{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 20.0,
  "eval_steps": 500,
  "global_step": 371940,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.05,
      "learning_rate": 3.125e-06,
      "loss": 7.5485,
      "step": 1000
    },
    {
      "epoch": 0.11,
      "learning_rate": 6.25e-06,
      "loss": 5.8151,
      "step": 2000
    },
    {
      "epoch": 0.16,
      "learning_rate": 9.375000000000001e-06,
      "loss": 5.3768,
      "step": 3000
    },
    {
      "epoch": 0.22,
      "learning_rate": 1.25e-05,
      "loss": 5.1553,
      "step": 4000
    },
    {
      "epoch": 0.27,
      "learning_rate": 1.5625e-05,
      "loss": 5.0099,
      "step": 5000
    },
    {
      "epoch": 0.32,
      "learning_rate": 1.8750000000000002e-05,
      "loss": 4.8767,
      "step": 6000
    },
    {
      "epoch": 0.38,
      "learning_rate": 2.1875e-05,
      "loss": 4.7604,
      "step": 7000
    },
    {
      "epoch": 0.43,
      "learning_rate": 2.5e-05,
      "loss": 4.6562,
      "step": 8000
    },
    {
      "epoch": 0.48,
      "learning_rate": 2.8125000000000003e-05,
      "loss": 4.5602,
      "step": 9000
    },
    {
      "epoch": 0.54,
      "learning_rate": 3.125e-05,
      "loss": 4.4821,
      "step": 10000
    },
    {
      "epoch": 0.59,
      "learning_rate": 3.4375e-05,
      "loss": 4.4129,
      "step": 11000
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 4.3503,
      "step": 12000
    },
    {
      "epoch": 0.7,
      "learning_rate": 4.061875e-05,
      "loss": 4.2942,
      "step": 13000
    },
    {
      "epoch": 0.75,
      "learning_rate": 4.374375e-05,
      "loss": 4.2367,
      "step": 14000
    },
    {
      "epoch": 0.81,
      "learning_rate": 4.686875e-05,
      "loss": 4.1833,
      "step": 15000
    },
    {
      "epoch": 0.86,
      "learning_rate": 4.999375e-05,
      "loss": 4.1412,
      "step": 16000
    },
    {
      "epoch": 0.91,
      "learning_rate": 5.3115625000000005e-05,
      "loss": 4.0988,
      "step": 17000
    },
    {
      "epoch": 0.97,
      "learning_rate": 5.62375e-05,
      "loss": 4.0532,
      "step": 18000
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.30846132523601394,
      "eval_loss": 4.257880210876465,
      "eval_runtime": 153.3032,
      "eval_samples_per_second": 377.8,
      "eval_steps_per_second": 5.903,
      "step": 18597
    },
    {
      "epoch": 1.02,
      "learning_rate": 5.93625e-05,
      "loss": 4.0065,
      "step": 19000
    },
    {
      "epoch": 1.08,
      "learning_rate": 6.24875e-05,
      "loss": 3.9678,
      "step": 20000
    },
    {
      "epoch": 1.13,
      "learning_rate": 6.56125e-05,
      "loss": 3.9301,
      "step": 21000
    },
    {
      "epoch": 1.18,
      "learning_rate": 6.8734375e-05,
      "loss": 3.8874,
      "step": 22000
    },
    {
      "epoch": 1.24,
      "learning_rate": 7.185625e-05,
      "loss": 3.8598,
      "step": 23000
    },
    {
      "epoch": 1.29,
      "learning_rate": 7.498125e-05,
      "loss": 3.8228,
      "step": 24000
    },
    {
      "epoch": 1.34,
      "learning_rate": 7.8103125e-05,
      "loss": 3.7974,
      "step": 25000
    },
    {
      "epoch": 1.4,
      "learning_rate": 8.1228125e-05,
      "loss": 3.7705,
      "step": 26000
    },
    {
      "epoch": 1.45,
      "learning_rate": 8.435e-05,
      "loss": 3.7446,
      "step": 27000
    },
    {
      "epoch": 1.51,
      "learning_rate": 8.747500000000001e-05,
      "loss": 3.7237,
      "step": 28000
    },
    {
      "epoch": 1.56,
      "learning_rate": 9.06e-05,
      "loss": 3.7083,
      "step": 29000
    },
    {
      "epoch": 1.61,
      "learning_rate": 9.3721875e-05,
      "loss": 3.6864,
      "step": 30000
    },
    {
      "epoch": 1.67,
      "learning_rate": 9.6846875e-05,
      "loss": 3.6687,
      "step": 31000
    },
    {
      "epoch": 1.72,
      "learning_rate": 9.996875e-05,
      "loss": 3.6513,
      "step": 32000
    },
    {
      "epoch": 1.77,
      "learning_rate": 9.970877213625934e-05,
      "loss": 3.6299,
      "step": 33000
    },
    {
      "epoch": 1.83,
      "learning_rate": 9.941460257692534e-05,
      "loss": 3.6153,
      "step": 34000
    },
    {
      "epoch": 1.88,
      "learning_rate": 9.912072718715068e-05,
      "loss": 3.5913,
      "step": 35000
    },
    {
      "epoch": 1.94,
      "learning_rate": 9.882655762781668e-05,
      "loss": 3.5765,
      "step": 36000
    },
    {
      "epoch": 1.99,
      "learning_rate": 9.8532682238042e-05,
      "loss": 3.566,
      "step": 37000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.3620144970824847,
      "eval_loss": 3.760495662689209,
      "eval_runtime": 155.0908,
      "eval_samples_per_second": 373.446,
      "eval_steps_per_second": 5.835,
      "step": 37194
    },
    {
      "epoch": 2.04,
      "learning_rate": 9.823851267870802e-05,
      "loss": 3.5234,
      "step": 38000
    },
    {
      "epoch": 2.1,
      "learning_rate": 9.794463728893335e-05,
      "loss": 3.5119,
      "step": 39000
    },
    {
      "epoch": 2.15,
      "learning_rate": 9.765046772959934e-05,
      "loss": 3.5002,
      "step": 40000
    },
    {
      "epoch": 2.2,
      "learning_rate": 9.735629817026534e-05,
      "loss": 3.4914,
      "step": 41000
    },
    {
      "epoch": 2.26,
      "learning_rate": 9.706212861093135e-05,
      "loss": 3.4834,
      "step": 42000
    },
    {
      "epoch": 2.31,
      "learning_rate": 9.676825322115668e-05,
      "loss": 3.4774,
      "step": 43000
    },
    {
      "epoch": 2.37,
      "learning_rate": 9.647437783138201e-05,
      "loss": 3.4655,
      "step": 44000
    },
    {
      "epoch": 2.42,
      "learning_rate": 9.618020827204801e-05,
      "loss": 3.4555,
      "step": 45000
    },
    {
      "epoch": 2.47,
      "learning_rate": 9.588603871271401e-05,
      "loss": 3.4512,
      "step": 46000
    },
    {
      "epoch": 2.53,
      "learning_rate": 9.559186915338e-05,
      "loss": 3.4414,
      "step": 47000
    },
    {
      "epoch": 2.58,
      "learning_rate": 9.529799376360535e-05,
      "loss": 3.4313,
      "step": 48000
    },
    {
      "epoch": 2.63,
      "learning_rate": 9.500382420427135e-05,
      "loss": 3.4242,
      "step": 49000
    },
    {
      "epoch": 2.69,
      "learning_rate": 9.470965464493736e-05,
      "loss": 3.4185,
      "step": 50000
    },
    {
      "epoch": 2.74,
      "learning_rate": 9.441577925516269e-05,
      "loss": 3.416,
      "step": 51000
    },
    {
      "epoch": 2.8,
      "learning_rate": 9.412160969582868e-05,
      "loss": 3.4054,
      "step": 52000
    },
    {
      "epoch": 2.85,
      "learning_rate": 9.382744013649468e-05,
      "loss": 3.3961,
      "step": 53000
    },
    {
      "epoch": 2.9,
      "learning_rate": 9.353356474672001e-05,
      "loss": 3.3925,
      "step": 54000
    },
    {
      "epoch": 2.96,
      "learning_rate": 9.323939518738601e-05,
      "loss": 3.3886,
      "step": 55000
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.3805629265473003,
      "eval_loss": 3.596226453781128,
      "eval_runtime": 155.0494,
      "eval_samples_per_second": 373.545,
      "eval_steps_per_second": 5.837,
      "step": 55791
    },
    {
      "epoch": 3.01,
      "learning_rate": 9.294551979761135e-05,
      "loss": 3.3707,
      "step": 56000
    },
    {
      "epoch": 3.07,
      "learning_rate": 9.265164440783668e-05,
      "loss": 3.3276,
      "step": 57000
    },
    {
      "epoch": 3.12,
      "learning_rate": 9.235747484850268e-05,
      "loss": 3.3305,
      "step": 58000
    },
    {
      "epoch": 3.17,
      "learning_rate": 9.206330528916868e-05,
      "loss": 3.3319,
      "step": 59000
    },
    {
      "epoch": 3.23,
      "learning_rate": 9.176913572983467e-05,
      "loss": 3.3226,
      "step": 60000
    },
    {
      "epoch": 3.28,
      "learning_rate": 9.147526034006e-05,
      "loss": 3.3256,
      "step": 61000
    },
    {
      "epoch": 3.33,
      "learning_rate": 9.118109078072601e-05,
      "loss": 3.3253,
      "step": 62000
    },
    {
      "epoch": 3.39,
      "learning_rate": 9.088721539095136e-05,
      "loss": 3.3165,
      "step": 63000
    },
    {
      "epoch": 3.44,
      "learning_rate": 9.059304583161736e-05,
      "loss": 3.3167,
      "step": 64000
    },
    {
      "epoch": 3.5,
      "learning_rate": 9.029917044184268e-05,
      "loss": 3.3141,
      "step": 65000
    },
    {
      "epoch": 3.55,
      "learning_rate": 9.000529505206801e-05,
      "loss": 3.3114,
      "step": 66000
    },
    {
      "epoch": 3.6,
      "learning_rate": 8.971112549273401e-05,
      "loss": 3.3068,
      "step": 67000
    },
    {
      "epoch": 3.66,
      "learning_rate": 8.941695593340001e-05,
      "loss": 3.3044,
      "step": 68000
    },
    {
      "epoch": 3.71,
      "learning_rate": 8.912278637406601e-05,
      "loss": 3.2989,
      "step": 69000
    },
    {
      "epoch": 3.76,
      "learning_rate": 8.882891098429135e-05,
      "loss": 3.3017,
      "step": 70000
    },
    {
      "epoch": 3.82,
      "learning_rate": 8.853503559451668e-05,
      "loss": 3.2947,
      "step": 71000
    },
    {
      "epoch": 3.87,
      "learning_rate": 8.824086603518268e-05,
      "loss": 3.2908,
      "step": 72000
    },
    {
      "epoch": 3.93,
      "learning_rate": 8.794669647584869e-05,
      "loss": 3.29,
      "step": 73000
    },
    {
      "epoch": 3.98,
      "learning_rate": 8.765252691651469e-05,
      "loss": 3.2899,
      "step": 74000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.38936209692947077,
      "eval_loss": 3.5175118446350098,
      "eval_runtime": 154.67,
      "eval_samples_per_second": 374.462,
      "eval_steps_per_second": 5.851,
      "step": 74388
    },
    {
      "epoch": 4.03,
      "learning_rate": 8.735865152674002e-05,
      "loss": 3.2502,
      "step": 75000
    },
    {
      "epoch": 4.09,
      "learning_rate": 8.706448196740601e-05,
      "loss": 3.2342,
      "step": 76000
    },
    {
      "epoch": 4.14,
      "learning_rate": 8.677031240807203e-05,
      "loss": 3.2327,
      "step": 77000
    },
    {
      "epoch": 4.19,
      "learning_rate": 8.647614284873802e-05,
      "loss": 3.2342,
      "step": 78000
    },
    {
      "epoch": 4.25,
      "learning_rate": 8.618226745896335e-05,
      "loss": 3.2326,
      "step": 79000
    },
    {
      "epoch": 4.3,
      "learning_rate": 8.588809789962935e-05,
      "loss": 3.2345,
      "step": 80000
    },
    {
      "epoch": 4.36,
      "learning_rate": 8.559392834029535e-05,
      "loss": 3.2361,
      "step": 81000
    },
    {
      "epoch": 4.41,
      "learning_rate": 8.529975878096135e-05,
      "loss": 3.2337,
      "step": 82000
    },
    {
      "epoch": 4.46,
      "learning_rate": 8.500588339118668e-05,
      "loss": 3.2295,
      "step": 83000
    },
    {
      "epoch": 4.52,
      "learning_rate": 8.471171383185267e-05,
      "loss": 3.2325,
      "step": 84000
    },
    {
      "epoch": 4.57,
      "learning_rate": 8.441754427251868e-05,
      "loss": 3.2275,
      "step": 85000
    },
    {
      "epoch": 4.62,
      "learning_rate": 8.412337471318468e-05,
      "loss": 3.2247,
      "step": 86000
    },
    {
      "epoch": 4.68,
      "learning_rate": 8.382949932341003e-05,
      "loss": 3.2267,
      "step": 87000
    },
    {
      "epoch": 4.73,
      "learning_rate": 8.353532976407602e-05,
      "loss": 3.2255,
      "step": 88000
    },
    {
      "epoch": 4.79,
      "learning_rate": 8.324145437430135e-05,
      "loss": 3.2242,
      "step": 89000
    },
    {
      "epoch": 4.84,
      "learning_rate": 8.294728481496735e-05,
      "loss": 3.2247,
      "step": 90000
    },
    {
      "epoch": 4.89,
      "learning_rate": 8.265340942519268e-05,
      "loss": 3.2213,
      "step": 91000
    },
    {
      "epoch": 4.95,
      "learning_rate": 8.235923986585868e-05,
      "loss": 3.2214,
      "step": 92000
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.3939114645245801,
      "eval_loss": 3.4618098735809326,
      "eval_runtime": 153.3377,
      "eval_samples_per_second": 377.715,
      "eval_steps_per_second": 5.902,
      "step": 92985
    },
    {
      "epoch": 5.0,
      "learning_rate": 8.206536447608402e-05,
      "loss": 3.2176,
      "step": 93000
    },
    {
      "epoch": 5.05,
      "learning_rate": 8.177119491675002e-05,
      "loss": 3.1605,
      "step": 94000
    },
    {
      "epoch": 5.11,
      "learning_rate": 8.147702535741602e-05,
      "loss": 3.1621,
      "step": 95000
    },
    {
      "epoch": 5.16,
      "learning_rate": 8.118314996764134e-05,
      "loss": 3.1684,
      "step": 96000
    },
    {
      "epoch": 5.22,
      "learning_rate": 8.088927457786669e-05,
      "loss": 3.1668,
      "step": 97000
    },
    {
      "epoch": 5.27,
      "learning_rate": 8.059510501853269e-05,
      "loss": 3.1728,
      "step": 98000
    },
    {
      "epoch": 5.32,
      "learning_rate": 8.030122962875803e-05,
      "loss": 3.1745,
      "step": 99000
    },
    {
      "epoch": 5.38,
      "learning_rate": 8.000706006942403e-05,
      "loss": 3.1689,
      "step": 100000
    },
    {
      "epoch": 5.43,
      "learning_rate": 7.971289051009002e-05,
      "loss": 3.1721,
      "step": 101000
    },
    {
      "epoch": 5.48,
      "learning_rate": 7.941872095075602e-05,
      "loss": 3.168,
      "step": 102000
    },
    {
      "epoch": 5.54,
      "learning_rate": 7.912484556098135e-05,
      "loss": 3.172,
      "step": 103000
    },
    {
      "epoch": 5.59,
      "learning_rate": 7.883067600164735e-05,
      "loss": 3.1735,
      "step": 104000
    },
    {
      "epoch": 5.65,
      "learning_rate": 7.853650644231335e-05,
      "loss": 3.1725,
      "step": 105000
    },
    {
      "epoch": 5.7,
      "learning_rate": 7.824263105253869e-05,
      "loss": 3.1731,
      "step": 106000
    },
    {
      "epoch": 5.75,
      "learning_rate": 7.794846149320469e-05,
      "loss": 3.1738,
      "step": 107000
    },
    {
      "epoch": 5.81,
      "learning_rate": 7.765458610343002e-05,
      "loss": 3.1708,
      "step": 108000
    },
    {
      "epoch": 5.86,
      "learning_rate": 7.736041654409603e-05,
      "loss": 3.1697,
      "step": 109000
    },
    {
      "epoch": 5.91,
      "learning_rate": 7.706624698476203e-05,
      "loss": 3.169,
      "step": 110000
    },
    {
      "epoch": 5.97,
      "learning_rate": 7.677207742542802e-05,
      "loss": 3.1702,
      "step": 111000
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.3978669427043171,
      "eval_loss": 3.425184726715088,
      "eval_runtime": 153.9864,
      "eval_samples_per_second": 376.124,
      "eval_steps_per_second": 5.877,
      "step": 111582
    },
    {
      "epoch": 6.02,
      "learning_rate": 7.647820203565335e-05,
      "loss": 3.1396,
      "step": 112000
    },
    {
      "epoch": 6.08,
      "learning_rate": 7.618403247631935e-05,
      "loss": 3.1105,
      "step": 113000
    },
    {
      "epoch": 6.13,
      "learning_rate": 7.589015708654469e-05,
      "loss": 3.115,
      "step": 114000
    },
    {
      "epoch": 6.18,
      "learning_rate": 7.559598752721069e-05,
      "loss": 3.1222,
      "step": 115000
    },
    {
      "epoch": 6.24,
      "learning_rate": 7.530181796787669e-05,
      "loss": 3.1212,
      "step": 116000
    },
    {
      "epoch": 6.29,
      "learning_rate": 7.500794257810202e-05,
      "loss": 3.1252,
      "step": 117000
    },
    {
      "epoch": 6.35,
      "learning_rate": 7.471377301876802e-05,
      "loss": 3.1216,
      "step": 118000
    },
    {
      "epoch": 6.4,
      "learning_rate": 7.441960345943401e-05,
      "loss": 3.1252,
      "step": 119000
    },
    {
      "epoch": 6.45,
      "learning_rate": 7.412543390010001e-05,
      "loss": 3.1233,
      "step": 120000
    },
    {
      "epoch": 6.51,
      "learning_rate": 7.383126434076602e-05,
      "loss": 3.1243,
      "step": 121000
    },
    {
      "epoch": 6.56,
      "learning_rate": 7.353738895099135e-05,
      "loss": 3.1264,
      "step": 122000
    },
    {
      "epoch": 6.61,
      "learning_rate": 7.324321939165736e-05,
      "loss": 3.1294,
      "step": 123000
    },
    {
      "epoch": 6.67,
      "learning_rate": 7.29493440018827e-05,
      "loss": 3.1263,
      "step": 124000
    },
    {
      "epoch": 6.72,
      "learning_rate": 7.265517444254869e-05,
      "loss": 3.1298,
      "step": 125000
    },
    {
      "epoch": 6.78,
      "learning_rate": 7.236100488321469e-05,
      "loss": 3.1298,
      "step": 126000
    },
    {
      "epoch": 6.83,
      "learning_rate": 7.206712949344002e-05,
      "loss": 3.1292,
      "step": 127000
    },
    {
      "epoch": 6.88,
      "learning_rate": 7.177325410366536e-05,
      "loss": 3.1258,
      "step": 128000
    },
    {
      "epoch": 6.94,
      "learning_rate": 7.147908454433136e-05,
      "loss": 3.1263,
      "step": 129000
    },
    {
      "epoch": 6.99,
      "learning_rate": 7.118491498499736e-05,
      "loss": 3.1294,
      "step": 130000
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.39953184852917645,
      "eval_loss": 3.425506114959717,
      "eval_runtime": 153.9146,
      "eval_samples_per_second": 376.3,
      "eval_steps_per_second": 5.88,
      "step": 130179
    },
    {
      "epoch": 7.04,
      "learning_rate": 7.089074542566336e-05,
      "loss": 3.0759,
      "step": 131000
    },
    {
      "epoch": 7.1,
      "learning_rate": 7.059687003588868e-05,
      "loss": 3.0733,
      "step": 132000
    },
    {
      "epoch": 7.15,
      "learning_rate": 7.03027004765547e-05,
      "loss": 3.0779,
      "step": 133000
    },
    {
      "epoch": 7.21,
      "learning_rate": 7.000911925633936e-05,
      "loss": 3.0819,
      "step": 134000
    },
    {
      "epoch": 7.26,
      "learning_rate": 6.971494969700537e-05,
      "loss": 3.0831,
      "step": 135000
    },
    {
      "epoch": 7.31,
      "learning_rate": 6.942078013767136e-05,
      "loss": 3.0837,
      "step": 136000
    },
    {
      "epoch": 7.37,
      "learning_rate": 6.912661057833736e-05,
      "loss": 3.0831,
      "step": 137000
    },
    {
      "epoch": 7.42,
      "learning_rate": 6.883244101900336e-05,
      "loss": 3.0849,
      "step": 138000
    },
    {
      "epoch": 7.47,
      "learning_rate": 6.853856562922869e-05,
      "loss": 3.0915,
      "step": 139000
    },
    {
      "epoch": 7.53,
      "learning_rate": 6.824439606989469e-05,
      "loss": 3.0902,
      "step": 140000
    },
    {
      "epoch": 7.58,
      "learning_rate": 6.795022651056069e-05,
      "loss": 3.0879,
      "step": 141000
    },
    {
      "epoch": 7.64,
      "learning_rate": 6.765635112078601e-05,
      "loss": 3.0852,
      "step": 142000
    },
    {
      "epoch": 7.69,
      "learning_rate": 6.736247573101136e-05,
      "loss": 3.0939,
      "step": 143000
    },
    {
      "epoch": 7.74,
      "learning_rate": 6.706830617167736e-05,
      "loss": 3.0892,
      "step": 144000
    },
    {
      "epoch": 7.8,
      "learning_rate": 6.67744307819027e-05,
      "loss": 3.0892,
      "step": 145000
    },
    {
      "epoch": 7.85,
      "learning_rate": 6.64802612225687e-05,
      "loss": 3.0942,
      "step": 146000
    },
    {
      "epoch": 7.9,
      "learning_rate": 6.61860916632347e-05,
      "loss": 3.0951,
      "step": 147000
    },
    {
      "epoch": 7.96,
      "learning_rate": 6.589192210390069e-05,
      "loss": 3.0899,
      "step": 148000
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.40102234972481554,
      "eval_loss": 3.418954372406006,
      "eval_runtime": 153.2262,
      "eval_samples_per_second": 377.99,
      "eval_steps_per_second": 5.906,
      "step": 148776
    },
    {
      "epoch": 8.01,
      "learning_rate": 6.559775254456669e-05,
      "loss": 3.0764,
      "step": 149000
    },
    {
      "epoch": 8.07,
      "learning_rate": 6.530387715479203e-05,
      "loss": 3.0342,
      "step": 150000
    },
    {
      "epoch": 8.12,
      "learning_rate": 6.500970759545803e-05,
      "loss": 3.0423,
      "step": 151000
    },
    {
      "epoch": 8.17,
      "learning_rate": 6.471583220568336e-05,
      "loss": 3.0441,
      "step": 152000
    },
    {
      "epoch": 8.23,
      "learning_rate": 6.442195681590869e-05,
      "loss": 3.0435,
      "step": 153000
    },
    {
      "epoch": 8.28,
      "learning_rate": 6.412778725657469e-05,
      "loss": 3.0519,
      "step": 154000
    },
    {
      "epoch": 8.33,
      "learning_rate": 6.383391186680003e-05,
      "loss": 3.05,
      "step": 155000
    },
    {
      "epoch": 8.39,
      "learning_rate": 6.353974230746603e-05,
      "loss": 3.0542,
      "step": 156000
    },
    {
      "epoch": 8.44,
      "learning_rate": 6.324557274813202e-05,
      "loss": 3.0513,
      "step": 157000
    },
    {
      "epoch": 8.5,
      "learning_rate": 6.295169735835737e-05,
      "loss": 3.0561,
      "step": 158000
    },
    {
      "epoch": 8.55,
      "learning_rate": 6.265752779902337e-05,
      "loss": 3.0625,
      "step": 159000
    },
    {
      "epoch": 8.6,
      "learning_rate": 6.236335823968936e-05,
      "loss": 3.057,
      "step": 160000
    },
    {
      "epoch": 8.66,
      "learning_rate": 6.206918868035536e-05,
      "loss": 3.0619,
      "step": 161000
    },
    {
      "epoch": 8.71,
      "learning_rate": 6.177531329058069e-05,
      "loss": 3.0584,
      "step": 162000
    },
    {
      "epoch": 8.76,
      "learning_rate": 6.148114373124669e-05,
      "loss": 3.0562,
      "step": 163000
    },
    {
      "epoch": 8.82,
      "learning_rate": 6.118726834147203e-05,
      "loss": 3.0594,
      "step": 164000
    },
    {
      "epoch": 8.87,
      "learning_rate": 6.089309878213802e-05,
      "loss": 3.0593,
      "step": 165000
    },
    {
      "epoch": 8.93,
      "learning_rate": 6.059922339236336e-05,
      "loss": 3.0619,
      "step": 166000
    },
    {
      "epoch": 8.98,
      "learning_rate": 6.030534800258869e-05,
      "loss": 3.0639,
      "step": 167000
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.40267046003453427,
      "eval_loss": 3.40413498878479,
      "eval_runtime": 154.0554,
      "eval_samples_per_second": 375.956,
      "eval_steps_per_second": 5.875,
      "step": 167373
    },
    {
      "epoch": 9.03,
      "learning_rate": 6.001117844325469e-05,
      "loss": 3.0252,
      "step": 168000
    },
    {
      "epoch": 9.09,
      "learning_rate": 5.97170088839207e-05,
      "loss": 3.0072,
      "step": 169000
    },
    {
      "epoch": 9.14,
      "learning_rate": 5.94228393245867e-05,
      "loss": 3.0135,
      "step": 170000
    },
    {
      "epoch": 9.2,
      "learning_rate": 5.912896393481203e-05,
      "loss": 3.0186,
      "step": 171000
    },
    {
      "epoch": 9.25,
      "learning_rate": 5.8835088545037366e-05,
      "loss": 3.0199,
      "step": 172000
    },
    {
      "epoch": 9.3,
      "learning_rate": 5.8540918985703364e-05,
      "loss": 3.0205,
      "step": 173000
    },
    {
      "epoch": 9.36,
      "learning_rate": 5.824674942636936e-05,
      "loss": 3.0232,
      "step": 174000
    },
    {
      "epoch": 9.41,
      "learning_rate": 5.795257986703536e-05,
      "loss": 3.0294,
      "step": 175000
    },
    {
      "epoch": 9.46,
      "learning_rate": 5.7658704477260696e-05,
      "loss": 3.0248,
      "step": 176000
    },
    {
      "epoch": 9.52,
      "learning_rate": 5.7364829087486024e-05,
      "loss": 3.0275,
      "step": 177000
    },
    {
      "epoch": 9.57,
      "learning_rate": 5.707065952815203e-05,
      "loss": 3.0257,
      "step": 178000
    },
    {
      "epoch": 9.63,
      "learning_rate": 5.677678413837736e-05,
      "loss": 3.0285,
      "step": 179000
    },
    {
      "epoch": 9.68,
      "learning_rate": 5.6482614579043356e-05,
      "loss": 3.0275,
      "step": 180000
    },
    {
      "epoch": 9.73,
      "learning_rate": 5.618844501970937e-05,
      "loss": 3.0305,
      "step": 181000
    },
    {
      "epoch": 9.79,
      "learning_rate": 5.5894275460375366e-05,
      "loss": 3.0357,
      "step": 182000
    },
    {
      "epoch": 9.84,
      "learning_rate": 5.56004000706007e-05,
      "loss": 3.0356,
      "step": 183000
    },
    {
      "epoch": 9.89,
      "learning_rate": 5.53062305112667e-05,
      "loss": 3.0363,
      "step": 184000
    },
    {
      "epoch": 9.95,
      "learning_rate": 5.5012355121492035e-05,
      "loss": 3.0329,
      "step": 185000
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.4029401960076926,
      "eval_loss": 3.4231081008911133,
      "eval_runtime": 153.3432,
      "eval_samples_per_second": 377.702,
      "eval_steps_per_second": 5.902,
      "step": 185970
    },
    {
      "epoch": 10.0,
      "learning_rate": 5.471818556215803e-05,
      "loss": 3.0316,
      "step": 186000
    },
    {
      "epoch": 10.06,
      "learning_rate": 5.442431017238336e-05,
      "loss": 2.9811,
      "step": 187000
    },
    {
      "epoch": 10.11,
      "learning_rate": 5.413014061304936e-05,
      "loss": 2.9815,
      "step": 188000
    },
    {
      "epoch": 10.16,
      "learning_rate": 5.383597105371536e-05,
      "loss": 2.9895,
      "step": 189000
    },
    {
      "epoch": 10.22,
      "learning_rate": 5.3542095663940694e-05,
      "loss": 2.9893,
      "step": 190000
    },
    {
      "epoch": 10.27,
      "learning_rate": 5.324792610460669e-05,
      "loss": 2.9913,
      "step": 191000
    },
    {
      "epoch": 10.32,
      "learning_rate": 5.295375654527269e-05,
      "loss": 2.9966,
      "step": 192000
    },
    {
      "epoch": 10.38,
      "learning_rate": 5.26595869859387e-05,
      "loss": 2.9958,
      "step": 193000
    },
    {
      "epoch": 10.43,
      "learning_rate": 5.236571159616404e-05,
      "loss": 3.0016,
      "step": 194000
    },
    {
      "epoch": 10.49,
      "learning_rate": 5.2071542036830035e-05,
      "loss": 3.0028,
      "step": 195000
    },
    {
      "epoch": 10.54,
      "learning_rate": 5.1777372477496033e-05,
      "loss": 3.005,
      "step": 196000
    },
    {
      "epoch": 10.59,
      "learning_rate": 5.14837912572807e-05,
      "loss": 2.9997,
      "step": 197000
    },
    {
      "epoch": 10.65,
      "learning_rate": 5.11896216979467e-05,
      "loss": 3.0042,
      "step": 198000
    },
    {
      "epoch": 10.7,
      "learning_rate": 5.0895452138612696e-05,
      "loss": 3.01,
      "step": 199000
    },
    {
      "epoch": 10.75,
      "learning_rate": 5.060157674883803e-05,
      "loss": 3.0098,
      "step": 200000
    },
    {
      "epoch": 10.81,
      "learning_rate": 5.030740718950403e-05,
      "loss": 3.0122,
      "step": 201000
    },
    {
      "epoch": 10.86,
      "learning_rate": 5.001323763017003e-05,
      "loss": 3.008,
      "step": 202000
    },
    {
      "epoch": 10.92,
      "learning_rate": 4.971936224039537e-05,
      "loss": 3.0088,
      "step": 203000
    },
    {
      "epoch": 10.97,
      "learning_rate": 4.942519268106137e-05,
      "loss": 3.0093,
      "step": 204000
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.40446972998051856,
      "eval_loss": 3.410041093826294,
      "eval_runtime": 153.1573,
      "eval_samples_per_second": 378.16,
      "eval_steps_per_second": 5.909,
      "step": 204567
    },
    {
      "epoch": 11.02,
      "learning_rate": 4.9131023121727366e-05,
      "loss": 2.9805,
      "step": 205000
    },
    {
      "epoch": 11.08,
      "learning_rate": 4.8836853562393364e-05,
      "loss": 2.9595,
      "step": 206000
    },
    {
      "epoch": 11.13,
      "learning_rate": 4.854268400305936e-05,
      "loss": 2.9631,
      "step": 207000
    },
    {
      "epoch": 11.18,
      "learning_rate": 4.8248808613284705e-05,
      "loss": 2.9678,
      "step": 208000
    },
    {
      "epoch": 11.24,
      "learning_rate": 4.79546390539507e-05,
      "loss": 2.9658,
      "step": 209000
    },
    {
      "epoch": 11.29,
      "learning_rate": 4.766076366417603e-05,
      "loss": 2.9768,
      "step": 210000
    },
    {
      "epoch": 11.35,
      "learning_rate": 4.736659410484203e-05,
      "loss": 2.9732,
      "step": 211000
    },
    {
      "epoch": 11.4,
      "learning_rate": 4.7072424545508035e-05,
      "loss": 2.9822,
      "step": 212000
    },
    {
      "epoch": 11.45,
      "learning_rate": 4.6778549155733364e-05,
      "loss": 2.9789,
      "step": 213000
    },
    {
      "epoch": 11.51,
      "learning_rate": 4.648437959639937e-05,
      "loss": 2.9786,
      "step": 214000
    },
    {
      "epoch": 11.56,
      "learning_rate": 4.6190210037065367e-05,
      "loss": 2.9792,
      "step": 215000
    },
    {
      "epoch": 11.61,
      "learning_rate": 4.5896040477731365e-05,
      "loss": 2.983,
      "step": 216000
    },
    {
      "epoch": 11.67,
      "learning_rate": 4.560187091839736e-05,
      "loss": 2.9819,
      "step": 217000
    },
    {
      "epoch": 11.72,
      "learning_rate": 4.53079955286227e-05,
      "loss": 2.9833,
      "step": 218000
    },
    {
      "epoch": 11.78,
      "learning_rate": 4.5013825969288696e-05,
      "loss": 2.9847,
      "step": 219000
    },
    {
      "epoch": 11.83,
      "learning_rate": 4.471995057951404e-05,
      "loss": 2.9858,
      "step": 220000
    },
    {
      "epoch": 11.88,
      "learning_rate": 4.442607518973937e-05,
      "loss": 2.9812,
      "step": 221000
    },
    {
      "epoch": 11.94,
      "learning_rate": 4.4131905630405366e-05,
      "loss": 2.9871,
      "step": 222000
    },
    {
      "epoch": 11.99,
      "learning_rate": 4.3837736071071364e-05,
      "loss": 2.9859,
      "step": 223000
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.40492038724277163,
      "eval_loss": 3.409681797027588,
      "eval_runtime": 153.4512,
      "eval_samples_per_second": 377.436,
      "eval_steps_per_second": 5.898,
      "step": 223164
    },
    {
      "epoch": 12.04,
      "learning_rate": 4.354356651173737e-05,
      "loss": 2.943,
      "step": 224000
    },
    {
      "epoch": 12.1,
      "learning_rate": 4.32496911219627e-05,
      "loss": 2.9399,
      "step": 225000
    },
    {
      "epoch": 12.15,
      "learning_rate": 4.29555215626287e-05,
      "loss": 2.9462,
      "step": 226000
    },
    {
      "epoch": 12.21,
      "learning_rate": 4.26613520032947e-05,
      "loss": 2.9433,
      "step": 227000
    },
    {
      "epoch": 12.26,
      "learning_rate": 4.23671824439607e-05,
      "loss": 2.9478,
      "step": 228000
    },
    {
      "epoch": 12.31,
      "learning_rate": 4.2073012884626704e-05,
      "loss": 2.9522,
      "step": 229000
    },
    {
      "epoch": 12.37,
      "learning_rate": 4.177913749485203e-05,
      "loss": 2.9532,
      "step": 230000
    },
    {
      "epoch": 12.42,
      "learning_rate": 4.148526210507737e-05,
      "loss": 2.9551,
      "step": 231000
    },
    {
      "epoch": 12.48,
      "learning_rate": 4.119109254574337e-05,
      "loss": 2.961,
      "step": 232000
    },
    {
      "epoch": 12.53,
      "learning_rate": 4.089692298640937e-05,
      "loss": 2.9528,
      "step": 233000
    },
    {
      "epoch": 12.58,
      "learning_rate": 4.060275342707537e-05,
      "loss": 2.9639,
      "step": 234000
    },
    {
      "epoch": 12.64,
      "learning_rate": 4.0308878037300705e-05,
      "loss": 2.9596,
      "step": 235000
    },
    {
      "epoch": 12.69,
      "learning_rate": 4.00147084779667e-05,
      "loss": 2.9624,
      "step": 236000
    },
    {
      "epoch": 12.74,
      "learning_rate": 3.972083308819203e-05,
      "loss": 2.9609,
      "step": 237000
    },
    {
      "epoch": 12.8,
      "learning_rate": 3.9426663528858036e-05,
      "loss": 2.9616,
      "step": 238000
    },
    {
      "epoch": 12.85,
      "learning_rate": 3.9132493969524034e-05,
      "loss": 2.9694,
      "step": 239000
    },
    {
      "epoch": 12.91,
      "learning_rate": 3.88389127493087e-05,
      "loss": 2.9629,
      "step": 240000
    },
    {
      "epoch": 12.96,
      "learning_rate": 3.8544743189974706e-05,
      "loss": 2.9662,
      "step": 241000
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.4052818267285978,
      "eval_loss": 3.4042649269104004,
      "eval_runtime": 153.6035,
      "eval_samples_per_second": 377.062,
      "eval_steps_per_second": 5.892,
      "step": 241761
    },
    {
      "epoch": 13.01,
      "learning_rate": 3.8250573630640704e-05,
      "loss": 2.9532,
      "step": 242000
    },
    {
      "epoch": 13.07,
      "learning_rate": 3.79564040713067e-05,
      "loss": 2.9159,
      "step": 243000
    },
    {
      "epoch": 13.12,
      "learning_rate": 3.766252868153204e-05,
      "loss": 2.9206,
      "step": 244000
    },
    {
      "epoch": 13.17,
      "learning_rate": 3.7368359122198036e-05,
      "loss": 2.9259,
      "step": 245000
    },
    {
      "epoch": 13.23,
      "learning_rate": 3.707448373242337e-05,
      "loss": 2.9251,
      "step": 246000
    },
    {
      "epoch": 13.28,
      "learning_rate": 3.678060834264871e-05,
      "loss": 2.9312,
      "step": 247000
    },
    {
      "epoch": 13.34,
      "learning_rate": 3.6486438783314705e-05,
      "loss": 2.9362,
      "step": 248000
    },
    {
      "epoch": 13.39,
      "learning_rate": 3.61922692239807e-05,
      "loss": 2.932,
      "step": 249000
    },
    {
      "epoch": 13.44,
      "learning_rate": 3.589809966464671e-05,
      "loss": 2.9368,
      "step": 250000
    },
    {
      "epoch": 13.5,
      "learning_rate": 3.5603930105312706e-05,
      "loss": 2.9376,
      "step": 251000
    },
    {
      "epoch": 13.55,
      "learning_rate": 3.531034888509737e-05,
      "loss": 2.938,
      "step": 252000
    },
    {
      "epoch": 13.6,
      "learning_rate": 3.501617932576337e-05,
      "loss": 2.942,
      "step": 253000
    },
    {
      "epoch": 13.66,
      "learning_rate": 3.472200976642937e-05,
      "loss": 2.9405,
      "step": 254000
    },
    {
      "epoch": 13.71,
      "learning_rate": 3.4427840207095367e-05,
      "loss": 2.9403,
      "step": 255000
    },
    {
      "epoch": 13.77,
      "learning_rate": 3.413396481732071e-05,
      "loss": 2.9459,
      "step": 256000
    },
    {
      "epoch": 13.82,
      "learning_rate": 3.383979525798671e-05,
      "loss": 2.9432,
      "step": 257000
    },
    {
      "epoch": 13.87,
      "learning_rate": 3.3545625698652705e-05,
      "loss": 2.9483,
      "step": 258000
    },
    {
      "epoch": 13.93,
      "learning_rate": 3.325175030887804e-05,
      "loss": 2.9437,
      "step": 259000
    },
    {
      "epoch": 13.98,
      "learning_rate": 3.295758074954404e-05,
      "loss": 2.9424,
      "step": 260000
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.4056935855777852,
      "eval_loss": 3.4045979976654053,
      "eval_runtime": 153.3807,
      "eval_samples_per_second": 377.61,
      "eval_steps_per_second": 5.9,
      "step": 260358
    },
    {
      "epoch": 14.03,
      "learning_rate": 3.2663705359769374e-05,
      "loss": 2.9161,
      "step": 261000
    },
    {
      "epoch": 14.09,
      "learning_rate": 3.236953580043537e-05,
      "loss": 2.9053,
      "step": 262000
    },
    {
      "epoch": 14.14,
      "learning_rate": 3.207566041066071e-05,
      "loss": 2.9064,
      "step": 263000
    },
    {
      "epoch": 14.2,
      "learning_rate": 3.1781490851326706e-05,
      "loss": 2.9123,
      "step": 264000
    },
    {
      "epoch": 14.25,
      "learning_rate": 3.148761546155204e-05,
      "loss": 2.9084,
      "step": 265000
    },
    {
      "epoch": 14.3,
      "learning_rate": 3.119344590221804e-05,
      "loss": 2.9113,
      "step": 266000
    },
    {
      "epoch": 14.36,
      "learning_rate": 3.0899570512443376e-05,
      "loss": 2.913,
      "step": 267000
    },
    {
      "epoch": 14.41,
      "learning_rate": 3.060569512266871e-05,
      "loss": 2.921,
      "step": 268000
    },
    {
      "epoch": 14.46,
      "learning_rate": 3.031152556333471e-05,
      "loss": 2.9186,
      "step": 269000
    },
    {
      "epoch": 14.52,
      "learning_rate": 3.0017356004000707e-05,
      "loss": 2.9163,
      "step": 270000
    },
    {
      "epoch": 14.57,
      "learning_rate": 2.9723186444666705e-05,
      "loss": 2.9189,
      "step": 271000
    },
    {
      "epoch": 14.63,
      "learning_rate": 2.9429311054892038e-05,
      "loss": 2.9215,
      "step": 272000
    },
    {
      "epoch": 14.68,
      "learning_rate": 2.9135141495558042e-05,
      "loss": 2.9208,
      "step": 273000
    },
    {
      "epoch": 14.73,
      "learning_rate": 2.8841266105783378e-05,
      "loss": 2.9262,
      "step": 274000
    },
    {
      "epoch": 14.79,
      "learning_rate": 2.8547096546449376e-05,
      "loss": 2.9249,
      "step": 275000
    },
    {
      "epoch": 14.84,
      "learning_rate": 2.8252926987115374e-05,
      "loss": 2.9224,
      "step": 276000
    },
    {
      "epoch": 14.89,
      "learning_rate": 2.7959051597340706e-05,
      "loss": 2.9233,
      "step": 277000
    },
    {
      "epoch": 14.95,
      "learning_rate": 2.7664882038006708e-05,
      "loss": 2.928,
      "step": 278000
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.40587793315196863,
      "eval_loss": 3.4078826904296875,
      "eval_runtime": 153.8126,
      "eval_samples_per_second": 376.549,
      "eval_steps_per_second": 5.884,
      "step": 278955
    },
    {
      "epoch": 15.0,
      "learning_rate": 2.7371006648232044e-05,
      "loss": 2.9273,
      "step": 279000
    },
    {
      "epoch": 15.06,
      "learning_rate": 2.7076837088898045e-05,
      "loss": 2.8835,
      "step": 280000
    },
    {
      "epoch": 15.11,
      "learning_rate": 2.678325586868271e-05,
      "loss": 2.8879,
      "step": 281000
    },
    {
      "epoch": 15.16,
      "learning_rate": 2.648908630934871e-05,
      "loss": 2.8894,
      "step": 282000
    },
    {
      "epoch": 15.22,
      "learning_rate": 2.6194916750014708e-05,
      "loss": 2.8945,
      "step": 283000
    },
    {
      "epoch": 15.27,
      "learning_rate": 2.590074719068071e-05,
      "loss": 2.8966,
      "step": 284000
    },
    {
      "epoch": 15.33,
      "learning_rate": 2.5606871800906045e-05,
      "loss": 2.8984,
      "step": 285000
    },
    {
      "epoch": 15.38,
      "learning_rate": 2.531299641113138e-05,
      "loss": 2.8989,
      "step": 286000
    },
    {
      "epoch": 15.43,
      "learning_rate": 2.501882685179738e-05,
      "loss": 2.9001,
      "step": 287000
    },
    {
      "epoch": 15.49,
      "learning_rate": 2.472495146202271e-05,
      "loss": 2.9013,
      "step": 288000
    },
    {
      "epoch": 15.54,
      "learning_rate": 2.443078190268871e-05,
      "loss": 2.9033,
      "step": 289000
    },
    {
      "epoch": 15.59,
      "learning_rate": 2.413661234335471e-05,
      "loss": 2.9021,
      "step": 290000
    },
    {
      "epoch": 15.65,
      "learning_rate": 2.384244278402071e-05,
      "loss": 2.9055,
      "step": 291000
    },
    {
      "epoch": 15.7,
      "learning_rate": 2.354827322468671e-05,
      "loss": 2.904,
      "step": 292000
    },
    {
      "epoch": 15.76,
      "learning_rate": 2.3254397834912045e-05,
      "loss": 2.9048,
      "step": 293000
    },
    {
      "epoch": 15.81,
      "learning_rate": 2.2960522445137378e-05,
      "loss": 2.9072,
      "step": 294000
    },
    {
      "epoch": 15.86,
      "learning_rate": 2.2666352885803376e-05,
      "loss": 2.9067,
      "step": 295000
    },
    {
      "epoch": 15.92,
      "learning_rate": 2.2372183326469377e-05,
      "loss": 2.9076,
      "step": 296000
    },
    {
      "epoch": 15.97,
      "learning_rate": 2.2078307936694713e-05,
      "loss": 2.908,
      "step": 297000
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.4060550922456719,
      "eval_loss": 3.411921262741089,
      "eval_runtime": 153.4871,
      "eval_samples_per_second": 377.348,
      "eval_steps_per_second": 5.896,
      "step": 297552
    },
    {
      "epoch": 16.02,
      "learning_rate": 2.178413837736071e-05,
      "loss": 2.893,
      "step": 298000
    },
    {
      "epoch": 16.08,
      "learning_rate": 2.1490262987586046e-05,
      "loss": 2.8702,
      "step": 299000
    },
    {
      "epoch": 16.13,
      "learning_rate": 2.1196093428252045e-05,
      "loss": 2.8782,
      "step": 300000
    },
    {
      "epoch": 16.19,
      "learning_rate": 2.0901923868918043e-05,
      "loss": 2.8769,
      "step": 301000
    },
    {
      "epoch": 16.24,
      "learning_rate": 2.0607754309584047e-05,
      "loss": 2.88,
      "step": 302000
    },
    {
      "epoch": 16.29,
      "learning_rate": 2.0313584750250046e-05,
      "loss": 2.8792,
      "step": 303000
    },
    {
      "epoch": 16.35,
      "learning_rate": 2.0019709360475378e-05,
      "loss": 2.8793,
      "step": 304000
    },
    {
      "epoch": 16.4,
      "learning_rate": 1.9725833970700713e-05,
      "loss": 2.8812,
      "step": 305000
    },
    {
      "epoch": 16.45,
      "learning_rate": 1.943166441136671e-05,
      "loss": 2.8852,
      "step": 306000
    },
    {
      "epoch": 16.51,
      "learning_rate": 1.9137494852032713e-05,
      "loss": 2.8845,
      "step": 307000
    },
    {
      "epoch": 16.56,
      "learning_rate": 1.8843325292698714e-05,
      "loss": 2.8871,
      "step": 308000
    },
    {
      "epoch": 16.62,
      "learning_rate": 1.8549155733364713e-05,
      "loss": 2.8845,
      "step": 309000
    },
    {
      "epoch": 16.67,
      "learning_rate": 1.8255280343590045e-05,
      "loss": 2.8865,
      "step": 310000
    },
    {
      "epoch": 16.72,
      "learning_rate": 1.7961110784256046e-05,
      "loss": 2.8871,
      "step": 311000
    },
    {
      "epoch": 16.78,
      "learning_rate": 1.7667529564040713e-05,
      "loss": 2.8893,
      "step": 312000
    },
    {
      "epoch": 16.83,
      "learning_rate": 1.7373360004706714e-05,
      "loss": 2.8926,
      "step": 313000
    },
    {
      "epoch": 16.88,
      "learning_rate": 1.7079190445372716e-05,
      "loss": 2.8909,
      "step": 314000
    },
    {
      "epoch": 16.94,
      "learning_rate": 1.6785315055598048e-05,
      "loss": 2.8898,
      "step": 315000
    },
    {
      "epoch": 16.99,
      "learning_rate": 1.6491145496264046e-05,
      "loss": 2.8912,
      "step": 316000
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.40615673870330293,
      "eval_loss": 3.4119112491607666,
      "eval_runtime": 153.7299,
      "eval_samples_per_second": 376.752,
      "eval_steps_per_second": 5.887,
      "step": 316149
    },
    {
      "epoch": 17.05,
      "learning_rate": 1.6196975936930047e-05,
      "loss": 2.8649,
      "step": 317000
    },
    {
      "epoch": 17.1,
      "learning_rate": 1.5902806377596045e-05,
      "loss": 2.8592,
      "step": 318000
    },
    {
      "epoch": 17.15,
      "learning_rate": 1.5608636818262047e-05,
      "loss": 2.8635,
      "step": 319000
    },
    {
      "epoch": 17.21,
      "learning_rate": 1.531446725892805e-05,
      "loss": 2.8652,
      "step": 320000
    },
    {
      "epoch": 17.26,
      "learning_rate": 1.502059186915338e-05,
      "loss": 2.8603,
      "step": 321000
    },
    {
      "epoch": 17.31,
      "learning_rate": 1.4726422309819379e-05,
      "loss": 2.8673,
      "step": 322000
    },
    {
      "epoch": 17.37,
      "learning_rate": 1.4432546920044714e-05,
      "loss": 2.8641,
      "step": 323000
    },
    {
      "epoch": 17.42,
      "learning_rate": 1.4138377360710714e-05,
      "loss": 2.8724,
      "step": 324000
    },
    {
      "epoch": 17.48,
      "learning_rate": 1.3844207801376712e-05,
      "loss": 2.8704,
      "step": 325000
    },
    {
      "epoch": 17.53,
      "learning_rate": 1.3550626581161382e-05,
      "loss": 2.8711,
      "step": 326000
    },
    {
      "epoch": 17.58,
      "learning_rate": 1.3256457021827382e-05,
      "loss": 2.8714,
      "step": 327000
    },
    {
      "epoch": 17.64,
      "learning_rate": 1.296228746249338e-05,
      "loss": 2.8715,
      "step": 328000
    },
    {
      "epoch": 17.69,
      "learning_rate": 1.2668117903159383e-05,
      "loss": 2.8702,
      "step": 329000
    },
    {
      "epoch": 17.74,
      "learning_rate": 1.2373948343825381e-05,
      "loss": 2.8737,
      "step": 330000
    },
    {
      "epoch": 17.8,
      "learning_rate": 1.2080072954050715e-05,
      "loss": 2.8738,
      "step": 331000
    },
    {
      "epoch": 17.85,
      "learning_rate": 1.1785903394716716e-05,
      "loss": 2.8758,
      "step": 332000
    },
    {
      "epoch": 17.91,
      "learning_rate": 1.1492028004942049e-05,
      "loss": 2.8721,
      "step": 333000
    },
    {
      "epoch": 17.96,
      "learning_rate": 1.1197858445608048e-05,
      "loss": 2.8716,
      "step": 334000
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.4063988628495701,
      "eval_loss": 3.415924549102783,
      "eval_runtime": 153.4402,
      "eval_samples_per_second": 377.463,
      "eval_steps_per_second": 5.898,
      "step": 334746
    },
    {
      "epoch": 18.01,
      "learning_rate": 1.0904277225392716e-05,
      "loss": 2.8667,
      "step": 335000
    },
    {
      "epoch": 18.07,
      "learning_rate": 1.0610107666058716e-05,
      "loss": 2.8473,
      "step": 336000
    },
    {
      "epoch": 18.12,
      "learning_rate": 1.0315938106724716e-05,
      "loss": 2.8493,
      "step": 337000
    },
    {
      "epoch": 18.17,
      "learning_rate": 1.0021768547390718e-05,
      "loss": 2.8574,
      "step": 338000
    },
    {
      "epoch": 18.23,
      "learning_rate": 9.727598988056716e-06,
      "loss": 2.8522,
      "step": 339000
    },
    {
      "epoch": 18.28,
      "learning_rate": 9.43372359828205e-06,
      "loss": 2.8504,
      "step": 340000
    },
    {
      "epoch": 18.34,
      "learning_rate": 9.139554038948051e-06,
      "loss": 2.854,
      "step": 341000
    },
    {
      "epoch": 18.39,
      "learning_rate": 8.845384479614049e-06,
      "loss": 2.8497,
      "step": 342000
    },
    {
      "epoch": 18.44,
      "learning_rate": 8.551509089839383e-06,
      "loss": 2.8555,
      "step": 343000
    },
    {
      "epoch": 18.5,
      "learning_rate": 8.257339530505384e-06,
      "loss": 2.8554,
      "step": 344000
    },
    {
      "epoch": 18.55,
      "learning_rate": 7.963464140730717e-06,
      "loss": 2.8551,
      "step": 345000
    },
    {
      "epoch": 18.61,
      "learning_rate": 7.669294581396718e-06,
      "loss": 2.8556,
      "step": 346000
    },
    {
      "epoch": 18.66,
      "learning_rate": 7.375125022062718e-06,
      "loss": 2.8563,
      "step": 347000
    },
    {
      "epoch": 18.71,
      "learning_rate": 7.080955462728717e-06,
      "loss": 2.8591,
      "step": 348000
    },
    {
      "epoch": 18.77,
      "learning_rate": 6.786785903394717e-06,
      "loss": 2.8552,
      "step": 349000
    },
    {
      "epoch": 18.82,
      "learning_rate": 6.4929105136200514e-06,
      "loss": 2.8633,
      "step": 350000
    },
    {
      "epoch": 18.87,
      "learning_rate": 6.198740954286051e-06,
      "loss": 2.8542,
      "step": 351000
    },
    {
      "epoch": 18.93,
      "learning_rate": 5.904865564511384e-06,
      "loss": 2.8612,
      "step": 352000
    },
    {
      "epoch": 18.98,
      "learning_rate": 5.610990174736719e-06,
      "loss": 2.8589,
      "step": 353000
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.406526172854336,
      "eval_loss": 3.422281503677368,
      "eval_runtime": 153.5913,
      "eval_samples_per_second": 377.092,
      "eval_steps_per_second": 5.892,
      "step": 353343
    },
    {
      "epoch": 19.04,
      "learning_rate": 5.316820615402719e-06,
      "loss": 2.8448,
      "step": 354000
    },
    {
      "epoch": 19.09,
      "learning_rate": 5.022945225628052e-06,
      "loss": 2.8385,
      "step": 355000
    },
    {
      "epoch": 19.14,
      "learning_rate": 4.7287756662940525e-06,
      "loss": 2.8445,
      "step": 356000
    },
    {
      "epoch": 19.2,
      "learning_rate": 4.434606106960052e-06,
      "loss": 2.8427,
      "step": 357000
    },
    {
      "epoch": 19.25,
      "learning_rate": 4.140436547626052e-06,
      "loss": 2.8386,
      "step": 358000
    },
    {
      "epoch": 19.3,
      "learning_rate": 3.846266988292051e-06,
      "loss": 2.8364,
      "step": 359000
    },
    {
      "epoch": 19.36,
      "learning_rate": 3.552391598517386e-06,
      "loss": 2.8417,
      "step": 360000
    },
    {
      "epoch": 19.41,
      "learning_rate": 3.258222039183385e-06,
      "loss": 2.8461,
      "step": 361000
    },
    {
      "epoch": 19.47,
      "learning_rate": 2.9640524798493854e-06,
      "loss": 2.8412,
      "step": 362000
    },
    {
      "epoch": 19.52,
      "learning_rate": 2.6704712596340534e-06,
      "loss": 2.8461,
      "step": 363000
    },
    {
      "epoch": 19.57,
      "learning_rate": 2.3763017003000528e-06,
      "loss": 2.8483,
      "step": 364000
    },
    {
      "epoch": 19.63,
      "learning_rate": 2.082132140966053e-06,
      "loss": 2.8418,
      "step": 365000
    },
    {
      "epoch": 19.68,
      "learning_rate": 1.787962581632053e-06,
      "loss": 2.842,
      "step": 366000
    },
    {
      "epoch": 19.73,
      "learning_rate": 1.4937930222980528e-06,
      "loss": 2.8432,
      "step": 367000
    },
    {
      "epoch": 19.79,
      "learning_rate": 1.1996234629640526e-06,
      "loss": 2.8444,
      "step": 368000
    },
    {
      "epoch": 19.84,
      "learning_rate": 9.060422427487204e-07,
      "loss": 2.8428,
      "step": 369000
    },
    {
      "epoch": 19.9,
      "learning_rate": 6.118726834147203e-07,
      "loss": 2.8443,
      "step": 370000
    },
    {
      "epoch": 19.95,
      "learning_rate": 3.1770312408072017e-07,
      "loss": 2.8424,
      "step": 371000
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.40654968657553286,
      "eval_loss": 3.424689292907715,
      "eval_runtime": 153.6195,
      "eval_samples_per_second": 377.022,
      "eval_steps_per_second": 5.891,
      "step": 371940
    },
    {
      "epoch": 20.0,
      "step": 371940,
      "total_flos": 1.56702845389824e+18,
      "train_loss": 3.1534978556531716,
      "train_runtime": 80977.43,
      "train_samples_per_second": 146.973,
      "train_steps_per_second": 4.593
    }
  ],
  "logging_steps": 1000,
  "max_steps": 371940,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "save_steps": 5000,
  "total_flos": 1.56702845389824e+18,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}