{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 20.0,
  "eval_steps": 500,
  "global_step": 371720,
  "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": 7.5535,
      "step": 1000
    },
    {
      "epoch": 0.11,
      "learning_rate": 6.25e-06,
      "loss": 5.819,
      "step": 2000
    },
    {
      "epoch": 0.16,
      "learning_rate": 9.375000000000001e-06,
      "loss": 5.3809,
      "step": 3000
    },
    {
      "epoch": 0.22,
      "learning_rate": 1.25e-05,
      "loss": 5.1534,
      "step": 4000
    },
    {
      "epoch": 0.27,
      "learning_rate": 1.5625e-05,
      "loss": 5.0157,
      "step": 5000
    },
    {
      "epoch": 0.32,
      "learning_rate": 1.8750000000000002e-05,
      "loss": 4.8786,
      "step": 6000
    },
    {
      "epoch": 0.38,
      "learning_rate": 2.1875e-05,
      "loss": 4.756,
      "step": 7000
    },
    {
      "epoch": 0.43,
      "learning_rate": 2.5e-05,
      "loss": 4.648,
      "step": 8000
    },
    {
      "epoch": 0.48,
      "learning_rate": 2.8125000000000003e-05,
      "loss": 4.5608,
      "step": 9000
    },
    {
      "epoch": 0.54,
      "learning_rate": 3.125e-05,
      "loss": 4.4865,
      "step": 10000
    },
    {
      "epoch": 0.59,
      "learning_rate": 3.4375e-05,
      "loss": 4.4109,
      "step": 11000
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 4.3492,
      "step": 12000
    },
    {
      "epoch": 0.7,
      "learning_rate": 4.0621875e-05,
      "loss": 4.2874,
      "step": 13000
    },
    {
      "epoch": 0.75,
      "learning_rate": 4.3746875000000005e-05,
      "loss": 4.241,
      "step": 14000
    },
    {
      "epoch": 0.81,
      "learning_rate": 4.6865625e-05,
      "loss": 4.18,
      "step": 15000
    },
    {
      "epoch": 0.86,
      "learning_rate": 4.9990625000000004e-05,
      "loss": 4.1416,
      "step": 16000
    },
    {
      "epoch": 0.91,
      "learning_rate": 5.3115625000000005e-05,
      "loss": 4.0952,
      "step": 17000
    },
    {
      "epoch": 0.97,
      "learning_rate": 5.6240625e-05,
      "loss": 4.0553,
      "step": 18000
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.3104167562396543,
      "eval_loss": 4.24772834777832,
      "eval_runtime": 153.1386,
      "eval_samples_per_second": 378.22,
      "eval_steps_per_second": 5.91,
      "step": 18586
    },
    {
      "epoch": 1.02,
      "learning_rate": 5.93625e-05,
      "loss": 4.0014,
      "step": 19000
    },
    {
      "epoch": 1.08,
      "learning_rate": 6.24875e-05,
      "loss": 3.9636,
      "step": 20000
    },
    {
      "epoch": 1.13,
      "learning_rate": 6.560937500000001e-05,
      "loss": 3.924,
      "step": 21000
    },
    {
      "epoch": 1.18,
      "learning_rate": 6.8734375e-05,
      "loss": 3.8853,
      "step": 22000
    },
    {
      "epoch": 1.24,
      "learning_rate": 7.185625e-05,
      "loss": 3.854,
      "step": 23000
    },
    {
      "epoch": 1.29,
      "learning_rate": 7.4978125e-05,
      "loss": 3.8229,
      "step": 24000
    },
    {
      "epoch": 1.35,
      "learning_rate": 7.8103125e-05,
      "loss": 3.8036,
      "step": 25000
    },
    {
      "epoch": 1.4,
      "learning_rate": 8.122500000000001e-05,
      "loss": 3.7745,
      "step": 26000
    },
    {
      "epoch": 1.45,
      "learning_rate": 8.435e-05,
      "loss": 3.7541,
      "step": 27000
    },
    {
      "epoch": 1.51,
      "learning_rate": 8.7471875e-05,
      "loss": 3.733,
      "step": 28000
    },
    {
      "epoch": 1.56,
      "learning_rate": 9.0596875e-05,
      "loss": 3.7088,
      "step": 29000
    },
    {
      "epoch": 1.61,
      "learning_rate": 9.371875e-05,
      "loss": 3.6882,
      "step": 30000
    },
    {
      "epoch": 1.67,
      "learning_rate": 9.684375e-05,
      "loss": 3.6656,
      "step": 31000
    },
    {
      "epoch": 1.72,
      "learning_rate": 9.996875e-05,
      "loss": 3.6542,
      "step": 32000
    },
    {
      "epoch": 1.78,
      "learning_rate": 9.970858353938538e-05,
      "loss": 3.6379,
      "step": 33000
    },
    {
      "epoch": 1.83,
      "learning_rate": 9.941451783821971e-05,
      "loss": 3.6178,
      "step": 34000
    },
    {
      "epoch": 1.88,
      "learning_rate": 9.912015777699282e-05,
      "loss": 3.5973,
      "step": 35000
    },
    {
      "epoch": 1.94,
      "learning_rate": 9.882609207582717e-05,
      "loss": 3.5872,
      "step": 36000
    },
    {
      "epoch": 1.99,
      "learning_rate": 9.853202637466149e-05,
      "loss": 3.572,
      "step": 37000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.36217867930004083,
      "eval_loss": 3.7583320140838623,
      "eval_runtime": 154.7165,
      "eval_samples_per_second": 374.362,
      "eval_steps_per_second": 5.849,
      "step": 37172
    },
    {
      "epoch": 2.04,
      "learning_rate": 9.823766631343459e-05,
      "loss": 3.5276,
      "step": 38000
    },
    {
      "epoch": 2.1,
      "learning_rate": 9.79433062522077e-05,
      "loss": 3.5189,
      "step": 39000
    },
    {
      "epoch": 2.15,
      "learning_rate": 9.764894619098082e-05,
      "loss": 3.5037,
      "step": 40000
    },
    {
      "epoch": 2.21,
      "learning_rate": 9.735517484987638e-05,
      "loss": 3.4951,
      "step": 41000
    },
    {
      "epoch": 2.26,
      "learning_rate": 9.706081478864948e-05,
      "loss": 3.4851,
      "step": 42000
    },
    {
      "epoch": 2.31,
      "learning_rate": 9.676645472742259e-05,
      "loss": 3.4802,
      "step": 43000
    },
    {
      "epoch": 2.37,
      "learning_rate": 9.647238902625692e-05,
      "loss": 3.47,
      "step": 44000
    },
    {
      "epoch": 2.42,
      "learning_rate": 9.617802896503003e-05,
      "loss": 3.4631,
      "step": 45000
    },
    {
      "epoch": 2.47,
      "learning_rate": 9.588396326386436e-05,
      "loss": 3.4513,
      "step": 46000
    },
    {
      "epoch": 2.53,
      "learning_rate": 9.558960320263747e-05,
      "loss": 3.4484,
      "step": 47000
    },
    {
      "epoch": 2.58,
      "learning_rate": 9.529524314141057e-05,
      "loss": 3.4426,
      "step": 48000
    },
    {
      "epoch": 2.64,
      "learning_rate": 9.500088308018369e-05,
      "loss": 3.432,
      "step": 49000
    },
    {
      "epoch": 2.69,
      "learning_rate": 9.470681737901802e-05,
      "loss": 3.4219,
      "step": 50000
    },
    {
      "epoch": 2.74,
      "learning_rate": 9.441275167785236e-05,
      "loss": 3.4248,
      "step": 51000
    },
    {
      "epoch": 2.8,
      "learning_rate": 9.411839161662546e-05,
      "loss": 3.4131,
      "step": 52000
    },
    {
      "epoch": 2.85,
      "learning_rate": 9.382403155539857e-05,
      "loss": 3.4077,
      "step": 53000
    },
    {
      "epoch": 2.91,
      "learning_rate": 9.35299658542329e-05,
      "loss": 3.4006,
      "step": 54000
    },
    {
      "epoch": 2.96,
      "learning_rate": 9.323560579300601e-05,
      "loss": 3.394,
      "step": 55000
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.37962814669905626,
      "eval_loss": 3.585693120956421,
      "eval_runtime": 155.2145,
      "eval_samples_per_second": 373.161,
      "eval_steps_per_second": 5.831,
      "step": 55758
    },
    {
      "epoch": 3.01,
      "learning_rate": 9.294154009184034e-05,
      "loss": 3.3788,
      "step": 56000
    },
    {
      "epoch": 3.07,
      "learning_rate": 9.264718003061346e-05,
      "loss": 3.3428,
      "step": 57000
    },
    {
      "epoch": 3.12,
      "learning_rate": 9.235281996938656e-05,
      "loss": 3.3405,
      "step": 58000
    },
    {
      "epoch": 3.17,
      "learning_rate": 9.20587542682209e-05,
      "loss": 3.3383,
      "step": 59000
    },
    {
      "epoch": 3.23,
      "learning_rate": 9.1764394206994e-05,
      "loss": 3.3338,
      "step": 60000
    },
    {
      "epoch": 3.28,
      "learning_rate": 9.147032850582832e-05,
      "loss": 3.3269,
      "step": 61000
    },
    {
      "epoch": 3.34,
      "learning_rate": 9.117596844460144e-05,
      "loss": 3.3275,
      "step": 62000
    },
    {
      "epoch": 3.39,
      "learning_rate": 9.088160838337455e-05,
      "loss": 3.3227,
      "step": 63000
    },
    {
      "epoch": 3.44,
      "learning_rate": 9.058754268220888e-05,
      "loss": 3.3202,
      "step": 64000
    },
    {
      "epoch": 3.5,
      "learning_rate": 9.029318262098199e-05,
      "loss": 3.3206,
      "step": 65000
    },
    {
      "epoch": 3.55,
      "learning_rate": 8.999911691981633e-05,
      "loss": 3.322,
      "step": 66000
    },
    {
      "epoch": 3.6,
      "learning_rate": 8.970475685858944e-05,
      "loss": 3.3167,
      "step": 67000
    },
    {
      "epoch": 3.66,
      "learning_rate": 8.941039679736254e-05,
      "loss": 3.3095,
      "step": 68000
    },
    {
      "epoch": 3.71,
      "learning_rate": 8.911633109619687e-05,
      "loss": 3.3077,
      "step": 69000
    },
    {
      "epoch": 3.77,
      "learning_rate": 8.882197103496997e-05,
      "loss": 3.2993,
      "step": 70000
    },
    {
      "epoch": 3.82,
      "learning_rate": 8.852790533380432e-05,
      "loss": 3.2999,
      "step": 71000
    },
    {
      "epoch": 3.87,
      "learning_rate": 8.823354527257742e-05,
      "loss": 3.2961,
      "step": 72000
    },
    {
      "epoch": 3.93,
      "learning_rate": 8.793947957141176e-05,
      "loss": 3.3002,
      "step": 73000
    },
    {
      "epoch": 3.98,
      "learning_rate": 8.764541387024609e-05,
      "loss": 3.2886,
      "step": 74000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.3882697454693983,
      "eval_loss": 3.4992237091064453,
      "eval_runtime": 154.0405,
      "eval_samples_per_second": 376.005,
      "eval_steps_per_second": 5.875,
      "step": 74344
    },
    {
      "epoch": 4.04,
      "learning_rate": 8.73510538090192e-05,
      "loss": 3.2529,
      "step": 75000
    },
    {
      "epoch": 4.09,
      "learning_rate": 8.70566937477923e-05,
      "loss": 3.2373,
      "step": 76000
    },
    {
      "epoch": 4.14,
      "learning_rate": 8.676262804662663e-05,
      "loss": 3.2418,
      "step": 77000
    },
    {
      "epoch": 4.2,
      "learning_rate": 8.646826798539974e-05,
      "loss": 3.2383,
      "step": 78000
    },
    {
      "epoch": 4.25,
      "learning_rate": 8.617390792417285e-05,
      "loss": 3.2458,
      "step": 79000
    },
    {
      "epoch": 4.3,
      "learning_rate": 8.587984222300719e-05,
      "loss": 3.2389,
      "step": 80000
    },
    {
      "epoch": 4.36,
      "learning_rate": 8.55854821617803e-05,
      "loss": 3.2418,
      "step": 81000
    },
    {
      "epoch": 4.41,
      "learning_rate": 8.52911221005534e-05,
      "loss": 3.235,
      "step": 82000
    },
    {
      "epoch": 4.47,
      "learning_rate": 8.499676203932651e-05,
      "loss": 3.2401,
      "step": 83000
    },
    {
      "epoch": 4.52,
      "learning_rate": 8.470269633816084e-05,
      "loss": 3.2353,
      "step": 84000
    },
    {
      "epoch": 4.57,
      "learning_rate": 8.440863063699518e-05,
      "loss": 3.2349,
      "step": 85000
    },
    {
      "epoch": 4.63,
      "learning_rate": 8.411427057576828e-05,
      "loss": 3.2328,
      "step": 86000
    },
    {
      "epoch": 4.68,
      "learning_rate": 8.381991051454139e-05,
      "loss": 3.2338,
      "step": 87000
    },
    {
      "epoch": 4.73,
      "learning_rate": 8.352555045331449e-05,
      "loss": 3.2302,
      "step": 88000
    },
    {
      "epoch": 4.79,
      "learning_rate": 8.323119039208761e-05,
      "loss": 3.2293,
      "step": 89000
    },
    {
      "epoch": 4.84,
      "learning_rate": 8.293712469092194e-05,
      "loss": 3.2334,
      "step": 90000
    },
    {
      "epoch": 4.9,
      "learning_rate": 8.264305898975628e-05,
      "loss": 3.228,
      "step": 91000
    },
    {
      "epoch": 4.95,
      "learning_rate": 8.234869892852938e-05,
      "loss": 3.2289,
      "step": 92000
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.393233858051895,
      "eval_loss": 3.47290301322937,
      "eval_runtime": 154.1839,
      "eval_samples_per_second": 375.655,
      "eval_steps_per_second": 5.87,
      "step": 92930
    },
    {
      "epoch": 5.0,
      "learning_rate": 8.205433886730249e-05,
      "loss": 3.221,
      "step": 93000
    },
    {
      "epoch": 5.06,
      "learning_rate": 8.17599788060756e-05,
      "loss": 3.1688,
      "step": 94000
    },
    {
      "epoch": 5.11,
      "learning_rate": 8.146591310490993e-05,
      "loss": 3.1701,
      "step": 95000
    },
    {
      "epoch": 5.17,
      "learning_rate": 8.117155304368303e-05,
      "loss": 3.1741,
      "step": 96000
    },
    {
      "epoch": 5.22,
      "learning_rate": 8.087748734251737e-05,
      "loss": 3.1762,
      "step": 97000
    },
    {
      "epoch": 5.27,
      "learning_rate": 8.058342164135171e-05,
      "loss": 3.1771,
      "step": 98000
    },
    {
      "epoch": 5.33,
      "learning_rate": 8.028906158012482e-05,
      "loss": 3.1782,
      "step": 99000
    },
    {
      "epoch": 5.38,
      "learning_rate": 7.999470151889792e-05,
      "loss": 3.1799,
      "step": 100000
    },
    {
      "epoch": 5.43,
      "learning_rate": 7.970063581773224e-05,
      "loss": 3.1764,
      "step": 101000
    },
    {
      "epoch": 5.49,
      "learning_rate": 7.940627575650536e-05,
      "loss": 3.1783,
      "step": 102000
    },
    {
      "epoch": 5.54,
      "learning_rate": 7.91122100553397e-05,
      "loss": 3.1785,
      "step": 103000
    },
    {
      "epoch": 5.6,
      "learning_rate": 7.88178499941128e-05,
      "loss": 3.1769,
      "step": 104000
    },
    {
      "epoch": 5.65,
      "learning_rate": 7.852348993288591e-05,
      "loss": 3.1755,
      "step": 105000
    },
    {
      "epoch": 5.7,
      "learning_rate": 7.822912987165901e-05,
      "loss": 3.178,
      "step": 106000
    },
    {
      "epoch": 5.76,
      "learning_rate": 7.793506417049336e-05,
      "loss": 3.182,
      "step": 107000
    },
    {
      "epoch": 5.81,
      "learning_rate": 7.764070410926647e-05,
      "loss": 3.1761,
      "step": 108000
    },
    {
      "epoch": 5.86,
      "learning_rate": 7.734663840810079e-05,
      "loss": 3.1755,
      "step": 109000
    },
    {
      "epoch": 5.92,
      "learning_rate": 7.705227834687389e-05,
      "loss": 3.1752,
      "step": 110000
    },
    {
      "epoch": 5.97,
      "learning_rate": 7.675791828564701e-05,
      "loss": 3.176,
      "step": 111000
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.39765455371154634,
      "eval_loss": 3.4186227321624756,
      "eval_runtime": 154.3341,
      "eval_samples_per_second": 375.29,
      "eval_steps_per_second": 5.864,
      "step": 111516
    },
    {
      "epoch": 6.03,
      "learning_rate": 7.646385258448134e-05,
      "loss": 3.1505,
      "step": 112000
    },
    {
      "epoch": 6.08,
      "learning_rate": 7.616949252325445e-05,
      "loss": 3.1141,
      "step": 113000
    },
    {
      "epoch": 6.13,
      "learning_rate": 7.587513246202755e-05,
      "loss": 3.1226,
      "step": 114000
    },
    {
      "epoch": 6.19,
      "learning_rate": 7.558106676086189e-05,
      "loss": 3.1282,
      "step": 115000
    },
    {
      "epoch": 6.24,
      "learning_rate": 7.5286706699635e-05,
      "loss": 3.1249,
      "step": 116000
    },
    {
      "epoch": 6.3,
      "learning_rate": 7.499234663840811e-05,
      "loss": 3.1295,
      "step": 117000
    },
    {
      "epoch": 6.35,
      "learning_rate": 7.469798657718122e-05,
      "loss": 3.1289,
      "step": 118000
    },
    {
      "epoch": 6.4,
      "learning_rate": 7.440392087601554e-05,
      "loss": 3.1265,
      "step": 119000
    },
    {
      "epoch": 6.46,
      "learning_rate": 7.411014953491111e-05,
      "loss": 3.1321,
      "step": 120000
    },
    {
      "epoch": 6.51,
      "learning_rate": 7.381578947368422e-05,
      "loss": 3.1311,
      "step": 121000
    },
    {
      "epoch": 6.56,
      "learning_rate": 7.352142941245732e-05,
      "loss": 3.1348,
      "step": 122000
    },
    {
      "epoch": 6.62,
      "learning_rate": 7.322706935123043e-05,
      "loss": 3.1311,
      "step": 123000
    },
    {
      "epoch": 6.67,
      "learning_rate": 7.293300365006476e-05,
      "loss": 3.1353,
      "step": 124000
    },
    {
      "epoch": 6.73,
      "learning_rate": 7.263864358883787e-05,
      "loss": 3.1356,
      "step": 125000
    },
    {
      "epoch": 6.78,
      "learning_rate": 7.234428352761097e-05,
      "loss": 3.131,
      "step": 126000
    },
    {
      "epoch": 6.83,
      "learning_rate": 7.204992346638408e-05,
      "loss": 3.1327,
      "step": 127000
    },
    {
      "epoch": 6.89,
      "learning_rate": 7.175585776521841e-05,
      "loss": 3.1357,
      "step": 128000
    },
    {
      "epoch": 6.94,
      "learning_rate": 7.146149770399153e-05,
      "loss": 3.1342,
      "step": 129000
    },
    {
      "epoch": 6.99,
      "learning_rate": 7.116743200282586e-05,
      "loss": 3.1344,
      "step": 130000
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.39899452082034526,
      "eval_loss": 3.414994239807129,
      "eval_runtime": 153.8887,
      "eval_samples_per_second": 376.376,
      "eval_steps_per_second": 5.881,
      "step": 130102
    },
    {
      "epoch": 7.05,
      "learning_rate": 7.087307194159897e-05,
      "loss": 3.0824,
      "step": 131000
    },
    {
      "epoch": 7.1,
      "learning_rate": 7.057930060049452e-05,
      "loss": 3.0799,
      "step": 132000
    },
    {
      "epoch": 7.16,
      "learning_rate": 7.028494053926764e-05,
      "loss": 3.0821,
      "step": 133000
    },
    {
      "epoch": 7.21,
      "learning_rate": 6.999058047804074e-05,
      "loss": 3.086,
      "step": 134000
    },
    {
      "epoch": 7.26,
      "learning_rate": 6.969622041681385e-05,
      "loss": 3.0877,
      "step": 135000
    },
    {
      "epoch": 7.32,
      "learning_rate": 6.940186035558695e-05,
      "loss": 3.0904,
      "step": 136000
    },
    {
      "epoch": 7.37,
      "learning_rate": 6.910750029436006e-05,
      "loss": 3.0903,
      "step": 137000
    },
    {
      "epoch": 7.42,
      "learning_rate": 6.88134345931944e-05,
      "loss": 3.0914,
      "step": 138000
    },
    {
      "epoch": 7.48,
      "learning_rate": 6.851907453196751e-05,
      "loss": 3.0967,
      "step": 139000
    },
    {
      "epoch": 7.53,
      "learning_rate": 6.822500883080184e-05,
      "loss": 3.0948,
      "step": 140000
    },
    {
      "epoch": 7.59,
      "learning_rate": 6.793064876957495e-05,
      "loss": 3.096,
      "step": 141000
    },
    {
      "epoch": 7.64,
      "learning_rate": 6.763658306840927e-05,
      "loss": 3.0955,
      "step": 142000
    },
    {
      "epoch": 7.69,
      "learning_rate": 6.734222300718239e-05,
      "loss": 3.0921,
      "step": 143000
    },
    {
      "epoch": 7.75,
      "learning_rate": 6.704815730601672e-05,
      "loss": 3.0961,
      "step": 144000
    },
    {
      "epoch": 7.8,
      "learning_rate": 6.675379724478983e-05,
      "loss": 3.0984,
      "step": 145000
    },
    {
      "epoch": 7.86,
      "learning_rate": 6.645943718356293e-05,
      "loss": 3.0988,
      "step": 146000
    },
    {
      "epoch": 7.91,
      "learning_rate": 6.616537148239728e-05,
      "loss": 3.1037,
      "step": 147000
    },
    {
      "epoch": 7.96,
      "learning_rate": 6.587101142117039e-05,
      "loss": 3.0979,
      "step": 148000
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.40087367252832296,
      "eval_loss": 3.419072151184082,
      "eval_runtime": 154.6845,
      "eval_samples_per_second": 374.44,
      "eval_steps_per_second": 5.851,
      "step": 148688
    },
    {
      "epoch": 8.02,
      "learning_rate": 6.557665135994349e-05,
      "loss": 3.0805,
      "step": 149000
    },
    {
      "epoch": 8.07,
      "learning_rate": 6.528258565877781e-05,
      "loss": 3.0389,
      "step": 150000
    },
    {
      "epoch": 8.12,
      "learning_rate": 6.498822559755092e-05,
      "loss": 3.0492,
      "step": 151000
    },
    {
      "epoch": 8.18,
      "learning_rate": 6.469386553632404e-05,
      "loss": 3.0475,
      "step": 152000
    },
    {
      "epoch": 8.23,
      "learning_rate": 6.439979983515837e-05,
      "loss": 3.0489,
      "step": 153000
    },
    {
      "epoch": 8.29,
      "learning_rate": 6.410543977393147e-05,
      "loss": 3.0604,
      "step": 154000
    },
    {
      "epoch": 8.34,
      "learning_rate": 6.381137407276581e-05,
      "loss": 3.0577,
      "step": 155000
    },
    {
      "epoch": 8.39,
      "learning_rate": 6.351701401153893e-05,
      "loss": 3.062,
      "step": 156000
    },
    {
      "epoch": 8.45,
      "learning_rate": 6.322294831037325e-05,
      "loss": 3.0588,
      "step": 157000
    },
    {
      "epoch": 8.5,
      "learning_rate": 6.292858824914635e-05,
      "loss": 3.0608,
      "step": 158000
    },
    {
      "epoch": 8.55,
      "learning_rate": 6.263422818791946e-05,
      "loss": 3.0633,
      "step": 159000
    },
    {
      "epoch": 8.61,
      "learning_rate": 6.234016248675379e-05,
      "loss": 3.0647,
      "step": 160000
    },
    {
      "epoch": 8.66,
      "learning_rate": 6.204580242552691e-05,
      "loss": 3.0677,
      "step": 161000
    },
    {
      "epoch": 8.72,
      "learning_rate": 6.175173672436124e-05,
      "loss": 3.0643,
      "step": 162000
    },
    {
      "epoch": 8.77,
      "learning_rate": 6.145737666313435e-05,
      "loss": 3.0646,
      "step": 163000
    },
    {
      "epoch": 8.82,
      "learning_rate": 6.116301660190745e-05,
      "loss": 3.0663,
      "step": 164000
    },
    {
      "epoch": 8.88,
      "learning_rate": 6.0868950900741795e-05,
      "loss": 3.0662,
      "step": 165000
    },
    {
      "epoch": 8.93,
      "learning_rate": 6.05745908395149e-05,
      "loss": 3.0679,
      "step": 166000
    },
    {
      "epoch": 8.99,
      "learning_rate": 6.0280525138349234e-05,
      "loss": 3.0701,
      "step": 167000
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.4015922270352774,
      "eval_loss": 3.4137425422668457,
      "eval_runtime": 154.7841,
      "eval_samples_per_second": 374.199,
      "eval_steps_per_second": 5.847,
      "step": 167274
    },
    {
      "epoch": 9.04,
      "learning_rate": 5.998645943718356e-05,
      "loss": 3.0238,
      "step": 168000
    },
    {
      "epoch": 9.09,
      "learning_rate": 5.969209937595668e-05,
      "loss": 3.0168,
      "step": 169000
    },
    {
      "epoch": 9.15,
      "learning_rate": 5.9397739314729785e-05,
      "loss": 3.0225,
      "step": 170000
    },
    {
      "epoch": 9.2,
      "learning_rate": 5.910337925350289e-05,
      "loss": 3.0203,
      "step": 171000
    },
    {
      "epoch": 9.25,
      "learning_rate": 5.8809019192275995e-05,
      "loss": 3.0246,
      "step": 172000
    },
    {
      "epoch": 9.31,
      "learning_rate": 5.851495349111032e-05,
      "loss": 3.0252,
      "step": 173000
    },
    {
      "epoch": 9.36,
      "learning_rate": 5.822059342988344e-05,
      "loss": 3.031,
      "step": 174000
    },
    {
      "epoch": 9.42,
      "learning_rate": 5.7926527728717775e-05,
      "loss": 3.0336,
      "step": 175000
    },
    {
      "epoch": 9.47,
      "learning_rate": 5.763216766749088e-05,
      "loss": 3.0331,
      "step": 176000
    },
    {
      "epoch": 9.52,
      "learning_rate": 5.733810196632521e-05,
      "loss": 3.0401,
      "step": 177000
    },
    {
      "epoch": 9.58,
      "learning_rate": 5.704374190509831e-05,
      "loss": 3.0333,
      "step": 178000
    },
    {
      "epoch": 9.63,
      "learning_rate": 5.674938184387143e-05,
      "loss": 3.0338,
      "step": 179000
    },
    {
      "epoch": 9.68,
      "learning_rate": 5.645531614270576e-05,
      "loss": 3.0361,
      "step": 180000
    },
    {
      "epoch": 9.74,
      "learning_rate": 5.616095608147886e-05,
      "loss": 3.0346,
      "step": 181000
    },
    {
      "epoch": 9.79,
      "learning_rate": 5.58668903803132e-05,
      "loss": 3.0408,
      "step": 182000
    },
    {
      "epoch": 9.85,
      "learning_rate": 5.5572530319086316e-05,
      "loss": 3.0356,
      "step": 183000
    },
    {
      "epoch": 9.9,
      "learning_rate": 5.527817025785942e-05,
      "loss": 3.041,
      "step": 184000
    },
    {
      "epoch": 9.95,
      "learning_rate": 5.4983810196632527e-05,
      "loss": 3.0392,
      "step": 185000
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.40293656082722445,
      "eval_loss": 3.4200661182403564,
      "eval_runtime": 154.4914,
      "eval_samples_per_second": 374.908,
      "eval_steps_per_second": 5.858,
      "step": 185860
    },
    {
      "epoch": 10.01,
      "learning_rate": 5.4689744495466853e-05,
      "loss": 3.0355,
      "step": 186000
    },
    {
      "epoch": 10.06,
      "learning_rate": 5.4395678794301194e-05,
      "loss": 2.9831,
      "step": 187000
    },
    {
      "epoch": 10.12,
      "learning_rate": 5.41013187330743e-05,
      "loss": 2.9913,
      "step": 188000
    },
    {
      "epoch": 10.17,
      "learning_rate": 5.3806958671847404e-05,
      "loss": 2.9957,
      "step": 189000
    },
    {
      "epoch": 10.22,
      "learning_rate": 5.351259861062051e-05,
      "loss": 2.9974,
      "step": 190000
    },
    {
      "epoch": 10.28,
      "learning_rate": 5.3218238549393615e-05,
      "loss": 2.9992,
      "step": 191000
    },
    {
      "epoch": 10.33,
      "learning_rate": 5.292387848816672e-05,
      "loss": 3.0018,
      "step": 192000
    },
    {
      "epoch": 10.38,
      "learning_rate": 5.262981278700107e-05,
      "loss": 3.0046,
      "step": 193000
    },
    {
      "epoch": 10.44,
      "learning_rate": 5.233545272577417e-05,
      "loss": 3.007,
      "step": 194000
    },
    {
      "epoch": 10.49,
      "learning_rate": 5.20413870246085e-05,
      "loss": 3.0037,
      "step": 195000
    },
    {
      "epoch": 10.55,
      "learning_rate": 5.1747026963381605e-05,
      "loss": 3.0089,
      "step": 196000
    },
    {
      "epoch": 10.6,
      "learning_rate": 5.1452666902154724e-05,
      "loss": 3.0153,
      "step": 197000
    },
    {
      "epoch": 10.65,
      "learning_rate": 5.115889556105028e-05,
      "loss": 3.0095,
      "step": 198000
    },
    {
      "epoch": 10.71,
      "learning_rate": 5.0864535499823385e-05,
      "loss": 3.0102,
      "step": 199000
    },
    {
      "epoch": 10.76,
      "learning_rate": 5.057046979865772e-05,
      "loss": 3.0119,
      "step": 200000
    },
    {
      "epoch": 10.81,
      "learning_rate": 5.027610973743083e-05,
      "loss": 3.0145,
      "step": 201000
    },
    {
      "epoch": 10.87,
      "learning_rate": 4.9982044036265164e-05,
      "loss": 3.0158,
      "step": 202000
    },
    {
      "epoch": 10.92,
      "learning_rate": 4.968768397503827e-05,
      "loss": 3.0162,
      "step": 203000
    },
    {
      "epoch": 10.98,
      "learning_rate": 4.939332391381138e-05,
      "loss": 3.0154,
      "step": 204000
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.4038809736225466,
      "eval_loss": 3.4057369232177734,
      "eval_runtime": 154.5917,
      "eval_samples_per_second": 374.664,
      "eval_steps_per_second": 5.854,
      "step": 204446
    },
    {
      "epoch": 11.03,
      "learning_rate": 4.909896385258449e-05,
      "loss": 2.9868,
      "step": 205000
    },
    {
      "epoch": 11.08,
      "learning_rate": 4.880519251148005e-05,
      "loss": 2.9666,
      "step": 206000
    },
    {
      "epoch": 11.14,
      "learning_rate": 4.8510832450253154e-05,
      "loss": 2.9701,
      "step": 207000
    },
    {
      "epoch": 11.19,
      "learning_rate": 4.821647238902626e-05,
      "loss": 2.9688,
      "step": 208000
    },
    {
      "epoch": 11.25,
      "learning_rate": 4.7922112327799365e-05,
      "loss": 2.9722,
      "step": 209000
    },
    {
      "epoch": 11.3,
      "learning_rate": 4.762775226657247e-05,
      "loss": 2.9804,
      "step": 210000
    },
    {
      "epoch": 11.35,
      "learning_rate": 4.733368656540681e-05,
      "loss": 2.9799,
      "step": 211000
    },
    {
      "epoch": 11.41,
      "learning_rate": 4.7039620864241144e-05,
      "loss": 2.9816,
      "step": 212000
    },
    {
      "epoch": 11.46,
      "learning_rate": 4.674526080301425e-05,
      "loss": 2.9846,
      "step": 213000
    },
    {
      "epoch": 11.51,
      "learning_rate": 4.6450900741787355e-05,
      "loss": 2.9832,
      "step": 214000
    },
    {
      "epoch": 11.57,
      "learning_rate": 4.615654068056046e-05,
      "loss": 2.9891,
      "step": 215000
    },
    {
      "epoch": 11.62,
      "learning_rate": 4.58624749793948e-05,
      "loss": 2.9879,
      "step": 216000
    },
    {
      "epoch": 11.68,
      "learning_rate": 4.5568114918167906e-05,
      "loss": 2.9877,
      "step": 217000
    },
    {
      "epoch": 11.73,
      "learning_rate": 4.527375485694101e-05,
      "loss": 2.9923,
      "step": 218000
    },
    {
      "epoch": 11.78,
      "learning_rate": 4.4979394795714116e-05,
      "loss": 2.9913,
      "step": 219000
    },
    {
      "epoch": 11.84,
      "learning_rate": 4.468532909454846e-05,
      "loss": 2.9899,
      "step": 220000
    },
    {
      "epoch": 11.89,
      "learning_rate": 4.439096903332156e-05,
      "loss": 2.9956,
      "step": 221000
    },
    {
      "epoch": 11.94,
      "learning_rate": 4.4096903332155896e-05,
      "loss": 2.9935,
      "step": 222000
    },
    {
      "epoch": 12.0,
      "learning_rate": 4.3802543270929e-05,
      "loss": 2.9892,
      "step": 223000
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.40456110131779777,
      "eval_loss": 3.4152183532714844,
      "eval_runtime": 154.8814,
      "eval_samples_per_second": 373.964,
      "eval_steps_per_second": 5.843,
      "step": 223032
    },
    {
      "epoch": 12.05,
      "learning_rate": 4.3508183209702106e-05,
      "loss": 2.9416,
      "step": 224000
    },
    {
      "epoch": 12.11,
      "learning_rate": 4.321382314847522e-05,
      "loss": 2.9463,
      "step": 225000
    },
    {
      "epoch": 12.16,
      "learning_rate": 4.291975744730955e-05,
      "loss": 2.949,
      "step": 226000
    },
    {
      "epoch": 12.21,
      "learning_rate": 4.262539738608266e-05,
      "loss": 2.9534,
      "step": 227000
    },
    {
      "epoch": 12.27,
      "learning_rate": 4.233103732485576e-05,
      "loss": 2.9555,
      "step": 228000
    },
    {
      "epoch": 12.32,
      "learning_rate": 4.203667726362887e-05,
      "loss": 2.9579,
      "step": 229000
    },
    {
      "epoch": 12.37,
      "learning_rate": 4.174231720240198e-05,
      "loss": 2.959,
      "step": 230000
    },
    {
      "epoch": 12.43,
      "learning_rate": 4.1448251501236314e-05,
      "loss": 2.9637,
      "step": 231000
    },
    {
      "epoch": 12.48,
      "learning_rate": 4.1153891440009426e-05,
      "loss": 2.9632,
      "step": 232000
    },
    {
      "epoch": 12.54,
      "learning_rate": 4.085982573884375e-05,
      "loss": 2.9566,
      "step": 233000
    },
    {
      "epoch": 12.59,
      "learning_rate": 4.0565465677616865e-05,
      "loss": 2.969,
      "step": 234000
    },
    {
      "epoch": 12.64,
      "learning_rate": 4.02713999764512e-05,
      "loss": 2.9709,
      "step": 235000
    },
    {
      "epoch": 12.7,
      "learning_rate": 3.9977039915224304e-05,
      "loss": 2.9668,
      "step": 236000
    },
    {
      "epoch": 12.75,
      "learning_rate": 3.968267985399741e-05,
      "loss": 2.9688,
      "step": 237000
    },
    {
      "epoch": 12.81,
      "learning_rate": 3.9388319792770515e-05,
      "loss": 2.9711,
      "step": 238000
    },
    {
      "epoch": 12.86,
      "learning_rate": 3.909395973154363e-05,
      "loss": 2.9714,
      "step": 239000
    },
    {
      "epoch": 12.91,
      "learning_rate": 3.879989403037796e-05,
      "loss": 2.9712,
      "step": 240000
    },
    {
      "epoch": 12.97,
      "learning_rate": 3.850553396915107e-05,
      "loss": 2.9688,
      "step": 241000
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.40468968334157407,
      "eval_loss": 3.41487979888916,
      "eval_runtime": 154.8371,
      "eval_samples_per_second": 374.07,
      "eval_steps_per_second": 5.845,
      "step": 241618
    },
    {
      "epoch": 13.02,
      "learning_rate": 3.82114682679854e-05,
      "loss": 2.953,
      "step": 242000
    },
    {
      "epoch": 13.07,
      "learning_rate": 3.791710820675851e-05,
      "loss": 2.9247,
      "step": 243000
    },
    {
      "epoch": 13.13,
      "learning_rate": 3.762274814553162e-05,
      "loss": 2.9257,
      "step": 244000
    },
    {
      "epoch": 13.18,
      "learning_rate": 3.732838808430472e-05,
      "loss": 2.9324,
      "step": 245000
    },
    {
      "epoch": 13.24,
      "learning_rate": 3.7034616743200284e-05,
      "loss": 2.9346,
      "step": 246000
    },
    {
      "epoch": 13.29,
      "learning_rate": 3.674055104203462e-05,
      "loss": 2.939,
      "step": 247000
    },
    {
      "epoch": 13.34,
      "learning_rate": 3.644619098080772e-05,
      "loss": 2.9406,
      "step": 248000
    },
    {
      "epoch": 13.4,
      "learning_rate": 3.615183091958083e-05,
      "loss": 2.9345,
      "step": 249000
    },
    {
      "epoch": 13.45,
      "learning_rate": 3.585747085835394e-05,
      "loss": 2.9428,
      "step": 250000
    },
    {
      "epoch": 13.5,
      "learning_rate": 3.5563110797127046e-05,
      "loss": 2.9415,
      "step": 251000
    },
    {
      "epoch": 13.56,
      "learning_rate": 3.5269045095961386e-05,
      "loss": 2.9452,
      "step": 252000
    },
    {
      "epoch": 13.61,
      "learning_rate": 3.497468503473449e-05,
      "loss": 2.9481,
      "step": 253000
    },
    {
      "epoch": 13.67,
      "learning_rate": 3.4680619333568825e-05,
      "loss": 2.9436,
      "step": 254000
    },
    {
      "epoch": 13.72,
      "learning_rate": 3.438625927234193e-05,
      "loss": 2.9522,
      "step": 255000
    },
    {
      "epoch": 13.77,
      "learning_rate": 3.4091899211115036e-05,
      "loss": 2.9483,
      "step": 256000
    },
    {
      "epoch": 13.83,
      "learning_rate": 3.379753914988815e-05,
      "loss": 2.9487,
      "step": 257000
    },
    {
      "epoch": 13.88,
      "learning_rate": 3.3503473448722475e-05,
      "loss": 2.9497,
      "step": 258000
    },
    {
      "epoch": 13.94,
      "learning_rate": 3.320911338749559e-05,
      "loss": 2.9529,
      "step": 259000
    },
    {
      "epoch": 13.99,
      "learning_rate": 3.291475332626869e-05,
      "loss": 2.9542,
      "step": 260000
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.4050978002665692,
      "eval_loss": 3.4116663932800293,
      "eval_runtime": 154.7798,
      "eval_samples_per_second": 374.209,
      "eval_steps_per_second": 5.847,
      "step": 260204
    },
    {
      "epoch": 14.04,
      "learning_rate": 3.26203932650418e-05,
      "loss": 2.9081,
      "step": 261000
    },
    {
      "epoch": 14.1,
      "learning_rate": 3.232632756387614e-05,
      "loss": 2.907,
      "step": 262000
    },
    {
      "epoch": 14.15,
      "learning_rate": 3.203196750264924e-05,
      "loss": 2.9141,
      "step": 263000
    },
    {
      "epoch": 14.2,
      "learning_rate": 3.173790180148358e-05,
      "loss": 2.9199,
      "step": 264000
    },
    {
      "epoch": 14.26,
      "learning_rate": 3.144354174025668e-05,
      "loss": 2.9172,
      "step": 265000
    },
    {
      "epoch": 14.31,
      "learning_rate": 3.1149181679029794e-05,
      "loss": 2.9217,
      "step": 266000
    },
    {
      "epoch": 14.37,
      "learning_rate": 3.08548216178029e-05,
      "loss": 2.9203,
      "step": 267000
    },
    {
      "epoch": 14.42,
      "learning_rate": 3.0560461556576005e-05,
      "loss": 2.9242,
      "step": 268000
    },
    {
      "epoch": 14.47,
      "learning_rate": 3.026639585541034e-05,
      "loss": 2.9249,
      "step": 269000
    },
    {
      "epoch": 14.53,
      "learning_rate": 2.9972330154244672e-05,
      "loss": 2.9265,
      "step": 270000
    },
    {
      "epoch": 14.58,
      "learning_rate": 2.967797009301778e-05,
      "loss": 2.9256,
      "step": 271000
    },
    {
      "epoch": 14.63,
      "learning_rate": 2.9383610031790886e-05,
      "loss": 2.9258,
      "step": 272000
    },
    {
      "epoch": 14.69,
      "learning_rate": 2.9089544330625223e-05,
      "loss": 2.9281,
      "step": 273000
    },
    {
      "epoch": 14.74,
      "learning_rate": 2.879518426939833e-05,
      "loss": 2.9301,
      "step": 274000
    },
    {
      "epoch": 14.8,
      "learning_rate": 2.850082420817144e-05,
      "loss": 2.93,
      "step": 275000
    },
    {
      "epoch": 14.85,
      "learning_rate": 2.8207052867066996e-05,
      "loss": 2.9288,
      "step": 276000
    },
    {
      "epoch": 14.9,
      "learning_rate": 2.7912692805840108e-05,
      "loss": 2.9305,
      "step": 277000
    },
    {
      "epoch": 14.96,
      "learning_rate": 2.7618332744613213e-05,
      "loss": 2.9338,
      "step": 278000
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.4051717651611239,
      "eval_loss": 3.423478603363037,
      "eval_runtime": 154.6212,
      "eval_samples_per_second": 374.593,
      "eval_steps_per_second": 5.853,
      "step": 278790
    },
    {
      "epoch": 15.01,
      "learning_rate": 2.732397268338632e-05,
      "loss": 2.924,
      "step": 279000
    },
    {
      "epoch": 15.07,
      "learning_rate": 2.7029612622159427e-05,
      "loss": 2.8929,
      "step": 280000
    },
    {
      "epoch": 15.12,
      "learning_rate": 2.6735252560932533e-05,
      "loss": 2.8938,
      "step": 281000
    },
    {
      "epoch": 15.17,
      "learning_rate": 2.644118685976687e-05,
      "loss": 2.8924,
      "step": 282000
    },
    {
      "epoch": 15.23,
      "learning_rate": 2.6146826798539975e-05,
      "loss": 2.9013,
      "step": 283000
    },
    {
      "epoch": 15.28,
      "learning_rate": 2.585246673731308e-05,
      "loss": 2.9019,
      "step": 284000
    },
    {
      "epoch": 15.33,
      "learning_rate": 2.5558106676086193e-05,
      "loss": 2.9016,
      "step": 285000
    },
    {
      "epoch": 15.39,
      "learning_rate": 2.5264040974920523e-05,
      "loss": 2.9055,
      "step": 286000
    },
    {
      "epoch": 15.44,
      "learning_rate": 2.496968091369363e-05,
      "loss": 2.9097,
      "step": 287000
    },
    {
      "epoch": 15.5,
      "learning_rate": 2.4675615212527965e-05,
      "loss": 2.9043,
      "step": 288000
    },
    {
      "epoch": 15.55,
      "learning_rate": 2.438125515130107e-05,
      "loss": 2.9101,
      "step": 289000
    },
    {
      "epoch": 15.6,
      "learning_rate": 2.4087189450135404e-05,
      "loss": 2.9096,
      "step": 290000
    },
    {
      "epoch": 15.66,
      "learning_rate": 2.3792829388908513e-05,
      "loss": 2.9117,
      "step": 291000
    },
    {
      "epoch": 15.71,
      "learning_rate": 2.349846932768162e-05,
      "loss": 2.9111,
      "step": 292000
    },
    {
      "epoch": 15.76,
      "learning_rate": 2.3204403626515955e-05,
      "loss": 2.9152,
      "step": 293000
    },
    {
      "epoch": 15.82,
      "learning_rate": 2.2910043565289064e-05,
      "loss": 2.9101,
      "step": 294000
    },
    {
      "epoch": 15.87,
      "learning_rate": 2.2615683504062173e-05,
      "loss": 2.9091,
      "step": 295000
    },
    {
      "epoch": 15.93,
      "learning_rate": 2.2321617802896506e-05,
      "loss": 2.9124,
      "step": 296000
    },
    {
      "epoch": 15.98,
      "learning_rate": 2.202725774166961e-05,
      "loss": 2.9143,
      "step": 297000
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.40591625104800394,
      "eval_loss": 3.413038492202759,
      "eval_runtime": 154.4253,
      "eval_samples_per_second": 375.068,
      "eval_steps_per_second": 5.86,
      "step": 297376
    },
    {
      "epoch": 16.03,
      "learning_rate": 2.1732897680442717e-05,
      "loss": 2.8928,
      "step": 298000
    },
    {
      "epoch": 16.09,
      "learning_rate": 2.1438537619215826e-05,
      "loss": 2.876,
      "step": 299000
    },
    {
      "epoch": 16.14,
      "learning_rate": 2.1144766278111384e-05,
      "loss": 2.8803,
      "step": 300000
    },
    {
      "epoch": 16.19,
      "learning_rate": 2.0850406216884493e-05,
      "loss": 2.8851,
      "step": 301000
    },
    {
      "epoch": 16.25,
      "learning_rate": 2.05560461556576e-05,
      "loss": 2.8827,
      "step": 302000
    },
    {
      "epoch": 16.3,
      "learning_rate": 2.026168609443071e-05,
      "loss": 2.8884,
      "step": 303000
    },
    {
      "epoch": 16.36,
      "learning_rate": 1.9967326033203816e-05,
      "loss": 2.8884,
      "step": 304000
    },
    {
      "epoch": 16.41,
      "learning_rate": 1.9673554692099378e-05,
      "loss": 2.8903,
      "step": 305000
    },
    {
      "epoch": 16.46,
      "learning_rate": 1.9379194630872486e-05,
      "loss": 2.8856,
      "step": 306000
    },
    {
      "epoch": 16.52,
      "learning_rate": 1.908512892970682e-05,
      "loss": 2.889,
      "step": 307000
    },
    {
      "epoch": 16.57,
      "learning_rate": 1.8790768868479925e-05,
      "loss": 2.8894,
      "step": 308000
    },
    {
      "epoch": 16.63,
      "learning_rate": 1.849640880725303e-05,
      "loss": 2.8956,
      "step": 309000
    },
    {
      "epoch": 16.68,
      "learning_rate": 1.820204874602614e-05,
      "loss": 2.8938,
      "step": 310000
    },
    {
      "epoch": 16.73,
      "learning_rate": 1.7907983044860473e-05,
      "loss": 2.8925,
      "step": 311000
    },
    {
      "epoch": 16.79,
      "learning_rate": 1.7613917343694807e-05,
      "loss": 2.8977,
      "step": 312000
    },
    {
      "epoch": 16.84,
      "learning_rate": 1.7319557282467915e-05,
      "loss": 2.8926,
      "step": 313000
    },
    {
      "epoch": 16.89,
      "learning_rate": 1.702549158130225e-05,
      "loss": 2.8979,
      "step": 314000
    },
    {
      "epoch": 16.95,
      "learning_rate": 1.6731131520075358e-05,
      "loss": 2.8967,
      "step": 315000
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.40593438957800376,
      "eval_loss": 3.416489839553833,
      "eval_runtime": 154.3348,
      "eval_samples_per_second": 375.288,
      "eval_steps_per_second": 5.864,
      "step": 315962
    },
    {
      "epoch": 17.0,
      "learning_rate": 1.6436771458848466e-05,
      "loss": 2.9,
      "step": 316000
    },
    {
      "epoch": 17.06,
      "learning_rate": 1.6142411397621572e-05,
      "loss": 2.8652,
      "step": 317000
    },
    {
      "epoch": 17.11,
      "learning_rate": 1.5848345696455905e-05,
      "loss": 2.865,
      "step": 318000
    },
    {
      "epoch": 17.16,
      "learning_rate": 1.5553985635229014e-05,
      "loss": 2.8698,
      "step": 319000
    },
    {
      "epoch": 17.22,
      "learning_rate": 1.5259919934063348e-05,
      "loss": 2.8714,
      "step": 320000
    },
    {
      "epoch": 17.27,
      "learning_rate": 1.4965559872836455e-05,
      "loss": 2.8742,
      "step": 321000
    },
    {
      "epoch": 17.32,
      "learning_rate": 1.4671494171670788e-05,
      "loss": 2.8741,
      "step": 322000
    },
    {
      "epoch": 17.38,
      "learning_rate": 1.4377134110443895e-05,
      "loss": 2.8707,
      "step": 323000
    },
    {
      "epoch": 17.43,
      "learning_rate": 1.408306840927823e-05,
      "loss": 2.8746,
      "step": 324000
    },
    {
      "epoch": 17.49,
      "learning_rate": 1.3788708348051338e-05,
      "loss": 2.875,
      "step": 325000
    },
    {
      "epoch": 17.54,
      "learning_rate": 1.3494642646885671e-05,
      "loss": 2.8757,
      "step": 326000
    },
    {
      "epoch": 17.59,
      "learning_rate": 1.320028258565878e-05,
      "loss": 2.8735,
      "step": 327000
    },
    {
      "epoch": 17.65,
      "learning_rate": 1.2905922524431886e-05,
      "loss": 2.8775,
      "step": 328000
    },
    {
      "epoch": 17.7,
      "learning_rate": 1.2611562463204993e-05,
      "loss": 2.8785,
      "step": 329000
    },
    {
      "epoch": 17.76,
      "learning_rate": 1.2317496762039326e-05,
      "loss": 2.875,
      "step": 330000
    },
    {
      "epoch": 17.81,
      "learning_rate": 1.202343106087366e-05,
      "loss": 2.8824,
      "step": 331000
    },
    {
      "epoch": 17.86,
      "learning_rate": 1.1729070999646769e-05,
      "loss": 2.8759,
      "step": 332000
    },
    {
      "epoch": 17.92,
      "learning_rate": 1.1435005298481102e-05,
      "loss": 2.8784,
      "step": 333000
    },
    {
      "epoch": 17.97,
      "learning_rate": 1.1140645237254211e-05,
      "loss": 2.8824,
      "step": 334000
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.40593082905174455,
      "eval_loss": 3.4298744201660156,
      "eval_runtime": 155.0508,
      "eval_samples_per_second": 373.555,
      "eval_steps_per_second": 5.837,
      "step": 334548
    },
    {
      "epoch": 18.02,
      "learning_rate": 1.0846285176027316e-05,
      "loss": 2.8675,
      "step": 335000
    },
    {
      "epoch": 18.08,
      "learning_rate": 1.055221947486165e-05,
      "loss": 2.8523,
      "step": 336000
    },
    {
      "epoch": 18.13,
      "learning_rate": 1.0257859413634759e-05,
      "loss": 2.8576,
      "step": 337000
    },
    {
      "epoch": 18.19,
      "learning_rate": 9.963499352407866e-06,
      "loss": 2.8585,
      "step": 338000
    },
    {
      "epoch": 18.24,
      "learning_rate": 9.669433651242201e-06,
      "loss": 2.8548,
      "step": 339000
    },
    {
      "epoch": 18.29,
      "learning_rate": 9.375073590015306e-06,
      "loss": 2.8582,
      "step": 340000
    },
    {
      "epoch": 18.35,
      "learning_rate": 9.080713528788415e-06,
      "loss": 2.8603,
      "step": 341000
    },
    {
      "epoch": 18.4,
      "learning_rate": 8.786353467561522e-06,
      "loss": 2.8625,
      "step": 342000
    },
    {
      "epoch": 18.45,
      "learning_rate": 8.492287766395856e-06,
      "loss": 2.8614,
      "step": 343000
    },
    {
      "epoch": 18.51,
      "learning_rate": 8.198222065230191e-06,
      "loss": 2.8586,
      "step": 344000
    },
    {
      "epoch": 18.56,
      "learning_rate": 7.903862004003296e-06,
      "loss": 2.8623,
      "step": 345000
    },
    {
      "epoch": 18.62,
      "learning_rate": 7.609501942776405e-06,
      "loss": 2.8674,
      "step": 346000
    },
    {
      "epoch": 18.67,
      "learning_rate": 7.315436241610739e-06,
      "loss": 2.865,
      "step": 347000
    },
    {
      "epoch": 18.72,
      "learning_rate": 7.021076180383846e-06,
      "loss": 2.8652,
      "step": 348000
    },
    {
      "epoch": 18.78,
      "learning_rate": 6.7267161191569535e-06,
      "loss": 2.8601,
      "step": 349000
    },
    {
      "epoch": 18.83,
      "learning_rate": 6.43235605793006e-06,
      "loss": 2.8656,
      "step": 350000
    },
    {
      "epoch": 18.89,
      "learning_rate": 6.138290356764394e-06,
      "loss": 2.8614,
      "step": 351000
    },
    {
      "epoch": 18.94,
      "learning_rate": 5.843930295537501e-06,
      "loss": 2.8603,
      "step": 352000
    },
    {
      "epoch": 18.99,
      "learning_rate": 5.549570234310609e-06,
      "loss": 2.863,
      "step": 353000
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.4061107363974461,
      "eval_loss": 3.431204080581665,
      "eval_runtime": 154.248,
      "eval_samples_per_second": 375.499,
      "eval_steps_per_second": 5.867,
      "step": 353134
    },
    {
      "epoch": 19.05,
      "learning_rate": 5.255210173083716e-06,
      "loss": 2.8509,
      "step": 354000
    },
    {
      "epoch": 19.1,
      "learning_rate": 4.961144471918051e-06,
      "loss": 2.8452,
      "step": 355000
    },
    {
      "epoch": 19.15,
      "learning_rate": 4.667078770752384e-06,
      "loss": 2.8463,
      "step": 356000
    },
    {
      "epoch": 19.21,
      "learning_rate": 4.372718709525492e-06,
      "loss": 2.8453,
      "step": 357000
    },
    {
      "epoch": 19.26,
      "learning_rate": 4.078358648298599e-06,
      "loss": 2.8513,
      "step": 358000
    },
    {
      "epoch": 19.32,
      "learning_rate": 3.783998587071706e-06,
      "loss": 2.8505,
      "step": 359000
    },
    {
      "epoch": 19.37,
      "learning_rate": 3.4896385258448136e-06,
      "loss": 2.8512,
      "step": 360000
    },
    {
      "epoch": 19.42,
      "learning_rate": 3.1955728246791473e-06,
      "loss": 2.8498,
      "step": 361000
    },
    {
      "epoch": 19.48,
      "learning_rate": 2.901212763452255e-06,
      "loss": 2.8496,
      "step": 362000
    },
    {
      "epoch": 19.53,
      "learning_rate": 2.606852702225362e-06,
      "loss": 2.8445,
      "step": 363000
    },
    {
      "epoch": 19.58,
      "learning_rate": 2.312787001059696e-06,
      "loss": 2.8434,
      "step": 364000
    },
    {
      "epoch": 19.64,
      "learning_rate": 2.0184269398328037e-06,
      "loss": 2.8488,
      "step": 365000
    },
    {
      "epoch": 19.69,
      "learning_rate": 1.7243612386671377e-06,
      "loss": 2.8486,
      "step": 366000
    },
    {
      "epoch": 19.75,
      "learning_rate": 1.4302955375014718e-06,
      "loss": 2.8488,
      "step": 367000
    },
    {
      "epoch": 19.8,
      "learning_rate": 1.135935476274579e-06,
      "loss": 2.8485,
      "step": 368000
    },
    {
      "epoch": 19.85,
      "learning_rate": 8.418697751089133e-07,
      "loss": 2.8454,
      "step": 369000
    },
    {
      "epoch": 19.91,
      "learning_rate": 5.475097138820205e-07,
      "loss": 2.8519,
      "step": 370000
    },
    {
      "epoch": 19.96,
      "learning_rate": 2.5314965265512776e-07,
      "loss": 2.8521,
      "step": 371000
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.40612524722144594,
      "eval_loss": 3.434004545211792,
      "eval_runtime": 154.5618,
      "eval_samples_per_second": 374.737,
      "eval_steps_per_second": 5.855,
      "step": 371720
    },
    {
      "epoch": 20.0,
      "step": 371720,
      "total_flos": 1.56614628708864e+18,
      "train_loss": 3.159074606280223,
      "train_runtime": 81093.9235,
      "train_samples_per_second": 146.679,
      "train_steps_per_second": 4.584
    }
  ],
  "logging_steps": 1000,
  "max_steps": 371720,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "save_steps": 5000,
  "total_flos": 1.56614628708864e+18,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}