{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 20.0,
  "eval_steps": 500,
  "global_step": 372020,
  "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": 7.5393,
      "step": 1000
    },
    {
      "epoch": 0.11,
      "learning_rate": 6.25e-06,
      "loss": 5.8087,
      "step": 2000
    },
    {
      "epoch": 0.16,
      "learning_rate": 9.375000000000001e-06,
      "loss": 5.3765,
      "step": 3000
    },
    {
      "epoch": 0.22,
      "learning_rate": 1.25e-05,
      "loss": 5.1483,
      "step": 4000
    },
    {
      "epoch": 0.27,
      "learning_rate": 1.5625e-05,
      "loss": 5.0077,
      "step": 5000
    },
    {
      "epoch": 0.32,
      "learning_rate": 1.8750000000000002e-05,
      "loss": 4.8717,
      "step": 6000
    },
    {
      "epoch": 0.38,
      "learning_rate": 2.1875e-05,
      "loss": 4.7466,
      "step": 7000
    },
    {
      "epoch": 0.43,
      "learning_rate": 2.5e-05,
      "loss": 4.6453,
      "step": 8000
    },
    {
      "epoch": 0.48,
      "learning_rate": 2.8125000000000003e-05,
      "loss": 4.5496,
      "step": 9000
    },
    {
      "epoch": 0.54,
      "learning_rate": 3.125e-05,
      "loss": 4.476,
      "step": 10000
    },
    {
      "epoch": 0.59,
      "learning_rate": 3.4371875e-05,
      "loss": 4.4003,
      "step": 11000
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.7496875e-05,
      "loss": 4.3406,
      "step": 12000
    },
    {
      "epoch": 0.7,
      "learning_rate": 4.0621875e-05,
      "loss": 4.2829,
      "step": 13000
    },
    {
      "epoch": 0.75,
      "learning_rate": 4.374375e-05,
      "loss": 4.2271,
      "step": 14000
    },
    {
      "epoch": 0.81,
      "learning_rate": 4.686875e-05,
      "loss": 4.18,
      "step": 15000
    },
    {
      "epoch": 0.86,
      "learning_rate": 4.999375e-05,
      "loss": 4.1378,
      "step": 16000
    },
    {
      "epoch": 0.91,
      "learning_rate": 5.3115625000000005e-05,
      "loss": 4.0856,
      "step": 17000
    },
    {
      "epoch": 0.97,
      "learning_rate": 5.6240625e-05,
      "loss": 4.0459,
      "step": 18000
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.311928837844229,
      "eval_loss": 4.251152515411377,
      "eval_runtime": 149.1416,
      "eval_samples_per_second": 388.356,
      "eval_steps_per_second": 6.068,
      "step": 18601
    },
    {
      "epoch": 1.02,
      "learning_rate": 5.93625e-05,
      "loss": 4.0009,
      "step": 19000
    },
    {
      "epoch": 1.08,
      "learning_rate": 6.24875e-05,
      "loss": 3.949,
      "step": 20000
    },
    {
      "epoch": 1.13,
      "learning_rate": 6.560937500000001e-05,
      "loss": 3.9066,
      "step": 21000
    },
    {
      "epoch": 1.18,
      "learning_rate": 6.873125000000001e-05,
      "loss": 3.8743,
      "step": 22000
    },
    {
      "epoch": 1.24,
      "learning_rate": 7.185625e-05,
      "loss": 3.8439,
      "step": 23000
    },
    {
      "epoch": 1.29,
      "learning_rate": 7.498125e-05,
      "loss": 3.8201,
      "step": 24000
    },
    {
      "epoch": 1.34,
      "learning_rate": 7.8103125e-05,
      "loss": 3.794,
      "step": 25000
    },
    {
      "epoch": 1.4,
      "learning_rate": 8.1228125e-05,
      "loss": 3.7704,
      "step": 26000
    },
    {
      "epoch": 1.45,
      "learning_rate": 8.435e-05,
      "loss": 3.758,
      "step": 27000
    },
    {
      "epoch": 1.51,
      "learning_rate": 8.747500000000001e-05,
      "loss": 3.7289,
      "step": 28000
    },
    {
      "epoch": 1.56,
      "learning_rate": 9.0596875e-05,
      "loss": 3.7091,
      "step": 29000
    },
    {
      "epoch": 1.61,
      "learning_rate": 9.3721875e-05,
      "loss": 3.6833,
      "step": 30000
    },
    {
      "epoch": 1.67,
      "learning_rate": 9.6846875e-05,
      "loss": 3.6705,
      "step": 31000
    },
    {
      "epoch": 1.72,
      "learning_rate": 9.996875e-05,
      "loss": 3.6547,
      "step": 32000
    },
    {
      "epoch": 1.77,
      "learning_rate": 9.970884065643197e-05,
      "loss": 3.6393,
      "step": 33000
    },
    {
      "epoch": 1.83,
      "learning_rate": 9.941474030939357e-05,
      "loss": 3.6141,
      "step": 34000
    },
    {
      "epoch": 1.88,
      "learning_rate": 9.91209340627022e-05,
      "loss": 3.6059,
      "step": 35000
    },
    {
      "epoch": 1.94,
      "learning_rate": 9.882683371566379e-05,
      "loss": 3.5886,
      "step": 36000
    },
    {
      "epoch": 1.99,
      "learning_rate": 9.853302746897242e-05,
      "loss": 3.5647,
      "step": 37000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.3623092767160393,
      "eval_loss": 3.735292434692383,
      "eval_runtime": 148.561,
      "eval_samples_per_second": 389.874,
      "eval_steps_per_second": 6.092,
      "step": 37202
    },
    {
      "epoch": 2.04,
      "learning_rate": 9.823892712193401e-05,
      "loss": 3.5298,
      "step": 38000
    },
    {
      "epoch": 2.1,
      "learning_rate": 9.79448267748956e-05,
      "loss": 3.5171,
      "step": 39000
    },
    {
      "epoch": 2.15,
      "learning_rate": 9.765102052820424e-05,
      "loss": 3.5025,
      "step": 40000
    },
    {
      "epoch": 2.2,
      "learning_rate": 9.735692018116582e-05,
      "loss": 3.4981,
      "step": 41000
    },
    {
      "epoch": 2.26,
      "learning_rate": 9.706281983412741e-05,
      "loss": 3.4842,
      "step": 42000
    },
    {
      "epoch": 2.31,
      "learning_rate": 9.676930768778307e-05,
      "loss": 3.4806,
      "step": 43000
    },
    {
      "epoch": 2.37,
      "learning_rate": 9.647520734074467e-05,
      "loss": 3.4678,
      "step": 44000
    },
    {
      "epoch": 2.42,
      "learning_rate": 9.618110699370626e-05,
      "loss": 3.4584,
      "step": 45000
    },
    {
      "epoch": 2.47,
      "learning_rate": 9.588700664666785e-05,
      "loss": 3.4574,
      "step": 46000
    },
    {
      "epoch": 2.53,
      "learning_rate": 9.559320039997648e-05,
      "loss": 3.4439,
      "step": 47000
    },
    {
      "epoch": 2.58,
      "learning_rate": 9.529939415328511e-05,
      "loss": 3.4414,
      "step": 48000
    },
    {
      "epoch": 2.63,
      "learning_rate": 9.50052938062467e-05,
      "loss": 3.4277,
      "step": 49000
    },
    {
      "epoch": 2.69,
      "learning_rate": 9.471148755955532e-05,
      "loss": 3.4244,
      "step": 50000
    },
    {
      "epoch": 2.74,
      "learning_rate": 9.441738721251692e-05,
      "loss": 3.4171,
      "step": 51000
    },
    {
      "epoch": 2.8,
      "learning_rate": 9.412358096582554e-05,
      "loss": 3.4125,
      "step": 52000
    },
    {
      "epoch": 2.85,
      "learning_rate": 9.382948061878713e-05,
      "loss": 3.403,
      "step": 53000
    },
    {
      "epoch": 2.9,
      "learning_rate": 9.353538027174873e-05,
      "loss": 3.3939,
      "step": 54000
    },
    {
      "epoch": 2.96,
      "learning_rate": 9.324127992471031e-05,
      "loss": 3.3872,
      "step": 55000
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.3792640996839865,
      "eval_loss": 3.5880677700042725,
      "eval_runtime": 148.7713,
      "eval_samples_per_second": 389.322,
      "eval_steps_per_second": 6.083,
      "step": 55803
    },
    {
      "epoch": 3.01,
      "learning_rate": 9.294747367801894e-05,
      "loss": 3.3801,
      "step": 56000
    },
    {
      "epoch": 3.06,
      "learning_rate": 9.265337333098053e-05,
      "loss": 3.3386,
      "step": 57000
    },
    {
      "epoch": 3.12,
      "learning_rate": 9.235956708428917e-05,
      "loss": 3.3305,
      "step": 58000
    },
    {
      "epoch": 3.17,
      "learning_rate": 9.206576083759779e-05,
      "loss": 3.3335,
      "step": 59000
    },
    {
      "epoch": 3.23,
      "learning_rate": 9.177166049055938e-05,
      "loss": 3.3319,
      "step": 60000
    },
    {
      "epoch": 3.28,
      "learning_rate": 9.147756014352097e-05,
      "loss": 3.3238,
      "step": 61000
    },
    {
      "epoch": 3.33,
      "learning_rate": 9.118345979648256e-05,
      "loss": 3.3269,
      "step": 62000
    },
    {
      "epoch": 3.39,
      "learning_rate": 9.088965354979119e-05,
      "loss": 3.3222,
      "step": 63000
    },
    {
      "epoch": 3.44,
      "learning_rate": 9.059584730309983e-05,
      "loss": 3.3139,
      "step": 64000
    },
    {
      "epoch": 3.49,
      "learning_rate": 9.030174695606142e-05,
      "loss": 3.3167,
      "step": 65000
    },
    {
      "epoch": 3.55,
      "learning_rate": 9.0007646609023e-05,
      "loss": 3.3146,
      "step": 66000
    },
    {
      "epoch": 3.6,
      "learning_rate": 8.971384036233164e-05,
      "loss": 3.3048,
      "step": 67000
    },
    {
      "epoch": 3.66,
      "learning_rate": 8.941974001529322e-05,
      "loss": 3.3056,
      "step": 68000
    },
    {
      "epoch": 3.71,
      "learning_rate": 8.912563966825482e-05,
      "loss": 3.3045,
      "step": 69000
    },
    {
      "epoch": 3.76,
      "learning_rate": 8.88315393212164e-05,
      "loss": 3.2999,
      "step": 70000
    },
    {
      "epoch": 3.82,
      "learning_rate": 8.853773307452503e-05,
      "loss": 3.2956,
      "step": 71000
    },
    {
      "epoch": 3.87,
      "learning_rate": 8.824363272748663e-05,
      "loss": 3.2906,
      "step": 72000
    },
    {
      "epoch": 3.92,
      "learning_rate": 8.794982648079525e-05,
      "loss": 3.2975,
      "step": 73000
    },
    {
      "epoch": 3.98,
      "learning_rate": 8.765602023410388e-05,
      "loss": 3.2888,
      "step": 74000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.3881653481522884,
      "eval_loss": 3.532650947570801,
      "eval_runtime": 149.2718,
      "eval_samples_per_second": 388.017,
      "eval_steps_per_second": 6.063,
      "step": 74404
    },
    {
      "epoch": 4.03,
      "learning_rate": 8.736191988706547e-05,
      "loss": 3.2528,
      "step": 75000
    },
    {
      "epoch": 4.09,
      "learning_rate": 8.706781954002707e-05,
      "loss": 3.2346,
      "step": 76000
    },
    {
      "epoch": 4.14,
      "learning_rate": 8.677371919298866e-05,
      "loss": 3.2292,
      "step": 77000
    },
    {
      "epoch": 4.19,
      "learning_rate": 8.647991294629728e-05,
      "loss": 3.2355,
      "step": 78000
    },
    {
      "epoch": 4.25,
      "learning_rate": 8.618581259925888e-05,
      "loss": 3.2388,
      "step": 79000
    },
    {
      "epoch": 4.3,
      "learning_rate": 8.589171225222045e-05,
      "loss": 3.2365,
      "step": 80000
    },
    {
      "epoch": 4.35,
      "learning_rate": 8.559761190518206e-05,
      "loss": 3.2357,
      "step": 81000
    },
    {
      "epoch": 4.41,
      "learning_rate": 8.530380565849068e-05,
      "loss": 3.2318,
      "step": 82000
    },
    {
      "epoch": 4.46,
      "learning_rate": 8.500970531145227e-05,
      "loss": 3.2359,
      "step": 83000
    },
    {
      "epoch": 4.52,
      "learning_rate": 8.471560496441387e-05,
      "loss": 3.2307,
      "step": 84000
    },
    {
      "epoch": 4.57,
      "learning_rate": 8.442179871772249e-05,
      "loss": 3.2319,
      "step": 85000
    },
    {
      "epoch": 4.62,
      "learning_rate": 8.412769837068408e-05,
      "loss": 3.2307,
      "step": 86000
    },
    {
      "epoch": 4.68,
      "learning_rate": 8.38338921239927e-05,
      "loss": 3.2261,
      "step": 87000
    },
    {
      "epoch": 4.73,
      "learning_rate": 8.354008587730134e-05,
      "loss": 3.2261,
      "step": 88000
    },
    {
      "epoch": 4.78,
      "learning_rate": 8.324598553026293e-05,
      "loss": 3.2257,
      "step": 89000
    },
    {
      "epoch": 4.84,
      "learning_rate": 8.295188518322452e-05,
      "loss": 3.2275,
      "step": 90000
    },
    {
      "epoch": 4.89,
      "learning_rate": 8.265778483618612e-05,
      "loss": 3.2228,
      "step": 91000
    },
    {
      "epoch": 4.95,
      "learning_rate": 8.236397858949474e-05,
      "loss": 3.2233,
      "step": 92000
    },
    {
      "epoch": 5.0,
      "learning_rate": 8.207017234280336e-05,
      "loss": 3.2221,
      "step": 93000
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.3930816287593783,
      "eval_loss": 3.4746081829071045,
      "eval_runtime": 149.0373,
      "eval_samples_per_second": 388.628,
      "eval_steps_per_second": 6.072,
      "step": 93005
    },
    {
      "epoch": 5.05,
      "learning_rate": 8.177607199576495e-05,
      "loss": 3.1624,
      "step": 94000
    },
    {
      "epoch": 5.11,
      "learning_rate": 8.148197164872655e-05,
      "loss": 3.1656,
      "step": 95000
    },
    {
      "epoch": 5.16,
      "learning_rate": 8.118787130168814e-05,
      "loss": 3.1654,
      "step": 96000
    },
    {
      "epoch": 5.21,
      "learning_rate": 8.089377095464973e-05,
      "loss": 3.1689,
      "step": 97000
    },
    {
      "epoch": 5.27,
      "learning_rate": 8.059996470795837e-05,
      "loss": 3.176,
      "step": 98000
    },
    {
      "epoch": 5.32,
      "learning_rate": 8.030586436091994e-05,
      "loss": 3.1761,
      "step": 99000
    },
    {
      "epoch": 5.38,
      "learning_rate": 8.001205811422858e-05,
      "loss": 3.1726,
      "step": 100000
    },
    {
      "epoch": 5.43,
      "learning_rate": 7.971795776719017e-05,
      "loss": 3.1727,
      "step": 101000
    },
    {
      "epoch": 5.48,
      "learning_rate": 7.94241515204988e-05,
      "loss": 3.1712,
      "step": 102000
    },
    {
      "epoch": 5.54,
      "learning_rate": 7.913005117346039e-05,
      "loss": 3.1718,
      "step": 103000
    },
    {
      "epoch": 5.59,
      "learning_rate": 7.883595082642198e-05,
      "loss": 3.1734,
      "step": 104000
    },
    {
      "epoch": 5.64,
      "learning_rate": 7.85421445797306e-05,
      "loss": 3.1744,
      "step": 105000
    },
    {
      "epoch": 5.7,
      "learning_rate": 7.82480442326922e-05,
      "loss": 3.1698,
      "step": 106000
    },
    {
      "epoch": 5.75,
      "learning_rate": 7.795423798600083e-05,
      "loss": 3.17,
      "step": 107000
    },
    {
      "epoch": 5.81,
      "learning_rate": 7.766013763896241e-05,
      "loss": 3.1727,
      "step": 108000
    },
    {
      "epoch": 5.86,
      "learning_rate": 7.736633139227105e-05,
      "loss": 3.1676,
      "step": 109000
    },
    {
      "epoch": 5.91,
      "learning_rate": 7.707223104523264e-05,
      "loss": 3.1727,
      "step": 110000
    },
    {
      "epoch": 5.97,
      "learning_rate": 7.677842479854128e-05,
      "loss": 3.1699,
      "step": 111000
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.3965298304920782,
      "eval_loss": 3.4427359104156494,
      "eval_runtime": 148.4478,
      "eval_samples_per_second": 390.171,
      "eval_steps_per_second": 6.096,
      "step": 111606
    },
    {
      "epoch": 6.02,
      "learning_rate": 7.648432445150285e-05,
      "loss": 3.1429,
      "step": 112000
    },
    {
      "epoch": 6.07,
      "learning_rate": 7.619022410446445e-05,
      "loss": 3.1111,
      "step": 113000
    },
    {
      "epoch": 6.13,
      "learning_rate": 7.589612375742604e-05,
      "loss": 3.1168,
      "step": 114000
    },
    {
      "epoch": 6.18,
      "learning_rate": 7.560202341038763e-05,
      "loss": 3.1164,
      "step": 115000
    },
    {
      "epoch": 6.24,
      "learning_rate": 7.530821716369626e-05,
      "loss": 3.1205,
      "step": 116000
    },
    {
      "epoch": 6.29,
      "learning_rate": 7.501441091700489e-05,
      "loss": 3.1255,
      "step": 117000
    },
    {
      "epoch": 6.34,
      "learning_rate": 7.472031056996648e-05,
      "loss": 3.1246,
      "step": 118000
    },
    {
      "epoch": 6.4,
      "learning_rate": 7.44265043232751e-05,
      "loss": 3.1239,
      "step": 119000
    },
    {
      "epoch": 6.45,
      "learning_rate": 7.41324039762367e-05,
      "loss": 3.1257,
      "step": 120000
    },
    {
      "epoch": 6.51,
      "learning_rate": 7.383830362919829e-05,
      "loss": 3.1243,
      "step": 121000
    },
    {
      "epoch": 6.56,
      "learning_rate": 7.354420328215988e-05,
      "loss": 3.1281,
      "step": 122000
    },
    {
      "epoch": 6.61,
      "learning_rate": 7.325010293512146e-05,
      "loss": 3.1252,
      "step": 123000
    },
    {
      "epoch": 6.67,
      "learning_rate": 7.295629668843009e-05,
      "loss": 3.1271,
      "step": 124000
    },
    {
      "epoch": 6.72,
      "learning_rate": 7.266219634139169e-05,
      "loss": 3.1266,
      "step": 125000
    },
    {
      "epoch": 6.77,
      "learning_rate": 7.236809599435328e-05,
      "loss": 3.127,
      "step": 126000
    },
    {
      "epoch": 6.83,
      "learning_rate": 7.20742897476619e-05,
      "loss": 3.1273,
      "step": 127000
    },
    {
      "epoch": 6.88,
      "learning_rate": 7.17801894006235e-05,
      "loss": 3.1282,
      "step": 128000
    },
    {
      "epoch": 6.94,
      "learning_rate": 7.148638315393212e-05,
      "loss": 3.1269,
      "step": 129000
    },
    {
      "epoch": 6.99,
      "learning_rate": 7.119228280689371e-05,
      "loss": 3.1314,
      "step": 130000
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.3991414429133435,
      "eval_loss": 3.4234976768493652,
      "eval_runtime": 148.917,
      "eval_samples_per_second": 388.941,
      "eval_steps_per_second": 6.077,
      "step": 130207
    },
    {
      "epoch": 7.04,
      "learning_rate": 7.089818245985531e-05,
      "loss": 3.0765,
      "step": 131000
    },
    {
      "epoch": 7.1,
      "learning_rate": 7.060467031351097e-05,
      "loss": 3.072,
      "step": 132000
    },
    {
      "epoch": 7.15,
      "learning_rate": 7.031056996647256e-05,
      "loss": 3.0734,
      "step": 133000
    },
    {
      "epoch": 7.2,
      "learning_rate": 7.001646961943415e-05,
      "loss": 3.0802,
      "step": 134000
    },
    {
      "epoch": 7.26,
      "learning_rate": 6.972236927239575e-05,
      "loss": 3.0808,
      "step": 135000
    },
    {
      "epoch": 7.31,
      "learning_rate": 6.942856302570437e-05,
      "loss": 3.0862,
      "step": 136000
    },
    {
      "epoch": 7.37,
      "learning_rate": 6.913446267866596e-05,
      "loss": 3.0842,
      "step": 137000
    },
    {
      "epoch": 7.42,
      "learning_rate": 6.884036233162755e-05,
      "loss": 3.0843,
      "step": 138000
    },
    {
      "epoch": 7.47,
      "learning_rate": 6.854626198458914e-05,
      "loss": 3.0856,
      "step": 139000
    },
    {
      "epoch": 7.53,
      "learning_rate": 6.825245573789777e-05,
      "loss": 3.0913,
      "step": 140000
    },
    {
      "epoch": 7.58,
      "learning_rate": 6.795835539085936e-05,
      "loss": 3.0894,
      "step": 141000
    },
    {
      "epoch": 7.63,
      "learning_rate": 6.766454914416799e-05,
      "loss": 3.0897,
      "step": 142000
    },
    {
      "epoch": 7.69,
      "learning_rate": 6.737074289747662e-05,
      "loss": 3.087,
      "step": 143000
    },
    {
      "epoch": 7.74,
      "learning_rate": 6.707664255043822e-05,
      "loss": 3.0885,
      "step": 144000
    },
    {
      "epoch": 7.8,
      "learning_rate": 6.67825422033998e-05,
      "loss": 3.089,
      "step": 145000
    },
    {
      "epoch": 7.85,
      "learning_rate": 6.648873595670844e-05,
      "loss": 3.0957,
      "step": 146000
    },
    {
      "epoch": 7.9,
      "learning_rate": 6.619463560967002e-05,
      "loss": 3.0895,
      "step": 147000
    },
    {
      "epoch": 7.96,
      "learning_rate": 6.590053526263161e-05,
      "loss": 3.0928,
      "step": 148000
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.40097370316228476,
      "eval_loss": 3.4091949462890625,
      "eval_runtime": 148.5017,
      "eval_samples_per_second": 390.029,
      "eval_steps_per_second": 6.094,
      "step": 148808
    },
    {
      "epoch": 8.01,
      "learning_rate": 6.560643491559321e-05,
      "loss": 3.0802,
      "step": 149000
    },
    {
      "epoch": 8.06,
      "learning_rate": 6.531262866890184e-05,
      "loss": 3.0332,
      "step": 150000
    },
    {
      "epoch": 8.12,
      "learning_rate": 6.501852832186342e-05,
      "loss": 3.0399,
      "step": 151000
    },
    {
      "epoch": 8.17,
      "learning_rate": 6.472472207517205e-05,
      "loss": 3.0454,
      "step": 152000
    },
    {
      "epoch": 8.23,
      "learning_rate": 6.443062172813365e-05,
      "loss": 3.0476,
      "step": 153000
    },
    {
      "epoch": 8.28,
      "learning_rate": 6.413652138109524e-05,
      "loss": 3.047,
      "step": 154000
    },
    {
      "epoch": 8.33,
      "learning_rate": 6.384271513440386e-05,
      "loss": 3.0517,
      "step": 155000
    },
    {
      "epoch": 8.39,
      "learning_rate": 6.354861478736546e-05,
      "loss": 3.0514,
      "step": 156000
    },
    {
      "epoch": 8.44,
      "learning_rate": 6.325480854067408e-05,
      "loss": 3.0516,
      "step": 157000
    },
    {
      "epoch": 8.49,
      "learning_rate": 6.296070819363567e-05,
      "loss": 3.0569,
      "step": 158000
    },
    {
      "epoch": 8.55,
      "learning_rate": 6.26669019469443e-05,
      "loss": 3.0594,
      "step": 159000
    },
    {
      "epoch": 8.6,
      "learning_rate": 6.23728015999059e-05,
      "loss": 3.0547,
      "step": 160000
    },
    {
      "epoch": 8.66,
      "learning_rate": 6.207870125286747e-05,
      "loss": 3.0574,
      "step": 161000
    },
    {
      "epoch": 8.71,
      "learning_rate": 6.178518910652314e-05,
      "loss": 3.0571,
      "step": 162000
    },
    {
      "epoch": 8.76,
      "learning_rate": 6.149108875948473e-05,
      "loss": 3.0583,
      "step": 163000
    },
    {
      "epoch": 8.82,
      "learning_rate": 6.119698841244633e-05,
      "loss": 3.0592,
      "step": 164000
    },
    {
      "epoch": 8.87,
      "learning_rate": 6.0902888065407914e-05,
      "loss": 3.0582,
      "step": 165000
    },
    {
      "epoch": 8.92,
      "learning_rate": 6.060908181871655e-05,
      "loss": 3.058,
      "step": 166000
    },
    {
      "epoch": 8.98,
      "learning_rate": 6.0314981471678146e-05,
      "loss": 3.0595,
      "step": 167000
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.40245810671367455,
      "eval_loss": 3.4073524475097656,
      "eval_runtime": 148.6704,
      "eval_samples_per_second": 389.587,
      "eval_steps_per_second": 6.087,
      "step": 167409
    },
    {
      "epoch": 9.03,
      "learning_rate": 6.002088112463973e-05,
      "loss": 3.029,
      "step": 168000
    },
    {
      "epoch": 9.09,
      "learning_rate": 5.9727074877948364e-05,
      "loss": 3.0066,
      "step": 169000
    },
    {
      "epoch": 9.14,
      "learning_rate": 5.9432974530909945e-05,
      "loss": 3.0108,
      "step": 170000
    },
    {
      "epoch": 9.19,
      "learning_rate": 5.913887418387154e-05,
      "loss": 3.0115,
      "step": 171000
    },
    {
      "epoch": 9.25,
      "learning_rate": 5.884506793718016e-05,
      "loss": 3.0159,
      "step": 172000
    },
    {
      "epoch": 9.3,
      "learning_rate": 5.855096759014176e-05,
      "loss": 3.0203,
      "step": 173000
    },
    {
      "epoch": 9.35,
      "learning_rate": 5.825716134345038e-05,
      "loss": 3.0182,
      "step": 174000
    },
    {
      "epoch": 9.41,
      "learning_rate": 5.7963060996411976e-05,
      "loss": 3.0229,
      "step": 175000
    },
    {
      "epoch": 9.46,
      "learning_rate": 5.76692547497206e-05,
      "loss": 3.0247,
      "step": 176000
    },
    {
      "epoch": 9.52,
      "learning_rate": 5.7375154402682194e-05,
      "loss": 3.0254,
      "step": 177000
    },
    {
      "epoch": 9.57,
      "learning_rate": 5.708105405564379e-05,
      "loss": 3.0298,
      "step": 178000
    },
    {
      "epoch": 9.62,
      "learning_rate": 5.6786953708605383e-05,
      "loss": 3.0312,
      "step": 179000
    },
    {
      "epoch": 9.68,
      "learning_rate": 5.649314746191401e-05,
      "loss": 3.0288,
      "step": 180000
    },
    {
      "epoch": 9.73,
      "learning_rate": 5.61990471148756e-05,
      "loss": 3.0296,
      "step": 181000
    },
    {
      "epoch": 9.78,
      "learning_rate": 5.590494676783718e-05,
      "loss": 3.0301,
      "step": 182000
    },
    {
      "epoch": 9.84,
      "learning_rate": 5.561114052114582e-05,
      "loss": 3.0331,
      "step": 183000
    },
    {
      "epoch": 9.89,
      "learning_rate": 5.531704017410741e-05,
      "loss": 3.0303,
      "step": 184000
    },
    {
      "epoch": 9.95,
      "learning_rate": 5.5022939827068995e-05,
      "loss": 3.0361,
      "step": 185000
    },
    {
      "epoch": 10.0,
      "learning_rate": 5.4729133580377625e-05,
      "loss": 3.0344,
      "step": 186000
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.40234524474923145,
      "eval_loss": 3.4222097396850586,
      "eval_runtime": 148.7314,
      "eval_samples_per_second": 389.427,
      "eval_steps_per_second": 6.085,
      "step": 186010
    },
    {
      "epoch": 10.05,
      "learning_rate": 5.443503323333921e-05,
      "loss": 2.9769,
      "step": 187000
    },
    {
      "epoch": 10.11,
      "learning_rate": 5.414122698664784e-05,
      "loss": 2.9789,
      "step": 188000
    },
    {
      "epoch": 10.16,
      "learning_rate": 5.384712663960944e-05,
      "loss": 2.9856,
      "step": 189000
    },
    {
      "epoch": 10.21,
      "learning_rate": 5.3553026292571026e-05,
      "loss": 2.991,
      "step": 190000
    },
    {
      "epoch": 10.27,
      "learning_rate": 5.3259220045879656e-05,
      "loss": 2.9899,
      "step": 191000
    },
    {
      "epoch": 10.32,
      "learning_rate": 5.296541379918829e-05,
      "loss": 2.9931,
      "step": 192000
    },
    {
      "epoch": 10.38,
      "learning_rate": 5.2671313452149874e-05,
      "loss": 2.9999,
      "step": 193000
    },
    {
      "epoch": 10.43,
      "learning_rate": 5.237721310511147e-05,
      "loss": 2.9974,
      "step": 194000
    },
    {
      "epoch": 10.48,
      "learning_rate": 5.2083112758073064e-05,
      "loss": 3.0014,
      "step": 195000
    },
    {
      "epoch": 10.54,
      "learning_rate": 5.178930651138169e-05,
      "loss": 2.9975,
      "step": 196000
    },
    {
      "epoch": 10.59,
      "learning_rate": 5.149520616434328e-05,
      "loss": 3.0047,
      "step": 197000
    },
    {
      "epoch": 10.64,
      "learning_rate": 5.120110581730486e-05,
      "loss": 3.0028,
      "step": 198000
    },
    {
      "epoch": 10.7,
      "learning_rate": 5.090700547026646e-05,
      "loss": 3.007,
      "step": 199000
    },
    {
      "epoch": 10.75,
      "learning_rate": 5.061319922357508e-05,
      "loss": 3.0077,
      "step": 200000
    },
    {
      "epoch": 10.81,
      "learning_rate": 5.0319098876536675e-05,
      "loss": 3.0085,
      "step": 201000
    },
    {
      "epoch": 10.86,
      "learning_rate": 5.002529262984531e-05,
      "loss": 3.0058,
      "step": 202000
    },
    {
      "epoch": 10.91,
      "learning_rate": 4.97311922828069e-05,
      "loss": 3.0082,
      "step": 203000
    },
    {
      "epoch": 10.97,
      "learning_rate": 4.943709193576849e-05,
      "loss": 3.0028,
      "step": 204000
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.40429614442031947,
      "eval_loss": 3.4034204483032227,
      "eval_runtime": 148.9465,
      "eval_samples_per_second": 388.864,
      "eval_steps_per_second": 6.076,
      "step": 204611
    },
    {
      "epoch": 11.02,
      "learning_rate": 4.914328568907712e-05,
      "loss": 2.9882,
      "step": 205000
    },
    {
      "epoch": 11.07,
      "learning_rate": 4.8849185342038706e-05,
      "loss": 2.957,
      "step": 206000
    },
    {
      "epoch": 11.13,
      "learning_rate": 4.8555084995000294e-05,
      "loss": 2.9639,
      "step": 207000
    },
    {
      "epoch": 11.18,
      "learning_rate": 4.8261278748308924e-05,
      "loss": 2.9612,
      "step": 208000
    },
    {
      "epoch": 11.24,
      "learning_rate": 4.796717840127052e-05,
      "loss": 2.9667,
      "step": 209000
    },
    {
      "epoch": 11.29,
      "learning_rate": 4.767337215457914e-05,
      "loss": 2.9693,
      "step": 210000
    },
    {
      "epoch": 11.34,
      "learning_rate": 4.737927180754074e-05,
      "loss": 2.9716,
      "step": 211000
    },
    {
      "epoch": 11.4,
      "learning_rate": 4.7085171460502325e-05,
      "loss": 2.9728,
      "step": 212000
    },
    {
      "epoch": 11.45,
      "learning_rate": 4.6791365213810955e-05,
      "loss": 2.975,
      "step": 213000
    },
    {
      "epoch": 11.5,
      "learning_rate": 4.649726486677254e-05,
      "loss": 2.9785,
      "step": 214000
    },
    {
      "epoch": 11.56,
      "learning_rate": 4.620316451973414e-05,
      "loss": 2.9788,
      "step": 215000
    },
    {
      "epoch": 11.61,
      "learning_rate": 4.590935827304276e-05,
      "loss": 2.9775,
      "step": 216000
    },
    {
      "epoch": 11.67,
      "learning_rate": 4.561555202635139e-05,
      "loss": 2.9811,
      "step": 217000
    },
    {
      "epoch": 11.72,
      "learning_rate": 4.532145167931298e-05,
      "loss": 2.9812,
      "step": 218000
    },
    {
      "epoch": 11.77,
      "learning_rate": 4.5027351332274574e-05,
      "loss": 2.9832,
      "step": 219000
    },
    {
      "epoch": 11.83,
      "learning_rate": 4.473325098523616e-05,
      "loss": 2.9856,
      "step": 220000
    },
    {
      "epoch": 11.88,
      "learning_rate": 4.443944473854479e-05,
      "loss": 2.9822,
      "step": 221000
    },
    {
      "epoch": 11.93,
      "learning_rate": 4.414534439150638e-05,
      "loss": 2.9838,
      "step": 222000
    },
    {
      "epoch": 11.99,
      "learning_rate": 4.3851538144815016e-05,
      "loss": 2.9831,
      "step": 223000
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.4042906356815788,
      "eval_loss": 3.4022037982940674,
      "eval_runtime": 148.5304,
      "eval_samples_per_second": 389.954,
      "eval_steps_per_second": 6.093,
      "step": 223212
    },
    {
      "epoch": 12.04,
      "learning_rate": 4.3557437797776604e-05,
      "loss": 2.9446,
      "step": 224000
    },
    {
      "epoch": 12.1,
      "learning_rate": 4.3263631551085234e-05,
      "loss": 2.9355,
      "step": 225000
    },
    {
      "epoch": 12.15,
      "learning_rate": 4.296953120404682e-05,
      "loss": 2.9396,
      "step": 226000
    },
    {
      "epoch": 12.2,
      "learning_rate": 4.267543085700842e-05,
      "loss": 2.9385,
      "step": 227000
    },
    {
      "epoch": 12.26,
      "learning_rate": 4.238162461031705e-05,
      "loss": 2.9467,
      "step": 228000
    },
    {
      "epoch": 12.31,
      "learning_rate": 4.2087524263278635e-05,
      "loss": 2.9519,
      "step": 229000
    },
    {
      "epoch": 12.36,
      "learning_rate": 4.179342391624022e-05,
      "loss": 2.9528,
      "step": 230000
    },
    {
      "epoch": 12.42,
      "learning_rate": 4.149961766954885e-05,
      "loss": 2.9494,
      "step": 231000
    },
    {
      "epoch": 12.47,
      "learning_rate": 4.120551732251045e-05,
      "loss": 2.9564,
      "step": 232000
    },
    {
      "epoch": 12.53,
      "learning_rate": 4.091171107581907e-05,
      "loss": 2.959,
      "step": 233000
    },
    {
      "epoch": 12.58,
      "learning_rate": 4.0617610728780666e-05,
      "loss": 2.9567,
      "step": 234000
    },
    {
      "epoch": 12.63,
      "learning_rate": 4.032380448208929e-05,
      "loss": 2.957,
      "step": 235000
    },
    {
      "epoch": 12.69,
      "learning_rate": 4.0029704135050884e-05,
      "loss": 2.9584,
      "step": 236000
    },
    {
      "epoch": 12.74,
      "learning_rate": 3.973589788835951e-05,
      "loss": 2.9599,
      "step": 237000
    },
    {
      "epoch": 12.8,
      "learning_rate": 3.94417975413211e-05,
      "loss": 2.9602,
      "step": 238000
    },
    {
      "epoch": 12.85,
      "learning_rate": 3.914769719428269e-05,
      "loss": 2.9628,
      "step": 239000
    },
    {
      "epoch": 12.9,
      "learning_rate": 3.885389094759132e-05,
      "loss": 2.9628,
      "step": 240000
    },
    {
      "epoch": 12.96,
      "learning_rate": 3.856008470089995e-05,
      "loss": 2.9626,
      "step": 241000
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.40541079067867664,
      "eval_loss": 3.405994415283203,
      "eval_runtime": 148.2628,
      "eval_samples_per_second": 390.658,
      "eval_steps_per_second": 6.104,
      "step": 241813
    },
    {
      "epoch": 13.01,
      "learning_rate": 3.826598435386154e-05,
      "loss": 2.9557,
      "step": 242000
    },
    {
      "epoch": 13.06,
      "learning_rate": 3.7971884006823126e-05,
      "loss": 2.9125,
      "step": 243000
    },
    {
      "epoch": 13.12,
      "learning_rate": 3.767778365978472e-05,
      "loss": 2.9231,
      "step": 244000
    },
    {
      "epoch": 13.17,
      "learning_rate": 3.738397741309335e-05,
      "loss": 2.9228,
      "step": 245000
    },
    {
      "epoch": 13.23,
      "learning_rate": 3.708987706605494e-05,
      "loss": 2.9258,
      "step": 246000
    },
    {
      "epoch": 13.28,
      "learning_rate": 3.6795776719016527e-05,
      "loss": 2.9261,
      "step": 247000
    },
    {
      "epoch": 13.33,
      "learning_rate": 3.6501970472325157e-05,
      "loss": 2.9313,
      "step": 248000
    },
    {
      "epoch": 13.39,
      "learning_rate": 3.6207870125286745e-05,
      "loss": 2.931,
      "step": 249000
    },
    {
      "epoch": 13.44,
      "learning_rate": 3.591376977824834e-05,
      "loss": 2.9389,
      "step": 250000
    },
    {
      "epoch": 13.49,
      "learning_rate": 3.561996353155697e-05,
      "loss": 2.9307,
      "step": 251000
    },
    {
      "epoch": 13.55,
      "learning_rate": 3.532586318451856e-05,
      "loss": 2.9358,
      "step": 252000
    },
    {
      "epoch": 13.6,
      "learning_rate": 3.503205693782719e-05,
      "loss": 2.939,
      "step": 253000
    },
    {
      "epoch": 13.66,
      "learning_rate": 3.473825069113582e-05,
      "loss": 2.9415,
      "step": 254000
    },
    {
      "epoch": 13.71,
      "learning_rate": 3.444415034409741e-05,
      "loss": 2.9364,
      "step": 255000
    },
    {
      "epoch": 13.76,
      "learning_rate": 3.4150049997059e-05,
      "loss": 2.9383,
      "step": 256000
    },
    {
      "epoch": 13.82,
      "learning_rate": 3.385594965002059e-05,
      "loss": 2.9436,
      "step": 257000
    },
    {
      "epoch": 13.87,
      "learning_rate": 3.3561849302982176e-05,
      "loss": 2.9437,
      "step": 258000
    },
    {
      "epoch": 13.92,
      "learning_rate": 3.326804305629081e-05,
      "loss": 2.9352,
      "step": 259000
    },
    {
      "epoch": 13.98,
      "learning_rate": 3.29739427092524e-05,
      "loss": 2.9442,
      "step": 260000
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.4059943139067438,
      "eval_loss": 3.400751829147339,
      "eval_runtime": 149.6101,
      "eval_samples_per_second": 387.14,
      "eval_steps_per_second": 6.049,
      "step": 260414
    },
    {
      "epoch": 14.03,
      "learning_rate": 3.268013646256103e-05,
      "loss": 2.9138,
      "step": 261000
    },
    {
      "epoch": 14.09,
      "learning_rate": 3.238603611552262e-05,
      "loss": 2.8972,
      "step": 262000
    },
    {
      "epoch": 14.14,
      "learning_rate": 3.2091935768484213e-05,
      "loss": 2.9006,
      "step": 263000
    },
    {
      "epoch": 14.19,
      "learning_rate": 3.179812952179284e-05,
      "loss": 2.9065,
      "step": 264000
    },
    {
      "epoch": 14.25,
      "learning_rate": 3.150402917475443e-05,
      "loss": 2.9111,
      "step": 265000
    },
    {
      "epoch": 14.3,
      "learning_rate": 3.120992882771602e-05,
      "loss": 2.9121,
      "step": 266000
    },
    {
      "epoch": 14.35,
      "learning_rate": 3.091612258102465e-05,
      "loss": 2.9129,
      "step": 267000
    },
    {
      "epoch": 14.41,
      "learning_rate": 3.062202223398624e-05,
      "loss": 2.9146,
      "step": 268000
    },
    {
      "epoch": 14.46,
      "learning_rate": 3.032792188694783e-05,
      "loss": 2.9164,
      "step": 269000
    },
    {
      "epoch": 14.52,
      "learning_rate": 3.0033821539909417e-05,
      "loss": 2.9158,
      "step": 270000
    },
    {
      "epoch": 14.57,
      "learning_rate": 2.974001529321805e-05,
      "loss": 2.9162,
      "step": 271000
    },
    {
      "epoch": 14.62,
      "learning_rate": 2.9445914946179638e-05,
      "loss": 2.9174,
      "step": 272000
    },
    {
      "epoch": 14.68,
      "learning_rate": 2.915181459914123e-05,
      "loss": 2.9214,
      "step": 273000
    },
    {
      "epoch": 14.73,
      "learning_rate": 2.8858008352449856e-05,
      "loss": 2.9206,
      "step": 274000
    },
    {
      "epoch": 14.78,
      "learning_rate": 2.856390800541145e-05,
      "loss": 2.9198,
      "step": 275000
    },
    {
      "epoch": 14.84,
      "learning_rate": 2.8270101758720074e-05,
      "loss": 2.9213,
      "step": 276000
    },
    {
      "epoch": 14.89,
      "learning_rate": 2.797600141168167e-05,
      "loss": 2.9212,
      "step": 277000
    },
    {
      "epoch": 14.95,
      "learning_rate": 2.7682195164990292e-05,
      "loss": 2.9243,
      "step": 278000
    },
    {
      "epoch": 15.0,
      "learning_rate": 2.7388094817951887e-05,
      "loss": 2.9257,
      "step": 279000
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.4065318861921448,
      "eval_loss": 3.4015512466430664,
      "eval_runtime": 148.8909,
      "eval_samples_per_second": 389.01,
      "eval_steps_per_second": 6.078,
      "step": 279015
    },
    {
      "epoch": 15.05,
      "learning_rate": 2.7093994470913475e-05,
      "loss": 2.8796,
      "step": 280000
    },
    {
      "epoch": 15.11,
      "learning_rate": 2.6800188224222105e-05,
      "loss": 2.8853,
      "step": 281000
    },
    {
      "epoch": 15.16,
      "learning_rate": 2.6506087877183693e-05,
      "loss": 2.89,
      "step": 282000
    },
    {
      "epoch": 15.21,
      "learning_rate": 2.6212281630492326e-05,
      "loss": 2.8875,
      "step": 283000
    },
    {
      "epoch": 15.27,
      "learning_rate": 2.5918181283453918e-05,
      "loss": 2.8879,
      "step": 284000
    },
    {
      "epoch": 15.32,
      "learning_rate": 2.5624375036762544e-05,
      "loss": 2.8914,
      "step": 285000
    },
    {
      "epoch": 15.38,
      "learning_rate": 2.5330274689724136e-05,
      "loss": 2.8944,
      "step": 286000
    },
    {
      "epoch": 15.43,
      "learning_rate": 2.5036174342685727e-05,
      "loss": 2.8939,
      "step": 287000
    },
    {
      "epoch": 15.48,
      "learning_rate": 2.4742073995647315e-05,
      "loss": 2.896,
      "step": 288000
    },
    {
      "epoch": 15.54,
      "learning_rate": 2.4448267748955945e-05,
      "loss": 2.9011,
      "step": 289000
    },
    {
      "epoch": 15.59,
      "learning_rate": 2.4154167401917536e-05,
      "loss": 2.9003,
      "step": 290000
    },
    {
      "epoch": 15.64,
      "learning_rate": 2.3860361155226163e-05,
      "loss": 2.9049,
      "step": 291000
    },
    {
      "epoch": 15.7,
      "learning_rate": 2.3566260808187754e-05,
      "loss": 2.9057,
      "step": 292000
    },
    {
      "epoch": 15.75,
      "learning_rate": 2.3272160461149346e-05,
      "loss": 2.9047,
      "step": 293000
    },
    {
      "epoch": 15.81,
      "learning_rate": 2.2978354214457976e-05,
      "loss": 2.9002,
      "step": 294000
    },
    {
      "epoch": 15.86,
      "learning_rate": 2.2684547967766602e-05,
      "loss": 2.9049,
      "step": 295000
    },
    {
      "epoch": 15.91,
      "learning_rate": 2.2390447620728194e-05,
      "loss": 2.9054,
      "step": 296000
    },
    {
      "epoch": 15.97,
      "learning_rate": 2.2096641374036824e-05,
      "loss": 2.909,
      "step": 297000
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.40646134746436785,
      "eval_loss": 3.403698205947876,
      "eval_runtime": 149.7316,
      "eval_samples_per_second": 386.826,
      "eval_steps_per_second": 6.044,
      "step": 297616
    },
    {
      "epoch": 16.02,
      "learning_rate": 2.1802541026998412e-05,
      "loss": 2.8931,
      "step": 298000
    },
    {
      "epoch": 16.07,
      "learning_rate": 2.1508440679960003e-05,
      "loss": 2.8729,
      "step": 299000
    },
    {
      "epoch": 16.13,
      "learning_rate": 2.1214340332921594e-05,
      "loss": 2.8687,
      "step": 300000
    },
    {
      "epoch": 16.18,
      "learning_rate": 2.0920239985883182e-05,
      "loss": 2.8725,
      "step": 301000
    },
    {
      "epoch": 16.24,
      "learning_rate": 2.0626433739191812e-05,
      "loss": 2.8742,
      "step": 302000
    },
    {
      "epoch": 16.29,
      "learning_rate": 2.0332627492500442e-05,
      "loss": 2.8779,
      "step": 303000
    },
    {
      "epoch": 16.34,
      "learning_rate": 2.0038527145462034e-05,
      "loss": 2.8811,
      "step": 304000
    },
    {
      "epoch": 16.4,
      "learning_rate": 1.974472089877066e-05,
      "loss": 2.8811,
      "step": 305000
    },
    {
      "epoch": 16.45,
      "learning_rate": 1.9450620551732252e-05,
      "loss": 2.8793,
      "step": 306000
    },
    {
      "epoch": 16.5,
      "learning_rate": 1.9156520204693843e-05,
      "loss": 2.8838,
      "step": 307000
    },
    {
      "epoch": 16.56,
      "learning_rate": 1.8862419857655434e-05,
      "loss": 2.8826,
      "step": 308000
    },
    {
      "epoch": 16.61,
      "learning_rate": 1.8568319510617026e-05,
      "loss": 2.8826,
      "step": 309000
    },
    {
      "epoch": 16.67,
      "learning_rate": 1.8274513263925652e-05,
      "loss": 2.8872,
      "step": 310000
    },
    {
      "epoch": 16.72,
      "learning_rate": 1.7980412916887244e-05,
      "loss": 2.8843,
      "step": 311000
    },
    {
      "epoch": 16.77,
      "learning_rate": 1.7686312569848835e-05,
      "loss": 2.8839,
      "step": 312000
    },
    {
      "epoch": 16.83,
      "learning_rate": 1.7392506323157462e-05,
      "loss": 2.8817,
      "step": 313000
    },
    {
      "epoch": 16.88,
      "learning_rate": 1.7098405976119053e-05,
      "loss": 2.8868,
      "step": 314000
    },
    {
      "epoch": 16.93,
      "learning_rate": 1.680459972942768e-05,
      "loss": 2.8839,
      "step": 315000
    },
    {
      "epoch": 16.99,
      "learning_rate": 1.651049938238927e-05,
      "loss": 2.8892,
      "step": 316000
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.4063379382806286,
      "eval_loss": 3.4124531745910645,
      "eval_runtime": 150.2283,
      "eval_samples_per_second": 385.547,
      "eval_steps_per_second": 6.024,
      "step": 316217
    },
    {
      "epoch": 17.04,
      "learning_rate": 1.6216399035350863e-05,
      "loss": 2.8636,
      "step": 317000
    },
    {
      "epoch": 17.1,
      "learning_rate": 1.5922298688312454e-05,
      "loss": 2.858,
      "step": 318000
    },
    {
      "epoch": 17.15,
      "learning_rate": 1.5628198341274045e-05,
      "loss": 2.8557,
      "step": 319000
    },
    {
      "epoch": 17.2,
      "learning_rate": 1.533468619492971e-05,
      "loss": 2.861,
      "step": 320000
    },
    {
      "epoch": 17.26,
      "learning_rate": 1.50405858478913e-05,
      "loss": 2.8635,
      "step": 321000
    },
    {
      "epoch": 17.31,
      "learning_rate": 1.4746485500852892e-05,
      "loss": 2.8637,
      "step": 322000
    },
    {
      "epoch": 17.36,
      "learning_rate": 1.4452385153814483e-05,
      "loss": 2.861,
      "step": 323000
    },
    {
      "epoch": 17.42,
      "learning_rate": 1.415857890712311e-05,
      "loss": 2.859,
      "step": 324000
    },
    {
      "epoch": 17.47,
      "learning_rate": 1.3864772660431741e-05,
      "loss": 2.8677,
      "step": 325000
    },
    {
      "epoch": 17.53,
      "learning_rate": 1.3570672313393331e-05,
      "loss": 2.8665,
      "step": 326000
    },
    {
      "epoch": 17.58,
      "learning_rate": 1.327686606670196e-05,
      "loss": 2.8679,
      "step": 327000
    },
    {
      "epoch": 17.63,
      "learning_rate": 1.298276571966355e-05,
      "loss": 2.8702,
      "step": 328000
    },
    {
      "epoch": 17.69,
      "learning_rate": 1.2688665372625142e-05,
      "loss": 2.8693,
      "step": 329000
    },
    {
      "epoch": 17.74,
      "learning_rate": 1.2394565025586732e-05,
      "loss": 2.8686,
      "step": 330000
    },
    {
      "epoch": 17.79,
      "learning_rate": 1.2100464678548321e-05,
      "loss": 2.87,
      "step": 331000
    },
    {
      "epoch": 17.85,
      "learning_rate": 1.180665843185695e-05,
      "loss": 2.8666,
      "step": 332000
    },
    {
      "epoch": 17.9,
      "learning_rate": 1.1512558084818541e-05,
      "loss": 2.8713,
      "step": 333000
    },
    {
      "epoch": 17.96,
      "learning_rate": 1.1218457737780132e-05,
      "loss": 2.872,
      "step": 334000
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.40656608068018146,
      "eval_loss": 3.413238286972046,
      "eval_runtime": 149.8799,
      "eval_samples_per_second": 386.443,
      "eval_steps_per_second": 6.038,
      "step": 334818
    },
    {
      "epoch": 18.01,
      "learning_rate": 1.0924651491088759e-05,
      "loss": 2.8666,
      "step": 335000
    },
    {
      "epoch": 18.06,
      "learning_rate": 1.063055114405035e-05,
      "loss": 2.8457,
      "step": 336000
    },
    {
      "epoch": 18.12,
      "learning_rate": 1.033674489735898e-05,
      "loss": 2.8431,
      "step": 337000
    },
    {
      "epoch": 18.17,
      "learning_rate": 1.004264455032057e-05,
      "loss": 2.8491,
      "step": 338000
    },
    {
      "epoch": 18.22,
      "learning_rate": 9.748838303629198e-06,
      "loss": 2.8476,
      "step": 339000
    },
    {
      "epoch": 18.28,
      "learning_rate": 9.45473795659079e-06,
      "loss": 2.8528,
      "step": 340000
    },
    {
      "epoch": 18.33,
      "learning_rate": 9.16063760955238e-06,
      "loss": 2.8486,
      "step": 341000
    },
    {
      "epoch": 18.39,
      "learning_rate": 8.866831362861008e-06,
      "loss": 2.8534,
      "step": 342000
    },
    {
      "epoch": 18.44,
      "learning_rate": 8.572731015822599e-06,
      "loss": 2.8534,
      "step": 343000
    },
    {
      "epoch": 18.49,
      "learning_rate": 8.27863066878419e-06,
      "loss": 2.8541,
      "step": 344000
    },
    {
      "epoch": 18.55,
      "learning_rate": 7.984824422092819e-06,
      "loss": 2.8533,
      "step": 345000
    },
    {
      "epoch": 18.6,
      "learning_rate": 7.690724075054408e-06,
      "loss": 2.852,
      "step": 346000
    },
    {
      "epoch": 18.65,
      "learning_rate": 7.396623728016e-06,
      "loss": 2.852,
      "step": 347000
    },
    {
      "epoch": 18.71,
      "learning_rate": 7.102523380977589e-06,
      "loss": 2.8528,
      "step": 348000
    },
    {
      "epoch": 18.76,
      "learning_rate": 6.808717134286219e-06,
      "loss": 2.8498,
      "step": 349000
    },
    {
      "epoch": 18.82,
      "learning_rate": 6.514910887594848e-06,
      "loss": 2.8542,
      "step": 350000
    },
    {
      "epoch": 18.87,
      "learning_rate": 6.220810540556438e-06,
      "loss": 2.8544,
      "step": 351000
    },
    {
      "epoch": 18.92,
      "learning_rate": 5.927004293865067e-06,
      "loss": 2.8527,
      "step": 352000
    },
    {
      "epoch": 18.98,
      "learning_rate": 5.632903946826658e-06,
      "loss": 2.8568,
      "step": 353000
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.4069465867317325,
      "eval_loss": 3.4157752990722656,
      "eval_runtime": 149.1511,
      "eval_samples_per_second": 388.331,
      "eval_steps_per_second": 6.068,
      "step": 353419
    },
    {
      "epoch": 19.03,
      "learning_rate": 5.3388035997882485e-06,
      "loss": 2.8422,
      "step": 354000
    },
    {
      "epoch": 19.08,
      "learning_rate": 5.044703252749838e-06,
      "loss": 2.8379,
      "step": 355000
    },
    {
      "epoch": 19.14,
      "learning_rate": 4.7506029057114295e-06,
      "loss": 2.8354,
      "step": 356000
    },
    {
      "epoch": 19.19,
      "learning_rate": 4.45650255867302e-06,
      "loss": 2.839,
      "step": 357000
    },
    {
      "epoch": 19.25,
      "learning_rate": 4.162990412328686e-06,
      "loss": 2.8412,
      "step": 358000
    },
    {
      "epoch": 19.3,
      "learning_rate": 3.8688900652902775e-06,
      "loss": 2.8385,
      "step": 359000
    },
    {
      "epoch": 19.35,
      "learning_rate": 3.575083818598906e-06,
      "loss": 2.8411,
      "step": 360000
    },
    {
      "epoch": 19.41,
      "learning_rate": 3.280983471560497e-06,
      "loss": 2.8399,
      "step": 361000
    },
    {
      "epoch": 19.46,
      "learning_rate": 2.9871772248691256e-06,
      "loss": 2.8385,
      "step": 362000
    },
    {
      "epoch": 19.52,
      "learning_rate": 2.693076877830716e-06,
      "loss": 2.8384,
      "step": 363000
    },
    {
      "epoch": 19.57,
      "learning_rate": 2.398976530792306e-06,
      "loss": 2.8395,
      "step": 364000
    },
    {
      "epoch": 19.62,
      "learning_rate": 2.1051702841009354e-06,
      "loss": 2.8384,
      "step": 365000
    },
    {
      "epoch": 19.68,
      "learning_rate": 1.8110699370625259e-06,
      "loss": 2.8395,
      "step": 366000
    },
    {
      "epoch": 19.73,
      "learning_rate": 1.5169695900241161e-06,
      "loss": 2.8366,
      "step": 367000
    },
    {
      "epoch": 19.78,
      "learning_rate": 1.2231633433327451e-06,
      "loss": 2.844,
      "step": 368000
    },
    {
      "epoch": 19.84,
      "learning_rate": 9.290629962943357e-07,
      "loss": 2.8368,
      "step": 369000
    },
    {
      "epoch": 19.89,
      "learning_rate": 6.349626492559261e-07,
      "loss": 2.8411,
      "step": 370000
    },
    {
      "epoch": 19.95,
      "learning_rate": 3.408623022175166e-07,
      "loss": 2.8385,
      "step": 371000
    },
    {
      "epoch": 20.0,
      "learning_rate": 4.6761955179107116e-08,
      "loss": 2.8385,
      "step": 372000
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.40685132585936323,
      "eval_loss": 3.420332908630371,
      "eval_runtime": 148.5396,
      "eval_samples_per_second": 389.93,
      "eval_steps_per_second": 6.093,
      "step": 372020
    },
    {
      "epoch": 20.0,
      "step": 372020,
      "total_flos": 1.56740238729216e+18,
      "train_loss": 3.1518569721931295,
      "train_runtime": 47850.3742,
      "train_samples_per_second": 248.782,
      "train_steps_per_second": 7.775
    }
  ],
  "logging_steps": 1000,
  "max_steps": 372020,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "save_steps": 5000,
  "total_flos": 1.56740238729216e+18,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}