{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 18.0,
  "eval_steps": 500,
  "global_step": 9522,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.19,
      "grad_norm": 6.645576477050781,
      "learning_rate": 1.9810964083175805e-05,
      "loss": 0.5588,
      "step": 100
    },
    {
      "epoch": 0.38,
      "grad_norm": 4.662398815155029,
      "learning_rate": 1.962192816635161e-05,
      "loss": 0.3737,
      "step": 200
    },
    {
      "epoch": 0.57,
      "grad_norm": 5.282985687255859,
      "learning_rate": 1.9432892249527412e-05,
      "loss": 0.2863,
      "step": 300
    },
    {
      "epoch": 0.76,
      "grad_norm": 1.5854848623275757,
      "learning_rate": 1.9243856332703215e-05,
      "loss": 0.323,
      "step": 400
    },
    {
      "epoch": 0.95,
      "grad_norm": 19.09914779663086,
      "learning_rate": 1.905482041587902e-05,
      "loss": 0.2956,
      "step": 500
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.20477035641670227,
      "eval_runtime": 4.245,
      "eval_samples_per_second": 91.638,
      "eval_steps_per_second": 6.596,
      "step": 529
    },
    {
      "epoch": 1.13,
      "grad_norm": 26.59105682373047,
      "learning_rate": 1.8865784499054822e-05,
      "loss": 0.2269,
      "step": 600
    },
    {
      "epoch": 1.32,
      "grad_norm": 9.394390106201172,
      "learning_rate": 1.8676748582230626e-05,
      "loss": 0.2393,
      "step": 700
    },
    {
      "epoch": 1.51,
      "grad_norm": 13.779593467712402,
      "learning_rate": 1.848771266540643e-05,
      "loss": 0.2135,
      "step": 800
    },
    {
      "epoch": 1.7,
      "grad_norm": 0.6053704619407654,
      "learning_rate": 1.8298676748582232e-05,
      "loss": 0.2241,
      "step": 900
    },
    {
      "epoch": 1.89,
      "grad_norm": 16.190296173095703,
      "learning_rate": 1.8109640831758036e-05,
      "loss": 0.2355,
      "step": 1000
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.21921110153198242,
      "eval_runtime": 4.2197,
      "eval_samples_per_second": 92.187,
      "eval_steps_per_second": 6.636,
      "step": 1058
    },
    {
      "epoch": 2.08,
      "grad_norm": 0.10688259452581406,
      "learning_rate": 1.792060491493384e-05,
      "loss": 0.1909,
      "step": 1100
    },
    {
      "epoch": 2.27,
      "grad_norm": 29.609710693359375,
      "learning_rate": 1.7731568998109643e-05,
      "loss": 0.1637,
      "step": 1200
    },
    {
      "epoch": 2.46,
      "grad_norm": 0.16163112223148346,
      "learning_rate": 1.7542533081285446e-05,
      "loss": 0.1828,
      "step": 1300
    },
    {
      "epoch": 2.65,
      "grad_norm": 2.127821207046509,
      "learning_rate": 1.735349716446125e-05,
      "loss": 0.1827,
      "step": 1400
    },
    {
      "epoch": 2.84,
      "grad_norm": 0.030160142108798027,
      "learning_rate": 1.7164461247637053e-05,
      "loss": 0.1528,
      "step": 1500
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.2590792179107666,
      "eval_runtime": 4.2371,
      "eval_samples_per_second": 91.808,
      "eval_steps_per_second": 6.608,
      "step": 1587
    },
    {
      "epoch": 3.02,
      "grad_norm": 0.0637928918004036,
      "learning_rate": 1.6975425330812856e-05,
      "loss": 0.1574,
      "step": 1600
    },
    {
      "epoch": 3.21,
      "grad_norm": 0.01419603731483221,
      "learning_rate": 1.678638941398866e-05,
      "loss": 0.0918,
      "step": 1700
    },
    {
      "epoch": 3.4,
      "grad_norm": 9.31396770477295,
      "learning_rate": 1.6597353497164463e-05,
      "loss": 0.1175,
      "step": 1800
    },
    {
      "epoch": 3.59,
      "grad_norm": 0.023506687954068184,
      "learning_rate": 1.6408317580340267e-05,
      "loss": 0.1094,
      "step": 1900
    },
    {
      "epoch": 3.78,
      "grad_norm": 13.4344482421875,
      "learning_rate": 1.621928166351607e-05,
      "loss": 0.1103,
      "step": 2000
    },
    {
      "epoch": 3.97,
      "grad_norm": 30.01926040649414,
      "learning_rate": 1.6030245746691873e-05,
      "loss": 0.1398,
      "step": 2100
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.33959057927131653,
      "eval_runtime": 4.2467,
      "eval_samples_per_second": 91.601,
      "eval_steps_per_second": 6.593,
      "step": 2116
    },
    {
      "epoch": 4.16,
      "grad_norm": 0.12888406217098236,
      "learning_rate": 1.5841209829867677e-05,
      "loss": 0.0536,
      "step": 2200
    },
    {
      "epoch": 4.35,
      "grad_norm": 0.06800971180200577,
      "learning_rate": 1.565217391304348e-05,
      "loss": 0.1039,
      "step": 2300
    },
    {
      "epoch": 4.54,
      "grad_norm": 2.495684862136841,
      "learning_rate": 1.5463137996219284e-05,
      "loss": 0.0886,
      "step": 2400
    },
    {
      "epoch": 4.73,
      "grad_norm": 3.7909159660339355,
      "learning_rate": 1.5274102079395087e-05,
      "loss": 0.0766,
      "step": 2500
    },
    {
      "epoch": 4.91,
      "grad_norm": 0.025901077315211296,
      "learning_rate": 1.5085066162570889e-05,
      "loss": 0.0916,
      "step": 2600
    },
    {
      "epoch": 5.0,
      "eval_loss": 0.3040069341659546,
      "eval_runtime": 4.2484,
      "eval_samples_per_second": 91.564,
      "eval_steps_per_second": 6.591,
      "step": 2645
    },
    {
      "epoch": 5.1,
      "grad_norm": 0.011547481641173363,
      "learning_rate": 1.4896030245746694e-05,
      "loss": 0.0644,
      "step": 2700
    },
    {
      "epoch": 5.29,
      "grad_norm": 0.026049258187413216,
      "learning_rate": 1.4706994328922497e-05,
      "loss": 0.0542,
      "step": 2800
    },
    {
      "epoch": 5.48,
      "grad_norm": 0.35288381576538086,
      "learning_rate": 1.45179584120983e-05,
      "loss": 0.0512,
      "step": 2900
    },
    {
      "epoch": 5.67,
      "grad_norm": 0.023430563509464264,
      "learning_rate": 1.4328922495274103e-05,
      "loss": 0.0491,
      "step": 3000
    },
    {
      "epoch": 5.86,
      "grad_norm": 0.0032355256844311953,
      "learning_rate": 1.4139886578449906e-05,
      "loss": 0.0326,
      "step": 3100
    },
    {
      "epoch": 6.0,
      "eval_loss": 0.3605019450187683,
      "eval_runtime": 4.2462,
      "eval_samples_per_second": 91.611,
      "eval_steps_per_second": 6.594,
      "step": 3174
    },
    {
      "epoch": 6.05,
      "grad_norm": 0.1259104311466217,
      "learning_rate": 1.395085066162571e-05,
      "loss": 0.0665,
      "step": 3200
    },
    {
      "epoch": 6.24,
      "grad_norm": 0.08823427557945251,
      "learning_rate": 1.3761814744801514e-05,
      "loss": 0.0252,
      "step": 3300
    },
    {
      "epoch": 6.43,
      "grad_norm": 101.22339630126953,
      "learning_rate": 1.3572778827977318e-05,
      "loss": 0.0416,
      "step": 3400
    },
    {
      "epoch": 6.62,
      "grad_norm": 0.033835213631391525,
      "learning_rate": 1.338374291115312e-05,
      "loss": 0.0465,
      "step": 3500
    },
    {
      "epoch": 6.81,
      "grad_norm": 0.15451325476169586,
      "learning_rate": 1.3194706994328923e-05,
      "loss": 0.0358,
      "step": 3600
    },
    {
      "epoch": 6.99,
      "grad_norm": 0.0031314522493630648,
      "learning_rate": 1.3005671077504726e-05,
      "loss": 0.0344,
      "step": 3700
    },
    {
      "epoch": 7.0,
      "eval_loss": 0.41286924481391907,
      "eval_runtime": 4.2519,
      "eval_samples_per_second": 91.488,
      "eval_steps_per_second": 6.585,
      "step": 3703
    },
    {
      "epoch": 7.18,
      "grad_norm": 0.0014451502356678247,
      "learning_rate": 1.281663516068053e-05,
      "loss": 0.0148,
      "step": 3800
    },
    {
      "epoch": 7.37,
      "grad_norm": 17.679460525512695,
      "learning_rate": 1.2627599243856335e-05,
      "loss": 0.0309,
      "step": 3900
    },
    {
      "epoch": 7.56,
      "grad_norm": 0.03577155992388725,
      "learning_rate": 1.2438563327032138e-05,
      "loss": 0.0235,
      "step": 4000
    },
    {
      "epoch": 7.75,
      "grad_norm": 28.639484405517578,
      "learning_rate": 1.224952741020794e-05,
      "loss": 0.0296,
      "step": 4100
    },
    {
      "epoch": 7.94,
      "grad_norm": 0.00975144561380148,
      "learning_rate": 1.2060491493383744e-05,
      "loss": 0.0218,
      "step": 4200
    },
    {
      "epoch": 8.0,
      "eval_loss": 0.3556581735610962,
      "eval_runtime": 4.245,
      "eval_samples_per_second": 91.637,
      "eval_steps_per_second": 6.596,
      "step": 4232
    },
    {
      "epoch": 8.13,
      "grad_norm": 0.019052820280194283,
      "learning_rate": 1.1871455576559547e-05,
      "loss": 0.0192,
      "step": 4300
    },
    {
      "epoch": 8.32,
      "grad_norm": 12.919537544250488,
      "learning_rate": 1.168241965973535e-05,
      "loss": 0.0269,
      "step": 4400
    },
    {
      "epoch": 8.51,
      "grad_norm": 0.001653732149861753,
      "learning_rate": 1.1493383742911156e-05,
      "loss": 0.0208,
      "step": 4500
    },
    {
      "epoch": 8.7,
      "grad_norm": 0.002103479579091072,
      "learning_rate": 1.1304347826086957e-05,
      "loss": 0.0263,
      "step": 4600
    },
    {
      "epoch": 8.88,
      "grad_norm": 0.001875244197435677,
      "learning_rate": 1.111531190926276e-05,
      "loss": 0.0086,
      "step": 4700
    },
    {
      "epoch": 9.0,
      "eval_loss": 0.3807547688484192,
      "eval_runtime": 4.2581,
      "eval_samples_per_second": 91.355,
      "eval_steps_per_second": 6.576,
      "step": 4761
    },
    {
      "epoch": 9.07,
      "grad_norm": 0.0033042938448488712,
      "learning_rate": 1.0926275992438564e-05,
      "loss": 0.0361,
      "step": 4800
    },
    {
      "epoch": 9.26,
      "grad_norm": 0.0017488696612417698,
      "learning_rate": 1.0737240075614367e-05,
      "loss": 0.0194,
      "step": 4900
    },
    {
      "epoch": 9.45,
      "grad_norm": 0.50156569480896,
      "learning_rate": 1.0548204158790173e-05,
      "loss": 0.0338,
      "step": 5000
    },
    {
      "epoch": 9.64,
      "grad_norm": 0.0015125612262636423,
      "learning_rate": 1.0359168241965973e-05,
      "loss": 0.0195,
      "step": 5100
    },
    {
      "epoch": 9.83,
      "grad_norm": 0.0021789916791021824,
      "learning_rate": 1.0170132325141778e-05,
      "loss": 0.0114,
      "step": 5200
    },
    {
      "epoch": 10.0,
      "eval_loss": 0.5712518095970154,
      "eval_runtime": 4.2506,
      "eval_samples_per_second": 91.516,
      "eval_steps_per_second": 6.587,
      "step": 5290
    },
    {
      "epoch": 10.02,
      "grad_norm": 0.000900177750736475,
      "learning_rate": 9.981096408317581e-06,
      "loss": 0.0229,
      "step": 5300
    },
    {
      "epoch": 10.21,
      "grad_norm": 0.004578994121402502,
      "learning_rate": 9.792060491493385e-06,
      "loss": 0.0226,
      "step": 5400
    },
    {
      "epoch": 10.4,
      "grad_norm": 0.0009541900944896042,
      "learning_rate": 9.603024574669188e-06,
      "loss": 0.0197,
      "step": 5500
    },
    {
      "epoch": 10.59,
      "grad_norm": 0.000505660951603204,
      "learning_rate": 9.413988657844991e-06,
      "loss": 0.016,
      "step": 5600
    },
    {
      "epoch": 10.78,
      "grad_norm": 0.001706490176729858,
      "learning_rate": 9.224952741020795e-06,
      "loss": 0.0188,
      "step": 5700
    },
    {
      "epoch": 10.96,
      "grad_norm": 0.015518520027399063,
      "learning_rate": 9.035916824196598e-06,
      "loss": 0.0214,
      "step": 5800
    },
    {
      "epoch": 11.0,
      "eval_loss": 0.48657986521720886,
      "eval_runtime": 4.2449,
      "eval_samples_per_second": 91.64,
      "eval_steps_per_second": 6.596,
      "step": 5819
    },
    {
      "epoch": 11.15,
      "grad_norm": 0.006599494256079197,
      "learning_rate": 8.846880907372402e-06,
      "loss": 0.0188,
      "step": 5900
    },
    {
      "epoch": 11.34,
      "grad_norm": 0.002344762207940221,
      "learning_rate": 8.657844990548205e-06,
      "loss": 0.0374,
      "step": 6000
    },
    {
      "epoch": 11.53,
      "grad_norm": 0.0021403091959655285,
      "learning_rate": 8.468809073724009e-06,
      "loss": 0.0331,
      "step": 6100
    },
    {
      "epoch": 11.72,
      "grad_norm": 0.638529360294342,
      "learning_rate": 8.279773156899812e-06,
      "loss": 0.0058,
      "step": 6200
    },
    {
      "epoch": 11.91,
      "grad_norm": 0.0011522163404151797,
      "learning_rate": 8.090737240075615e-06,
      "loss": 0.0169,
      "step": 6300
    },
    {
      "epoch": 12.0,
      "eval_loss": 0.514002799987793,
      "eval_runtime": 4.2661,
      "eval_samples_per_second": 91.185,
      "eval_steps_per_second": 6.563,
      "step": 6348
    },
    {
      "epoch": 12.1,
      "grad_norm": 0.0009117820300161839,
      "learning_rate": 7.901701323251419e-06,
      "loss": 0.0139,
      "step": 6400
    },
    {
      "epoch": 12.29,
      "grad_norm": 0.05164305865764618,
      "learning_rate": 7.712665406427222e-06,
      "loss": 0.0152,
      "step": 6500
    },
    {
      "epoch": 12.48,
      "grad_norm": 22.255159378051758,
      "learning_rate": 7.523629489603025e-06,
      "loss": 0.018,
      "step": 6600
    },
    {
      "epoch": 12.67,
      "grad_norm": 0.0009817625395953655,
      "learning_rate": 7.334593572778829e-06,
      "loss": 0.0092,
      "step": 6700
    },
    {
      "epoch": 12.85,
      "grad_norm": 0.00901391077786684,
      "learning_rate": 7.1455576559546324e-06,
      "loss": 0.0247,
      "step": 6800
    },
    {
      "epoch": 13.0,
      "eval_loss": 0.40500932931900024,
      "eval_runtime": 4.2563,
      "eval_samples_per_second": 91.395,
      "eval_steps_per_second": 6.579,
      "step": 6877
    },
    {
      "epoch": 13.04,
      "grad_norm": 0.002569295698776841,
      "learning_rate": 6.956521739130435e-06,
      "loss": 0.0228,
      "step": 6900
    },
    {
      "epoch": 13.23,
      "grad_norm": 0.0035759450402110815,
      "learning_rate": 6.767485822306239e-06,
      "loss": 0.0166,
      "step": 7000
    },
    {
      "epoch": 13.42,
      "grad_norm": 0.0027633486315608025,
      "learning_rate": 6.578449905482042e-06,
      "loss": 0.0185,
      "step": 7100
    },
    {
      "epoch": 13.61,
      "grad_norm": 0.0013614584458991885,
      "learning_rate": 6.389413988657845e-06,
      "loss": 0.0088,
      "step": 7200
    },
    {
      "epoch": 13.8,
      "grad_norm": 0.002298623789101839,
      "learning_rate": 6.2003780718336495e-06,
      "loss": 0.0152,
      "step": 7300
    },
    {
      "epoch": 13.99,
      "grad_norm": 0.0007343721226789057,
      "learning_rate": 6.011342155009452e-06,
      "loss": 0.0082,
      "step": 7400
    },
    {
      "epoch": 14.0,
      "eval_loss": 0.48581066727638245,
      "eval_runtime": 4.2624,
      "eval_samples_per_second": 91.264,
      "eval_steps_per_second": 6.569,
      "step": 7406
    },
    {
      "epoch": 14.18,
      "grad_norm": 0.009072878398001194,
      "learning_rate": 5.8223062381852555e-06,
      "loss": 0.0116,
      "step": 7500
    },
    {
      "epoch": 14.37,
      "grad_norm": 0.0018677357584238052,
      "learning_rate": 5.63327032136106e-06,
      "loss": 0.0083,
      "step": 7600
    },
    {
      "epoch": 14.56,
      "grad_norm": 0.003356189001351595,
      "learning_rate": 5.444234404536862e-06,
      "loss": 0.0148,
      "step": 7700
    },
    {
      "epoch": 14.74,
      "grad_norm": 0.011417761445045471,
      "learning_rate": 5.255198487712666e-06,
      "loss": 0.014,
      "step": 7800
    },
    {
      "epoch": 14.93,
      "grad_norm": 0.0009720730595290661,
      "learning_rate": 5.066162570888468e-06,
      "loss": 0.0058,
      "step": 7900
    },
    {
      "epoch": 15.0,
      "eval_loss": 0.5540414452552795,
      "eval_runtime": 4.242,
      "eval_samples_per_second": 91.701,
      "eval_steps_per_second": 6.601,
      "step": 7935
    },
    {
      "epoch": 15.12,
      "grad_norm": 0.0007459772750735283,
      "learning_rate": 4.877126654064273e-06,
      "loss": 0.0061,
      "step": 8000
    },
    {
      "epoch": 15.31,
      "grad_norm": 0.0017996943788602948,
      "learning_rate": 4.688090737240076e-06,
      "loss": 0.0133,
      "step": 8100
    },
    {
      "epoch": 15.5,
      "grad_norm": 0.0010721386643126607,
      "learning_rate": 4.4990548204158795e-06,
      "loss": 0.0075,
      "step": 8200
    },
    {
      "epoch": 15.69,
      "grad_norm": 0.0009099834132939577,
      "learning_rate": 4.310018903591683e-06,
      "loss": 0.0026,
      "step": 8300
    },
    {
      "epoch": 15.88,
      "grad_norm": 0.00020547083113342524,
      "learning_rate": 4.120982986767486e-06,
      "loss": 0.0,
      "step": 8400
    },
    {
      "epoch": 16.0,
      "eval_loss": 0.5253579020500183,
      "eval_runtime": 4.2546,
      "eval_samples_per_second": 91.43,
      "eval_steps_per_second": 6.581,
      "step": 8464
    },
    {
      "epoch": 16.07,
      "grad_norm": 0.0004376711731310934,
      "learning_rate": 3.93194706994329e-06,
      "loss": 0.0073,
      "step": 8500
    },
    {
      "epoch": 16.26,
      "grad_norm": 0.0015534511767327785,
      "learning_rate": 3.742911153119093e-06,
      "loss": 0.0053,
      "step": 8600
    },
    {
      "epoch": 16.45,
      "grad_norm": 0.0005960518028587103,
      "learning_rate": 3.5538752362948966e-06,
      "loss": 0.0,
      "step": 8700
    },
    {
      "epoch": 16.64,
      "grad_norm": 0.00040510049439035356,
      "learning_rate": 3.3648393194706996e-06,
      "loss": 0.0,
      "step": 8800
    },
    {
      "epoch": 16.82,
      "grad_norm": 0.0001634809304960072,
      "learning_rate": 3.175803402646503e-06,
      "loss": 0.0,
      "step": 8900
    },
    {
      "epoch": 17.0,
      "eval_loss": 0.574542760848999,
      "eval_runtime": 4.2742,
      "eval_samples_per_second": 91.011,
      "eval_steps_per_second": 6.551,
      "step": 8993
    },
    {
      "epoch": 17.01,
      "grad_norm": 0.0002437944058328867,
      "learning_rate": 2.986767485822307e-06,
      "loss": 0.0104,
      "step": 9000
    },
    {
      "epoch": 17.2,
      "grad_norm": 0.00035620402195490897,
      "learning_rate": 2.79773156899811e-06,
      "loss": 0.0001,
      "step": 9100
    },
    {
      "epoch": 17.39,
      "grad_norm": 0.00021548435324802995,
      "learning_rate": 2.6086956521739132e-06,
      "loss": 0.0037,
      "step": 9200
    },
    {
      "epoch": 17.58,
      "grad_norm": 0.00014057717635296285,
      "learning_rate": 2.4196597353497167e-06,
      "loss": 0.0,
      "step": 9300
    },
    {
      "epoch": 17.77,
      "grad_norm": 0.0010042821522802114,
      "learning_rate": 2.23062381852552e-06,
      "loss": 0.0109,
      "step": 9400
    },
    {
      "epoch": 17.96,
      "grad_norm": 0.00016069511184468865,
      "learning_rate": 2.0415879017013235e-06,
      "loss": 0.0,
      "step": 9500
    },
    {
      "epoch": 18.0,
      "eval_loss": 0.5317645072937012,
      "eval_runtime": 4.2532,
      "eval_samples_per_second": 91.461,
      "eval_steps_per_second": 6.583,
      "step": 9522
    }
  ],
  "logging_steps": 100,
  "max_steps": 10580,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "save_steps": 500,
  "total_flos": 9.72342785672254e+16,
  "train_batch_size": 14,
  "trial_name": null,
  "trial_params": null
}