{
  "best_metric": 4.3424248695373535,
  "best_model_checkpoint": "/media/user/Expansion/pythia-70m-wikipedia-paragraphs-50epochs/checkpoint-11344",
  "epoch": 50.0,
  "eval_steps": 500,
  "global_step": 35450,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.7052186177715092,
      "grad_norm": 84.19266510009766,
      "learning_rate": 4.929478138222849e-05,
      "loss": 7.2253,
      "num_input_tokens_seen": 4096000,
      "step": 500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.1339588143789194,
      "eval_loss": 6.6662068367004395,
      "eval_runtime": 2.3283,
      "eval_samples_per_second": 129.28,
      "eval_steps_per_second": 16.321,
      "num_input_tokens_seen": 5800960,
      "step": 709
    },
    {
      "epoch": 1.4104372355430184,
      "grad_norm": 34.27415466308594,
      "learning_rate": 4.858956276445699e-05,
      "loss": 6.5693,
      "num_input_tokens_seen": 8184832,
      "step": 1000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.13123735479324378,
      "eval_loss": 6.502222061157227,
      "eval_runtime": 2.5457,
      "eval_samples_per_second": 118.237,
      "eval_steps_per_second": 14.927,
      "num_input_tokens_seen": 11601920,
      "step": 1418
    },
    {
      "epoch": 2.1156558533145273,
      "grad_norm": 62.75070571899414,
      "learning_rate": 4.788434414668547e-05,
      "loss": 6.235,
      "num_input_tokens_seen": 12273664,
      "step": 1500
    },
    {
      "epoch": 2.8208744710860367,
      "grad_norm": 83.26922607421875,
      "learning_rate": 4.717912552891397e-05,
      "loss": 5.9786,
      "num_input_tokens_seen": 16369664,
      "step": 2000
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.15382417032829637,
      "eval_loss": 6.0094099044799805,
      "eval_runtime": 2.2872,
      "eval_samples_per_second": 131.601,
      "eval_steps_per_second": 16.614,
      "num_input_tokens_seen": 17402880,
      "step": 2127
    },
    {
      "epoch": 3.5260930888575457,
      "grad_norm": 22.89128875732422,
      "learning_rate": 4.647390691114246e-05,
      "loss": 5.3855,
      "num_input_tokens_seen": 20458496,
      "step": 2500
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.19935828112872375,
      "eval_loss": 5.0559468269348145,
      "eval_runtime": 2.3199,
      "eval_samples_per_second": 129.746,
      "eval_steps_per_second": 16.38,
      "num_input_tokens_seen": 23203840,
      "step": 2836
    },
    {
      "epoch": 4.231311706629055,
      "grad_norm": 45.5040168762207,
      "learning_rate": 4.576868829337095e-05,
      "loss": 5.0498,
      "num_input_tokens_seen": 24547328,
      "step": 3000
    },
    {
      "epoch": 4.9365303244005645,
      "grad_norm": 33.329322814941406,
      "learning_rate": 4.506346967559944e-05,
      "loss": 4.9003,
      "num_input_tokens_seen": 28643328,
      "step": 3500
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.20717517041598063,
      "eval_loss": 4.983646869659424,
      "eval_runtime": 2.3314,
      "eval_samples_per_second": 129.105,
      "eval_steps_per_second": 16.299,
      "num_input_tokens_seen": 29004800,
      "step": 3545
    },
    {
      "epoch": 5.6417489421720735,
      "grad_norm": 23.64671516418457,
      "learning_rate": 4.4358251057827926e-05,
      "loss": 4.726,
      "num_input_tokens_seen": 32732160,
      "step": 4000
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.24000805396154234,
      "eval_loss": 4.574463844299316,
      "eval_runtime": 2.395,
      "eval_samples_per_second": 125.68,
      "eval_steps_per_second": 15.867,
      "num_input_tokens_seen": 34805760,
      "step": 4254
    },
    {
      "epoch": 6.346967559943582,
      "grad_norm": 64.83744049072266,
      "learning_rate": 4.3653032440056416e-05,
      "loss": 4.5269,
      "num_input_tokens_seen": 36820992,
      "step": 4500
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.2230005553336386,
      "eval_loss": 4.76109504699707,
      "eval_runtime": 2.3456,
      "eval_samples_per_second": 128.328,
      "eval_steps_per_second": 16.201,
      "num_input_tokens_seen": 40606720,
      "step": 4963
    },
    {
      "epoch": 7.052186177715091,
      "grad_norm": 24.155548095703125,
      "learning_rate": 4.294781382228491e-05,
      "loss": 4.5745,
      "num_input_tokens_seen": 40909824,
      "step": 5000
    },
    {
      "epoch": 7.757404795486601,
      "grad_norm": 16.52519416809082,
      "learning_rate": 4.22425952045134e-05,
      "loss": 4.5249,
      "num_input_tokens_seen": 45005824,
      "step": 5500
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.23566605937198584,
      "eval_loss": 4.611110687255859,
      "eval_runtime": 2.3827,
      "eval_samples_per_second": 126.326,
      "eval_steps_per_second": 15.948,
      "num_input_tokens_seen": 46407680,
      "step": 5672
    },
    {
      "epoch": 8.46262341325811,
      "grad_norm": 19.4478759765625,
      "learning_rate": 4.153737658674189e-05,
      "loss": 4.3857,
      "num_input_tokens_seen": 49094656,
      "step": 6000
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.24934155616826284,
      "eval_loss": 4.472122669219971,
      "eval_runtime": 2.3575,
      "eval_samples_per_second": 127.679,
      "eval_steps_per_second": 16.119,
      "num_input_tokens_seen": 52208640,
      "step": 6381
    },
    {
      "epoch": 9.16784203102962,
      "grad_norm": 12.421043395996094,
      "learning_rate": 4.083215796897038e-05,
      "loss": 4.2982,
      "num_input_tokens_seen": 53183488,
      "step": 6500
    },
    {
      "epoch": 9.873060648801129,
      "grad_norm": 19.52003288269043,
      "learning_rate": 4.012693935119887e-05,
      "loss": 4.2523,
      "num_input_tokens_seen": 57279488,
      "step": 7000
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.2492506243443978,
      "eval_loss": 4.475858688354492,
      "eval_runtime": 2.3615,
      "eval_samples_per_second": 127.463,
      "eval_steps_per_second": 16.092,
      "num_input_tokens_seen": 58009600,
      "step": 7090
    },
    {
      "epoch": 10.578279266572638,
      "grad_norm": 13.172361373901367,
      "learning_rate": 3.942172073342737e-05,
      "loss": 4.2256,
      "num_input_tokens_seen": 61368320,
      "step": 7500
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.2573857750151824,
      "eval_loss": 4.4382004737854,
      "eval_runtime": 2.3391,
      "eval_samples_per_second": 128.68,
      "eval_steps_per_second": 16.245,
      "num_input_tokens_seen": 63810560,
      "step": 7799
    },
    {
      "epoch": 11.283497884344147,
      "grad_norm": 20.001747131347656,
      "learning_rate": 3.871650211565586e-05,
      "loss": 4.151,
      "num_input_tokens_seen": 65457152,
      "step": 8000
    },
    {
      "epoch": 11.988716502115656,
      "grad_norm": 12.548118591308594,
      "learning_rate": 3.801128349788435e-05,
      "loss": 4.0897,
      "num_input_tokens_seen": 69553152,
      "step": 8500
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.26011697729627214,
      "eval_loss": 4.386884689331055,
      "eval_runtime": 2.3571,
      "eval_samples_per_second": 127.697,
      "eval_steps_per_second": 16.121,
      "num_input_tokens_seen": 69611520,
      "step": 8508
    },
    {
      "epoch": 12.693935119887165,
      "grad_norm": 13.00560188293457,
      "learning_rate": 3.7306064880112836e-05,
      "loss": 4.0155,
      "num_input_tokens_seen": 73641984,
      "step": 9000
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.2617082842139106,
      "eval_loss": 4.3910322189331055,
      "eval_runtime": 2.3687,
      "eval_samples_per_second": 127.077,
      "eval_steps_per_second": 16.043,
      "num_input_tokens_seen": 75412480,
      "step": 9217
    },
    {
      "epoch": 13.399153737658674,
      "grad_norm": 17.632902145385742,
      "learning_rate": 3.6600846262341326e-05,
      "loss": 4.0051,
      "num_input_tokens_seen": 77730816,
      "step": 9500
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.2632378873939264,
      "eval_loss": 4.3866143226623535,
      "eval_runtime": 2.3505,
      "eval_samples_per_second": 128.057,
      "eval_steps_per_second": 16.167,
      "num_input_tokens_seen": 81213440,
      "step": 9926
    },
    {
      "epoch": 14.104372355430183,
      "grad_norm": 8.833736419677734,
      "learning_rate": 3.5895627644569816e-05,
      "loss": 3.9612,
      "num_input_tokens_seen": 81819648,
      "step": 10000
    },
    {
      "epoch": 14.809590973201692,
      "grad_norm": 12.257335662841797,
      "learning_rate": 3.519040902679831e-05,
      "loss": 3.9221,
      "num_input_tokens_seen": 85915648,
      "step": 10500
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.2629553492269171,
      "eval_loss": 4.392822742462158,
      "eval_runtime": 2.3838,
      "eval_samples_per_second": 126.269,
      "eval_steps_per_second": 15.941,
      "num_input_tokens_seen": 87014400,
      "step": 10635
    },
    {
      "epoch": 15.514809590973202,
      "grad_norm": 10.67398738861084,
      "learning_rate": 3.44851904090268e-05,
      "loss": 3.8844,
      "num_input_tokens_seen": 90004480,
      "step": 11000
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.2727662435089292,
      "eval_loss": 4.3424248695373535,
      "eval_runtime": 2.3672,
      "eval_samples_per_second": 127.154,
      "eval_steps_per_second": 16.053,
      "num_input_tokens_seen": 92815360,
      "step": 11344
    },
    {
      "epoch": 16.22002820874471,
      "grad_norm": 9.101444244384766,
      "learning_rate": 3.377997179125529e-05,
      "loss": 3.8528,
      "num_input_tokens_seen": 94093312,
      "step": 11500
    },
    {
      "epoch": 16.92524682651622,
      "grad_norm": 9.585593223571777,
      "learning_rate": 3.307475317348378e-05,
      "loss": 3.8165,
      "num_input_tokens_seen": 98189312,
      "step": 12000
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.2663425596658905,
      "eval_loss": 4.406184196472168,
      "eval_runtime": 2.3686,
      "eval_samples_per_second": 127.081,
      "eval_steps_per_second": 16.043,
      "num_input_tokens_seen": 98616320,
      "step": 12053
    },
    {
      "epoch": 17.63046544428773,
      "grad_norm": 12.549637794494629,
      "learning_rate": 3.236953455571227e-05,
      "loss": 3.7716,
      "num_input_tokens_seen": 102278144,
      "step": 12500
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.26832682196523155,
      "eval_loss": 4.423529624938965,
      "eval_runtime": 2.3586,
      "eval_samples_per_second": 127.62,
      "eval_steps_per_second": 16.111,
      "num_input_tokens_seen": 104417280,
      "step": 12762
    },
    {
      "epoch": 18.33568406205924,
      "grad_norm": 10.797262191772461,
      "learning_rate": 3.166431593794077e-05,
      "loss": 3.7558,
      "num_input_tokens_seen": 106366976,
      "step": 13000
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.26874575786803845,
      "eval_loss": 4.390767574310303,
      "eval_runtime": 2.3418,
      "eval_samples_per_second": 128.535,
      "eval_steps_per_second": 16.227,
      "num_input_tokens_seen": 110218240,
      "step": 13471
    },
    {
      "epoch": 19.040902679830747,
      "grad_norm": 10.686503410339355,
      "learning_rate": 3.095909732016926e-05,
      "loss": 3.7383,
      "num_input_tokens_seen": 110455808,
      "step": 13500
    },
    {
      "epoch": 19.746121297602258,
      "grad_norm": 8.79494857788086,
      "learning_rate": 3.0253878702397743e-05,
      "loss": 3.6958,
      "num_input_tokens_seen": 114551808,
      "step": 14000
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.26914196081487907,
      "eval_loss": 4.424190521240234,
      "eval_runtime": 2.4033,
      "eval_samples_per_second": 125.243,
      "eval_steps_per_second": 15.811,
      "num_input_tokens_seen": 116019200,
      "step": 14180
    },
    {
      "epoch": 20.451339915373765,
      "grad_norm": 8.323198318481445,
      "learning_rate": 2.9548660084626233e-05,
      "loss": 3.684,
      "num_input_tokens_seen": 118640640,
      "step": 14500
    },
    {
      "epoch": 21.0,
      "eval_accuracy": 0.26918742672681156,
      "eval_loss": 4.429287433624268,
      "eval_runtime": 2.3485,
      "eval_samples_per_second": 128.164,
      "eval_steps_per_second": 16.18,
      "num_input_tokens_seen": 121820160,
      "step": 14889
    },
    {
      "epoch": 21.156558533145276,
      "grad_norm": 13.749457359313965,
      "learning_rate": 2.8843441466854726e-05,
      "loss": 3.6566,
      "num_input_tokens_seen": 122729472,
      "step": 15000
    },
    {
      "epoch": 21.861777150916783,
      "grad_norm": 8.61412239074707,
      "learning_rate": 2.813822284908322e-05,
      "loss": 3.6349,
      "num_input_tokens_seen": 126825472,
      "step": 15500
    },
    {
      "epoch": 22.0,
      "eval_accuracy": 0.26987591053607557,
      "eval_loss": 4.465507984161377,
      "eval_runtime": 2.4176,
      "eval_samples_per_second": 124.505,
      "eval_steps_per_second": 15.718,
      "num_input_tokens_seen": 127621120,
      "step": 15598
    },
    {
      "epoch": 22.566995768688294,
      "grad_norm": 8.218836784362793,
      "learning_rate": 2.743300423131171e-05,
      "loss": 3.6091,
      "num_input_tokens_seen": 130914304,
      "step": 16000
    },
    {
      "epoch": 23.0,
      "eval_accuracy": 0.2670992423430533,
      "eval_loss": 4.5008721351623535,
      "eval_runtime": 2.366,
      "eval_samples_per_second": 127.22,
      "eval_steps_per_second": 16.061,
      "num_input_tokens_seen": 133422080,
      "step": 16307
    },
    {
      "epoch": 23.2722143864598,
      "grad_norm": 10.475564956665039,
      "learning_rate": 2.6727785613540195e-05,
      "loss": 3.602,
      "num_input_tokens_seen": 135003136,
      "step": 16500
    },
    {
      "epoch": 23.97743300423131,
      "grad_norm": 9.387394905090332,
      "learning_rate": 2.6022566995768688e-05,
      "loss": 3.5949,
      "num_input_tokens_seen": 139099136,
      "step": 17000
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 0.26471877709687164,
      "eval_loss": 4.58220911026001,
      "eval_runtime": 2.3413,
      "eval_samples_per_second": 128.561,
      "eval_steps_per_second": 16.23,
      "num_input_tokens_seen": 139223040,
      "step": 17016
    },
    {
      "epoch": 24.682651622002822,
      "grad_norm": 9.367217063903809,
      "learning_rate": 2.531734837799718e-05,
      "loss": 3.5476,
      "num_input_tokens_seen": 143187968,
      "step": 17500
    },
    {
      "epoch": 25.0,
      "eval_accuracy": 0.2674499793779614,
      "eval_loss": 4.5724711418151855,
      "eval_runtime": 2.3479,
      "eval_samples_per_second": 128.2,
      "eval_steps_per_second": 16.185,
      "num_input_tokens_seen": 145024000,
      "step": 17725
    },
    {
      "epoch": 25.38787023977433,
      "grad_norm": 9.069378852844238,
      "learning_rate": 2.461212976022567e-05,
      "loss": 3.5392,
      "num_input_tokens_seen": 147276800,
      "step": 18000
    },
    {
      "epoch": 26.0,
      "eval_accuracy": 0.266644583223728,
      "eval_loss": 4.611934185028076,
      "eval_runtime": 2.372,
      "eval_samples_per_second": 126.897,
      "eval_steps_per_second": 16.02,
      "num_input_tokens_seen": 150824960,
      "step": 18434
    },
    {
      "epoch": 26.09308885754584,
      "grad_norm": 7.782838821411133,
      "learning_rate": 2.390691114245416e-05,
      "loss": 3.5423,
      "num_input_tokens_seen": 151365632,
      "step": 18500
    },
    {
      "epoch": 26.798307475317348,
      "grad_norm": 7.378510475158691,
      "learning_rate": 2.3201692524682653e-05,
      "loss": 3.524,
      "num_input_tokens_seen": 155461632,
      "step": 19000
    },
    {
      "epoch": 27.0,
      "eval_accuracy": 0.26156214378269893,
      "eval_loss": 4.708793640136719,
      "eval_runtime": 2.4833,
      "eval_samples_per_second": 121.209,
      "eval_steps_per_second": 15.302,
      "num_input_tokens_seen": 156625920,
      "step": 19143
    },
    {
      "epoch": 27.50352609308886,
      "grad_norm": 8.874760627746582,
      "learning_rate": 2.2496473906911146e-05,
      "loss": 3.4999,
      "num_input_tokens_seen": 159550464,
      "step": 19500
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.26340676078110437,
      "eval_loss": 4.730751037597656,
      "eval_runtime": 2.3629,
      "eval_samples_per_second": 127.388,
      "eval_steps_per_second": 16.082,
      "num_input_tokens_seen": 162426880,
      "step": 19852
    },
    {
      "epoch": 28.208744710860366,
      "grad_norm": 8.382628440856934,
      "learning_rate": 2.1791255289139633e-05,
      "loss": 3.5073,
      "num_input_tokens_seen": 163639296,
      "step": 20000
    },
    {
      "epoch": 28.913963328631876,
      "grad_norm": 7.747939109802246,
      "learning_rate": 2.1086036671368126e-05,
      "loss": 3.4888,
      "num_input_tokens_seen": 167735296,
      "step": 20500
    },
    {
      "epoch": 29.0,
      "eval_accuracy": 0.26255589871493845,
      "eval_loss": 4.776463031768799,
      "eval_runtime": 2.3641,
      "eval_samples_per_second": 127.324,
      "eval_steps_per_second": 16.074,
      "num_input_tokens_seen": 168227840,
      "step": 20561
    },
    {
      "epoch": 29.619181946403383,
      "grad_norm": 10.21652603149414,
      "learning_rate": 2.0380818053596615e-05,
      "loss": 3.4665,
      "num_input_tokens_seen": 171824128,
      "step": 21000
    },
    {
      "epoch": 30.0,
      "eval_accuracy": 0.2627962185351533,
      "eval_loss": 4.861045837402344,
      "eval_runtime": 2.3278,
      "eval_samples_per_second": 129.308,
      "eval_steps_per_second": 16.325,
      "num_input_tokens_seen": 174028800,
      "step": 21270
    },
    {
      "epoch": 30.324400564174894,
      "grad_norm": 7.920512676239014,
      "learning_rate": 1.9675599435825108e-05,
      "loss": 3.4628,
      "num_input_tokens_seen": 175912960,
      "step": 21500
    },
    {
      "epoch": 31.0,
      "eval_accuracy": 0.26106526631657917,
      "eval_loss": 4.910233974456787,
      "eval_runtime": 2.3245,
      "eval_samples_per_second": 129.491,
      "eval_steps_per_second": 16.348,
      "num_input_tokens_seen": 179829760,
      "step": 21979
    },
    {
      "epoch": 31.029619181946405,
      "grad_norm": 9.639541625976562,
      "learning_rate": 1.8970380818053598e-05,
      "loss": 3.4774,
      "num_input_tokens_seen": 180001792,
      "step": 22000
    },
    {
      "epoch": 31.734837799717912,
      "grad_norm": 8.425528526306152,
      "learning_rate": 1.8265162200282088e-05,
      "loss": 3.4478,
      "num_input_tokens_seen": 184097792,
      "step": 22500
    },
    {
      "epoch": 32.0,
      "eval_accuracy": 0.2584022629033882,
      "eval_loss": 4.99893856048584,
      "eval_runtime": 2.4209,
      "eval_samples_per_second": 124.332,
      "eval_steps_per_second": 15.696,
      "num_input_tokens_seen": 185630720,
      "step": 22688
    },
    {
      "epoch": 32.44005641748942,
      "grad_norm": 12.94858455657959,
      "learning_rate": 1.755994358251058e-05,
      "loss": 3.4539,
      "num_input_tokens_seen": 188186624,
      "step": 23000
    },
    {
      "epoch": 33.0,
      "eval_accuracy": 0.2561939186095225,
      "eval_loss": 5.060764789581299,
      "eval_runtime": 2.3544,
      "eval_samples_per_second": 127.845,
      "eval_steps_per_second": 16.14,
      "num_input_tokens_seen": 191431680,
      "step": 23397
    },
    {
      "epoch": 33.14527503526093,
      "grad_norm": 10.570552825927734,
      "learning_rate": 1.685472496473907e-05,
      "loss": 3.4532,
      "num_input_tokens_seen": 192275456,
      "step": 23500
    },
    {
      "epoch": 33.85049365303244,
      "grad_norm": 9.403538703918457,
      "learning_rate": 1.614950634696756e-05,
      "loss": 3.4486,
      "num_input_tokens_seen": 196371456,
      "step": 24000
    },
    {
      "epoch": 34.0,
      "eval_accuracy": 0.2573338139729738,
      "eval_loss": 5.155101299285889,
      "eval_runtime": 2.3548,
      "eval_samples_per_second": 127.825,
      "eval_steps_per_second": 16.137,
      "num_input_tokens_seen": 197232640,
      "step": 24106
    },
    {
      "epoch": 34.55571227080395,
      "grad_norm": 8.42175006866455,
      "learning_rate": 1.5444287729196053e-05,
      "loss": 3.4265,
      "num_input_tokens_seen": 200460288,
      "step": 24500
    },
    {
      "epoch": 35.0,
      "eval_accuracy": 0.2559276182682034,
      "eval_loss": 5.19606876373291,
      "eval_runtime": 2.3467,
      "eval_samples_per_second": 128.266,
      "eval_steps_per_second": 16.193,
      "num_input_tokens_seen": 203033600,
      "step": 24815
    },
    {
      "epoch": 35.26093088857546,
      "grad_norm": 9.095492362976074,
      "learning_rate": 1.4739069111424541e-05,
      "loss": 3.4376,
      "num_input_tokens_seen": 204549120,
      "step": 25000
    },
    {
      "epoch": 35.966149506346966,
      "grad_norm": 9.631875038146973,
      "learning_rate": 1.4033850493653034e-05,
      "loss": 3.4427,
      "num_input_tokens_seen": 208645120,
      "step": 25500
    },
    {
      "epoch": 36.0,
      "eval_accuracy": 0.2559828268755501,
      "eval_loss": 5.287012100219727,
      "eval_runtime": 2.3609,
      "eval_samples_per_second": 127.496,
      "eval_steps_per_second": 16.096,
      "num_input_tokens_seen": 208834560,
      "step": 25524
    },
    {
      "epoch": 36.67136812411848,
      "grad_norm": 9.127022743225098,
      "learning_rate": 1.3328631875881522e-05,
      "loss": 3.4203,
      "num_input_tokens_seen": 212733952,
      "step": 26000
    },
    {
      "epoch": 37.0,
      "eval_accuracy": 0.25666481555453796,
      "eval_loss": 5.381103515625,
      "eval_runtime": 2.408,
      "eval_samples_per_second": 125.0,
      "eval_steps_per_second": 15.781,
      "num_input_tokens_seen": 214635520,
      "step": 26233
    },
    {
      "epoch": 37.37658674188999,
      "grad_norm": 8.888460159301758,
      "learning_rate": 1.2623413258110015e-05,
      "loss": 3.4309,
      "num_input_tokens_seen": 216822784,
      "step": 26500
    },
    {
      "epoch": 38.0,
      "eval_accuracy": 0.2557100314039549,
      "eval_loss": 5.395905017852783,
      "eval_runtime": 2.3536,
      "eval_samples_per_second": 127.891,
      "eval_steps_per_second": 16.146,
      "num_input_tokens_seen": 220436480,
      "step": 26942
    },
    {
      "epoch": 38.081805359661494,
      "grad_norm": 9.691740036010742,
      "learning_rate": 1.1918194640338505e-05,
      "loss": 3.442,
      "num_input_tokens_seen": 220911616,
      "step": 27000
    },
    {
      "epoch": 38.787023977433,
      "grad_norm": 9.186783790588379,
      "learning_rate": 1.1212976022566998e-05,
      "loss": 3.4265,
      "num_input_tokens_seen": 225007616,
      "step": 27500
    },
    {
      "epoch": 39.0,
      "eval_accuracy": 0.25483643638182274,
      "eval_loss": 5.532092571258545,
      "eval_runtime": 2.3449,
      "eval_samples_per_second": 128.361,
      "eval_steps_per_second": 16.205,
      "num_input_tokens_seen": 226237440,
      "step": 27651
    },
    {
      "epoch": 39.492242595204516,
      "grad_norm": 8.784052848815918,
      "learning_rate": 1.0507757404795487e-05,
      "loss": 3.4267,
      "num_input_tokens_seen": 229096448,
      "step": 28000
    },
    {
      "epoch": 40.0,
      "eval_accuracy": 0.2528132033008252,
      "eval_loss": 5.587483882904053,
      "eval_runtime": 2.4783,
      "eval_samples_per_second": 121.456,
      "eval_steps_per_second": 15.333,
      "num_input_tokens_seen": 232038400,
      "step": 28360
    },
    {
      "epoch": 40.19746121297602,
      "grad_norm": 9.762134552001953,
      "learning_rate": 9.802538787023979e-06,
      "loss": 3.4331,
      "num_input_tokens_seen": 233185280,
      "step": 28500
    },
    {
      "epoch": 40.90267983074753,
      "grad_norm": 9.590001106262207,
      "learning_rate": 9.097320169252468e-06,
      "loss": 3.456,
      "num_input_tokens_seen": 237281280,
      "step": 29000
    },
    {
      "epoch": 41.0,
      "eval_accuracy": 0.25092312039048725,
      "eval_loss": 5.615477561950684,
      "eval_runtime": 2.4777,
      "eval_samples_per_second": 121.485,
      "eval_steps_per_second": 15.337,
      "num_input_tokens_seen": 237839360,
      "step": 29069
    },
    {
      "epoch": 41.60789844851904,
      "grad_norm": 9.302165985107422,
      "learning_rate": 8.39210155148096e-06,
      "loss": 3.445,
      "num_input_tokens_seen": 241370112,
      "step": 29500
    },
    {
      "epoch": 42.0,
      "eval_accuracy": 0.2526670628696135,
      "eval_loss": 5.626107215881348,
      "eval_runtime": 2.4409,
      "eval_samples_per_second": 123.313,
      "eval_steps_per_second": 15.568,
      "num_input_tokens_seen": 243640320,
      "step": 29778
    },
    {
      "epoch": 42.31311706629055,
      "grad_norm": 11.168075561523438,
      "learning_rate": 7.686882933709451e-06,
      "loss": 3.4674,
      "num_input_tokens_seen": 245458944,
      "step": 30000
    },
    {
      "epoch": 43.0,
      "eval_accuracy": 0.2473248182175414,
      "eval_loss": 5.813519477844238,
      "eval_runtime": 2.4619,
      "eval_samples_per_second": 122.263,
      "eval_steps_per_second": 15.435,
      "num_input_tokens_seen": 249441280,
      "step": 30487
    },
    {
      "epoch": 43.01833568406206,
      "grad_norm": 13.5941743850708,
      "learning_rate": 6.9816643159379414e-06,
      "loss": 3.4873,
      "num_input_tokens_seen": 249547776,
      "step": 30500
    },
    {
      "epoch": 43.723554301833566,
      "grad_norm": 10.031736373901367,
      "learning_rate": 6.276445698166432e-06,
      "loss": 3.4888,
      "num_input_tokens_seen": 253643776,
      "step": 31000
    },
    {
      "epoch": 44.0,
      "eval_accuracy": 0.24821465106536375,
      "eval_loss": 5.875204086303711,
      "eval_runtime": 2.3416,
      "eval_samples_per_second": 128.546,
      "eval_steps_per_second": 16.228,
      "num_input_tokens_seen": 255242240,
      "step": 31196
    },
    {
      "epoch": 44.42877291960508,
      "grad_norm": 11.021130561828613,
      "learning_rate": 5.5712270803949224e-06,
      "loss": 3.5004,
      "num_input_tokens_seen": 257732608,
      "step": 31500
    },
    {
      "epoch": 45.0,
      "eval_accuracy": 0.24905577043611551,
      "eval_loss": 5.897055625915527,
      "eval_runtime": 2.3591,
      "eval_samples_per_second": 127.589,
      "eval_steps_per_second": 16.108,
      "num_input_tokens_seen": 261043200,
      "step": 31905
    },
    {
      "epoch": 45.13399153737659,
      "grad_norm": 9.840156555175781,
      "learning_rate": 4.866008462623414e-06,
      "loss": 3.4993,
      "num_input_tokens_seen": 261821440,
      "step": 32000
    },
    {
      "epoch": 45.839210155148095,
      "grad_norm": 12.190547943115234,
      "learning_rate": 4.160789844851904e-06,
      "loss": 3.5048,
      "num_input_tokens_seen": 265917440,
      "step": 32500
    },
    {
      "epoch": 46.0,
      "eval_accuracy": 0.2463960145880626,
      "eval_loss": 5.995114803314209,
      "eval_runtime": 2.6148,
      "eval_samples_per_second": 115.115,
      "eval_steps_per_second": 14.533,
      "num_input_tokens_seen": 266844160,
      "step": 32614
    },
    {
      "epoch": 46.5444287729196,
      "grad_norm": 10.28553581237793,
      "learning_rate": 3.455571227080395e-06,
      "loss": 3.5265,
      "num_input_tokens_seen": 270006272,
      "step": 33000
    },
    {
      "epoch": 47.0,
      "eval_accuracy": 0.24732806578267943,
      "eval_loss": 5.983969688415527,
      "eval_runtime": 2.5908,
      "eval_samples_per_second": 116.181,
      "eval_steps_per_second": 14.667,
      "num_input_tokens_seen": 272645120,
      "step": 33323
    },
    {
      "epoch": 47.249647390691116,
      "grad_norm": 9.88039779663086,
      "learning_rate": 2.7503526093088857e-06,
      "loss": 3.5267,
      "num_input_tokens_seen": 274095104,
      "step": 33500
    },
    {
      "epoch": 47.95486600846262,
      "grad_norm": 18.002471923828125,
      "learning_rate": 2.0451339915373766e-06,
      "loss": 3.5454,
      "num_input_tokens_seen": 278191104,
      "step": 34000
    },
    {
      "epoch": 48.0,
      "eval_accuracy": 0.24613296181188155,
      "eval_loss": 6.077630043029785,
      "eval_runtime": 2.56,
      "eval_samples_per_second": 117.579,
      "eval_steps_per_second": 14.844,
      "num_input_tokens_seen": 278446080,
      "step": 34032
    },
    {
      "epoch": 48.66008462623413,
      "grad_norm": 9.459271430969238,
      "learning_rate": 1.3399153737658675e-06,
      "loss": 3.5405,
      "num_input_tokens_seen": 282279936,
      "step": 34500
    },
    {
      "epoch": 49.0,
      "eval_accuracy": 0.2454509731328936,
      "eval_loss": 6.075345993041992,
      "eval_runtime": 2.392,
      "eval_samples_per_second": 125.835,
      "eval_steps_per_second": 15.886,
      "num_input_tokens_seen": 284247040,
      "step": 34741
    },
    {
      "epoch": 49.365303244005645,
      "grad_norm": 10.473950386047363,
      "learning_rate": 6.346967559943583e-07,
      "loss": 3.5448,
      "num_input_tokens_seen": 286368768,
      "step": 35000
    },
    {
      "epoch": 50.0,
      "eval_accuracy": 0.2452431289640592,
      "eval_loss": 6.103971004486084,
      "eval_runtime": 2.5249,
      "eval_samples_per_second": 119.214,
      "eval_steps_per_second": 15.05,
      "num_input_tokens_seen": 290048000,
      "step": 35450
    },
    {
      "epoch": 50.0,
      "num_input_tokens_seen": 290048000,
      "step": 35450,
      "total_flos": 7.7740363481088e+16,
      "train_loss": 3.9178111260633375,
      "train_runtime": 4788.9263,
      "train_samples_per_second": 59.147,
      "train_steps_per_second": 7.402,
      "train_tokens_per_second": 60566.395
    }
  ],
  "logging_steps": 500,
  "max_steps": 35450,
  "num_input_tokens_seen": 290048000,
  "num_train_epochs": 50,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 7.7740363481088e+16,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}