
      "step": 275000
    },
    {
      "epoch": 2.302164285117406,
      "grad_norm": 4.817805767059326,
      "learning_rate": 2.1257476264168304e-05,
      "loss": 3.1442,
      "step": 275500
    },
    {
      "epoch": 2.3063424417147154,
      "grad_norm": 5.789041042327881,
      "learning_rate": 2.120519469655778e-05,
      "loss": 3.134,
      "step": 276000
    },
    {
      "epoch": 2.3063424417147154,
      "eval_loss": 3.3131399154663086,
      "eval_runtime": 40.4546,
      "eval_samples_per_second": 110.296,
      "eval_steps_per_second": 13.793,
      "step": 276000
    },
    {
      "epoch": 2.3105205983120247,
      "grad_norm": 5.171199321746826,
      "learning_rate": 2.1152913128947262e-05,
      "loss": 3.108,
      "step": 276500
    },
    {
      "epoch": 2.314698754909334,
      "grad_norm": 5.447643280029297,
      "learning_rate": 2.1100736124471957e-05,
      "loss": 3.1813,
      "step": 277000
    },
    {
      "epoch": 2.314698754909334,
      "eval_loss": 3.313253402709961,
      "eval_runtime": 40.5148,
      "eval_samples_per_second": 110.133,
      "eval_steps_per_second": 13.773,
      "step": 277000
    },
    {
      "epoch": 2.3188769115066434,
      "grad_norm": 4.172455787658691,
      "learning_rate": 2.1048454556861433e-05,
      "loss": 3.1188,
      "step": 277500
    },
    {
      "epoch": 2.3230550681039523,
      "grad_norm": 5.615136623382568,
      "learning_rate": 2.099617298925091e-05,
      "loss": 3.1557,
      "step": 278000
    },
    {
      "epoch": 2.3230550681039523,
      "eval_loss": 3.310053586959839,
      "eval_runtime": 40.4633,
      "eval_samples_per_second": 110.273,
      "eval_steps_per_second": 13.79,
      "step": 278000
    },
    {
      "epoch": 2.3272332247012617,
      "grad_norm": 3.739656448364258,
      "learning_rate": 2.0943891421640387e-05,
      "loss": 3.2203,
      "step": 278500
    },
    {
      "epoch": 2.331411381298571,
      "grad_norm": 6.395985126495361,
      "learning_rate": 2.0891609854029866e-05,
      "loss": 3.1628,
      "step": 279000
    },
    {
      "epoch": 2.331411381298571,
      "eval_loss": 3.306927442550659,
      "eval_runtime": 40.4619,
      "eval_samples_per_second": 110.277,
      "eval_steps_per_second": 13.791,
      "step": 279000
    },
    {
      "epoch": 2.3355895378958804,
      "grad_norm": 7.713507652282715,
      "learning_rate": 2.0839432849554564e-05,
      "loss": 3.0996,
      "step": 279500
    },
    {
      "epoch": 2.3397676944931898,
      "grad_norm": 5.105501651763916,
      "learning_rate": 2.078715128194404e-05,
      "loss": 3.0756,
      "step": 280000
    },
    {
      "epoch": 2.3397676944931898,
      "eval_loss": 3.3022100925445557,
      "eval_runtime": 40.4441,
      "eval_samples_per_second": 110.325,
      "eval_steps_per_second": 13.797,
      "step": 280000
    },
    {
      "epoch": 2.3439458510904987,
      "grad_norm": 4.082066535949707,
      "learning_rate": 2.0734869714333515e-05,
      "loss": 3.0481,
      "step": 280500
    },
    {
      "epoch": 2.348124007687808,
      "grad_norm": 5.3705596923828125,
      "learning_rate": 2.068258814672299e-05,
      "loss": 3.0449,
      "step": 281000
    },
    {
      "epoch": 2.348124007687808,
      "eval_loss": 3.301225185394287,
      "eval_runtime": 40.4433,
      "eval_samples_per_second": 110.327,
      "eval_steps_per_second": 13.797,
      "step": 281000
    },
    {
      "epoch": 2.3523021642851174,
      "grad_norm": 6.272683620452881,
      "learning_rate": 2.063041114224769e-05,
      "loss": 3.0657,
      "step": 281500
    },
    {
      "epoch": 2.3564803208824268,
      "grad_norm": 7.7794976234436035,
      "learning_rate": 2.0578129574637168e-05,
      "loss": 3.0655,
      "step": 282000
    },
    {
      "epoch": 2.3564803208824268,
      "eval_loss": 3.302551507949829,
      "eval_runtime": 40.4599,
      "eval_samples_per_second": 110.282,
      "eval_steps_per_second": 13.791,
      "step": 282000
    },
    {
      "epoch": 2.360658477479736,
      "grad_norm": 8.646065711975098,
      "learning_rate": 2.0525848007026644e-05,
      "loss": 3.0216,
      "step": 282500
    },
    {
      "epoch": 2.364836634077045,
      "grad_norm": 5.221062183380127,
      "learning_rate": 2.047356643941612e-05,
      "loss": 3.1117,
      "step": 283000
    },
    {
      "epoch": 2.364836634077045,
      "eval_loss": 3.297528028488159,
      "eval_runtime": 40.4511,
      "eval_samples_per_second": 110.306,
      "eval_steps_per_second": 13.794,
      "step": 283000
    },
    {
      "epoch": 2.3690147906743544,
      "grad_norm": 10.375117301940918,
      "learning_rate": 2.0421284871805595e-05,
      "loss": 3.1431,
      "step": 283500
    },
    {
      "epoch": 2.3731929472716637,
      "grad_norm": 8.012441635131836,
      "learning_rate": 2.0369107867330296e-05,
      "loss": 3.1707,
      "step": 284000
    },
    {
      "epoch": 2.3731929472716637,
      "eval_loss": 3.2980761528015137,
      "eval_runtime": 40.4552,
      "eval_samples_per_second": 110.295,
      "eval_steps_per_second": 13.793,
      "step": 284000
    },
    {
      "epoch": 2.377371103868973,
      "grad_norm": 3.3558294773101807,
      "learning_rate": 2.0316826299719772e-05,
      "loss": 3.1075,
      "step": 284500
    },
    {
      "epoch": 2.3815492604662825,
      "grad_norm": 4.599654674530029,
      "learning_rate": 2.0264544732109248e-05,
      "loss": 3.1749,
      "step": 285000
    },
    {
      "epoch": 2.3815492604662825,
      "eval_loss": 3.291957139968872,
      "eval_runtime": 40.4393,
      "eval_samples_per_second": 110.338,
      "eval_steps_per_second": 13.798,
      "step": 285000
    },
    {
      "epoch": 2.3857274170635914,
      "grad_norm": 5.334101676940918,
      "learning_rate": 2.0212263164498726e-05,
      "loss": 3.0942,
      "step": 285500
    },
    {
      "epoch": 2.3899055736609007,
      "grad_norm": 6.905644416809082,
      "learning_rate": 2.016008616002342e-05,
      "loss": 3.1453,
      "step": 286000
    },
    {
      "epoch": 2.3899055736609007,
      "eval_loss": 3.2937710285186768,
      "eval_runtime": 40.4861,
      "eval_samples_per_second": 110.211,
      "eval_steps_per_second": 13.783,
      "step": 286000
    },
    {
      "epoch": 2.39408373025821,
      "grad_norm": 5.987873554229736,
      "learning_rate": 2.01078045924129e-05,
      "loss": 3.1466,
      "step": 286500
    },
    {
      "epoch": 2.3982618868555194,
      "grad_norm": 6.2963361740112305,
      "learning_rate": 2.005552302480238e-05,
      "loss": 3.1464,
      "step": 287000
    },
    {
      "epoch": 2.3982618868555194,
      "eval_loss": 3.2935688495635986,
      "eval_runtime": 40.4586,
      "eval_samples_per_second": 110.286,
      "eval_steps_per_second": 13.792,
      "step": 287000
    },
    {
      "epoch": 2.402440043452829,
      "grad_norm": 4.570940971374512,
      "learning_rate": 2.0003241457191855e-05,
      "loss": 3.0698,
      "step": 287500
    },
    {
      "epoch": 2.4066182000501377,
      "grad_norm": 3.961766242980957,
      "learning_rate": 1.995106445271655e-05,
      "loss": 3.11,
      "step": 288000
    },
    {
      "epoch": 2.4066182000501377,
      "eval_loss": 3.3014984130859375,
      "eval_runtime": 40.4731,
      "eval_samples_per_second": 110.246,
      "eval_steps_per_second": 13.787,
      "step": 288000
    },
    {
      "epoch": 2.410796356647447,
      "grad_norm": 3.297183036804199,
      "learning_rate": 1.989878288510603e-05,
      "loss": 3.1153,
      "step": 288500
    },
    {
      "epoch": 2.4149745132447564,
      "grad_norm": 3.5967624187469482,
      "learning_rate": 1.9846501317495504e-05,
      "loss": 3.1049,
      "step": 289000
    },
    {
      "epoch": 2.4149745132447564,
      "eval_loss": 3.2914204597473145,
      "eval_runtime": 40.4566,
      "eval_samples_per_second": 110.291,
      "eval_steps_per_second": 13.793,
      "step": 289000
    },
    {
      "epoch": 2.419152669842066,
      "grad_norm": 5.414641857147217,
      "learning_rate": 1.9794219749884983e-05,
      "loss": 3.1524,
      "step": 289500
    },
    {
      "epoch": 2.423330826439375,
      "grad_norm": 9.401138305664062,
      "learning_rate": 1.974204274540968e-05,
      "loss": 3.1139,
      "step": 290000
    },
    {
      "epoch": 2.423330826439375,
      "eval_loss": 3.2959020137786865,
      "eval_runtime": 40.456,
      "eval_samples_per_second": 110.293,
      "eval_steps_per_second": 13.793,
      "step": 290000
    },
    {
      "epoch": 2.427508983036684,
      "grad_norm": 6.612576961517334,
      "learning_rate": 1.9689761177799157e-05,
      "loss": 3.1063,
      "step": 290500
    },
    {
      "epoch": 2.4316871396339934,
      "grad_norm": 5.513810157775879,
      "learning_rate": 1.9637479610188633e-05,
      "loss": 3.1631,
      "step": 291000
    },
    {
      "epoch": 2.4316871396339934,
      "eval_loss": 3.2913076877593994,
      "eval_runtime": 40.4983,
      "eval_samples_per_second": 110.177,
      "eval_steps_per_second": 13.778,
      "step": 291000
    },
    {
      "epoch": 2.4358652962313028,
      "grad_norm": 4.694360733032227,
      "learning_rate": 1.9585198042578108e-05,
      "loss": 3.1455,
      "step": 291500
    },
    {
      "epoch": 2.440043452828612,
      "grad_norm": 7.717827320098877,
      "learning_rate": 1.9532916474967587e-05,
      "loss": 3.1298,
      "step": 292000
    },
    {
      "epoch": 2.440043452828612,
      "eval_loss": 3.290560483932495,
      "eval_runtime": 40.466,
      "eval_samples_per_second": 110.265,
      "eval_steps_per_second": 13.789,
      "step": 292000
    },
    {
      "epoch": 2.444221609425921,
      "grad_norm": 5.080421447753906,
      "learning_rate": 1.9480634907357062e-05,
      "loss": 3.0543,
      "step": 292500
    },
    {
      "epoch": 2.4483997660232304,
      "grad_norm": 4.552901744842529,
      "learning_rate": 1.942845790288176e-05,
      "loss": 3.1352,
      "step": 293000
    },
    {
      "epoch": 2.4483997660232304,
      "eval_loss": 3.2912685871124268,
      "eval_runtime": 40.4768,
      "eval_samples_per_second": 110.236,
      "eval_steps_per_second": 13.786,
      "step": 293000
    },
    {
      "epoch": 2.4525779226205398,
      "grad_norm": 15.8800630569458,
      "learning_rate": 1.9376176335271236e-05,
      "loss": 3.1242,
      "step": 293500
    },
    {
      "epoch": 2.456756079217849,
      "grad_norm": 4.738433837890625,
      "learning_rate": 1.9323894767660712e-05,
      "loss": 3.0613,
      "step": 294000
    },
    {
      "epoch": 2.456756079217849,
      "eval_loss": 3.296588182449341,
      "eval_runtime": 40.4568,
      "eval_samples_per_second": 110.291,
      "eval_steps_per_second": 13.792,
      "step": 294000
    },
    {
      "epoch": 2.4609342358151585,
      "grad_norm": 7.438526630401611,
      "learning_rate": 1.9271613200050194e-05,
      "loss": 3.1276,
      "step": 294500
    },
    {
      "epoch": 2.4651123924124674,
      "grad_norm": 7.734090328216553,
      "learning_rate": 1.921943619557489e-05,
      "loss": 3.1559,
      "step": 295000
    },
    {
      "epoch": 2.4651123924124674,
      "eval_loss": 3.2927656173706055,
      "eval_runtime": 40.5018,
      "eval_samples_per_second": 110.168,
      "eval_steps_per_second": 13.777,
      "step": 295000
    },
    {
      "epoch": 2.4692905490097767,
      "grad_norm": 9.876496315002441,
      "learning_rate": 1.9167154627964365e-05,
      "loss": 3.119,
      "step": 295500
    },
    {
      "epoch": 2.473468705607086,
      "grad_norm": 11.522672653198242,
      "learning_rate": 1.9114873060353844e-05,
      "loss": 3.2179,
      "step": 296000
    },
    {
      "epoch": 2.473468705607086,
      "eval_loss": 3.295011281967163,
      "eval_runtime": 40.4852,
      "eval_samples_per_second": 110.213,
      "eval_steps_per_second": 13.783,
      "step": 296000
    },
    {
      "epoch": 2.4776468622043955,
      "grad_norm": 4.019059181213379,
      "learning_rate": 1.906259149274332e-05,
      "loss": 3.1037,
      "step": 296500
    },
    {
      "epoch": 2.481825018801705,
      "grad_norm": 3.5689988136291504,
      "learning_rate": 1.9010309925132798e-05,
      "loss": 3.1687,
      "step": 297000
    },
    {
      "epoch": 2.481825018801705,
      "eval_loss": 3.291651964187622,
      "eval_runtime": 40.4797,
      "eval_samples_per_second": 110.228,
      "eval_steps_per_second": 13.785,
      "step": 297000
    },
    {
      "epoch": 2.4860031753990137,
      "grad_norm": 9.938336372375488,
      "learning_rate": 1.8958132920657493e-05,
      "loss": 3.1607,
      "step": 297500
    },
    {
      "epoch": 2.490181331996323,
      "grad_norm": 5.011236667633057,
      "learning_rate": 1.8905851353046972e-05,
      "loss": 3.0375,
      "step": 298000
    },
    {
      "epoch": 2.490181331996323,
      "eval_loss": 3.2923643589019775,
      "eval_runtime": 40.4644,
      "eval_samples_per_second": 110.27,
      "eval_steps_per_second": 13.79,
      "step": 298000
    },
    {
      "epoch": 2.4943594885936324,
      "grad_norm": 4.473458766937256,
      "learning_rate": 1.8853569785436447e-05,
      "loss": 3.1018,
      "step": 298500
    },
    {
      "epoch": 2.498537645190942,
      "grad_norm": 8.452305793762207,
      "learning_rate": 1.8801288217825923e-05,
      "loss": 3.1135,
      "step": 299000
    },
    {
      "epoch": 2.498537645190942,
      "eval_loss": 3.2886199951171875,
      "eval_runtime": 40.4752,
      "eval_samples_per_second": 110.24,
      "eval_steps_per_second": 13.786,
      "step": 299000
    },
    {
      "epoch": 2.502715801788251,
      "grad_norm": 7.06476354598999,
      "learning_rate": 1.874911121335062e-05,
      "loss": 3.133,
      "step": 299500
    },
    {
      "epoch": 2.50689395838556,
      "grad_norm": 5.118690490722656,
      "learning_rate": 1.86968296457401e-05,
      "loss": 3.0742,
      "step": 300000
    },
    {
      "epoch": 2.50689395838556,
      "eval_loss": 3.2926464080810547,
      "eval_runtime": 40.4914,
      "eval_samples_per_second": 110.196,
      "eval_steps_per_second": 13.781,
      "step": 300000
    }
  ],
  "logging_steps": 500,
  "max_steps": 478680,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 2000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 6.357960662471148e+17,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}