{
  "best_metric": 1.9251400232315063,
  "best_model_checkpoint": "./model_tweets_2020_Q2_full/checkpoint-1216000",
  "epoch": 5.052631578947368,
  "eval_steps": 8000,
  "global_step": 2400000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "eval_loss": 2.222038507461548,
      "eval_runtime": 443.0178,
      "eval_samples_per_second": 902.898,
      "eval_steps_per_second": 56.431,
      "step": 8000
    },
    {
      "epoch": 0.03,
      "learning_rate": 4.0726666666666665e-07,
      "loss": 2.4154,
      "step": 16000
    },
    {
      "epoch": 0.03,
      "eval_loss": 2.142679214477539,
      "eval_runtime": 443.0058,
      "eval_samples_per_second": 902.923,
      "eval_steps_per_second": 56.433,
      "step": 16000
    },
    {
      "epoch": 0.05,
      "eval_loss": 2.1027772426605225,
      "eval_runtime": 442.7033,
      "eval_samples_per_second": 903.54,
      "eval_steps_per_second": 56.471,
      "step": 24000
    },
    {
      "epoch": 0.07,
      "learning_rate": 4.0453333333333336e-07,
      "loss": 2.2273,
      "step": 32000
    },
    {
      "epoch": 0.07,
      "eval_loss": 2.0823824405670166,
      "eval_runtime": 443.6086,
      "eval_samples_per_second": 901.696,
      "eval_steps_per_second": 56.356,
      "step": 32000
    },
    {
      "epoch": 0.08,
      "eval_loss": 2.0644640922546387,
      "eval_runtime": 443.4406,
      "eval_samples_per_second": 902.037,
      "eval_steps_per_second": 56.377,
      "step": 40000
    },
    {
      "epoch": 0.1,
      "learning_rate": 4.018e-07,
      "loss": 2.1774,
      "step": 48000
    },
    {
      "epoch": 0.1,
      "eval_loss": 2.0478036403656006,
      "eval_runtime": 444.714,
      "eval_samples_per_second": 899.454,
      "eval_steps_per_second": 56.216,
      "step": 48000
    },
    {
      "epoch": 0.12,
      "eval_loss": 2.0327365398406982,
      "eval_runtime": 445.1474,
      "eval_samples_per_second": 898.579,
      "eval_steps_per_second": 56.161,
      "step": 56000
    },
    {
      "epoch": 0.13,
      "learning_rate": 3.9906666666666667e-07,
      "loss": 2.1569,
      "step": 64000
    },
    {
      "epoch": 0.13,
      "eval_loss": 2.0247883796691895,
      "eval_runtime": 443.3798,
      "eval_samples_per_second": 902.161,
      "eval_steps_per_second": 56.385,
      "step": 64000
    },
    {
      "epoch": 0.15,
      "eval_loss": 2.020857334136963,
      "eval_runtime": 443.9842,
      "eval_samples_per_second": 900.933,
      "eval_steps_per_second": 56.308,
      "step": 72000
    },
    {
      "epoch": 0.17,
      "learning_rate": 3.963333333333333e-07,
      "loss": 2.1439,
      "step": 80000
    },
    {
      "epoch": 0.17,
      "eval_loss": 2.004868507385254,
      "eval_runtime": 446.7998,
      "eval_samples_per_second": 895.256,
      "eval_steps_per_second": 55.953,
      "step": 80000
    },
    {
      "epoch": 0.19,
      "eval_loss": 2.0112717151641846,
      "eval_runtime": 445.8029,
      "eval_samples_per_second": 897.257,
      "eval_steps_per_second": 56.079,
      "step": 88000
    },
    {
      "epoch": 0.2,
      "learning_rate": 3.936e-07,
      "loss": 2.1271,
      "step": 96000
    },
    {
      "epoch": 0.2,
      "eval_loss": 2.0037786960601807,
      "eval_runtime": 447.4745,
      "eval_samples_per_second": 893.906,
      "eval_steps_per_second": 55.869,
      "step": 96000
    },
    {
      "epoch": 0.22,
      "eval_loss": 2.0065057277679443,
      "eval_runtime": 449.3147,
      "eval_samples_per_second": 890.245,
      "eval_steps_per_second": 55.64,
      "step": 104000
    },
    {
      "epoch": 0.24,
      "learning_rate": 3.908666666666667e-07,
      "loss": 2.1211,
      "step": 112000
    },
    {
      "epoch": 0.24,
      "eval_loss": 1.998653769493103,
      "eval_runtime": 447.7781,
      "eval_samples_per_second": 893.3,
      "eval_steps_per_second": 55.831,
      "step": 112000
    },
    {
      "epoch": 0.25,
      "eval_loss": 1.9929168224334717,
      "eval_runtime": 444.7126,
      "eval_samples_per_second": 899.457,
      "eval_steps_per_second": 56.216,
      "step": 120000
    },
    {
      "epoch": 0.27,
      "learning_rate": 3.8813333333333334e-07,
      "loss": 2.1194,
      "step": 128000
    },
    {
      "epoch": 0.27,
      "eval_loss": 1.9922369718551636,
      "eval_runtime": 447.3304,
      "eval_samples_per_second": 894.194,
      "eval_steps_per_second": 55.887,
      "step": 128000
    },
    {
      "epoch": 0.29,
      "eval_loss": 1.9916871786117554,
      "eval_runtime": 444.9748,
      "eval_samples_per_second": 898.927,
      "eval_steps_per_second": 56.183,
      "step": 136000
    },
    {
      "epoch": 0.3,
      "learning_rate": 3.854e-07,
      "loss": 2.1118,
      "step": 144000
    },
    {
      "epoch": 0.3,
      "eval_loss": 1.9884706735610962,
      "eval_runtime": 444.0647,
      "eval_samples_per_second": 900.77,
      "eval_steps_per_second": 56.298,
      "step": 144000
    },
    {
      "epoch": 0.32,
      "eval_loss": 1.9870291948318481,
      "eval_runtime": 445.1345,
      "eval_samples_per_second": 898.605,
      "eval_steps_per_second": 56.163,
      "step": 152000
    },
    {
      "epoch": 0.34,
      "learning_rate": 3.8266666666666665e-07,
      "loss": 2.1047,
      "step": 160000
    },
    {
      "epoch": 0.34,
      "eval_loss": 1.9842537641525269,
      "eval_runtime": 444.6868,
      "eval_samples_per_second": 899.51,
      "eval_steps_per_second": 56.219,
      "step": 160000
    },
    {
      "epoch": 0.35,
      "eval_loss": 1.9826897382736206,
      "eval_runtime": 446.0995,
      "eval_samples_per_second": 896.661,
      "eval_steps_per_second": 56.041,
      "step": 168000
    },
    {
      "epoch": 0.37,
      "learning_rate": 3.799333333333333e-07,
      "loss": 2.1015,
      "step": 176000
    },
    {
      "epoch": 0.37,
      "eval_loss": 1.98258376121521,
      "eval_runtime": 447.5655,
      "eval_samples_per_second": 893.724,
      "eval_steps_per_second": 55.858,
      "step": 176000
    },
    {
      "epoch": 0.39,
      "eval_loss": 1.9773671627044678,
      "eval_runtime": 446.1508,
      "eval_samples_per_second": 896.558,
      "eval_steps_per_second": 56.035,
      "step": 184000
    },
    {
      "epoch": 0.4,
      "learning_rate": 3.772e-07,
      "loss": 2.1042,
      "step": 192000
    },
    {
      "epoch": 0.4,
      "eval_loss": 1.9771127700805664,
      "eval_runtime": 446.0195,
      "eval_samples_per_second": 896.822,
      "eval_steps_per_second": 56.051,
      "step": 192000
    },
    {
      "epoch": 0.42,
      "eval_loss": 1.9769738912582397,
      "eval_runtime": 447.4748,
      "eval_samples_per_second": 893.905,
      "eval_steps_per_second": 55.869,
      "step": 200000
    },
    {
      "epoch": 0.44,
      "learning_rate": 3.7446666666666667e-07,
      "loss": 2.0919,
      "step": 208000
    },
    {
      "epoch": 0.44,
      "eval_loss": 1.975176453590393,
      "eval_runtime": 445.3011,
      "eval_samples_per_second": 898.269,
      "eval_steps_per_second": 56.142,
      "step": 208000
    },
    {
      "epoch": 0.45,
      "eval_loss": 1.9774501323699951,
      "eval_runtime": 445.3222,
      "eval_samples_per_second": 898.226,
      "eval_steps_per_second": 56.139,
      "step": 216000
    },
    {
      "epoch": 0.47,
      "learning_rate": 3.7173333333333333e-07,
      "loss": 2.0953,
      "step": 224000
    },
    {
      "epoch": 0.47,
      "eval_loss": 1.9684234857559204,
      "eval_runtime": 445.4157,
      "eval_samples_per_second": 898.038,
      "eval_steps_per_second": 56.127,
      "step": 224000
    },
    {
      "epoch": 0.49,
      "eval_loss": 1.97478187084198,
      "eval_runtime": 446.0926,
      "eval_samples_per_second": 896.675,
      "eval_steps_per_second": 56.042,
      "step": 232000
    },
    {
      "epoch": 0.51,
      "learning_rate": 3.69e-07,
      "loss": 2.0848,
      "step": 240000
    },
    {
      "epoch": 0.51,
      "eval_loss": 1.9713643789291382,
      "eval_runtime": 445.9733,
      "eval_samples_per_second": 896.915,
      "eval_steps_per_second": 56.057,
      "step": 240000
    },
    {
      "epoch": 0.52,
      "eval_loss": 1.9780809879302979,
      "eval_runtime": 446.8053,
      "eval_samples_per_second": 895.245,
      "eval_steps_per_second": 55.953,
      "step": 248000
    },
    {
      "epoch": 0.54,
      "learning_rate": 3.6626666666666664e-07,
      "loss": 2.0882,
      "step": 256000
    },
    {
      "epoch": 0.54,
      "eval_loss": 1.9709409475326538,
      "eval_runtime": 446.8628,
      "eval_samples_per_second": 895.129,
      "eval_steps_per_second": 55.946,
      "step": 256000
    },
    {
      "epoch": 0.56,
      "eval_loss": 1.9659632444381714,
      "eval_runtime": 446.6679,
      "eval_samples_per_second": 895.52,
      "eval_steps_per_second": 55.97,
      "step": 264000
    },
    {
      "epoch": 0.57,
      "learning_rate": 3.6353333333333335e-07,
      "loss": 2.0922,
      "step": 272000
    },
    {
      "epoch": 0.57,
      "eval_loss": 1.9651156663894653,
      "eval_runtime": 447.5675,
      "eval_samples_per_second": 893.72,
      "eval_steps_per_second": 55.857,
      "step": 272000
    },
    {
      "epoch": 0.59,
      "eval_loss": 1.9677698612213135,
      "eval_runtime": 447.5478,
      "eval_samples_per_second": 893.759,
      "eval_steps_per_second": 55.86,
      "step": 280000
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.608e-07,
      "loss": 2.0938,
      "step": 288000
    },
    {
      "epoch": 0.61,
      "eval_loss": 1.9667097330093384,
      "eval_runtime": 445.5237,
      "eval_samples_per_second": 897.82,
      "eval_steps_per_second": 56.114,
      "step": 288000
    },
    {
      "epoch": 0.62,
      "eval_loss": 1.9630117416381836,
      "eval_runtime": 447.5815,
      "eval_samples_per_second": 893.692,
      "eval_steps_per_second": 55.856,
      "step": 296000
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.5806666666666666e-07,
      "loss": 2.095,
      "step": 304000
    },
    {
      "epoch": 0.64,
      "eval_loss": 1.964175820350647,
      "eval_runtime": 446.8848,
      "eval_samples_per_second": 895.085,
      "eval_steps_per_second": 55.943,
      "step": 304000
    },
    {
      "epoch": 0.66,
      "eval_loss": 1.9624465703964233,
      "eval_runtime": 446.5827,
      "eval_samples_per_second": 895.691,
      "eval_steps_per_second": 55.981,
      "step": 312000
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.553333333333333e-07,
      "loss": 2.0908,
      "step": 320000
    },
    {
      "epoch": 0.67,
      "eval_loss": 1.9602919816970825,
      "eval_runtime": 446.0632,
      "eval_samples_per_second": 896.734,
      "eval_steps_per_second": 56.046,
      "step": 320000
    },
    {
      "epoch": 0.69,
      "eval_loss": 1.9648582935333252,
      "eval_runtime": 446.2312,
      "eval_samples_per_second": 896.396,
      "eval_steps_per_second": 56.025,
      "step": 328000
    },
    {
      "epoch": 0.71,
      "learning_rate": 3.5259999999999997e-07,
      "loss": 2.0927,
      "step": 336000
    },
    {
      "epoch": 0.71,
      "eval_loss": 1.9641398191452026,
      "eval_runtime": 449.2021,
      "eval_samples_per_second": 890.468,
      "eval_steps_per_second": 55.654,
      "step": 336000
    },
    {
      "epoch": 0.72,
      "eval_loss": 1.9602792263031006,
      "eval_runtime": 447.5196,
      "eval_samples_per_second": 893.816,
      "eval_steps_per_second": 55.863,
      "step": 344000
    },
    {
      "epoch": 0.74,
      "learning_rate": 3.498666666666667e-07,
      "loss": 2.0931,
      "step": 352000
    },
    {
      "epoch": 0.74,
      "eval_loss": 1.9589898586273193,
      "eval_runtime": 447.4097,
      "eval_samples_per_second": 894.035,
      "eval_steps_per_second": 55.877,
      "step": 352000
    },
    {
      "epoch": 0.76,
      "eval_loss": 1.964429259300232,
      "eval_runtime": 447.2712,
      "eval_samples_per_second": 894.312,
      "eval_steps_per_second": 55.895,
      "step": 360000
    },
    {
      "epoch": 0.77,
      "learning_rate": 3.4713333333333333e-07,
      "loss": 2.087,
      "step": 368000
    },
    {
      "epoch": 0.77,
      "eval_loss": 1.963483452796936,
      "eval_runtime": 447.9748,
      "eval_samples_per_second": 892.907,
      "eval_steps_per_second": 55.807,
      "step": 368000
    },
    {
      "epoch": 0.79,
      "eval_loss": 1.9613640308380127,
      "eval_runtime": 448.1094,
      "eval_samples_per_second": 892.639,
      "eval_steps_per_second": 55.79,
      "step": 376000
    },
    {
      "epoch": 0.81,
      "learning_rate": 3.444e-07,
      "loss": 2.0792,
      "step": 384000
    },
    {
      "epoch": 0.81,
      "eval_loss": 1.959149956703186,
      "eval_runtime": 447.4535,
      "eval_samples_per_second": 893.948,
      "eval_steps_per_second": 55.872,
      "step": 384000
    },
    {
      "epoch": 0.83,
      "eval_loss": 1.9575395584106445,
      "eval_runtime": 448.4642,
      "eval_samples_per_second": 891.933,
      "eval_steps_per_second": 55.746,
      "step": 392000
    },
    {
      "epoch": 0.84,
      "learning_rate": 3.416666666666667e-07,
      "loss": 2.0899,
      "step": 400000
    },
    {
      "epoch": 0.84,
      "eval_loss": 1.959159016609192,
      "eval_runtime": 446.7586,
      "eval_samples_per_second": 895.338,
      "eval_steps_per_second": 55.959,
      "step": 400000
    },
    {
      "epoch": 0.86,
      "eval_loss": 1.9618536233901978,
      "eval_runtime": 449.0448,
      "eval_samples_per_second": 890.78,
      "eval_steps_per_second": 55.674,
      "step": 408000
    },
    {
      "epoch": 0.88,
      "learning_rate": 3.3893333333333335e-07,
      "loss": 2.0812,
      "step": 416000
    },
    {
      "epoch": 0.88,
      "eval_loss": 1.9582098722457886,
      "eval_runtime": 448.2032,
      "eval_samples_per_second": 892.452,
      "eval_steps_per_second": 55.778,
      "step": 416000
    },
    {
      "epoch": 0.89,
      "eval_loss": 1.958033800125122,
      "eval_runtime": 448.1857,
      "eval_samples_per_second": 892.487,
      "eval_steps_per_second": 55.78,
      "step": 424000
    },
    {
      "epoch": 0.91,
      "learning_rate": 3.3619999999999995e-07,
      "loss": 2.0948,
      "step": 432000
    },
    {
      "epoch": 0.91,
      "eval_loss": 1.9587032794952393,
      "eval_runtime": 447.736,
      "eval_samples_per_second": 893.384,
      "eval_steps_per_second": 55.836,
      "step": 432000
    },
    {
      "epoch": 0.93,
      "eval_loss": 1.9593387842178345,
      "eval_runtime": 448.6332,
      "eval_samples_per_second": 891.597,
      "eval_steps_per_second": 55.725,
      "step": 440000
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.3346666666666666e-07,
      "loss": 2.0895,
      "step": 448000
    },
    {
      "epoch": 0.94,
      "eval_loss": 1.960838794708252,
      "eval_runtime": 448.8111,
      "eval_samples_per_second": 891.244,
      "eval_steps_per_second": 55.703,
      "step": 448000
    },
    {
      "epoch": 0.96,
      "eval_loss": 1.956569790840149,
      "eval_runtime": 448.6177,
      "eval_samples_per_second": 891.628,
      "eval_steps_per_second": 55.727,
      "step": 456000
    },
    {
      "epoch": 0.98,
      "learning_rate": 3.307333333333333e-07,
      "loss": 2.0756,
      "step": 464000
    },
    {
      "epoch": 0.98,
      "eval_loss": 1.952539086341858,
      "eval_runtime": 448.9343,
      "eval_samples_per_second": 890.999,
      "eval_steps_per_second": 55.687,
      "step": 464000
    },
    {
      "epoch": 0.99,
      "eval_loss": 1.9540966749191284,
      "eval_runtime": 449.2991,
      "eval_samples_per_second": 890.275,
      "eval_steps_per_second": 55.642,
      "step": 472000
    },
    {
      "epoch": 1.01,
      "learning_rate": 3.28e-07,
      "loss": 2.0842,
      "step": 480000
    },
    {
      "epoch": 1.01,
      "eval_loss": 1.9600876569747925,
      "eval_runtime": 449.599,
      "eval_samples_per_second": 889.682,
      "eval_steps_per_second": 55.605,
      "step": 480000
    },
    {
      "epoch": 1.03,
      "eval_loss": 1.9564368724822998,
      "eval_runtime": 452.8797,
      "eval_samples_per_second": 883.237,
      "eval_steps_per_second": 55.202,
      "step": 488000
    },
    {
      "epoch": 1.04,
      "learning_rate": 3.252666666666667e-07,
      "loss": 2.0935,
      "step": 496000
    },
    {
      "epoch": 1.04,
      "eval_loss": 1.9522088766098022,
      "eval_runtime": 452.7891,
      "eval_samples_per_second": 883.414,
      "eval_steps_per_second": 55.213,
      "step": 496000
    },
    {
      "epoch": 1.06,
      "eval_loss": 1.95320463180542,
      "eval_runtime": 454.7973,
      "eval_samples_per_second": 879.513,
      "eval_steps_per_second": 54.97,
      "step": 504000
    },
    {
      "epoch": 1.08,
      "learning_rate": 3.2253333333333334e-07,
      "loss": 2.0836,
      "step": 512000
    },
    {
      "epoch": 1.08,
      "eval_loss": 1.9537328481674194,
      "eval_runtime": 451.2722,
      "eval_samples_per_second": 886.383,
      "eval_steps_per_second": 55.399,
      "step": 512000
    },
    {
      "epoch": 1.09,
      "eval_loss": 1.9553171396255493,
      "eval_runtime": 453.3773,
      "eval_samples_per_second": 882.267,
      "eval_steps_per_second": 55.142,
      "step": 520000
    },
    {
      "epoch": 1.11,
      "learning_rate": 3.198e-07,
      "loss": 2.0876,
      "step": 528000
    },
    {
      "epoch": 1.11,
      "eval_loss": 1.946892499923706,
      "eval_runtime": 451.8379,
      "eval_samples_per_second": 885.273,
      "eval_steps_per_second": 55.33,
      "step": 528000
    },
    {
      "epoch": 1.13,
      "eval_loss": 1.9497493505477905,
      "eval_runtime": 452.793,
      "eval_samples_per_second": 883.406,
      "eval_steps_per_second": 55.213,
      "step": 536000
    },
    {
      "epoch": 1.15,
      "learning_rate": 3.1706666666666665e-07,
      "loss": 2.0778,
      "step": 544000
    },
    {
      "epoch": 1.15,
      "eval_loss": 1.9541878700256348,
      "eval_runtime": 449.7452,
      "eval_samples_per_second": 889.392,
      "eval_steps_per_second": 55.587,
      "step": 544000
    },
    {
      "epoch": 1.16,
      "eval_loss": 1.951553463935852,
      "eval_runtime": 451.9652,
      "eval_samples_per_second": 885.024,
      "eval_steps_per_second": 55.314,
      "step": 552000
    },
    {
      "epoch": 1.18,
      "learning_rate": 3.1433333333333336e-07,
      "loss": 2.0829,
      "step": 560000
    },
    {
      "epoch": 1.18,
      "eval_loss": 1.9506231546401978,
      "eval_runtime": 448.9862,
      "eval_samples_per_second": 890.896,
      "eval_steps_per_second": 55.681,
      "step": 560000
    },
    {
      "epoch": 1.2,
      "eval_loss": 1.9504673480987549,
      "eval_runtime": 448.899,
      "eval_samples_per_second": 891.069,
      "eval_steps_per_second": 55.692,
      "step": 568000
    },
    {
      "epoch": 1.21,
      "learning_rate": 3.116e-07,
      "loss": 2.0864,
      "step": 576000
    },
    {
      "epoch": 1.21,
      "eval_loss": 1.9530742168426514,
      "eval_runtime": 450.2299,
      "eval_samples_per_second": 888.435,
      "eval_steps_per_second": 55.527,
      "step": 576000
    },
    {
      "epoch": 1.23,
      "eval_loss": 1.9455041885375977,
      "eval_runtime": 449.7657,
      "eval_samples_per_second": 889.352,
      "eval_steps_per_second": 55.584,
      "step": 584000
    },
    {
      "epoch": 1.25,
      "learning_rate": 3.0886666666666667e-07,
      "loss": 2.0893,
      "step": 592000
    },
    {
      "epoch": 1.25,
      "eval_loss": 1.9471276998519897,
      "eval_runtime": 450.0648,
      "eval_samples_per_second": 888.761,
      "eval_steps_per_second": 55.548,
      "step": 592000
    },
    {
      "epoch": 1.26,
      "eval_loss": 1.9539045095443726,
      "eval_runtime": 449.915,
      "eval_samples_per_second": 889.057,
      "eval_steps_per_second": 55.566,
      "step": 600000
    },
    {
      "epoch": 1.28,
      "learning_rate": 3.061333333333333e-07,
      "loss": 2.0808,
      "step": 608000
    },
    {
      "epoch": 1.28,
      "eval_loss": 1.9455146789550781,
      "eval_runtime": 450.3195,
      "eval_samples_per_second": 888.258,
      "eval_steps_per_second": 55.516,
      "step": 608000
    },
    {
      "epoch": 1.3,
      "eval_loss": 1.9497132301330566,
      "eval_runtime": 452.443,
      "eval_samples_per_second": 884.089,
      "eval_steps_per_second": 55.256,
      "step": 616000
    },
    {
      "epoch": 1.31,
      "learning_rate": 3.034e-07,
      "loss": 2.0838,
      "step": 624000
    },
    {
      "epoch": 1.31,
      "eval_loss": 1.946612000465393,
      "eval_runtime": 452.7896,
      "eval_samples_per_second": 883.413,
      "eval_steps_per_second": 55.213,
      "step": 624000
    },
    {
      "epoch": 1.33,
      "eval_loss": 1.9498077630996704,
      "eval_runtime": 453.0599,
      "eval_samples_per_second": 882.886,
      "eval_steps_per_second": 55.18,
      "step": 632000
    },
    {
      "epoch": 1.35,
      "learning_rate": 3.0066666666666663e-07,
      "loss": 2.0812,
      "step": 640000
    },
    {
      "epoch": 1.35,
      "eval_loss": 1.9510040283203125,
      "eval_runtime": 451.3079,
      "eval_samples_per_second": 886.313,
      "eval_steps_per_second": 55.395,
      "step": 640000
    },
    {
      "epoch": 1.36,
      "eval_loss": 1.9526185989379883,
      "eval_runtime": 453.3541,
      "eval_samples_per_second": 882.313,
      "eval_steps_per_second": 55.145,
      "step": 648000
    },
    {
      "epoch": 1.38,
      "learning_rate": 2.9793333333333334e-07,
      "loss": 2.0793,
      "step": 656000
    },
    {
      "epoch": 1.38,
      "eval_loss": 1.9471242427825928,
      "eval_runtime": 453.3637,
      "eval_samples_per_second": 882.294,
      "eval_steps_per_second": 55.143,
      "step": 656000
    },
    {
      "epoch": 1.4,
      "eval_loss": 1.9469006061553955,
      "eval_runtime": 450.6009,
      "eval_samples_per_second": 887.703,
      "eval_steps_per_second": 55.481,
      "step": 664000
    },
    {
      "epoch": 1.41,
      "learning_rate": 2.952e-07,
      "loss": 2.0789,
      "step": 672000
    },
    {
      "epoch": 1.41,
      "eval_loss": 1.9454900026321411,
      "eval_runtime": 447.7405,
      "eval_samples_per_second": 893.375,
      "eval_steps_per_second": 55.836,
      "step": 672000
    },
    {
      "epoch": 1.43,
      "eval_loss": 1.9469287395477295,
      "eval_runtime": 447.6789,
      "eval_samples_per_second": 893.498,
      "eval_steps_per_second": 55.844,
      "step": 680000
    },
    {
      "epoch": 1.45,
      "learning_rate": 2.9246666666666665e-07,
      "loss": 2.0883,
      "step": 688000
    },
    {
      "epoch": 1.45,
      "eval_loss": 1.94387948513031,
      "eval_runtime": 448.8787,
      "eval_samples_per_second": 891.109,
      "eval_steps_per_second": 55.694,
      "step": 688000
    },
    {
      "epoch": 1.47,
      "eval_loss": 1.9438937902450562,
      "eval_runtime": 448.7929,
      "eval_samples_per_second": 891.28,
      "eval_steps_per_second": 55.705,
      "step": 696000
    },
    {
      "epoch": 1.48,
      "learning_rate": 2.897333333333333e-07,
      "loss": 2.09,
      "step": 704000
    },
    {
      "epoch": 1.48,
      "eval_loss": 1.9415992498397827,
      "eval_runtime": 449.4189,
      "eval_samples_per_second": 890.038,
      "eval_steps_per_second": 55.627,
      "step": 704000
    },
    {
      "epoch": 1.5,
      "eval_loss": 1.9492276906967163,
      "eval_runtime": 448.9056,
      "eval_samples_per_second": 891.056,
      "eval_steps_per_second": 55.691,
      "step": 712000
    },
    {
      "epoch": 1.52,
      "learning_rate": 2.8699999999999996e-07,
      "loss": 2.0845,
      "step": 720000
    },
    {
      "epoch": 1.52,
      "eval_loss": 1.943044900894165,
      "eval_runtime": 449.7144,
      "eval_samples_per_second": 889.453,
      "eval_steps_per_second": 55.591,
      "step": 720000
    },
    {
      "epoch": 1.53,
      "eval_loss": 1.948350191116333,
      "eval_runtime": 450.4171,
      "eval_samples_per_second": 888.066,
      "eval_steps_per_second": 55.504,
      "step": 728000
    },
    {
      "epoch": 1.55,
      "learning_rate": 2.8426666666666667e-07,
      "loss": 2.0742,
      "step": 736000
    },
    {
      "epoch": 1.55,
      "eval_loss": 1.9455984830856323,
      "eval_runtime": 449.7084,
      "eval_samples_per_second": 889.465,
      "eval_steps_per_second": 55.592,
      "step": 736000
    },
    {
      "epoch": 1.57,
      "eval_loss": 1.9379788637161255,
      "eval_runtime": 450.7564,
      "eval_samples_per_second": 887.397,
      "eval_steps_per_second": 55.462,
      "step": 744000
    },
    {
      "epoch": 1.58,
      "learning_rate": 2.815333333333333e-07,
      "loss": 2.0839,
      "step": 752000
    },
    {
      "epoch": 1.58,
      "eval_loss": 1.9418160915374756,
      "eval_runtime": 448.5558,
      "eval_samples_per_second": 891.751,
      "eval_steps_per_second": 55.734,
      "step": 752000
    },
    {
      "epoch": 1.6,
      "eval_loss": 1.943367600440979,
      "eval_runtime": 449.0202,
      "eval_samples_per_second": 890.829,
      "eval_steps_per_second": 55.677,
      "step": 760000
    },
    {
      "epoch": 1.62,
      "learning_rate": 2.7880000000000003e-07,
      "loss": 2.0806,
      "step": 768000
    },
    {
      "epoch": 1.62,
      "eval_loss": 1.9450091123580933,
      "eval_runtime": 450.2134,
      "eval_samples_per_second": 888.468,
      "eval_steps_per_second": 55.529,
      "step": 768000
    },
    {
      "epoch": 1.63,
      "eval_loss": 1.9426227807998657,
      "eval_runtime": 450.7888,
      "eval_samples_per_second": 887.334,
      "eval_steps_per_second": 55.458,
      "step": 776000
    },
    {
      "epoch": 1.65,
      "learning_rate": 2.7606666666666664e-07,
      "loss": 2.0805,
      "step": 784000
    },
    {
      "epoch": 1.65,
      "eval_loss": 1.944136619567871,
      "eval_runtime": 450.1142,
      "eval_samples_per_second": 888.663,
      "eval_steps_per_second": 55.541,
      "step": 784000
    },
    {
      "epoch": 1.67,
      "eval_loss": 1.945942997932434,
      "eval_runtime": 449.1687,
      "eval_samples_per_second": 890.534,
      "eval_steps_per_second": 55.658,
      "step": 792000
    },
    {
      "epoch": 1.68,
      "learning_rate": 2.733333333333333e-07,
      "loss": 2.0833,
      "step": 800000
    },
    {
      "epoch": 1.68,
      "eval_loss": 1.9434839487075806,
      "eval_runtime": 450.7395,
      "eval_samples_per_second": 887.43,
      "eval_steps_per_second": 55.464,
      "step": 800000
    },
    {
      "epoch": 1.7,
      "eval_loss": 1.9455143213272095,
      "eval_runtime": 449.9182,
      "eval_samples_per_second": 889.051,
      "eval_steps_per_second": 55.566,
      "step": 808000
    },
    {
      "epoch": 1.72,
      "learning_rate": 2.706e-07,
      "loss": 2.0763,
      "step": 816000
    },
    {
      "epoch": 1.72,
      "eval_loss": 1.9420539140701294,
      "eval_runtime": 452.5501,
      "eval_samples_per_second": 883.88,
      "eval_steps_per_second": 55.242,
      "step": 816000
    },
    {
      "epoch": 1.73,
      "eval_loss": 1.9438146352767944,
      "eval_runtime": 450.6041,
      "eval_samples_per_second": 887.697,
      "eval_steps_per_second": 55.481,
      "step": 824000
    },
    {
      "epoch": 1.75,
      "learning_rate": 2.6786666666666666e-07,
      "loss": 2.0758,
      "step": 832000
    },
    {
      "epoch": 1.75,
      "eval_loss": 1.937113642692566,
      "eval_runtime": 452.047,
      "eval_samples_per_second": 884.864,
      "eval_steps_per_second": 55.304,
      "step": 832000
    },
    {
      "epoch": 1.77,
      "eval_loss": 1.9432255029678345,
      "eval_runtime": 451.3314,
      "eval_samples_per_second": 886.267,
      "eval_steps_per_second": 55.392,
      "step": 840000
    },
    {
      "epoch": 1.79,
      "learning_rate": 2.651333333333333e-07,
      "loss": 2.0888,
      "step": 848000
    },
    {
      "epoch": 1.79,
      "eval_loss": 1.941444993019104,
      "eval_runtime": 449.4111,
      "eval_samples_per_second": 890.054,
      "eval_steps_per_second": 55.628,
      "step": 848000
    },
    {
      "epoch": 1.8,
      "eval_loss": 1.9443625211715698,
      "eval_runtime": 449.8026,
      "eval_samples_per_second": 889.279,
      "eval_steps_per_second": 55.58,
      "step": 856000
    },
    {
      "epoch": 1.82,
      "learning_rate": 2.624e-07,
      "loss": 2.0786,
      "step": 864000
    },
    {
      "epoch": 1.82,
      "eval_loss": 1.9407739639282227,
      "eval_runtime": 450.2493,
      "eval_samples_per_second": 888.397,
      "eval_steps_per_second": 55.525,
      "step": 864000
    },
    {
      "epoch": 1.84,
      "eval_loss": 1.9396659135818481,
      "eval_runtime": 452.6466,
      "eval_samples_per_second": 883.692,
      "eval_steps_per_second": 55.231,
      "step": 872000
    },
    {
      "epoch": 1.85,
      "learning_rate": 2.596666666666667e-07,
      "loss": 2.079,
      "step": 880000
    },
    {
      "epoch": 1.85,
      "eval_loss": 1.9406061172485352,
      "eval_runtime": 451.5534,
      "eval_samples_per_second": 885.831,
      "eval_steps_per_second": 55.364,
      "step": 880000
    },
    {
      "epoch": 1.87,
      "eval_loss": 1.944212555885315,
      "eval_runtime": 452.1338,
      "eval_samples_per_second": 884.694,
      "eval_steps_per_second": 55.293,
      "step": 888000
    },
    {
      "epoch": 1.89,
      "learning_rate": 2.5693333333333333e-07,
      "loss": 2.0817,
      "step": 896000
    },
    {
      "epoch": 1.89,
      "eval_loss": 1.9403553009033203,
      "eval_runtime": 450.9134,
      "eval_samples_per_second": 887.088,
      "eval_steps_per_second": 55.443,
      "step": 896000
    },
    {
      "epoch": 1.9,
      "eval_loss": 1.945005178451538,
      "eval_runtime": 453.0225,
      "eval_samples_per_second": 882.958,
      "eval_steps_per_second": 55.185,
      "step": 904000
    },
    {
      "epoch": 1.92,
      "learning_rate": 2.542e-07,
      "loss": 2.0792,
      "step": 912000
    },
    {
      "epoch": 1.92,
      "eval_loss": 1.9379777908325195,
      "eval_runtime": 450.9233,
      "eval_samples_per_second": 887.069,
      "eval_steps_per_second": 55.442,
      "step": 912000
    },
    {
      "epoch": 1.94,
      "eval_loss": 1.938461422920227,
      "eval_runtime": 451.8903,
      "eval_samples_per_second": 885.171,
      "eval_steps_per_second": 55.323,
      "step": 920000
    },
    {
      "epoch": 1.95,
      "learning_rate": 2.5146666666666664e-07,
      "loss": 2.0741,
      "step": 928000
    },
    {
      "epoch": 1.95,
      "eval_loss": 1.9448977708816528,
      "eval_runtime": 452.6642,
      "eval_samples_per_second": 883.657,
      "eval_steps_per_second": 55.229,
      "step": 928000
    },
    {
      "epoch": 1.97,
      "eval_loss": 1.9413866996765137,
      "eval_runtime": 452.86,
      "eval_samples_per_second": 883.275,
      "eval_steps_per_second": 55.205,
      "step": 936000
    },
    {
      "epoch": 1.99,
      "learning_rate": 2.4873333333333335e-07,
      "loss": 2.0832,
      "step": 944000
    },
    {
      "epoch": 1.99,
      "eval_loss": 1.9401600360870361,
      "eval_runtime": 451.3974,
      "eval_samples_per_second": 886.137,
      "eval_steps_per_second": 55.384,
      "step": 944000
    },
    {
      "epoch": 2.0,
      "eval_loss": 1.940971851348877,
      "eval_runtime": 452.3973,
      "eval_samples_per_second": 884.178,
      "eval_steps_per_second": 55.261,
      "step": 952000
    },
    {
      "epoch": 2.02,
      "learning_rate": 2.46e-07,
      "loss": 2.0695,
      "step": 960000
    },
    {
      "epoch": 2.02,
      "eval_loss": 1.9370893239974976,
      "eval_runtime": 453.3847,
      "eval_samples_per_second": 882.253,
      "eval_steps_per_second": 55.141,
      "step": 960000
    },
    {
      "epoch": 2.04,
      "eval_loss": 1.934204339981079,
      "eval_runtime": 452.8123,
      "eval_samples_per_second": 883.368,
      "eval_steps_per_second": 55.211,
      "step": 968000
    },
    {
      "epoch": 2.05,
      "learning_rate": 2.4326666666666666e-07,
      "loss": 2.0813,
      "step": 976000
    },
    {
      "epoch": 2.05,
      "eval_loss": 1.9375728368759155,
      "eval_runtime": 454.1486,
      "eval_samples_per_second": 880.769,
      "eval_steps_per_second": 55.048,
      "step": 976000
    },
    {
      "epoch": 2.07,
      "eval_loss": 1.939716100692749,
      "eval_runtime": 451.3741,
      "eval_samples_per_second": 886.183,
      "eval_steps_per_second": 55.386,
      "step": 984000
    },
    {
      "epoch": 2.09,
      "learning_rate": 2.405333333333333e-07,
      "loss": 2.0804,
      "step": 992000
    },
    {
      "epoch": 2.09,
      "eval_loss": 1.9394439458847046,
      "eval_runtime": 453.2345,
      "eval_samples_per_second": 882.545,
      "eval_steps_per_second": 55.159,
      "step": 992000
    },
    {
      "epoch": 2.11,
      "eval_loss": 1.9370408058166504,
      "eval_runtime": 453.6157,
      "eval_samples_per_second": 881.804,
      "eval_steps_per_second": 55.113,
      "step": 1000000
    },
    {
      "epoch": 2.12,
      "learning_rate": 2.3779999999999997e-07,
      "loss": 2.0789,
      "step": 1008000
    },
    {
      "epoch": 2.12,
      "eval_loss": 1.93497896194458,
      "eval_runtime": 452.9534,
      "eval_samples_per_second": 883.093,
      "eval_steps_per_second": 55.193,
      "step": 1008000
    },
    {
      "epoch": 2.14,
      "eval_loss": 1.9327107667922974,
      "eval_runtime": 453.3718,
      "eval_samples_per_second": 882.278,
      "eval_steps_per_second": 55.142,
      "step": 1016000
    },
    {
      "epoch": 2.16,
      "learning_rate": 2.3506666666666668e-07,
      "loss": 2.0754,
      "step": 1024000
    },
    {
      "epoch": 2.16,
      "eval_loss": 1.9420740604400635,
      "eval_runtime": 454.3554,
      "eval_samples_per_second": 880.368,
      "eval_steps_per_second": 55.023,
      "step": 1024000
    },
    {
      "epoch": 2.17,
      "eval_loss": 1.937127947807312,
      "eval_runtime": 453.978,
      "eval_samples_per_second": 881.1,
      "eval_steps_per_second": 55.069,
      "step": 1032000
    },
    {
      "epoch": 2.19,
      "learning_rate": 2.3233333333333334e-07,
      "loss": 2.0774,
      "step": 1040000
    },
    {
      "epoch": 2.19,
      "eval_loss": 1.9410624504089355,
      "eval_runtime": 453.3501,
      "eval_samples_per_second": 882.32,
      "eval_steps_per_second": 55.145,
      "step": 1040000
    },
    {
      "epoch": 2.21,
      "eval_loss": 1.9337198734283447,
      "eval_runtime": 454.7842,
      "eval_samples_per_second": 879.538,
      "eval_steps_per_second": 54.971,
      "step": 1048000
    },
    {
      "epoch": 2.22,
      "learning_rate": 2.2960000000000002e-07,
      "loss": 2.0766,
      "step": 1056000
    },
    {
      "epoch": 2.22,
      "eval_loss": 1.9387423992156982,
      "eval_runtime": 455.5914,
      "eval_samples_per_second": 877.98,
      "eval_steps_per_second": 54.874,
      "step": 1056000
    },
    {
      "epoch": 2.24,
      "eval_loss": 1.9334228038787842,
      "eval_runtime": 454.6433,
      "eval_samples_per_second": 879.811,
      "eval_steps_per_second": 54.988,
      "step": 1064000
    },
    {
      "epoch": 2.26,
      "learning_rate": 2.2686666666666667e-07,
      "loss": 2.079,
      "step": 1072000
    },
    {
      "epoch": 2.26,
      "eval_loss": 1.938550353050232,
      "eval_runtime": 454.9961,
      "eval_samples_per_second": 879.128,
      "eval_steps_per_second": 54.946,
      "step": 1072000
    },
    {
      "epoch": 2.27,
      "eval_loss": 1.9334533214569092,
      "eval_runtime": 453.0142,
      "eval_samples_per_second": 882.975,
      "eval_steps_per_second": 55.186,
      "step": 1080000
    },
    {
      "epoch": 2.29,
      "learning_rate": 2.2413333333333333e-07,
      "loss": 2.068,
      "step": 1088000
    },
    {
      "epoch": 2.29,
      "eval_loss": 1.9363125562667847,
      "eval_runtime": 452.9254,
      "eval_samples_per_second": 883.148,
      "eval_steps_per_second": 55.197,
      "step": 1088000
    },
    {
      "epoch": 2.31,
      "eval_loss": 1.9420162439346313,
      "eval_runtime": 453.9753,
      "eval_samples_per_second": 881.105,
      "eval_steps_per_second": 55.069,
      "step": 1096000
    },
    {
      "epoch": 2.32,
      "learning_rate": 2.214e-07,
      "loss": 2.0786,
      "step": 1104000
    },
    {
      "epoch": 2.32,
      "eval_loss": 1.9330793619155884,
      "eval_runtime": 454.6549,
      "eval_samples_per_second": 879.788,
      "eval_steps_per_second": 54.987,
      "step": 1104000
    },
    {
      "epoch": 2.34,
      "eval_loss": 1.9327301979064941,
      "eval_runtime": 455.7252,
      "eval_samples_per_second": 877.722,
      "eval_steps_per_second": 54.858,
      "step": 1112000
    },
    {
      "epoch": 2.36,
      "learning_rate": 2.1866666666666667e-07,
      "loss": 2.0734,
      "step": 1120000
    },
    {
      "epoch": 2.36,
      "eval_loss": 1.939077615737915,
      "eval_runtime": 456.0008,
      "eval_samples_per_second": 877.191,
      "eval_steps_per_second": 54.824,
      "step": 1120000
    },
    {
      "epoch": 2.37,
      "eval_loss": 1.9362618923187256,
      "eval_runtime": 455.372,
      "eval_samples_per_second": 878.403,
      "eval_steps_per_second": 54.9,
      "step": 1128000
    },
    {
      "epoch": 2.39,
      "learning_rate": 2.1593333333333332e-07,
      "loss": 2.0787,
      "step": 1136000
    },
    {
      "epoch": 2.39,
      "eval_loss": 1.932115912437439,
      "eval_runtime": 454.5174,
      "eval_samples_per_second": 880.054,
      "eval_steps_per_second": 55.003,
      "step": 1136000
    },
    {
      "epoch": 2.41,
      "eval_loss": 1.9333146810531616,
      "eval_runtime": 458.093,
      "eval_samples_per_second": 873.185,
      "eval_steps_per_second": 54.574,
      "step": 1144000
    },
    {
      "epoch": 2.43,
      "learning_rate": 2.132e-07,
      "loss": 2.0731,
      "step": 1152000
    },
    {
      "epoch": 2.43,
      "eval_loss": 1.9368531703948975,
      "eval_runtime": 454.4244,
      "eval_samples_per_second": 880.234,
      "eval_steps_per_second": 55.015,
      "step": 1152000
    },
    {
      "epoch": 2.44,
      "eval_loss": 1.9357047080993652,
      "eval_runtime": 456.0905,
      "eval_samples_per_second": 877.019,
      "eval_steps_per_second": 54.814,
      "step": 1160000
    },
    {
      "epoch": 2.46,
      "learning_rate": 2.1046666666666666e-07,
      "loss": 2.0816,
      "step": 1168000
    },
    {
      "epoch": 2.46,
      "eval_loss": 1.9352905750274658,
      "eval_runtime": 457.1609,
      "eval_samples_per_second": 874.966,
      "eval_steps_per_second": 54.685,
      "step": 1168000
    },
    {
      "epoch": 2.48,
      "eval_loss": 1.9318699836730957,
      "eval_runtime": 457.231,
      "eval_samples_per_second": 874.831,
      "eval_steps_per_second": 54.677,
      "step": 1176000
    },
    {
      "epoch": 2.49,
      "learning_rate": 2.0773333333333334e-07,
      "loss": 2.0758,
      "step": 1184000
    },
    {
      "epoch": 2.49,
      "eval_loss": 1.9365841150283813,
      "eval_runtime": 455.9751,
      "eval_samples_per_second": 877.241,
      "eval_steps_per_second": 54.828,
      "step": 1184000
    },
    {
      "epoch": 2.51,
      "eval_loss": 1.9300509691238403,
      "eval_runtime": 460.0832,
      "eval_samples_per_second": 869.408,
      "eval_steps_per_second": 54.338,
      "step": 1192000
    },
    {
      "epoch": 2.53,
      "learning_rate": 2.05e-07,
      "loss": 2.0725,
      "step": 1200000
    },
    {
      "epoch": 2.53,
      "eval_loss": 1.9328936338424683,
      "eval_runtime": 456.956,
      "eval_samples_per_second": 875.358,
      "eval_steps_per_second": 54.71,
      "step": 1200000
    },
    {
      "epoch": 2.54,
      "eval_loss": 1.937027931213379,
      "eval_runtime": 455.5679,
      "eval_samples_per_second": 878.025,
      "eval_steps_per_second": 54.877,
      "step": 1208000
    },
    {
      "epoch": 2.56,
      "learning_rate": 2.0226666666666668e-07,
      "loss": 2.085,
      "step": 1216000
    },
    {
      "epoch": 2.56,
      "eval_loss": 1.9251400232315063,
      "eval_runtime": 457.3381,
      "eval_samples_per_second": 874.626,
      "eval_steps_per_second": 54.664,
      "step": 1216000
    },
    {
      "epoch": 2.58,
      "eval_loss": 1.9369462728500366,
      "eval_runtime": 459.1359,
      "eval_samples_per_second": 871.202,
      "eval_steps_per_second": 54.45,
      "step": 1224000
    },
    {
      "epoch": 2.59,
      "learning_rate": 1.9953333333333333e-07,
      "loss": 2.0809,
      "step": 1232000
    },
    {
      "epoch": 2.59,
      "eval_loss": 1.9377222061157227,
      "eval_runtime": 457.7839,
      "eval_samples_per_second": 873.775,
      "eval_steps_per_second": 54.611,
      "step": 1232000
    },
    {
      "epoch": 2.61,
      "eval_loss": 1.9397977590560913,
      "eval_runtime": 458.2173,
      "eval_samples_per_second": 872.948,
      "eval_steps_per_second": 54.559,
      "step": 1240000
    },
    {
      "epoch": 2.63,
      "learning_rate": 1.968e-07,
      "loss": 2.0742,
      "step": 1248000
    },
    {
      "epoch": 2.63,
      "eval_loss": 1.9367727041244507,
      "eval_runtime": 456.9366,
      "eval_samples_per_second": 875.395,
      "eval_steps_per_second": 54.712,
      "step": 1248000
    },
    {
      "epoch": 2.64,
      "eval_loss": 1.9389311075210571,
      "eval_runtime": 457.4588,
      "eval_samples_per_second": 874.396,
      "eval_steps_per_second": 54.65,
      "step": 1256000
    },
    {
      "epoch": 2.66,
      "learning_rate": 1.9406666666666667e-07,
      "loss": 2.0743,
      "step": 1264000
    },
    {
      "epoch": 2.66,
      "eval_loss": 1.9287104606628418,
      "eval_runtime": 458.2038,
      "eval_samples_per_second": 872.974,
      "eval_steps_per_second": 54.561,
      "step": 1264000
    },
    {
      "epoch": 2.68,
      "eval_loss": 1.9337459802627563,
      "eval_runtime": 456.8104,
      "eval_samples_per_second": 875.637,
      "eval_steps_per_second": 54.727,
      "step": 1272000
    },
    {
      "epoch": 2.69,
      "learning_rate": 1.9133333333333333e-07,
      "loss": 2.0822,
      "step": 1280000
    },
    {
      "epoch": 2.69,
      "eval_loss": 1.9323461055755615,
      "eval_runtime": 459.5048,
      "eval_samples_per_second": 870.502,
      "eval_steps_per_second": 54.406,
      "step": 1280000
    },
    {
      "epoch": 2.71,
      "eval_loss": 1.9348268508911133,
      "eval_runtime": 459.882,
      "eval_samples_per_second": 869.788,
      "eval_steps_per_second": 54.362,
      "step": 1288000
    },
    {
      "epoch": 2.73,
      "learning_rate": 1.886e-07,
      "loss": 2.0845,
      "step": 1296000
    },
    {
      "epoch": 2.73,
      "eval_loss": 1.932786464691162,
      "eval_runtime": 456.5549,
      "eval_samples_per_second": 876.127,
      "eval_steps_per_second": 54.758,
      "step": 1296000
    },
    {
      "epoch": 2.75,
      "eval_loss": 1.9324398040771484,
      "eval_runtime": 458.1511,
      "eval_samples_per_second": 873.074,
      "eval_steps_per_second": 54.567,
      "step": 1304000
    },
    {
      "epoch": 2.76,
      "learning_rate": 1.8586666666666666e-07,
      "loss": 2.0706,
      "step": 1312000
    },
    {
      "epoch": 2.76,
      "eval_loss": 1.9304131269454956,
      "eval_runtime": 458.1533,
      "eval_samples_per_second": 873.07,
      "eval_steps_per_second": 54.567,
      "step": 1312000
    },
    {
      "epoch": 2.78,
      "eval_loss": 1.9322303533554077,
      "eval_runtime": 458.9935,
      "eval_samples_per_second": 871.472,
      "eval_steps_per_second": 54.467,
      "step": 1320000
    },
    {
      "epoch": 2.8,
      "learning_rate": 1.8313333333333332e-07,
      "loss": 2.0813,
      "step": 1328000
    },
    {
      "epoch": 2.8,
      "eval_loss": 1.9320390224456787,
      "eval_runtime": 460.999,
      "eval_samples_per_second": 867.681,
      "eval_steps_per_second": 54.23,
      "step": 1328000
    },
    {
      "epoch": 2.81,
      "eval_loss": 1.9378987550735474,
      "eval_runtime": 457.836,
      "eval_samples_per_second": 873.675,
      "eval_steps_per_second": 54.605,
      "step": 1336000
    },
    {
      "epoch": 2.83,
      "learning_rate": 1.804e-07,
      "loss": 2.0768,
      "step": 1344000
    },
    {
      "epoch": 2.83,
      "eval_loss": 1.9283299446105957,
      "eval_runtime": 457.711,
      "eval_samples_per_second": 873.914,
      "eval_steps_per_second": 54.62,
      "step": 1344000
    },
    {
      "epoch": 2.85,
      "eval_loss": 1.9352092742919922,
      "eval_runtime": 459.6254,
      "eval_samples_per_second": 870.274,
      "eval_steps_per_second": 54.392,
      "step": 1352000
    },
    {
      "epoch": 2.86,
      "learning_rate": 1.7766666666666666e-07,
      "loss": 2.0776,
      "step": 1360000
    },
    {
      "epoch": 2.86,
      "eval_loss": 1.9265965223312378,
      "eval_runtime": 456.3209,
      "eval_samples_per_second": 876.576,
      "eval_steps_per_second": 54.786,
      "step": 1360000
    },
    {
      "epoch": 2.88,
      "eval_loss": 1.9339468479156494,
      "eval_runtime": 460.6096,
      "eval_samples_per_second": 868.414,
      "eval_steps_per_second": 54.276,
      "step": 1368000
    },
    {
      "epoch": 2.9,
      "learning_rate": 1.7493333333333334e-07,
      "loss": 2.0776,
      "step": 1376000
    },
    {
      "epoch": 2.9,
      "eval_loss": 1.9370609521865845,
      "eval_runtime": 459.3688,
      "eval_samples_per_second": 870.76,
      "eval_steps_per_second": 54.423,
      "step": 1376000
    },
    {
      "epoch": 2.91,
      "eval_loss": 1.9353028535842896,
      "eval_runtime": 459.3054,
      "eval_samples_per_second": 870.88,
      "eval_steps_per_second": 54.43,
      "step": 1384000
    },
    {
      "epoch": 2.93,
      "learning_rate": 1.722e-07,
      "loss": 2.072,
      "step": 1392000
    },
    {
      "epoch": 2.93,
      "eval_loss": 1.928996205329895,
      "eval_runtime": 459.5911,
      "eval_samples_per_second": 870.339,
      "eval_steps_per_second": 54.396,
      "step": 1392000
    },
    {
      "epoch": 2.95,
      "eval_loss": 1.9337087869644165,
      "eval_runtime": 457.9253,
      "eval_samples_per_second": 873.505,
      "eval_steps_per_second": 54.594,
      "step": 1400000
    },
    {
      "epoch": 2.96,
      "learning_rate": 1.6946666666666668e-07,
      "loss": 2.077,
      "step": 1408000
    },
    {
      "epoch": 2.96,
      "eval_loss": 1.931803584098816,
      "eval_runtime": 459.0247,
      "eval_samples_per_second": 871.413,
      "eval_steps_per_second": 54.463,
      "step": 1408000
    },
    {
      "epoch": 2.98,
      "eval_loss": 1.9326242208480835,
      "eval_runtime": 459.782,
      "eval_samples_per_second": 869.977,
      "eval_steps_per_second": 54.374,
      "step": 1416000
    },
    {
      "epoch": 3.0,
      "learning_rate": 1.6673333333333333e-07,
      "loss": 2.0777,
      "step": 1424000
    },
    {
      "epoch": 3.0,
      "eval_loss": 1.9337918758392334,
      "eval_runtime": 460.1196,
      "eval_samples_per_second": 869.339,
      "eval_steps_per_second": 54.334,
      "step": 1424000
    },
    {
      "epoch": 3.01,
      "eval_loss": 1.9306981563568115,
      "eval_runtime": 460.6247,
      "eval_samples_per_second": 868.386,
      "eval_steps_per_second": 54.274,
      "step": 1432000
    },
    {
      "epoch": 3.03,
      "learning_rate": 1.64e-07,
      "loss": 2.0846,
      "step": 1440000
    },
    {
      "epoch": 3.03,
      "eval_loss": 1.9305415153503418,
      "eval_runtime": 459.7751,
      "eval_samples_per_second": 869.991,
      "eval_steps_per_second": 54.374,
      "step": 1440000
    },
    {
      "epoch": 3.05,
      "eval_loss": 1.931223750114441,
      "eval_runtime": 457.3015,
      "eval_samples_per_second": 874.696,
      "eval_steps_per_second": 54.669,
      "step": 1448000
    },
    {
      "epoch": 3.07,
      "learning_rate": 1.6126666666666667e-07,
      "loss": 2.0744,
      "step": 1456000
    },
    {
      "epoch": 3.07,
      "eval_loss": 1.9331880807876587,
      "eval_runtime": 459.0917,
      "eval_samples_per_second": 871.286,
      "eval_steps_per_second": 54.455,
      "step": 1456000
    },
    {
      "epoch": 3.08,
      "eval_loss": 1.9313483238220215,
      "eval_runtime": 461.1089,
      "eval_samples_per_second": 867.474,
      "eval_steps_per_second": 54.217,
      "step": 1464000
    },
    {
      "epoch": 3.1,
      "learning_rate": 1.5853333333333332e-07,
      "loss": 2.0767,
      "step": 1472000
    },
    {
      "epoch": 3.1,
      "eval_loss": 1.931079387664795,
      "eval_runtime": 461.0181,
      "eval_samples_per_second": 867.645,
      "eval_steps_per_second": 54.228,
      "step": 1472000
    },
    {
      "epoch": 3.12,
      "eval_loss": 1.9322373867034912,
      "eval_runtime": 461.2145,
      "eval_samples_per_second": 867.275,
      "eval_steps_per_second": 54.205,
      "step": 1480000
    },
    {
      "epoch": 3.13,
      "learning_rate": 1.558e-07,
      "loss": 2.082,
      "step": 1488000
    },
    {
      "epoch": 3.13,
      "eval_loss": 1.9361698627471924,
      "eval_runtime": 460.1009,
      "eval_samples_per_second": 869.375,
      "eval_steps_per_second": 54.336,
      "step": 1488000
    },
    {
      "epoch": 3.15,
      "eval_loss": 1.932852864265442,
      "eval_runtime": 461.8912,
      "eval_samples_per_second": 866.005,
      "eval_steps_per_second": 54.125,
      "step": 1496000
    },
    {
      "epoch": 3.17,
      "learning_rate": 1.5306666666666666e-07,
      "loss": 2.0774,
      "step": 1504000
    },
    {
      "epoch": 3.17,
      "eval_loss": 1.933501124382019,
      "eval_runtime": 461.263,
      "eval_samples_per_second": 867.184,
      "eval_steps_per_second": 54.199,
      "step": 1504000
    },
    {
      "epoch": 3.18,
      "eval_loss": 1.9341800212860107,
      "eval_runtime": 462.3299,
      "eval_samples_per_second": 865.183,
      "eval_steps_per_second": 54.074,
      "step": 1512000
    },
    {
      "epoch": 3.2,
      "learning_rate": 1.5033333333333332e-07,
      "loss": 2.0793,
      "step": 1520000
    },
    {
      "epoch": 3.2,
      "eval_loss": 1.9325579404830933,
      "eval_runtime": 462.9892,
      "eval_samples_per_second": 863.951,
      "eval_steps_per_second": 53.997,
      "step": 1520000
    },
    {
      "epoch": 3.22,
      "eval_loss": 1.9313201904296875,
      "eval_runtime": 462.2771,
      "eval_samples_per_second": 865.282,
      "eval_steps_per_second": 54.08,
      "step": 1528000
    },
    {
      "epoch": 3.23,
      "learning_rate": 1.476e-07,
      "loss": 2.0834,
      "step": 1536000
    },
    {
      "epoch": 3.23,
      "eval_loss": 1.9301501512527466,
      "eval_runtime": 461.1678,
      "eval_samples_per_second": 867.363,
      "eval_steps_per_second": 54.21,
      "step": 1536000
    },
    {
      "epoch": 3.25,
      "eval_loss": 1.9299404621124268,
      "eval_runtime": 463.2734,
      "eval_samples_per_second": 863.421,
      "eval_steps_per_second": 53.964,
      "step": 1544000
    },
    {
      "epoch": 3.27,
      "learning_rate": 1.4486666666666665e-07,
      "loss": 2.0698,
      "step": 1552000
    },
    {
      "epoch": 3.27,
      "eval_loss": 1.9287976026535034,
      "eval_runtime": 462.9712,
      "eval_samples_per_second": 863.985,
      "eval_steps_per_second": 53.999,
      "step": 1552000
    },
    {
      "epoch": 3.28,
      "eval_loss": 1.931135892868042,
      "eval_runtime": 463.0622,
      "eval_samples_per_second": 863.815,
      "eval_steps_per_second": 53.988,
      "step": 1560000
    },
    {
      "epoch": 3.3,
      "learning_rate": 1.4213333333333334e-07,
      "loss": 2.0721,
      "step": 1568000
    },
    {
      "epoch": 3.3,
      "eval_loss": 1.926220178604126,
      "eval_runtime": 461.1578,
      "eval_samples_per_second": 867.382,
      "eval_steps_per_second": 54.211,
      "step": 1568000
    },
    {
      "epoch": 3.32,
      "eval_loss": 1.9320148229599,
      "eval_runtime": 460.1501,
      "eval_samples_per_second": 869.282,
      "eval_steps_per_second": 54.33,
      "step": 1576000
    },
    {
      "epoch": 3.33,
      "learning_rate": 1.3940000000000002e-07,
      "loss": 2.0742,
      "step": 1584000
    },
    {
      "epoch": 3.33,
      "eval_loss": 1.927839994430542,
      "eval_runtime": 463.0138,
      "eval_samples_per_second": 863.905,
      "eval_steps_per_second": 53.994,
      "step": 1584000
    },
    {
      "epoch": 3.35,
      "eval_loss": 1.9333292245864868,
      "eval_runtime": 461.8845,
      "eval_samples_per_second": 866.017,
      "eval_steps_per_second": 54.126,
      "step": 1592000
    },
    {
      "epoch": 3.37,
      "learning_rate": 1.3666666666666665e-07,
      "loss": 2.0774,
      "step": 1600000
    },
    {
      "epoch": 3.37,
      "eval_loss": 1.9251993894577026,
      "eval_runtime": 464.4513,
      "eval_samples_per_second": 861.231,
      "eval_steps_per_second": 53.827,
      "step": 1600000
    },
    {
      "epoch": 3.39,
      "eval_loss": 1.930081844329834,
      "eval_runtime": 465.4239,
      "eval_samples_per_second": 859.432,
      "eval_steps_per_second": 53.714,
      "step": 1608000
    },
    {
      "epoch": 3.4,
      "learning_rate": 1.3393333333333333e-07,
      "loss": 2.0766,
      "step": 1616000
    },
    {
      "epoch": 3.4,
      "eval_loss": 1.934423804283142,
      "eval_runtime": 468.7051,
      "eval_samples_per_second": 853.415,
      "eval_steps_per_second": 53.338,
      "step": 1616000
    },
    {
      "epoch": 3.42,
      "eval_loss": 1.9320313930511475,
      "eval_runtime": 464.1689,
      "eval_samples_per_second": 861.755,
      "eval_steps_per_second": 53.86,
      "step": 1624000
    },
    {
      "epoch": 3.44,
      "learning_rate": 1.312e-07,
      "loss": 2.0702,
      "step": 1632000
    },
    {
      "epoch": 3.44,
      "eval_loss": 1.9307453632354736,
      "eval_runtime": 464.689,
      "eval_samples_per_second": 860.791,
      "eval_steps_per_second": 53.799,
      "step": 1632000
    },
    {
      "epoch": 3.45,
      "eval_loss": 1.9304145574569702,
      "eval_runtime": 463.7779,
      "eval_samples_per_second": 862.482,
      "eval_steps_per_second": 53.905,
      "step": 1640000
    },
    {
      "epoch": 3.47,
      "learning_rate": 1.2846666666666667e-07,
      "loss": 2.0772,
      "step": 1648000
    },
    {
      "epoch": 3.47,
      "eval_loss": 1.9280321598052979,
      "eval_runtime": 465.1445,
      "eval_samples_per_second": 859.948,
      "eval_steps_per_second": 53.747,
      "step": 1648000
    },
    {
      "epoch": 3.49,
      "eval_loss": 1.9324473142623901,
      "eval_runtime": 466.7608,
      "eval_samples_per_second": 856.97,
      "eval_steps_per_second": 53.561,
      "step": 1656000
    },
    {
      "epoch": 3.5,
      "learning_rate": 1.2573333333333332e-07,
      "loss": 2.0757,
      "step": 1664000
    },
    {
      "epoch": 3.5,
      "eval_loss": 1.9342966079711914,
      "eval_runtime": 466.6862,
      "eval_samples_per_second": 857.107,
      "eval_steps_per_second": 53.569,
      "step": 1664000
    },
    {
      "epoch": 3.52,
      "eval_loss": 1.9311579465866089,
      "eval_runtime": 464.2418,
      "eval_samples_per_second": 861.62,
      "eval_steps_per_second": 53.851,
      "step": 1672000
    },
    {
      "epoch": 3.54,
      "learning_rate": 1.23e-07,
      "loss": 2.0747,
      "step": 1680000
    },
    {
      "epoch": 3.54,
      "eval_loss": 1.9303609132766724,
      "eval_runtime": 463.1295,
      "eval_samples_per_second": 863.689,
      "eval_steps_per_second": 53.981,
      "step": 1680000
    },
    {
      "epoch": 3.55,
      "eval_loss": 1.9359545707702637,
      "eval_runtime": 463.1345,
      "eval_samples_per_second": 863.68,
      "eval_steps_per_second": 53.98,
      "step": 1688000
    },
    {
      "epoch": 3.57,
      "learning_rate": 1.2026666666666666e-07,
      "loss": 2.068,
      "step": 1696000
    },
    {
      "epoch": 3.57,
      "eval_loss": 1.9296965599060059,
      "eval_runtime": 463.742,
      "eval_samples_per_second": 862.548,
      "eval_steps_per_second": 53.909,
      "step": 1696000
    },
    {
      "epoch": 3.59,
      "eval_loss": 1.9337373971939087,
      "eval_runtime": 466.0229,
      "eval_samples_per_second": 858.327,
      "eval_steps_per_second": 53.645,
      "step": 1704000
    },
    {
      "epoch": 3.6,
      "learning_rate": 1.1753333333333334e-07,
      "loss": 2.0825,
      "step": 1712000
    },
    {
      "epoch": 3.6,
      "eval_loss": 1.9293195009231567,
      "eval_runtime": 468.6282,
      "eval_samples_per_second": 853.555,
      "eval_steps_per_second": 53.347,
      "step": 1712000
    },
    {
      "epoch": 3.62,
      "eval_loss": 1.929545283317566,
      "eval_runtime": 467.7047,
      "eval_samples_per_second": 855.24,
      "eval_steps_per_second": 53.453,
      "step": 1720000
    },
    {
      "epoch": 3.64,
      "learning_rate": 1.1480000000000001e-07,
      "loss": 2.0811,
      "step": 1728000
    },
    {
      "epoch": 3.64,
      "eval_loss": 1.9315083026885986,
      "eval_runtime": 464.4272,
      "eval_samples_per_second": 861.276,
      "eval_steps_per_second": 53.83,
      "step": 1728000
    },
    {
      "epoch": 3.65,
      "eval_loss": 1.9279147386550903,
      "eval_runtime": 464.6795,
      "eval_samples_per_second": 860.808,
      "eval_steps_per_second": 53.801,
      "step": 1736000
    },
    {
      "epoch": 3.67,
      "learning_rate": 1.1206666666666666e-07,
      "loss": 2.0844,
      "step": 1744000
    },
    {
      "epoch": 3.67,
      "eval_loss": 1.9289209842681885,
      "eval_runtime": 466.6058,
      "eval_samples_per_second": 857.255,
      "eval_steps_per_second": 53.578,
      "step": 1744000
    },
    {
      "epoch": 3.69,
      "eval_loss": 1.9279465675354004,
      "eval_runtime": 464.5589,
      "eval_samples_per_second": 861.032,
      "eval_steps_per_second": 53.814,
      "step": 1752000
    },
    {
      "epoch": 3.71,
      "learning_rate": 1.0933333333333333e-07,
      "loss": 2.0827,
      "step": 1760000
    },
    {
      "epoch": 3.71,
      "eval_loss": 1.9282883405685425,
      "eval_runtime": 467.4378,
      "eval_samples_per_second": 855.729,
      "eval_steps_per_second": 53.483,
      "step": 1760000
    },
    {
      "epoch": 3.72,
      "eval_loss": 1.9295260906219482,
      "eval_runtime": 466.5213,
      "eval_samples_per_second": 857.41,
      "eval_steps_per_second": 53.588,
      "step": 1768000
    },
    {
      "epoch": 3.74,
      "learning_rate": 1.066e-07,
      "loss": 2.0684,
      "step": 1776000
    },
    {
      "epoch": 3.74,
      "eval_loss": 1.9280706644058228,
      "eval_runtime": 468.3975,
      "eval_samples_per_second": 853.975,
      "eval_steps_per_second": 53.373,
      "step": 1776000
    },
    {
      "epoch": 3.76,
      "eval_loss": 1.9329652786254883,
      "eval_runtime": 467.0852,
      "eval_samples_per_second": 856.375,
      "eval_steps_per_second": 53.523,
      "step": 1784000
    },
    {
      "epoch": 3.77,
      "learning_rate": 1.0386666666666667e-07,
      "loss": 2.0724,
      "step": 1792000
    },
    {
      "epoch": 3.77,
      "eval_loss": 1.9293663501739502,
      "eval_runtime": 466.1999,
      "eval_samples_per_second": 858.001,
      "eval_steps_per_second": 53.625,
      "step": 1792000
    },
    {
      "epoch": 3.79,
      "eval_loss": 1.9276474714279175,
      "eval_runtime": 466.9185,
      "eval_samples_per_second": 856.681,
      "eval_steps_per_second": 53.543,
      "step": 1800000
    },
    {
      "epoch": 3.81,
      "learning_rate": 1.0113333333333334e-07,
      "loss": 2.074,
      "step": 1808000
    },
    {
      "epoch": 3.81,
      "eval_loss": 1.9226585626602173,
      "eval_runtime": 466.1573,
      "eval_samples_per_second": 858.079,
      "eval_steps_per_second": 53.63,
      "step": 1808000
    },
    {
      "epoch": 3.82,
      "eval_loss": 1.931994080543518,
      "eval_runtime": 468.1658,
      "eval_samples_per_second": 854.398,
      "eval_steps_per_second": 53.4,
      "step": 1816000
    },
    {
      "epoch": 3.84,
      "learning_rate": 9.84e-08,
      "loss": 2.0801,
      "step": 1824000
    },
    {
      "epoch": 3.84,
      "eval_loss": 1.927461862564087,
      "eval_runtime": 468.6252,
      "eval_samples_per_second": 853.561,
      "eval_steps_per_second": 53.348,
      "step": 1824000
    },
    {
      "epoch": 3.86,
      "eval_loss": 1.9301567077636719,
      "eval_runtime": 468.5364,
      "eval_samples_per_second": 853.722,
      "eval_steps_per_second": 53.358,
      "step": 1832000
    },
    {
      "epoch": 3.87,
      "learning_rate": 9.566666666666666e-08,
      "loss": 2.0783,
      "step": 1840000
    },
    {
      "epoch": 3.87,
      "eval_loss": 1.9332624673843384,
      "eval_runtime": 468.1279,
      "eval_samples_per_second": 854.467,
      "eval_steps_per_second": 53.404,
      "step": 1840000
    },
    {
      "epoch": 3.89,
      "eval_loss": 1.9296493530273438,
      "eval_runtime": 468.9713,
      "eval_samples_per_second": 852.931,
      "eval_steps_per_second": 53.308,
      "step": 1848000
    },
    {
      "epoch": 3.91,
      "learning_rate": 9.293333333333333e-08,
      "loss": 2.0787,
      "step": 1856000
    },
    {
      "epoch": 3.91,
      "eval_loss": 1.9301527738571167,
      "eval_runtime": 465.9133,
      "eval_samples_per_second": 858.529,
      "eval_steps_per_second": 53.658,
      "step": 1856000
    },
    {
      "epoch": 3.92,
      "eval_loss": 1.9347186088562012,
      "eval_runtime": 467.3282,
      "eval_samples_per_second": 855.93,
      "eval_steps_per_second": 53.496,
      "step": 1864000
    },
    {
      "epoch": 3.94,
      "learning_rate": 9.02e-08,
      "loss": 2.0733,
      "step": 1872000
    },
    {
      "epoch": 3.94,
      "eval_loss": 1.9297592639923096,
      "eval_runtime": 467.0743,
      "eval_samples_per_second": 856.395,
      "eval_steps_per_second": 53.525,
      "step": 1872000
    },
    {
      "epoch": 3.96,
      "eval_loss": 1.9302401542663574,
      "eval_runtime": 469.1433,
      "eval_samples_per_second": 852.618,
      "eval_steps_per_second": 53.289,
      "step": 1880000
    },
    {
      "epoch": 3.97,
      "learning_rate": 8.746666666666667e-08,
      "loss": 2.0742,
      "step": 1888000
    },
    {
      "epoch": 3.97,
      "eval_loss": 1.9278994798660278,
      "eval_runtime": 468.7857,
      "eval_samples_per_second": 853.268,
      "eval_steps_per_second": 53.329,
      "step": 1888000
    },
    {
      "epoch": 3.99,
      "eval_loss": 1.9257820844650269,
      "eval_runtime": 467.4664,
      "eval_samples_per_second": 855.676,
      "eval_steps_per_second": 53.48,
      "step": 1896000
    },
    {
      "epoch": 4.01,
      "learning_rate": 8.473333333333334e-08,
      "loss": 2.0769,
      "step": 1904000
    },
    {
      "epoch": 4.01,
      "eval_loss": 1.9254648685455322,
      "eval_runtime": 469.7925,
      "eval_samples_per_second": 851.44,
      "eval_steps_per_second": 53.215,
      "step": 1904000
    },
    {
      "epoch": 4.03,
      "eval_loss": 1.9282057285308838,
      "eval_runtime": 467.1165,
      "eval_samples_per_second": 856.317,
      "eval_steps_per_second": 53.52,
      "step": 1912000
    },
    {
      "epoch": 4.04,
      "learning_rate": 8.2e-08,
      "loss": 2.0736,
      "step": 1920000
    },
    {
      "epoch": 4.04,
      "eval_loss": 1.9297821521759033,
      "eval_runtime": 467.8574,
      "eval_samples_per_second": 854.961,
      "eval_steps_per_second": 53.435,
      "step": 1920000
    },
    {
      "epoch": 4.06,
      "eval_loss": 1.9324908256530762,
      "eval_runtime": 469.8593,
      "eval_samples_per_second": 851.319,
      "eval_steps_per_second": 53.207,
      "step": 1928000
    },
    {
      "epoch": 4.08,
      "learning_rate": 7.926666666666666e-08,
      "loss": 2.0713,
      "step": 1936000
    },
    {
      "epoch": 4.08,
      "eval_loss": 1.929565668106079,
      "eval_runtime": 473.7904,
      "eval_samples_per_second": 844.255,
      "eval_steps_per_second": 52.766,
      "step": 1936000
    },
    {
      "epoch": 4.09,
      "eval_loss": 1.9292526245117188,
      "eval_runtime": 474.7388,
      "eval_samples_per_second": 842.569,
      "eval_steps_per_second": 52.661,
      "step": 1944000
    },
    {
      "epoch": 4.11,
      "learning_rate": 7.653333333333333e-08,
      "loss": 2.0825,
      "step": 1952000
    },
    {
      "epoch": 4.11,
      "eval_loss": 1.9344995021820068,
      "eval_runtime": 474.9247,
      "eval_samples_per_second": 842.239,
      "eval_steps_per_second": 52.64,
      "step": 1952000
    },
    {
      "epoch": 4.13,
      "eval_loss": 1.934618592262268,
      "eval_runtime": 472.2426,
      "eval_samples_per_second": 847.022,
      "eval_steps_per_second": 52.939,
      "step": 1960000
    },
    {
      "epoch": 4.14,
      "learning_rate": 7.38e-08,
      "loss": 2.0828,
      "step": 1968000
    },
    {
      "epoch": 4.14,
      "eval_loss": 1.9310798645019531,
      "eval_runtime": 468.8102,
      "eval_samples_per_second": 853.224,
      "eval_steps_per_second": 53.326,
      "step": 1968000
    },
    {
      "epoch": 4.16,
      "eval_loss": 1.9307339191436768,
      "eval_runtime": 470.4345,
      "eval_samples_per_second": 850.278,
      "eval_steps_per_second": 53.142,
      "step": 1976000
    },
    {
      "epoch": 4.18,
      "learning_rate": 7.106666666666667e-08,
      "loss": 2.0821,
      "step": 1984000
    },
    {
      "epoch": 4.18,
      "eval_loss": 1.9335579872131348,
      "eval_runtime": 469.9157,
      "eval_samples_per_second": 851.216,
      "eval_steps_per_second": 53.201,
      "step": 1984000
    },
    {
      "epoch": 4.19,
      "eval_loss": 1.9265415668487549,
      "eval_runtime": 469.8042,
      "eval_samples_per_second": 851.418,
      "eval_steps_per_second": 53.214,
      "step": 1992000
    },
    {
      "epoch": 4.21,
      "learning_rate": 6.833333333333332e-08,
      "loss": 2.0768,
      "step": 2000000
    },
    {
      "epoch": 4.21,
      "eval_loss": 1.9284056425094604,
      "eval_runtime": 473.143,
      "eval_samples_per_second": 845.41,
      "eval_steps_per_second": 52.838,
      "step": 2000000
    },
    {
      "epoch": 4.23,
      "eval_loss": 1.9290404319763184,
      "eval_runtime": 475.6044,
      "eval_samples_per_second": 841.035,
      "eval_steps_per_second": 52.565,
      "step": 2008000
    },
    {
      "epoch": 4.24,
      "learning_rate": 6.56e-08,
      "loss": 2.0695,
      "step": 2016000
    },
    {
      "epoch": 4.24,
      "eval_loss": 1.9305514097213745,
      "eval_runtime": 470.7659,
      "eval_samples_per_second": 849.679,
      "eval_steps_per_second": 53.105,
      "step": 2016000
    },
    {
      "epoch": 4.26,
      "eval_loss": 1.9299392700195312,
      "eval_runtime": 474.6989,
      "eval_samples_per_second": 842.639,
      "eval_steps_per_second": 52.665,
      "step": 2024000
    },
    {
      "epoch": 4.28,
      "learning_rate": 6.286666666666666e-08,
      "loss": 2.0698,
      "step": 2032000
    },
    {
      "epoch": 4.28,
      "eval_loss": 1.9230471849441528,
      "eval_runtime": 473.1742,
      "eval_samples_per_second": 845.355,
      "eval_steps_per_second": 52.835,
      "step": 2032000
    },
    {
      "epoch": 4.29,
      "eval_loss": 1.9271833896636963,
      "eval_runtime": 471.6682,
      "eval_samples_per_second": 848.054,
      "eval_steps_per_second": 53.003,
      "step": 2040000
    },
    {
      "epoch": 4.31,
      "learning_rate": 6.013333333333333e-08,
      "loss": 2.0776,
      "step": 2048000
    },
    {
      "epoch": 4.31,
      "eval_loss": 1.9306389093399048,
      "eval_runtime": 472.8901,
      "eval_samples_per_second": 845.863,
      "eval_steps_per_second": 52.866,
      "step": 2048000
    },
    {
      "epoch": 4.33,
      "eval_loss": 1.9242680072784424,
      "eval_runtime": 472.2094,
      "eval_samples_per_second": 847.082,
      "eval_steps_per_second": 52.943,
      "step": 2056000
    },
    {
      "epoch": 4.35,
      "learning_rate": 5.7400000000000004e-08,
      "loss": 2.0797,
      "step": 2064000
    },
    {
      "epoch": 4.35,
      "eval_loss": 1.9265525341033936,
      "eval_runtime": 472.6552,
      "eval_samples_per_second": 846.283,
      "eval_steps_per_second": 52.893,
      "step": 2064000
    },
    {
      "epoch": 4.36,
      "eval_loss": 1.9248952865600586,
      "eval_runtime": 470.1754,
      "eval_samples_per_second": 850.746,
      "eval_steps_per_second": 53.172,
      "step": 2072000
    },
    {
      "epoch": 4.38,
      "learning_rate": 5.4666666666666666e-08,
      "loss": 2.0808,
      "step": 2080000
    },
    {
      "epoch": 4.38,
      "eval_loss": 1.927932858467102,
      "eval_runtime": 471.7256,
      "eval_samples_per_second": 847.951,
      "eval_steps_per_second": 52.997,
      "step": 2080000
    },
    {
      "epoch": 4.4,
      "eval_loss": 1.926186203956604,
      "eval_runtime": 472.0307,
      "eval_samples_per_second": 847.402,
      "eval_steps_per_second": 52.963,
      "step": 2088000
    },
    {
      "epoch": 4.41,
      "learning_rate": 5.1933333333333335e-08,
      "loss": 2.0776,
      "step": 2096000
    },
    {
      "epoch": 4.41,
      "eval_loss": 1.9350157976150513,
      "eval_runtime": 468.711,
      "eval_samples_per_second": 853.404,
      "eval_steps_per_second": 53.338,
      "step": 2096000
    },
    {
      "epoch": 4.43,
      "eval_loss": 1.9297165870666504,
      "eval_runtime": 470.9356,
      "eval_samples_per_second": 849.373,
      "eval_steps_per_second": 53.086,
      "step": 2104000
    },
    {
      "epoch": 4.45,
      "learning_rate": 4.92e-08,
      "loss": 2.0805,
      "step": 2112000
    },
    {
      "epoch": 4.45,
      "eval_loss": 1.933711051940918,
      "eval_runtime": 472.0879,
      "eval_samples_per_second": 847.3,
      "eval_steps_per_second": 52.956,
      "step": 2112000
    },
    {
      "epoch": 4.46,
      "eval_loss": 1.9301713705062866,
      "eval_runtime": 472.6372,
      "eval_samples_per_second": 846.315,
      "eval_steps_per_second": 52.895,
      "step": 2120000
    },
    {
      "epoch": 4.48,
      "learning_rate": 4.6466666666666666e-08,
      "loss": 2.0791,
      "step": 2128000
    },
    {
      "epoch": 4.48,
      "eval_loss": 1.9336822032928467,
      "eval_runtime": 472.4005,
      "eval_samples_per_second": 846.739,
      "eval_steps_per_second": 52.921,
      "step": 2128000
    },
    {
      "epoch": 4.5,
      "eval_loss": 1.9298382997512817,
      "eval_runtime": 470.1621,
      "eval_samples_per_second": 850.77,
      "eval_steps_per_second": 53.173,
      "step": 2136000
    },
    {
      "epoch": 4.51,
      "learning_rate": 4.3733333333333335e-08,
      "loss": 2.0771,
      "step": 2144000
    },
    {
      "epoch": 4.51,
      "eval_loss": 1.9268031120300293,
      "eval_runtime": 469.03,
      "eval_samples_per_second": 852.824,
      "eval_steps_per_second": 53.301,
      "step": 2144000
    },
    {
      "epoch": 4.53,
      "eval_loss": 1.9369832277297974,
      "eval_runtime": 469.9036,
      "eval_samples_per_second": 851.238,
      "eval_steps_per_second": 53.202,
      "step": 2152000
    },
    {
      "epoch": 4.55,
      "learning_rate": 4.1e-08,
      "loss": 2.0807,
      "step": 2160000
    },
    {
      "epoch": 4.55,
      "eval_loss": 1.93067467212677,
      "eval_runtime": 471.0137,
      "eval_samples_per_second": 849.232,
      "eval_steps_per_second": 53.077,
      "step": 2160000
    },
    {
      "epoch": 4.56,
      "eval_loss": 1.9291709661483765,
      "eval_runtime": 470.7625,
      "eval_samples_per_second": 849.685,
      "eval_steps_per_second": 53.105,
      "step": 2168000
    },
    {
      "epoch": 4.58,
      "learning_rate": 3.8266666666666665e-08,
      "loss": 2.0856,
      "step": 2176000
    },
    {
      "epoch": 4.58,
      "eval_loss": 1.9300168752670288,
      "eval_runtime": 473.6712,
      "eval_samples_per_second": 844.468,
      "eval_steps_per_second": 52.779,
      "step": 2176000
    },
    {
      "epoch": 4.6,
      "eval_loss": 1.932855486869812,
      "eval_runtime": 472.2129,
      "eval_samples_per_second": 847.076,
      "eval_steps_per_second": 52.942,
      "step": 2184000
    },
    {
      "epoch": 4.61,
      "learning_rate": 3.5533333333333334e-08,
      "loss": 2.0744,
      "step": 2192000
    },
    {
      "epoch": 4.61,
      "eval_loss": 1.9319262504577637,
      "eval_runtime": 471.3448,
      "eval_samples_per_second": 848.636,
      "eval_steps_per_second": 53.04,
      "step": 2192000
    },
    {
      "epoch": 4.63,
      "eval_loss": 1.9351770877838135,
      "eval_runtime": 471.0763,
      "eval_samples_per_second": 849.119,
      "eval_steps_per_second": 53.07,
      "step": 2200000
    },
    {
      "epoch": 4.65,
      "learning_rate": 3.28e-08,
      "loss": 2.0839,
      "step": 2208000
    },
    {
      "epoch": 4.65,
      "eval_loss": 1.9368445873260498,
      "eval_runtime": 472.7057,
      "eval_samples_per_second": 846.193,
      "eval_steps_per_second": 52.887,
      "step": 2208000
    },
    {
      "epoch": 4.67,
      "eval_loss": 1.9342936277389526,
      "eval_runtime": 470.7852,
      "eval_samples_per_second": 849.644,
      "eval_steps_per_second": 53.103,
      "step": 2216000
    },
    {
      "epoch": 4.68,
      "learning_rate": 3.0066666666666665e-08,
      "loss": 2.0706,
      "step": 2224000
    },
    {
      "epoch": 4.68,
      "eval_loss": 1.929012656211853,
      "eval_runtime": 471.422,
      "eval_samples_per_second": 848.497,
      "eval_steps_per_second": 53.031,
      "step": 2224000
    },
    {
      "epoch": 4.7,
      "eval_loss": 1.934726357460022,
      "eval_runtime": 473.3032,
      "eval_samples_per_second": 845.124,
      "eval_steps_per_second": 52.82,
      "step": 2232000
    },
    {
      "epoch": 4.72,
      "learning_rate": 2.7333333333333333e-08,
      "loss": 2.0745,
      "step": 2240000
    },
    {
      "epoch": 4.72,
      "eval_loss": 1.9293938875198364,
      "eval_runtime": 473.9686,
      "eval_samples_per_second": 843.938,
      "eval_steps_per_second": 52.746,
      "step": 2240000
    },
    {
      "epoch": 4.73,
      "eval_loss": 1.9254814386367798,
      "eval_runtime": 474.8799,
      "eval_samples_per_second": 842.318,
      "eval_steps_per_second": 52.645,
      "step": 2248000
    },
    {
      "epoch": 4.75,
      "learning_rate": 2.46e-08,
      "loss": 2.0767,
      "step": 2256000
    },
    {
      "epoch": 4.75,
      "eval_loss": 1.927069902420044,
      "eval_runtime": 471.8653,
      "eval_samples_per_second": 847.7,
      "eval_steps_per_second": 52.981,
      "step": 2256000
    },
    {
      "epoch": 4.77,
      "eval_loss": 1.9295985698699951,
      "eval_runtime": 470.4126,
      "eval_samples_per_second": 850.317,
      "eval_steps_per_second": 53.145,
      "step": 2264000
    },
    {
      "epoch": 4.78,
      "learning_rate": 2.1866666666666667e-08,
      "loss": 2.0753,
      "step": 2272000
    },
    {
      "epoch": 4.78,
      "eval_loss": 1.9268025159835815,
      "eval_runtime": 473.0248,
      "eval_samples_per_second": 845.622,
      "eval_steps_per_second": 52.851,
      "step": 2272000
    },
    {
      "epoch": 4.8,
      "eval_loss": 1.9291967153549194,
      "eval_runtime": 472.2138,
      "eval_samples_per_second": 847.074,
      "eval_steps_per_second": 52.942,
      "step": 2280000
    },
    {
      "epoch": 4.82,
      "learning_rate": 1.9133333333333333e-08,
      "loss": 2.0716,
      "step": 2288000
    },
    {
      "epoch": 4.82,
      "eval_loss": 1.9310023784637451,
      "eval_runtime": 473.6905,
      "eval_samples_per_second": 844.433,
      "eval_steps_per_second": 52.777,
      "step": 2288000
    },
    {
      "epoch": 4.83,
      "eval_loss": 1.9267252683639526,
      "eval_runtime": 474.7542,
      "eval_samples_per_second": 842.541,
      "eval_steps_per_second": 52.659,
      "step": 2296000
    },
    {
      "epoch": 4.85,
      "learning_rate": 1.64e-08,
      "loss": 2.0778,
      "step": 2304000
    },
    {
      "epoch": 4.85,
      "eval_loss": 1.9301416873931885,
      "eval_runtime": 476.5948,
      "eval_samples_per_second": 839.287,
      "eval_steps_per_second": 52.455,
      "step": 2304000
    },
    {
      "epoch": 4.87,
      "eval_loss": 1.9280035495758057,
      "eval_runtime": 475.4076,
      "eval_samples_per_second": 841.383,
      "eval_steps_per_second": 52.586,
      "step": 2312000
    },
    {
      "epoch": 4.88,
      "learning_rate": 1.3666666666666667e-08,
      "loss": 2.0724,
      "step": 2320000
    },
    {
      "epoch": 4.88,
      "eval_loss": 1.9283393621444702,
      "eval_runtime": 475.7915,
      "eval_samples_per_second": 840.704,
      "eval_steps_per_second": 52.544,
      "step": 2320000
    },
    {
      "epoch": 4.9,
      "eval_loss": 1.9288876056671143,
      "eval_runtime": 478.8452,
      "eval_samples_per_second": 835.343,
      "eval_steps_per_second": 52.209,
      "step": 2328000
    },
    {
      "epoch": 4.92,
      "learning_rate": 1.0933333333333334e-08,
      "loss": 2.0811,
      "step": 2336000
    },
    {
      "epoch": 4.92,
      "eval_loss": 1.931495189666748,
      "eval_runtime": 475.8623,
      "eval_samples_per_second": 840.579,
      "eval_steps_per_second": 52.536,
      "step": 2336000
    },
    {
      "epoch": 4.93,
      "eval_loss": 1.9268301725387573,
      "eval_runtime": 473.6667,
      "eval_samples_per_second": 844.476,
      "eval_steps_per_second": 52.78,
      "step": 2344000
    },
    {
      "epoch": 4.95,
      "learning_rate": 8.2e-09,
      "loss": 2.0816,
      "step": 2352000
    },
    {
      "epoch": 4.95,
      "eval_loss": 1.9304386377334595,
      "eval_runtime": 475.5489,
      "eval_samples_per_second": 841.133,
      "eval_steps_per_second": 52.571,
      "step": 2352000
    },
    {
      "epoch": 4.97,
      "eval_loss": 1.9301530122756958,
      "eval_runtime": 475.0873,
      "eval_samples_per_second": 841.951,
      "eval_steps_per_second": 52.622,
      "step": 2360000
    },
    {
      "epoch": 4.99,
      "learning_rate": 5.466666666666667e-09,
      "loss": 2.0775,
      "step": 2368000
    },
    {
      "epoch": 4.99,
      "eval_loss": 1.929231882095337,
      "eval_runtime": 473.3176,
      "eval_samples_per_second": 845.098,
      "eval_steps_per_second": 52.819,
      "step": 2368000
    },
    {
      "epoch": 5.0,
      "eval_loss": 1.927372694015503,
      "eval_runtime": 472.3148,
      "eval_samples_per_second": 846.893,
      "eval_steps_per_second": 52.931,
      "step": 2376000
    },
    {
      "epoch": 5.02,
      "learning_rate": 2.7333333333333334e-09,
      "loss": 2.0807,
      "step": 2384000
    },
    {
      "epoch": 5.02,
      "eval_loss": 1.9316951036453247,
      "eval_runtime": 472.3854,
      "eval_samples_per_second": 846.766,
      "eval_steps_per_second": 52.923,
      "step": 2384000
    },
    {
      "epoch": 5.04,
      "eval_loss": 1.9297714233398438,
      "eval_runtime": 474.0973,
      "eval_samples_per_second": 843.709,
      "eval_steps_per_second": 52.732,
      "step": 2392000
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0,
      "loss": 2.0668,
      "step": 2400000
    },
    {
      "epoch": 5.05,
      "eval_loss": 1.9348891973495483,
      "eval_runtime": 472.3546,
      "eval_samples_per_second": 846.821,
      "eval_steps_per_second": 52.926,
      "step": 2400000
    },
    {
      "epoch": 5.05,
      "step": 2400000,
      "total_flos": 7.57317099323892e+17,
      "train_loss": 2.0860133251953124,
      "train_runtime": 289964.3288,
      "train_samples_per_second": 132.43,
      "train_steps_per_second": 8.277
    }
  ],
  "logging_steps": 16000,
  "max_steps": 2400000,
  "num_train_epochs": 6,
  "save_steps": 32000,
  "total_flos": 7.57317099323892e+17,
  "trial_name": null,
  "trial_params": null
}