{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.929814973846515,
  "global_step": 980000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "learning_rate": 1.4999999999999998e-06,
      "loss": 0.9292,
      "step": 500
    },
    {
      "epoch": 0.01,
      "learning_rate": 2.9999999999999997e-06,
      "loss": 0.7143,
      "step": 1000
    },
    {
      "epoch": 0.01,
      "eval_loss": 0.6849376559257507,
      "eval_runtime": 2.3913,
      "eval_samples_per_second": 960.554,
      "eval_steps_per_second": 15.054,
      "step": 1000
    },
    {
      "epoch": 0.02,
      "learning_rate": 4.499999999999999e-06,
      "loss": 0.6841,
      "step": 1500
    },
    {
      "epoch": 0.02,
      "learning_rate": 5.999999999999999e-06,
      "loss": 0.6817,
      "step": 2000
    },
    {
      "epoch": 0.02,
      "eval_loss": 0.678993284702301,
      "eval_runtime": 2.3276,
      "eval_samples_per_second": 986.872,
      "eval_steps_per_second": 15.467,
      "step": 2000
    },
    {
      "epoch": 0.03,
      "learning_rate": 7.499999999999999e-06,
      "loss": 0.6813,
      "step": 2500
    },
    {
      "epoch": 0.03,
      "learning_rate": 8.999999999999999e-06,
      "loss": 0.6806,
      "step": 3000
    },
    {
      "epoch": 0.03,
      "eval_loss": 0.6789333820343018,
      "eval_runtime": 2.3788,
      "eval_samples_per_second": 965.594,
      "eval_steps_per_second": 15.133,
      "step": 3000
    },
    {
      "epoch": 0.04,
      "learning_rate": 1.05e-05,
      "loss": 0.6806,
      "step": 3500
    },
    {
      "epoch": 0.04,
      "learning_rate": 1.1999999999999999e-05,
      "loss": 0.6804,
      "step": 4000
    },
    {
      "epoch": 0.04,
      "eval_loss": 0.6781470775604248,
      "eval_runtime": 2.452,
      "eval_samples_per_second": 936.772,
      "eval_steps_per_second": 14.682,
      "step": 4000
    },
    {
      "epoch": 0.05,
      "learning_rate": 1.3499999999999998e-05,
      "loss": 0.6799,
      "step": 4500
    },
    {
      "epoch": 0.06,
      "learning_rate": 1.4999999999999999e-05,
      "loss": 0.6797,
      "step": 5000
    },
    {
      "epoch": 0.06,
      "eval_loss": 0.6779675483703613,
      "eval_runtime": 2.3912,
      "eval_samples_per_second": 960.599,
      "eval_steps_per_second": 15.055,
      "step": 5000
    },
    {
      "epoch": 0.06,
      "learning_rate": 1.6499999999999998e-05,
      "loss": 0.6796,
      "step": 5500
    },
    {
      "epoch": 0.07,
      "learning_rate": 1.7999999999999997e-05,
      "loss": 0.6793,
      "step": 6000
    },
    {
      "epoch": 0.07,
      "eval_loss": 0.677739143371582,
      "eval_runtime": 2.3548,
      "eval_samples_per_second": 975.453,
      "eval_steps_per_second": 15.288,
      "step": 6000
    },
    {
      "epoch": 0.07,
      "learning_rate": 1.95e-05,
      "loss": 0.6791,
      "step": 6500
    },
    {
      "epoch": 0.08,
      "learning_rate": 2.1e-05,
      "loss": 0.6787,
      "step": 7000
    },
    {
      "epoch": 0.08,
      "eval_loss": 0.6776532530784607,
      "eval_runtime": 2.3673,
      "eval_samples_per_second": 970.284,
      "eval_steps_per_second": 15.207,
      "step": 7000
    },
    {
      "epoch": 0.08,
      "learning_rate": 2.2499999999999998e-05,
      "loss": 0.6784,
      "step": 7500
    },
    {
      "epoch": 0.09,
      "learning_rate": 2.3999999999999997e-05,
      "loss": 0.6775,
      "step": 8000
    },
    {
      "epoch": 0.09,
      "eval_loss": 0.6731997728347778,
      "eval_runtime": 2.3959,
      "eval_samples_per_second": 958.715,
      "eval_steps_per_second": 15.026,
      "step": 8000
    },
    {
      "epoch": 0.09,
      "learning_rate": 2.55e-05,
      "loss": 0.6689,
      "step": 8500
    },
    {
      "epoch": 0.1,
      "learning_rate": 2.6999999999999996e-05,
      "loss": 0.6553,
      "step": 9000
    },
    {
      "epoch": 0.1,
      "eval_loss": 0.645298957824707,
      "eval_runtime": 2.3807,
      "eval_samples_per_second": 964.84,
      "eval_steps_per_second": 15.122,
      "step": 9000
    },
    {
      "epoch": 0.11,
      "learning_rate": 2.8499999999999998e-05,
      "loss": 0.6445,
      "step": 9500
    },
    {
      "epoch": 0.11,
      "learning_rate": 2.9999999999999997e-05,
      "loss": 0.6351,
      "step": 10000
    },
    {
      "epoch": 0.11,
      "eval_loss": 0.6254131197929382,
      "eval_runtime": 2.3999,
      "eval_samples_per_second": 957.11,
      "eval_steps_per_second": 15.0,
      "step": 10000
    },
    {
      "epoch": 0.12,
      "learning_rate": 3.149999999999999e-05,
      "loss": 0.6293,
      "step": 10500
    },
    {
      "epoch": 0.12,
      "learning_rate": 3.2999999999999996e-05,
      "loss": 0.6245,
      "step": 11000
    },
    {
      "epoch": 0.12,
      "eval_loss": 0.6135886311531067,
      "eval_runtime": 2.3072,
      "eval_samples_per_second": 995.595,
      "eval_steps_per_second": 15.604,
      "step": 11000
    },
    {
      "epoch": 0.13,
      "learning_rate": 3.45e-05,
      "loss": 0.6197,
      "step": 11500
    },
    {
      "epoch": 0.13,
      "learning_rate": 3.5999999999999994e-05,
      "loss": 0.6148,
      "step": 12000
    },
    {
      "epoch": 0.13,
      "eval_loss": 0.6036174297332764,
      "eval_runtime": 2.3862,
      "eval_samples_per_second": 962.612,
      "eval_steps_per_second": 15.087,
      "step": 12000
    },
    {
      "epoch": 0.14,
      "learning_rate": 3.75e-05,
      "loss": 0.6096,
      "step": 12500
    },
    {
      "epoch": 0.14,
      "learning_rate": 3.9e-05,
      "loss": 0.6046,
      "step": 13000
    },
    {
      "epoch": 0.14,
      "eval_loss": 0.5924868583679199,
      "eval_runtime": 2.386,
      "eval_samples_per_second": 962.685,
      "eval_steps_per_second": 15.088,
      "step": 13000
    },
    {
      "epoch": 0.15,
      "learning_rate": 4.05e-05,
      "loss": 0.5998,
      "step": 13500
    },
    {
      "epoch": 0.16,
      "learning_rate": 4.2e-05,
      "loss": 0.5951,
      "step": 14000
    },
    {
      "epoch": 0.16,
      "eval_loss": 0.5810989141464233,
      "eval_runtime": 2.4081,
      "eval_samples_per_second": 953.875,
      "eval_steps_per_second": 14.95,
      "step": 14000
    },
    {
      "epoch": 0.16,
      "learning_rate": 4.3499999999999993e-05,
      "loss": 0.5908,
      "step": 14500
    },
    {
      "epoch": 0.17,
      "learning_rate": 4.4999999999999996e-05,
      "loss": 0.586,
      "step": 15000
    },
    {
      "epoch": 0.17,
      "eval_loss": 0.5703989863395691,
      "eval_runtime": 2.3852,
      "eval_samples_per_second": 963.035,
      "eval_steps_per_second": 15.093,
      "step": 15000
    },
    {
      "epoch": 0.17,
      "learning_rate": 4.65e-05,
      "loss": 0.5815,
      "step": 15500
    },
    {
      "epoch": 0.18,
      "learning_rate": 4.7999999999999994e-05,
      "loss": 0.5769,
      "step": 16000
    },
    {
      "epoch": 0.18,
      "eval_loss": 0.5616013407707214,
      "eval_runtime": 2.3872,
      "eval_samples_per_second": 962.217,
      "eval_steps_per_second": 15.08,
      "step": 16000
    },
    {
      "epoch": 0.18,
      "learning_rate": 4.95e-05,
      "loss": 0.5722,
      "step": 16500
    },
    {
      "epoch": 0.19,
      "learning_rate": 5.1e-05,
      "loss": 0.5673,
      "step": 17000
    },
    {
      "epoch": 0.19,
      "eval_loss": 0.5503653287887573,
      "eval_runtime": 2.3451,
      "eval_samples_per_second": 979.501,
      "eval_steps_per_second": 15.351,
      "step": 17000
    },
    {
      "epoch": 0.2,
      "learning_rate": 5.2499999999999995e-05,
      "loss": 0.5616,
      "step": 17500
    },
    {
      "epoch": 0.2,
      "learning_rate": 5.399999999999999e-05,
      "loss": 0.5553,
      "step": 18000
    },
    {
      "epoch": 0.2,
      "eval_loss": 0.5397240519523621,
      "eval_runtime": 2.4358,
      "eval_samples_per_second": 943.009,
      "eval_steps_per_second": 14.779,
      "step": 18000
    },
    {
      "epoch": 0.21,
      "learning_rate": 5.5499999999999994e-05,
      "loss": 0.5491,
      "step": 18500
    },
    {
      "epoch": 0.21,
      "learning_rate": 5.6999999999999996e-05,
      "loss": 0.5421,
      "step": 19000
    },
    {
      "epoch": 0.21,
      "eval_loss": 0.5266169905662537,
      "eval_runtime": 2.4081,
      "eval_samples_per_second": 953.853,
      "eval_steps_per_second": 14.949,
      "step": 19000
    },
    {
      "epoch": 0.22,
      "learning_rate": 5.85e-05,
      "loss": 0.5357,
      "step": 19500
    },
    {
      "epoch": 0.22,
      "learning_rate": 5.9999999999999995e-05,
      "loss": 0.5301,
      "step": 20000
    },
    {
      "epoch": 0.22,
      "eval_loss": 0.5150080919265747,
      "eval_runtime": 2.3725,
      "eval_samples_per_second": 968.19,
      "eval_steps_per_second": 15.174,
      "step": 20000
    },
    {
      "epoch": 0.23,
      "learning_rate": 6.149999999999999e-05,
      "loss": 0.5248,
      "step": 20500
    },
    {
      "epoch": 0.23,
      "learning_rate": 6.299999999999999e-05,
      "loss": 0.5203,
      "step": 21000
    },
    {
      "epoch": 0.23,
      "eval_loss": 0.5055692195892334,
      "eval_runtime": 2.4443,
      "eval_samples_per_second": 939.719,
      "eval_steps_per_second": 14.728,
      "step": 21000
    },
    {
      "epoch": 0.24,
      "learning_rate": 6.45e-05,
      "loss": 0.5151,
      "step": 21500
    },
    {
      "epoch": 0.25,
      "learning_rate": 6.599999999999999e-05,
      "loss": 0.5105,
      "step": 22000
    },
    {
      "epoch": 0.25,
      "eval_loss": 0.4966994822025299,
      "eval_runtime": 2.3756,
      "eval_samples_per_second": 966.927,
      "eval_steps_per_second": 15.154,
      "step": 22000
    },
    {
      "epoch": 0.25,
      "learning_rate": 6.75e-05,
      "loss": 0.5062,
      "step": 22500
    },
    {
      "epoch": 0.26,
      "learning_rate": 6.9e-05,
      "loss": 0.5023,
      "step": 23000
    },
    {
      "epoch": 0.26,
      "eval_loss": 0.4855397045612335,
      "eval_runtime": 2.4105,
      "eval_samples_per_second": 952.91,
      "eval_steps_per_second": 14.935,
      "step": 23000
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.049999999999999e-05,
      "loss": 0.4981,
      "step": 23500
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.199999999999999e-05,
      "loss": 0.4951,
      "step": 24000
    },
    {
      "epoch": 0.27,
      "eval_loss": 0.4793773889541626,
      "eval_runtime": 2.3882,
      "eval_samples_per_second": 961.816,
      "eval_steps_per_second": 15.074,
      "step": 24000
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.35e-05,
      "loss": 0.4904,
      "step": 24500
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.5e-05,
      "loss": 0.4874,
      "step": 25000
    },
    {
      "epoch": 0.28,
      "eval_loss": 0.4764867126941681,
      "eval_runtime": 2.4085,
      "eval_samples_per_second": 953.713,
      "eval_steps_per_second": 14.947,
      "step": 25000
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.649999999999999e-05,
      "loss": 0.4837,
      "step": 25500
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.8e-05,
      "loss": 0.4803,
      "step": 26000
    },
    {
      "epoch": 0.29,
      "eval_loss": 0.46588340401649475,
      "eval_runtime": 2.4213,
      "eval_samples_per_second": 948.654,
      "eval_steps_per_second": 14.868,
      "step": 26000
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.95e-05,
      "loss": 0.4768,
      "step": 26500
    },
    {
      "epoch": 0.3,
      "learning_rate": 8.1e-05,
      "loss": 0.4731,
      "step": 27000
    },
    {
      "epoch": 0.3,
      "eval_loss": 0.4568893611431122,
      "eval_runtime": 2.416,
      "eval_samples_per_second": 950.734,
      "eval_steps_per_second": 14.9,
      "step": 27000
    },
    {
      "epoch": 0.31,
      "learning_rate": 8.25e-05,
      "loss": 0.4701,
      "step": 27500
    },
    {
      "epoch": 0.31,
      "learning_rate": 8.4e-05,
      "loss": 0.4673,
      "step": 28000
    },
    {
      "epoch": 0.31,
      "eval_loss": 0.45133140683174133,
      "eval_runtime": 2.3585,
      "eval_samples_per_second": 973.912,
      "eval_steps_per_second": 15.264,
      "step": 28000
    },
    {
      "epoch": 0.32,
      "learning_rate": 8.549999999999999e-05,
      "loss": 0.4637,
      "step": 28500
    },
    {
      "epoch": 0.32,
      "learning_rate": 8.699999999999999e-05,
      "loss": 0.4607,
      "step": 29000
    },
    {
      "epoch": 0.32,
      "eval_loss": 0.44702839851379395,
      "eval_runtime": 2.4192,
      "eval_samples_per_second": 949.488,
      "eval_steps_per_second": 14.881,
      "step": 29000
    },
    {
      "epoch": 0.33,
      "learning_rate": 8.849999999999998e-05,
      "loss": 0.4579,
      "step": 29500
    },
    {
      "epoch": 0.33,
      "learning_rate": 8.999999999999999e-05,
      "loss": 0.4542,
      "step": 30000
    },
    {
      "epoch": 0.33,
      "eval_loss": 0.4402887225151062,
      "eval_runtime": 2.4131,
      "eval_samples_per_second": 951.902,
      "eval_steps_per_second": 14.919,
      "step": 30000
    },
    {
      "epoch": 0.34,
      "learning_rate": 9.149999999999999e-05,
      "loss": 0.4509,
      "step": 30500
    },
    {
      "epoch": 0.35,
      "learning_rate": 9.3e-05,
      "loss": 0.4477,
      "step": 31000
    },
    {
      "epoch": 0.35,
      "eval_loss": 0.4341259002685547,
      "eval_runtime": 2.3655,
      "eval_samples_per_second": 971.032,
      "eval_steps_per_second": 15.219,
      "step": 31000
    },
    {
      "epoch": 0.35,
      "learning_rate": 9.449999999999999e-05,
      "loss": 0.4443,
      "step": 31500
    },
    {
      "epoch": 0.36,
      "learning_rate": 9.599999999999999e-05,
      "loss": 0.4413,
      "step": 32000
    },
    {
      "epoch": 0.36,
      "eval_loss": 0.42718443274497986,
      "eval_runtime": 2.4008,
      "eval_samples_per_second": 956.762,
      "eval_steps_per_second": 14.995,
      "step": 32000
    },
    {
      "epoch": 0.36,
      "learning_rate": 9.75e-05,
      "loss": 0.4376,
      "step": 32500
    },
    {
      "epoch": 0.37,
      "learning_rate": 9.9e-05,
      "loss": 0.4341,
      "step": 33000
    },
    {
      "epoch": 0.37,
      "eval_loss": 0.41980886459350586,
      "eval_runtime": 2.3982,
      "eval_samples_per_second": 957.813,
      "eval_steps_per_second": 15.011,
      "step": 33000
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0001005,
      "loss": 0.4312,
      "step": 33500
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000102,
      "loss": 0.4289,
      "step": 34000
    },
    {
      "epoch": 0.38,
      "eval_loss": 0.4122560918331146,
      "eval_runtime": 2.425,
      "eval_samples_per_second": 947.221,
      "eval_steps_per_second": 14.845,
      "step": 34000
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00010349999999999998,
      "loss": 0.4257,
      "step": 34500
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00010499999999999999,
      "loss": 0.4224,
      "step": 35000
    },
    {
      "epoch": 0.39,
      "eval_loss": 0.40835943818092346,
      "eval_runtime": 2.4071,
      "eval_samples_per_second": 954.265,
      "eval_steps_per_second": 14.956,
      "step": 35000
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00010649999999999999,
      "loss": 0.4202,
      "step": 35500
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00010799999999999998,
      "loss": 0.4173,
      "step": 36000
    },
    {
      "epoch": 0.4,
      "eval_loss": 0.40327221155166626,
      "eval_runtime": 2.3753,
      "eval_samples_per_second": 967.051,
      "eval_steps_per_second": 15.156,
      "step": 36000
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00010949999999999999,
      "loss": 0.4142,
      "step": 36500
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00011099999999999999,
      "loss": 0.412,
      "step": 37000
    },
    {
      "epoch": 0.41,
      "eval_loss": 0.39642444252967834,
      "eval_runtime": 2.4394,
      "eval_samples_per_second": 941.616,
      "eval_steps_per_second": 14.758,
      "step": 37000
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0001125,
      "loss": 0.4098,
      "step": 37500
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00011399999999999999,
      "loss": 0.407,
      "step": 38000
    },
    {
      "epoch": 0.42,
      "eval_loss": 0.39146095514297485,
      "eval_runtime": 2.3792,
      "eval_samples_per_second": 965.449,
      "eval_steps_per_second": 15.131,
      "step": 38000
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00011549999999999999,
      "loss": 0.4033,
      "step": 38500
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000117,
      "loss": 0.402,
      "step": 39000
    },
    {
      "epoch": 0.43,
      "eval_loss": 0.3854062259197235,
      "eval_runtime": 2.4437,
      "eval_samples_per_second": 939.967,
      "eval_steps_per_second": 14.732,
      "step": 39000
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0001185,
      "loss": 0.3991,
      "step": 39500
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00011999999999999999,
      "loss": 0.3966,
      "step": 40000
    },
    {
      "epoch": 0.45,
      "eval_loss": 0.3808075189590454,
      "eval_runtime": 2.4355,
      "eval_samples_per_second": 943.135,
      "eval_steps_per_second": 14.781,
      "step": 40000
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0001215,
      "loss": 0.3943,
      "step": 40500
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00012299999999999998,
      "loss": 0.3929,
      "step": 41000
    },
    {
      "epoch": 0.46,
      "eval_loss": 0.37926527857780457,
      "eval_runtime": 2.3557,
      "eval_samples_per_second": 975.086,
      "eval_steps_per_second": 15.282,
      "step": 41000
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0001245,
      "loss": 0.39,
      "step": 41500
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00012599999999999997,
      "loss": 0.3873,
      "step": 42000
    },
    {
      "epoch": 0.47,
      "eval_loss": 0.37127774953842163,
      "eval_runtime": 2.3664,
      "eval_samples_per_second": 970.675,
      "eval_steps_per_second": 15.213,
      "step": 42000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00012749999999999998,
      "loss": 0.3861,
      "step": 42500
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000129,
      "loss": 0.3837,
      "step": 43000
    },
    {
      "epoch": 0.48,
      "eval_loss": 0.36950594186782837,
      "eval_runtime": 2.4197,
      "eval_samples_per_second": 949.3,
      "eval_steps_per_second": 14.878,
      "step": 43000
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0001305,
      "loss": 0.3812,
      "step": 43500
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00013199999999999998,
      "loss": 0.3793,
      "step": 44000
    },
    {
      "epoch": 0.49,
      "eval_loss": 0.3651977479457855,
      "eval_runtime": 2.3939,
      "eval_samples_per_second": 959.528,
      "eval_steps_per_second": 15.038,
      "step": 44000
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0001335,
      "loss": 0.3775,
      "step": 44500
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000135,
      "loss": 0.3756,
      "step": 45000
    },
    {
      "epoch": 0.5,
      "eval_loss": 0.3592735230922699,
      "eval_runtime": 2.3855,
      "eval_samples_per_second": 962.901,
      "eval_steps_per_second": 15.091,
      "step": 45000
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00013649999999999998,
      "loss": 0.3737,
      "step": 45500
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000138,
      "loss": 0.3718,
      "step": 46000
    },
    {
      "epoch": 0.51,
      "eval_loss": 0.3585481643676758,
      "eval_runtime": 2.3854,
      "eval_samples_per_second": 962.952,
      "eval_steps_per_second": 15.092,
      "step": 46000
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0001395,
      "loss": 0.3704,
      "step": 46500
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00014099999999999998,
      "loss": 0.3687,
      "step": 47000
    },
    {
      "epoch": 0.52,
      "eval_loss": 0.3562163710594177,
      "eval_runtime": 2.4137,
      "eval_samples_per_second": 951.637,
      "eval_steps_per_second": 14.915,
      "step": 47000
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0001425,
      "loss": 0.367,
      "step": 47500
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00014399999999999998,
      "loss": 0.3654,
      "step": 48000
    },
    {
      "epoch": 0.54,
      "eval_loss": 0.35154005885124207,
      "eval_runtime": 2.4671,
      "eval_samples_per_second": 931.04,
      "eval_steps_per_second": 14.592,
      "step": 48000
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00014549999999999999,
      "loss": 0.3638,
      "step": 48500
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000147,
      "loss": 0.3625,
      "step": 49000
    },
    {
      "epoch": 0.55,
      "eval_loss": 0.3474389910697937,
      "eval_runtime": 2.4177,
      "eval_samples_per_second": 950.09,
      "eval_steps_per_second": 14.89,
      "step": 49000
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00014849999999999998,
      "loss": 0.3612,
      "step": 49500
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00015,
      "loss": 0.3592,
      "step": 50000
    },
    {
      "epoch": 0.56,
      "eval_loss": 0.3449079096317291,
      "eval_runtime": 2.4174,
      "eval_samples_per_second": 950.212,
      "eval_steps_per_second": 14.892,
      "step": 50000
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00014999990431133645,
      "loss": 0.3579,
      "step": 50500
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0001499996172456075,
      "loss": 0.3563,
      "step": 51000
    },
    {
      "epoch": 0.57,
      "eval_loss": 0.3443795442581177,
      "eval_runtime": 2.3466,
      "eval_samples_per_second": 978.854,
      "eval_steps_per_second": 15.341,
      "step": 51000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00014999913880359787,
      "loss": 0.3547,
      "step": 51500
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00014999846898661572,
      "loss": 0.353,
      "step": 52000
    },
    {
      "epoch": 0.58,
      "eval_loss": 0.34043800830841064,
      "eval_runtime": 2.4046,
      "eval_samples_per_second": 955.259,
      "eval_steps_per_second": 14.971,
      "step": 52000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00014999760779649222,
      "loss": 0.3516,
      "step": 52500
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00014999655523558183,
      "loss": 0.3502,
      "step": 53000
    },
    {
      "epoch": 0.59,
      "eval_loss": 0.3358021378517151,
      "eval_runtime": 2.3861,
      "eval_samples_per_second": 962.643,
      "eval_steps_per_second": 15.087,
      "step": 53000
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00014999531130676229,
      "loss": 0.3491,
      "step": 53500
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00014999387601343436,
      "loss": 0.3473,
      "step": 54000
    },
    {
      "epoch": 0.6,
      "eval_loss": 0.3330630362033844,
      "eval_runtime": 2.3348,
      "eval_samples_per_second": 983.811,
      "eval_steps_per_second": 15.419,
      "step": 54000
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00014999224935952215,
      "loss": 0.3463,
      "step": 54500
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00014999043134947282,
      "loss": 0.3445,
      "step": 55000
    },
    {
      "epoch": 0.61,
      "eval_loss": 0.33030495047569275,
      "eval_runtime": 2.318,
      "eval_samples_per_second": 990.921,
      "eval_steps_per_second": 15.53,
      "step": 55000
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00014998842198825674,
      "loss": 0.3434,
      "step": 55500
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00014998622128136748,
      "loss": 0.342,
      "step": 56000
    },
    {
      "epoch": 0.62,
      "eval_loss": 0.3295433819293976,
      "eval_runtime": 2.3042,
      "eval_samples_per_second": 996.891,
      "eval_steps_per_second": 15.624,
      "step": 56000
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00014998382923482164,
      "loss": 0.3411,
      "step": 56500
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000149981245855159,
      "loss": 0.3396,
      "step": 57000
    },
    {
      "epoch": 0.64,
      "eval_loss": 0.32805606722831726,
      "eval_runtime": 2.4137,
      "eval_samples_per_second": 951.659,
      "eval_steps_per_second": 14.915,
      "step": 57000
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00014997847114944242,
      "loss": 0.3383,
      "step": 57500
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00014997550512525784,
      "loss": 0.3374,
      "step": 58000
    },
    {
      "epoch": 0.65,
      "eval_loss": 0.3256905674934387,
      "eval_runtime": 2.4211,
      "eval_samples_per_second": 948.748,
      "eval_steps_per_second": 14.869,
      "step": 58000
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00014997234779071426,
      "loss": 0.3365,
      "step": 58500
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001499689991544437,
      "loss": 0.3353,
      "step": 59000
    },
    {
      "epoch": 0.66,
      "eval_loss": 0.32548126578330994,
      "eval_runtime": 2.4309,
      "eval_samples_per_second": 944.932,
      "eval_steps_per_second": 14.81,
      "step": 59000
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001499654592256012,
      "loss": 0.3342,
      "step": 59500
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00014996172801386482,
      "loss": 0.3333,
      "step": 60000
    },
    {
      "epoch": 0.67,
      "eval_loss": 0.3190021514892578,
      "eval_runtime": 2.4381,
      "eval_samples_per_second": 942.115,
      "eval_steps_per_second": 14.765,
      "step": 60000
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00014995780552943551,
      "loss": 0.3321,
      "step": 60500
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00014995369178303722,
      "loss": 0.3311,
      "step": 61000
    },
    {
      "epoch": 0.68,
      "eval_loss": 0.3181557357311249,
      "eval_runtime": 2.3632,
      "eval_samples_per_second": 971.967,
      "eval_steps_per_second": 15.233,
      "step": 61000
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001499493867859168,
      "loss": 0.3298,
      "step": 61500
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001499448905498439,
      "loss": 0.3289,
      "step": 62000
    },
    {
      "epoch": 0.69,
      "eval_loss": 0.31774210929870605,
      "eval_runtime": 2.4377,
      "eval_samples_per_second": 942.269,
      "eval_steps_per_second": 14.768,
      "step": 62000
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014994020308711106,
      "loss": 0.3281,
      "step": 62500
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014993532441053364,
      "loss": 0.3272,
      "step": 63000
    },
    {
      "epoch": 0.7,
      "eval_loss": 0.31380537152290344,
      "eval_runtime": 2.4068,
      "eval_samples_per_second": 954.378,
      "eval_steps_per_second": 14.958,
      "step": 63000
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001499302545334498,
      "loss": 0.3262,
      "step": 63500
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001499249934697203,
      "loss": 0.3253,
      "step": 64000
    },
    {
      "epoch": 0.71,
      "eval_loss": 0.3134210705757141,
      "eval_runtime": 2.4456,
      "eval_samples_per_second": 939.234,
      "eval_steps_per_second": 14.72,
      "step": 64000
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014991954123372875,
      "loss": 0.3246,
      "step": 64500
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001499138978403813,
      "loss": 0.3242,
      "step": 65000
    },
    {
      "epoch": 0.72,
      "eval_loss": 0.3107437193393707,
      "eval_runtime": 2.3823,
      "eval_samples_per_second": 964.183,
      "eval_steps_per_second": 15.111,
      "step": 65000
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00014990806330510687,
      "loss": 0.3231,
      "step": 65500
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00014990203764385677,
      "loss": 0.3221,
      "step": 66000
    },
    {
      "epoch": 0.74,
      "eval_loss": 0.308339387178421,
      "eval_runtime": 2.4467,
      "eval_samples_per_second": 938.797,
      "eval_steps_per_second": 14.713,
      "step": 66000
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00014989582087310494,
      "loss": 0.3211,
      "step": 66500
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00014988941300984784,
      "loss": 0.3203,
      "step": 67000
    },
    {
      "epoch": 0.75,
      "eval_loss": 0.3079957365989685,
      "eval_runtime": 2.3779,
      "eval_samples_per_second": 965.976,
      "eval_steps_per_second": 15.139,
      "step": 67000
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00014988281407160426,
      "loss": 0.3194,
      "step": 67500
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001498760240764155,
      "loss": 0.3188,
      "step": 68000
    },
    {
      "epoch": 0.76,
      "eval_loss": 0.30548328161239624,
      "eval_runtime": 2.3902,
      "eval_samples_per_second": 961.017,
      "eval_steps_per_second": 15.062,
      "step": 68000
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00014986904304284512,
      "loss": 0.3181,
      "step": 68500
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.000149861870989979,
      "loss": 0.3169,
      "step": 69000
    },
    {
      "epoch": 0.77,
      "eval_loss": 0.3034500181674957,
      "eval_runtime": 2.3995,
      "eval_samples_per_second": 957.293,
      "eval_steps_per_second": 15.003,
      "step": 69000
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00014985450793742527,
      "loss": 0.3164,
      "step": 69500
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001498469539053142,
      "loss": 0.3157,
      "step": 70000
    },
    {
      "epoch": 0.78,
      "eval_loss": 0.3020155429840088,
      "eval_runtime": 2.3865,
      "eval_samples_per_second": 962.488,
      "eval_steps_per_second": 15.085,
      "step": 70000
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014983920891429827,
      "loss": 0.3158,
      "step": 70500
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014983127298555198,
      "loss": 0.3143,
      "step": 71000
    },
    {
      "epoch": 0.79,
      "eval_loss": 0.3028620779514313,
      "eval_runtime": 2.4295,
      "eval_samples_per_second": 945.481,
      "eval_steps_per_second": 14.818,
      "step": 71000
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014982314614077184,
      "loss": 0.3136,
      "step": 71500
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014981482840217632,
      "loss": 0.3136,
      "step": 72000
    },
    {
      "epoch": 0.8,
      "eval_loss": 0.2998126149177551,
      "eval_runtime": 2.3756,
      "eval_samples_per_second": 966.92,
      "eval_steps_per_second": 15.154,
      "step": 72000
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00014980631979250587,
      "loss": 0.3125,
      "step": 72500
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00014979762033502262,
      "loss": 0.3117,
      "step": 73000
    },
    {
      "epoch": 0.81,
      "eval_loss": 0.2993732988834381,
      "eval_runtime": 2.3811,
      "eval_samples_per_second": 964.682,
      "eval_steps_per_second": 15.119,
      "step": 73000
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001497887300535106,
      "loss": 0.3113,
      "step": 73500
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00014977964897227547,
      "loss": 0.3104,
      "step": 74000
    },
    {
      "epoch": 0.83,
      "eval_loss": 0.2992545962333679,
      "eval_runtime": 2.3838,
      "eval_samples_per_second": 963.585,
      "eval_steps_per_second": 15.102,
      "step": 74000
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001497703771161446,
      "loss": 0.3096,
      "step": 74500
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00014976091451046687,
      "loss": 0.3094,
      "step": 75000
    },
    {
      "epoch": 0.84,
      "eval_loss": 0.2973059415817261,
      "eval_runtime": 2.4216,
      "eval_samples_per_second": 948.528,
      "eval_steps_per_second": 14.866,
      "step": 75000
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00014975126118111268,
      "loss": 0.3086,
      "step": 75500
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00014974141715447386,
      "loss": 0.3075,
      "step": 76000
    },
    {
      "epoch": 0.85,
      "eval_loss": 0.2966085374355316,
      "eval_runtime": 2.4258,
      "eval_samples_per_second": 946.915,
      "eval_steps_per_second": 14.841,
      "step": 76000
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00014973138245746363,
      "loss": 0.3078,
      "step": 76500
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00014972115711751644,
      "loss": 0.3062,
      "step": 77000
    },
    {
      "epoch": 0.86,
      "eval_loss": 0.29561737179756165,
      "eval_runtime": 2.3702,
      "eval_samples_per_second": 969.126,
      "eval_steps_per_second": 15.189,
      "step": 77000
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00014971074116258796,
      "loss": 0.3063,
      "step": 77500
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.00014970013462115505,
      "loss": 0.3054,
      "step": 78000
    },
    {
      "epoch": 0.87,
      "eval_loss": 0.29380762577056885,
      "eval_runtime": 2.3743,
      "eval_samples_per_second": 967.43,
      "eval_steps_per_second": 15.162,
      "step": 78000
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.00014968933752221558,
      "loss": 0.3051,
      "step": 78500
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.00014967834989528843,
      "loss": 0.3044,
      "step": 79000
    },
    {
      "epoch": 0.88,
      "eval_loss": 0.29091179370880127,
      "eval_runtime": 2.4306,
      "eval_samples_per_second": 945.023,
      "eval_steps_per_second": 14.811,
      "step": 79000
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0001496671717704133,
      "loss": 0.3039,
      "step": 79500
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.00014965580317815078,
      "loss": 0.3031,
      "step": 80000
    },
    {
      "epoch": 0.89,
      "eval_loss": 0.2893276512622833,
      "eval_runtime": 2.3942,
      "eval_samples_per_second": 959.413,
      "eval_steps_per_second": 15.037,
      "step": 80000
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0001496442441495822,
      "loss": 0.303,
      "step": 80500
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.00014963249471630944,
      "loss": 0.3024,
      "step": 81000
    },
    {
      "epoch": 0.9,
      "eval_loss": 0.2880454361438751,
      "eval_runtime": 2.3598,
      "eval_samples_per_second": 973.408,
      "eval_steps_per_second": 15.256,
      "step": 81000
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.00014962055491045506,
      "loss": 0.3019,
      "step": 81500
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.000149608424764662,
      "loss": 0.3016,
      "step": 82000
    },
    {
      "epoch": 0.91,
      "eval_loss": 0.2889193892478943,
      "eval_runtime": 2.392,
      "eval_samples_per_second": 960.275,
      "eval_steps_per_second": 15.05,
      "step": 82000
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.00014959610431209363,
      "loss": 0.3008,
      "step": 82500
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0001495835935864336,
      "loss": 0.3001,
      "step": 83000
    },
    {
      "epoch": 0.93,
      "eval_loss": 0.2879210412502289,
      "eval_runtime": 2.3362,
      "eval_samples_per_second": 983.21,
      "eval_steps_per_second": 15.409,
      "step": 83000
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.00014957089262188571,
      "loss": 0.3001,
      "step": 83500
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.00014955800145317397,
      "loss": 0.2994,
      "step": 84000
    },
    {
      "epoch": 0.94,
      "eval_loss": 0.28926190733909607,
      "eval_runtime": 2.354,
      "eval_samples_per_second": 975.772,
      "eval_steps_per_second": 15.293,
      "step": 84000
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.00014954492011554234,
      "loss": 0.2992,
      "step": 84500
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.00014953164864475466,
      "loss": 0.2986,
      "step": 85000
    },
    {
      "epoch": 0.95,
      "eval_loss": 0.28680220246315,
      "eval_runtime": 2.4523,
      "eval_samples_per_second": 936.685,
      "eval_steps_per_second": 14.68,
      "step": 85000
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.00014951818707709463,
      "loss": 0.2981,
      "step": 85500
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0001495045354493657,
      "loss": 0.297,
      "step": 86000
    },
    {
      "epoch": 0.96,
      "eval_loss": 0.28654542565345764,
      "eval_runtime": 2.3431,
      "eval_samples_per_second": 980.338,
      "eval_steps_per_second": 15.364,
      "step": 86000
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.00014949069379889088,
      "loss": 0.2972,
      "step": 86500
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.00014947666216351272,
      "loss": 0.2965,
      "step": 87000
    },
    {
      "epoch": 0.97,
      "eval_loss": 0.2838832139968872,
      "eval_runtime": 2.3161,
      "eval_samples_per_second": 991.759,
      "eval_steps_per_second": 15.543,
      "step": 87000
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.00014946244058159313,
      "loss": 0.2961,
      "step": 87500
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.00014944802909201344,
      "loss": 0.2951,
      "step": 88000
    },
    {
      "epoch": 0.98,
      "eval_loss": 0.28492867946624756,
      "eval_runtime": 2.2962,
      "eval_samples_per_second": 1000.341,
      "eval_steps_per_second": 15.678,
      "step": 88000
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.00014943342773417407,
      "loss": 0.2954,
      "step": 88500
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.00014941863654799456,
      "loss": 0.2947,
      "step": 89000
    },
    {
      "epoch": 0.99,
      "eval_loss": 0.28069427609443665,
      "eval_runtime": 2.3864,
      "eval_samples_per_second": 962.549,
      "eval_steps_per_second": 15.086,
      "step": 89000
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0001494036555739135,
      "loss": 0.2943,
      "step": 89500
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00014938848485288825,
      "loss": 0.2943,
      "step": 90000
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.2816523611545563,
      "eval_runtime": 2.5066,
      "eval_samples_per_second": 916.368,
      "eval_steps_per_second": 14.362,
      "step": 90000
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000149373124426395,
      "loss": 0.2933,
      "step": 90500
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0001493575743364286,
      "loss": 0.2934,
      "step": 91000
    },
    {
      "epoch": 1.01,
      "eval_loss": 0.2812552750110626,
      "eval_runtime": 2.6367,
      "eval_samples_per_second": 871.176,
      "eval_steps_per_second": 13.654,
      "step": 91000
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00014934183462550238,
      "loss": 0.2926,
      "step": 91500
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00014932590533664808,
      "loss": 0.2923,
      "step": 92000
    },
    {
      "epoch": 1.03,
      "eval_loss": 0.2808503210544586,
      "eval_runtime": 2.4894,
      "eval_samples_per_second": 922.7,
      "eval_steps_per_second": 14.461,
      "step": 92000
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00014930978651341581,
      "loss": 0.292,
      "step": 92500
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0001492934781998738,
      "loss": 0.2918,
      "step": 93000
    },
    {
      "epoch": 1.04,
      "eval_loss": 0.2790589928627014,
      "eval_runtime": 2.5204,
      "eval_samples_per_second": 911.355,
      "eval_steps_per_second": 14.283,
      "step": 93000
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00014927698044060834,
      "loss": 0.291,
      "step": 93500
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0001492602932807237,
      "loss": 0.2911,
      "step": 94000
    },
    {
      "epoch": 1.05,
      "eval_loss": 0.2791271209716797,
      "eval_runtime": 2.5437,
      "eval_samples_per_second": 903.029,
      "eval_steps_per_second": 14.153,
      "step": 94000
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00014924341676584194,
      "loss": 0.29,
      "step": 94500
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00014922635094210277,
      "loss": 0.2904,
      "step": 95000
    },
    {
      "epoch": 1.06,
      "eval_loss": 0.2779350280761719,
      "eval_runtime": 2.6041,
      "eval_samples_per_second": 882.055,
      "eval_steps_per_second": 13.824,
      "step": 95000
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00014920909585616356,
      "loss": 0.2893,
      "step": 95500
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000149191651555199,
      "loss": 0.2897,
      "step": 96000
    },
    {
      "epoch": 1.07,
      "eval_loss": 0.278182715177536,
      "eval_runtime": 2.5817,
      "eval_samples_per_second": 889.72,
      "eval_steps_per_second": 13.944,
      "step": 96000
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00014917401808690116,
      "loss": 0.2884,
      "step": 96500
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0001491561954994793,
      "loss": 0.2888,
      "step": 97000
    },
    {
      "epoch": 1.08,
      "eval_loss": 0.27881118655204773,
      "eval_runtime": 2.5487,
      "eval_samples_per_second": 901.234,
      "eval_steps_per_second": 14.125,
      "step": 97000
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00014913818384165964,
      "loss": 0.2882,
      "step": 97500
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00014911998316268537,
      "loss": 0.2878,
      "step": 98000
    },
    {
      "epoch": 1.09,
      "eval_loss": 0.2754364311695099,
      "eval_runtime": 2.537,
      "eval_samples_per_second": 905.405,
      "eval_steps_per_second": 14.19,
      "step": 98000
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.00014910159351231653,
      "loss": 0.2873,
      "step": 98500
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.00014908301494082963,
      "loss": 0.2869,
      "step": 99000
    },
    {
      "epoch": 1.1,
      "eval_loss": 0.2750239074230194,
      "eval_runtime": 2.5521,
      "eval_samples_per_second": 900.043,
      "eval_steps_per_second": 14.106,
      "step": 99000
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0001490642474990178,
      "loss": 0.2865,
      "step": 99500
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.00014904529123819054,
      "loss": 0.2863,
      "step": 100000
    },
    {
      "epoch": 1.12,
      "eval_loss": 0.27601054310798645,
      "eval_runtime": 2.5872,
      "eval_samples_per_second": 887.823,
      "eval_steps_per_second": 13.915,
      "step": 100000
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.00014902614621017352,
      "loss": 0.2863,
      "step": 100500
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.00014900681246730852,
      "loss": 0.2856,
      "step": 101000
    },
    {
      "epoch": 1.13,
      "eval_loss": 0.27266713976860046,
      "eval_runtime": 2.477,
      "eval_samples_per_second": 927.344,
      "eval_steps_per_second": 14.534,
      "step": 101000
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.00014898729006245328,
      "loss": 0.2853,
      "step": 101500
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.00014896757904898125,
      "loss": 0.2852,
      "step": 102000
    },
    {
      "epoch": 1.14,
      "eval_loss": 0.27444031834602356,
      "eval_runtime": 2.5666,
      "eval_samples_per_second": 894.954,
      "eval_steps_per_second": 14.026,
      "step": 102000
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0001489476794807816,
      "loss": 0.2846,
      "step": 102500
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.00014892759141225904,
      "loss": 0.2846,
      "step": 103000
    },
    {
      "epoch": 1.15,
      "eval_loss": 0.2744878828525543,
      "eval_runtime": 2.61,
      "eval_samples_per_second": 880.071,
      "eval_steps_per_second": 13.793,
      "step": 103000
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.00014890731489833355,
      "loss": 0.2838,
      "step": 103500
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.00014888684999444035,
      "loss": 0.2838,
      "step": 104000
    },
    {
      "epoch": 1.16,
      "eval_loss": 0.2741919755935669,
      "eval_runtime": 2.5544,
      "eval_samples_per_second": 899.235,
      "eval_steps_per_second": 14.093,
      "step": 104000
    },
    {
      "epoch": 1.17,
      "learning_rate": 0.00014886619675652968,
      "loss": 0.2836,
      "step": 104500
    },
    {
      "epoch": 1.17,
      "learning_rate": 0.00014884535524106675,
      "loss": 0.2832,
      "step": 105000
    },
    {
      "epoch": 1.17,
      "eval_loss": 0.2710511386394501,
      "eval_runtime": 2.6471,
      "eval_samples_per_second": 867.747,
      "eval_steps_per_second": 13.6,
      "step": 105000
    },
    {
      "epoch": 1.18,
      "learning_rate": 0.00014882432550503144,
      "loss": 0.2826,
      "step": 105500
    },
    {
      "epoch": 1.18,
      "learning_rate": 0.00014880310760591824,
      "loss": 0.2822,
      "step": 106000
    },
    {
      "epoch": 1.18,
      "eval_loss": 0.2754230499267578,
      "eval_runtime": 2.637,
      "eval_samples_per_second": 871.051,
      "eval_steps_per_second": 13.652,
      "step": 106000
    },
    {
      "epoch": 1.19,
      "learning_rate": 0.0001487817016017361,
      "loss": 0.2824,
      "step": 106500
    },
    {
      "epoch": 1.19,
      "learning_rate": 0.0001487601075510082,
      "loss": 0.2819,
      "step": 107000
    },
    {
      "epoch": 1.19,
      "eval_loss": 0.2707637846469879,
      "eval_runtime": 2.5493,
      "eval_samples_per_second": 901.018,
      "eval_steps_per_second": 14.121,
      "step": 107000
    },
    {
      "epoch": 1.2,
      "learning_rate": 0.00014873832551277186,
      "loss": 0.2811,
      "step": 107500
    },
    {
      "epoch": 1.2,
      "learning_rate": 0.0001487163555465783,
      "loss": 0.2813,
      "step": 108000
    },
    {
      "epoch": 1.2,
      "eval_loss": 0.27113014459609985,
      "eval_runtime": 2.5222,
      "eval_samples_per_second": 910.7,
      "eval_steps_per_second": 14.273,
      "step": 108000
    },
    {
      "epoch": 1.21,
      "learning_rate": 0.00014869419771249264,
      "loss": 0.2815,
      "step": 108500
    },
    {
      "epoch": 1.22,
      "learning_rate": 0.0001486718520710935,
      "loss": 0.2807,
      "step": 109000
    },
    {
      "epoch": 1.22,
      "eval_loss": 0.270040363073349,
      "eval_runtime": 2.5255,
      "eval_samples_per_second": 909.511,
      "eval_steps_per_second": 14.254,
      "step": 109000
    },
    {
      "epoch": 1.22,
      "learning_rate": 0.00014864931868347302,
      "loss": 0.2802,
      "step": 109500
    },
    {
      "epoch": 1.23,
      "learning_rate": 0.00014862659761123663,
      "loss": 0.2802,
      "step": 110000
    },
    {
      "epoch": 1.23,
      "eval_loss": 0.26776713132858276,
      "eval_runtime": 2.447,
      "eval_samples_per_second": 938.683,
      "eval_steps_per_second": 14.712,
      "step": 110000
    },
    {
      "epoch": 1.23,
      "learning_rate": 0.0001486036889165029,
      "loss": 0.28,
      "step": 110500
    },
    {
      "epoch": 1.24,
      "learning_rate": 0.00014858059266190327,
      "loss": 0.2794,
      "step": 111000
    },
    {
      "epoch": 1.24,
      "eval_loss": 0.2689948081970215,
      "eval_runtime": 2.4933,
      "eval_samples_per_second": 921.261,
      "eval_steps_per_second": 14.439,
      "step": 111000
    },
    {
      "epoch": 1.24,
      "learning_rate": 0.00014855730891058204,
      "loss": 0.2795,
      "step": 111500
    },
    {
      "epoch": 1.25,
      "learning_rate": 0.00014853383772619612,
      "loss": 0.2791,
      "step": 112000
    },
    {
      "epoch": 1.25,
      "eval_loss": 0.26764577627182007,
      "eval_runtime": 2.6324,
      "eval_samples_per_second": 872.576,
      "eval_steps_per_second": 13.676,
      "step": 112000
    },
    {
      "epoch": 1.25,
      "learning_rate": 0.0001485101791729148,
      "loss": 0.2791,
      "step": 112500
    },
    {
      "epoch": 1.26,
      "learning_rate": 0.00014848633331541967,
      "loss": 0.2784,
      "step": 113000
    },
    {
      "epoch": 1.26,
      "eval_loss": 0.27024123072624207,
      "eval_runtime": 2.5135,
      "eval_samples_per_second": 913.848,
      "eval_steps_per_second": 14.322,
      "step": 113000
    },
    {
      "epoch": 1.27,
      "learning_rate": 0.00014846230021890443,
      "loss": 0.2781,
      "step": 113500
    },
    {
      "epoch": 1.27,
      "learning_rate": 0.0001484380799490746,
      "loss": 0.278,
      "step": 114000
    },
    {
      "epoch": 1.27,
      "eval_loss": 0.26657336950302124,
      "eval_runtime": 2.6002,
      "eval_samples_per_second": 883.406,
      "eval_steps_per_second": 13.845,
      "step": 114000
    },
    {
      "epoch": 1.28,
      "learning_rate": 0.0001484136725721475,
      "loss": 0.2776,
      "step": 114500
    },
    {
      "epoch": 1.28,
      "learning_rate": 0.00014838907815485194,
      "loss": 0.2775,
      "step": 115000
    },
    {
      "epoch": 1.28,
      "eval_loss": 0.2676675021648407,
      "eval_runtime": 2.608,
      "eval_samples_per_second": 880.739,
      "eval_steps_per_second": 13.803,
      "step": 115000
    },
    {
      "epoch": 1.29,
      "learning_rate": 0.00014836429676442814,
      "loss": 0.2774,
      "step": 115500
    },
    {
      "epoch": 1.29,
      "learning_rate": 0.00014833932846862748,
      "loss": 0.2767,
      "step": 116000
    },
    {
      "epoch": 1.29,
      "eval_loss": 0.2658541798591614,
      "eval_runtime": 2.528,
      "eval_samples_per_second": 908.623,
      "eval_steps_per_second": 14.24,
      "step": 116000
    },
    {
      "epoch": 1.3,
      "learning_rate": 0.0001483141733357123,
      "loss": 0.2767,
      "step": 116500
    },
    {
      "epoch": 1.3,
      "learning_rate": 0.00014828883143445582,
      "loss": 0.2762,
      "step": 117000
    },
    {
      "epoch": 1.3,
      "eval_loss": 0.2639394998550415,
      "eval_runtime": 2.5898,
      "eval_samples_per_second": 886.929,
      "eval_steps_per_second": 13.9,
      "step": 117000
    },
    {
      "epoch": 1.31,
      "learning_rate": 0.00014826330283414178,
      "loss": 0.2774,
      "step": 117500
    },
    {
      "epoch": 1.32,
      "learning_rate": 0.0001482375876045644,
      "loss": 0.2753,
      "step": 118000
    },
    {
      "epoch": 1.32,
      "eval_loss": 0.2651030123233795,
      "eval_runtime": 2.4743,
      "eval_samples_per_second": 928.341,
      "eval_steps_per_second": 14.55,
      "step": 118000
    },
    {
      "epoch": 1.32,
      "learning_rate": 0.0001482116858160282,
      "loss": 0.2756,
      "step": 118500
    },
    {
      "epoch": 1.33,
      "learning_rate": 0.0001481855975393476,
      "loss": 0.275,
      "step": 119000
    },
    {
      "epoch": 1.33,
      "eval_loss": 0.2663949429988861,
      "eval_runtime": 2.5801,
      "eval_samples_per_second": 890.26,
      "eval_steps_per_second": 13.953,
      "step": 119000
    },
    {
      "epoch": 1.33,
      "learning_rate": 0.000148159322845847,
      "loss": 0.2747,
      "step": 119500
    },
    {
      "epoch": 1.34,
      "learning_rate": 0.0001481328618073604,
      "loss": 0.2746,
      "step": 120000
    },
    {
      "epoch": 1.34,
      "eval_loss": 0.2646242678165436,
      "eval_runtime": 2.5527,
      "eval_samples_per_second": 899.828,
      "eval_steps_per_second": 14.103,
      "step": 120000
    },
    {
      "epoch": 1.34,
      "learning_rate": 0.00014810621449623125,
      "loss": 0.2749,
      "step": 120500
    },
    {
      "epoch": 1.35,
      "learning_rate": 0.0001480793809853123,
      "loss": 0.2749,
      "step": 121000
    },
    {
      "epoch": 1.35,
      "eval_loss": 0.26220399141311646,
      "eval_runtime": 2.518,
      "eval_samples_per_second": 912.241,
      "eval_steps_per_second": 14.297,
      "step": 121000
    },
    {
      "epoch": 1.36,
      "learning_rate": 0.00014805236134796536,
      "loss": 0.2741,
      "step": 121500
    },
    {
      "epoch": 1.36,
      "learning_rate": 0.00014802515565806107,
      "loss": 0.2736,
      "step": 122000
    },
    {
      "epoch": 1.36,
      "eval_loss": 0.26149359345436096,
      "eval_runtime": 2.5346,
      "eval_samples_per_second": 906.265,
      "eval_steps_per_second": 14.204,
      "step": 122000
    },
    {
      "epoch": 1.37,
      "learning_rate": 0.00014799776398997873,
      "loss": 0.2734,
      "step": 122500
    },
    {
      "epoch": 1.37,
      "learning_rate": 0.00014797018641860612,
      "loss": 0.2735,
      "step": 123000
    },
    {
      "epoch": 1.37,
      "eval_loss": 0.2607334852218628,
      "eval_runtime": 2.5063,
      "eval_samples_per_second": 916.502,
      "eval_steps_per_second": 14.364,
      "step": 123000
    },
    {
      "epoch": 1.38,
      "learning_rate": 0.00014794242301933928,
      "loss": 0.2736,
      "step": 123500
    },
    {
      "epoch": 1.38,
      "learning_rate": 0.0001479144738680823,
      "loss": 0.2729,
      "step": 124000
    },
    {
      "epoch": 1.38,
      "eval_loss": 0.2600594162940979,
      "eval_runtime": 2.5363,
      "eval_samples_per_second": 905.664,
      "eval_steps_per_second": 14.194,
      "step": 124000
    },
    {
      "epoch": 1.39,
      "learning_rate": 0.00014788633904124708,
      "loss": 0.2727,
      "step": 124500
    },
    {
      "epoch": 1.39,
      "learning_rate": 0.00014785801861575312,
      "loss": 0.2724,
      "step": 125000
    },
    {
      "epoch": 1.39,
      "eval_loss": 0.2614007294178009,
      "eval_runtime": 2.556,
      "eval_samples_per_second": 898.656,
      "eval_steps_per_second": 14.084,
      "step": 125000
    },
    {
      "epoch": 1.4,
      "learning_rate": 0.0001478295126690274,
      "loss": 0.2723,
      "step": 125500
    },
    {
      "epoch": 1.41,
      "learning_rate": 0.00014780082127900416,
      "loss": 0.2721,
      "step": 126000
    },
    {
      "epoch": 1.41,
      "eval_loss": 0.2622178792953491,
      "eval_runtime": 2.616,
      "eval_samples_per_second": 878.045,
      "eval_steps_per_second": 13.761,
      "step": 126000
    },
    {
      "epoch": 1.41,
      "learning_rate": 0.0001477719445241245,
      "loss": 0.2718,
      "step": 126500
    },
    {
      "epoch": 1.42,
      "learning_rate": 0.00014774288248333635,
      "loss": 0.2715,
      "step": 127000
    },
    {
      "epoch": 1.42,
      "eval_loss": 0.2606135308742523,
      "eval_runtime": 2.4928,
      "eval_samples_per_second": 921.454,
      "eval_steps_per_second": 14.442,
      "step": 127000
    },
    {
      "epoch": 1.42,
      "learning_rate": 0.00014771363523609428,
      "loss": 0.2715,
      "step": 127500
    },
    {
      "epoch": 1.43,
      "learning_rate": 0.00014768420286235908,
      "loss": 0.271,
      "step": 128000
    },
    {
      "epoch": 1.43,
      "eval_loss": 0.2602754831314087,
      "eval_runtime": 2.6006,
      "eval_samples_per_second": 883.262,
      "eval_steps_per_second": 13.843,
      "step": 128000
    },
    {
      "epoch": 1.43,
      "learning_rate": 0.0001476545854425978,
      "loss": 0.2708,
      "step": 128500
    },
    {
      "epoch": 1.44,
      "learning_rate": 0.00014762478305778328,
      "loss": 0.2711,
      "step": 129000
    },
    {
      "epoch": 1.44,
      "eval_loss": 0.25960421562194824,
      "eval_runtime": 2.5551,
      "eval_samples_per_second": 898.977,
      "eval_steps_per_second": 14.089,
      "step": 129000
    },
    {
      "epoch": 1.44,
      "learning_rate": 0.00014759479578939415,
      "loss": 0.2704,
      "step": 129500
    },
    {
      "epoch": 1.45,
      "learning_rate": 0.0001475646237194144,
      "loss": 0.2704,
      "step": 130000
    },
    {
      "epoch": 1.45,
      "eval_loss": 0.2571023404598236,
      "eval_runtime": 2.4929,
      "eval_samples_per_second": 921.433,
      "eval_steps_per_second": 14.441,
      "step": 130000
    },
    {
      "epoch": 1.46,
      "learning_rate": 0.00014753426693033336,
      "loss": 0.2706,
      "step": 130500
    },
    {
      "epoch": 1.46,
      "learning_rate": 0.00014750372550514533,
      "loss": 0.2698,
      "step": 131000
    },
    {
      "epoch": 1.46,
      "eval_loss": 0.2580989897251129,
      "eval_runtime": 2.5522,
      "eval_samples_per_second": 900.002,
      "eval_steps_per_second": 14.105,
      "step": 131000
    },
    {
      "epoch": 1.47,
      "learning_rate": 0.00014747299952734937,
      "loss": 0.2698,
      "step": 131500
    },
    {
      "epoch": 1.47,
      "learning_rate": 0.0001474420890809492,
      "loss": 0.2693,
      "step": 132000
    },
    {
      "epoch": 1.47,
      "eval_loss": 0.25762051343917847,
      "eval_runtime": 2.6493,
      "eval_samples_per_second": 867.024,
      "eval_steps_per_second": 13.589,
      "step": 132000
    },
    {
      "epoch": 1.48,
      "learning_rate": 0.00014741099425045272,
      "loss": 0.269,
      "step": 132500
    },
    {
      "epoch": 1.48,
      "learning_rate": 0.00014737971512087202,
      "loss": 0.2689,
      "step": 133000
    },
    {
      "epoch": 1.48,
      "eval_loss": 0.2582535445690155,
      "eval_runtime": 2.5702,
      "eval_samples_per_second": 893.707,
      "eval_steps_per_second": 14.007,
      "step": 133000
    },
    {
      "epoch": 1.49,
      "learning_rate": 0.00014734825177772313,
      "loss": 0.2686,
      "step": 133500
    },
    {
      "epoch": 1.49,
      "learning_rate": 0.00014731660430702552,
      "loss": 0.2684,
      "step": 134000
    },
    {
      "epoch": 1.49,
      "eval_loss": 0.2552024722099304,
      "eval_runtime": 2.5578,
      "eval_samples_per_second": 898.041,
      "eval_steps_per_second": 14.075,
      "step": 134000
    },
    {
      "epoch": 1.5,
      "learning_rate": 0.00014728477279530222,
      "loss": 0.268,
      "step": 134500
    },
    {
      "epoch": 1.51,
      "learning_rate": 0.00014725275732957937,
      "loss": 0.2682,
      "step": 135000
    },
    {
      "epoch": 1.51,
      "eval_loss": 0.2560090124607086,
      "eval_runtime": 2.5793,
      "eval_samples_per_second": 890.566,
      "eval_steps_per_second": 13.958,
      "step": 135000
    },
    {
      "epoch": 1.51,
      "learning_rate": 0.000147220557997386,
      "loss": 0.2678,
      "step": 135500
    },
    {
      "epoch": 1.52,
      "learning_rate": 0.00014718817488675387,
      "loss": 0.2673,
      "step": 136000
    },
    {
      "epoch": 1.52,
      "eval_loss": 0.2554505169391632,
      "eval_runtime": 2.5602,
      "eval_samples_per_second": 897.208,
      "eval_steps_per_second": 14.062,
      "step": 136000
    },
    {
      "epoch": 1.52,
      "learning_rate": 0.0001471556080862172,
      "loss": 0.2674,
      "step": 136500
    },
    {
      "epoch": 1.53,
      "learning_rate": 0.00014712285768481235,
      "loss": 0.2672,
      "step": 137000
    },
    {
      "epoch": 1.53,
      "eval_loss": 0.2564171552658081,
      "eval_runtime": 2.6088,
      "eval_samples_per_second": 880.495,
      "eval_steps_per_second": 13.8,
      "step": 137000
    },
    {
      "epoch": 1.53,
      "learning_rate": 0.00014708992377207767,
      "loss": 0.2666,
      "step": 137500
    },
    {
      "epoch": 1.54,
      "learning_rate": 0.00014705680643805323,
      "loss": 0.2671,
      "step": 138000
    },
    {
      "epoch": 1.54,
      "eval_loss": 0.2554504871368408,
      "eval_runtime": 2.607,
      "eval_samples_per_second": 881.099,
      "eval_steps_per_second": 13.809,
      "step": 138000
    },
    {
      "epoch": 1.54,
      "learning_rate": 0.00014702350577328052,
      "loss": 0.2668,
      "step": 138500
    },
    {
      "epoch": 1.55,
      "learning_rate": 0.00014699002186880232,
      "loss": 0.2663,
      "step": 139000
    },
    {
      "epoch": 1.55,
      "eval_loss": 0.2547665238380432,
      "eval_runtime": 2.5426,
      "eval_samples_per_second": 903.392,
      "eval_steps_per_second": 14.159,
      "step": 139000
    },
    {
      "epoch": 1.56,
      "learning_rate": 0.00014695635481616235,
      "loss": 0.2662,
      "step": 139500
    },
    {
      "epoch": 1.56,
      "learning_rate": 0.00014692250470740503,
      "loss": 0.2659,
      "step": 140000
    },
    {
      "epoch": 1.56,
      "eval_loss": 0.25474175810813904,
      "eval_runtime": 2.5555,
      "eval_samples_per_second": 898.837,
      "eval_steps_per_second": 14.087,
      "step": 140000
    },
    {
      "epoch": 1.57,
      "learning_rate": 0.00014688847163507525,
      "loss": 0.2656,
      "step": 140500
    },
    {
      "epoch": 1.57,
      "learning_rate": 0.00014685425569221819,
      "loss": 0.2662,
      "step": 141000
    },
    {
      "epoch": 1.57,
      "eval_loss": 0.25465691089630127,
      "eval_runtime": 2.5777,
      "eval_samples_per_second": 891.114,
      "eval_steps_per_second": 13.966,
      "step": 141000
    },
    {
      "epoch": 1.58,
      "learning_rate": 0.00014681985697237885,
      "loss": 0.2659,
      "step": 141500
    },
    {
      "epoch": 1.58,
      "learning_rate": 0.00014678527556960207,
      "loss": 0.265,
      "step": 142000
    },
    {
      "epoch": 1.58,
      "eval_loss": 0.25828590989112854,
      "eval_runtime": 2.5027,
      "eval_samples_per_second": 917.822,
      "eval_steps_per_second": 14.385,
      "step": 142000
    },
    {
      "epoch": 1.59,
      "learning_rate": 0.00014675051157843208,
      "loss": 0.2651,
      "step": 142500
    },
    {
      "epoch": 1.59,
      "learning_rate": 0.0001467155650939123,
      "loss": 0.2645,
      "step": 143000
    },
    {
      "epoch": 1.59,
      "eval_loss": 0.2552579939365387,
      "eval_runtime": 2.5709,
      "eval_samples_per_second": 893.469,
      "eval_steps_per_second": 14.003,
      "step": 143000
    },
    {
      "epoch": 1.6,
      "learning_rate": 0.00014668043621158508,
      "loss": 0.2647,
      "step": 143500
    },
    {
      "epoch": 1.61,
      "learning_rate": 0.00014664512502749141,
      "loss": 0.2646,
      "step": 144000
    },
    {
      "epoch": 1.61,
      "eval_loss": 0.2546204626560211,
      "eval_runtime": 2.5612,
      "eval_samples_per_second": 896.85,
      "eval_steps_per_second": 14.056,
      "step": 144000
    },
    {
      "epoch": 1.61,
      "learning_rate": 0.00014660963163817077,
      "loss": 0.2642,
      "step": 144500
    },
    {
      "epoch": 1.62,
      "learning_rate": 0.00014657395614066075,
      "loss": 0.2639,
      "step": 145000
    },
    {
      "epoch": 1.62,
      "eval_loss": 0.251667857170105,
      "eval_runtime": 2.6211,
      "eval_samples_per_second": 876.335,
      "eval_steps_per_second": 13.734,
      "step": 145000
    },
    {
      "epoch": 1.62,
      "learning_rate": 0.0001465380986324967,
      "loss": 0.2639,
      "step": 145500
    },
    {
      "epoch": 1.63,
      "learning_rate": 0.0001465020592117118,
      "loss": 0.2641,
      "step": 146000
    },
    {
      "epoch": 1.63,
      "eval_loss": 0.25381070375442505,
      "eval_runtime": 2.4892,
      "eval_samples_per_second": 922.778,
      "eval_steps_per_second": 14.462,
      "step": 146000
    },
    {
      "epoch": 1.63,
      "learning_rate": 0.00014646583797683636,
      "loss": 0.2638,
      "step": 146500
    },
    {
      "epoch": 1.64,
      "learning_rate": 0.0001464294350268979,
      "loss": 0.2633,
      "step": 147000
    },
    {
      "epoch": 1.64,
      "eval_loss": 0.2526147961616516,
      "eval_runtime": 2.5593,
      "eval_samples_per_second": 897.499,
      "eval_steps_per_second": 14.066,
      "step": 147000
    },
    {
      "epoch": 1.65,
      "learning_rate": 0.00014639285046142065,
      "loss": 0.2635,
      "step": 147500
    },
    {
      "epoch": 1.65,
      "learning_rate": 0.00014635608438042546,
      "loss": 0.2632,
      "step": 148000
    },
    {
      "epoch": 1.65,
      "eval_loss": 0.25286561250686646,
      "eval_runtime": 2.5815,
      "eval_samples_per_second": 889.785,
      "eval_steps_per_second": 13.945,
      "step": 148000
    },
    {
      "epoch": 1.66,
      "learning_rate": 0.00014631913688442936,
      "loss": 0.2627,
      "step": 148500
    },
    {
      "epoch": 1.66,
      "learning_rate": 0.00014628200807444543,
      "loss": 0.2626,
      "step": 149000
    },
    {
      "epoch": 1.66,
      "eval_loss": 0.25125652551651,
      "eval_runtime": 2.5016,
      "eval_samples_per_second": 918.2,
      "eval_steps_per_second": 14.391,
      "step": 149000
    },
    {
      "epoch": 1.67,
      "learning_rate": 0.0001462446980519824,
      "loss": 0.2626,
      "step": 149500
    },
    {
      "epoch": 1.67,
      "learning_rate": 0.0001462072069190444,
      "loss": 0.2619,
      "step": 150000
    },
    {
      "epoch": 1.67,
      "eval_loss": 0.25238075852394104,
      "eval_runtime": 2.514,
      "eval_samples_per_second": 913.666,
      "eval_steps_per_second": 14.32,
      "step": 150000
    },
    {
      "epoch": 1.68,
      "learning_rate": 0.00014616953477813085,
      "loss": 0.2626,
      "step": 150500
    },
    {
      "epoch": 1.68,
      "learning_rate": 0.00014613168173223585,
      "loss": 0.2618,
      "step": 151000
    },
    {
      "epoch": 1.68,
      "eval_loss": 0.25222572684288025,
      "eval_runtime": 2.5557,
      "eval_samples_per_second": 898.762,
      "eval_steps_per_second": 14.086,
      "step": 151000
    },
    {
      "epoch": 1.69,
      "learning_rate": 0.00014609364788484825,
      "loss": 0.2617,
      "step": 151500
    },
    {
      "epoch": 1.7,
      "learning_rate": 0.00014605543333995113,
      "loss": 0.2615,
      "step": 152000
    },
    {
      "epoch": 1.7,
      "eval_loss": 0.25088632106781006,
      "eval_runtime": 2.6299,
      "eval_samples_per_second": 873.405,
      "eval_steps_per_second": 13.689,
      "step": 152000
    },
    {
      "epoch": 1.7,
      "learning_rate": 0.00014601703820202154,
      "loss": 0.2615,
      "step": 152500
    },
    {
      "epoch": 1.71,
      "learning_rate": 0.00014597846257603038,
      "loss": 0.2617,
      "step": 153000
    },
    {
      "epoch": 1.71,
      "eval_loss": 0.2507455348968506,
      "eval_runtime": 2.5324,
      "eval_samples_per_second": 907.05,
      "eval_steps_per_second": 14.216,
      "step": 153000
    },
    {
      "epoch": 1.71,
      "learning_rate": 0.00014593970656744194,
      "loss": 0.261,
      "step": 153500
    },
    {
      "epoch": 1.72,
      "learning_rate": 0.0001459007702822136,
      "loss": 0.261,
      "step": 154000
    },
    {
      "epoch": 1.72,
      "eval_loss": 0.25084322690963745,
      "eval_runtime": 2.5322,
      "eval_samples_per_second": 907.132,
      "eval_steps_per_second": 14.217,
      "step": 154000
    },
    {
      "epoch": 1.72,
      "learning_rate": 0.00014586165382679577,
      "loss": 0.2612,
      "step": 154500
    },
    {
      "epoch": 1.73,
      "learning_rate": 0.00014582235730813128,
      "loss": 0.2614,
      "step": 155000
    },
    {
      "epoch": 1.73,
      "eval_loss": 0.2498544454574585,
      "eval_runtime": 2.5674,
      "eval_samples_per_second": 894.688,
      "eval_steps_per_second": 14.022,
      "step": 155000
    },
    {
      "epoch": 1.73,
      "learning_rate": 0.00014578288083365532,
      "loss": 0.2603,
      "step": 155500
    },
    {
      "epoch": 1.74,
      "learning_rate": 0.00014574322451129507,
      "loss": 0.2602,
      "step": 156000
    },
    {
      "epoch": 1.74,
      "eval_loss": 0.25135913491249084,
      "eval_runtime": 2.5422,
      "eval_samples_per_second": 903.538,
      "eval_steps_per_second": 14.161,
      "step": 156000
    },
    {
      "epoch": 1.75,
      "learning_rate": 0.00014570338844946943,
      "loss": 0.2603,
      "step": 156500
    },
    {
      "epoch": 1.75,
      "learning_rate": 0.00014566337275708863,
      "loss": 0.2601,
      "step": 157000
    },
    {
      "epoch": 1.75,
      "eval_loss": 0.24927592277526855,
      "eval_runtime": 2.5955,
      "eval_samples_per_second": 884.984,
      "eval_steps_per_second": 13.87,
      "step": 157000
    },
    {
      "epoch": 1.76,
      "learning_rate": 0.00014562317754355405,
      "loss": 0.26,
      "step": 157500
    },
    {
      "epoch": 1.76,
      "learning_rate": 0.0001455828029187579,
      "loss": 0.2599,
      "step": 158000
    },
    {
      "epoch": 1.76,
      "eval_loss": 0.2505401074886322,
      "eval_runtime": 2.6067,
      "eval_samples_per_second": 881.183,
      "eval_steps_per_second": 13.81,
      "step": 158000
    },
    {
      "epoch": 1.77,
      "learning_rate": 0.00014554224899308285,
      "loss": 0.2595,
      "step": 158500
    },
    {
      "epoch": 1.77,
      "learning_rate": 0.00014550151587740178,
      "loss": 0.2595,
      "step": 159000
    },
    {
      "epoch": 1.77,
      "eval_loss": 0.24686762690544128,
      "eval_runtime": 2.5852,
      "eval_samples_per_second": 888.524,
      "eval_steps_per_second": 13.925,
      "step": 159000
    },
    {
      "epoch": 1.78,
      "learning_rate": 0.00014546060368307744,
      "loss": 0.2589,
      "step": 159500
    },
    {
      "epoch": 1.78,
      "learning_rate": 0.00014541951252196225,
      "loss": 0.259,
      "step": 160000
    },
    {
      "epoch": 1.78,
      "eval_loss": 0.248440220952034,
      "eval_runtime": 2.56,
      "eval_samples_per_second": 897.273,
      "eval_steps_per_second": 14.063,
      "step": 160000
    },
    {
      "epoch": 1.79,
      "learning_rate": 0.00014537824250639785,
      "loss": 0.2589,
      "step": 160500
    },
    {
      "epoch": 1.8,
      "learning_rate": 0.00014533679374921493,
      "loss": 0.259,
      "step": 161000
    },
    {
      "epoch": 1.8,
      "eval_loss": 0.24788442254066467,
      "eval_runtime": 2.5373,
      "eval_samples_per_second": 905.276,
      "eval_steps_per_second": 14.188,
      "step": 161000
    },
    {
      "epoch": 1.8,
      "learning_rate": 0.00014529516636373275,
      "loss": 0.2581,
      "step": 161500
    },
    {
      "epoch": 1.81,
      "learning_rate": 0.00014525336046375905,
      "loss": 0.2582,
      "step": 162000
    },
    {
      "epoch": 1.81,
      "eval_loss": 0.2487788200378418,
      "eval_runtime": 2.5002,
      "eval_samples_per_second": 918.721,
      "eval_steps_per_second": 14.399,
      "step": 162000
    },
    {
      "epoch": 1.81,
      "learning_rate": 0.00014521137616358952,
      "loss": 0.2582,
      "step": 162500
    },
    {
      "epoch": 1.82,
      "learning_rate": 0.00014516921357800766,
      "loss": 0.2582,
      "step": 163000
    },
    {
      "epoch": 1.82,
      "eval_loss": 0.24880746006965637,
      "eval_runtime": 2.5819,
      "eval_samples_per_second": 889.642,
      "eval_steps_per_second": 13.943,
      "step": 163000
    },
    {
      "epoch": 1.82,
      "learning_rate": 0.00014512687282228432,
      "loss": 0.258,
      "step": 163500
    },
    {
      "epoch": 1.83,
      "learning_rate": 0.00014508435401217759,
      "loss": 0.2572,
      "step": 164000
    },
    {
      "epoch": 1.83,
      "eval_loss": 0.2471989244222641,
      "eval_runtime": 2.5023,
      "eval_samples_per_second": 917.96,
      "eval_steps_per_second": 14.387,
      "step": 164000
    },
    {
      "epoch": 1.83,
      "learning_rate": 0.0001450416572639322,
      "loss": 0.2572,
      "step": 164500
    },
    {
      "epoch": 1.84,
      "learning_rate": 0.00014499878269427948,
      "loss": 0.2571,
      "step": 165000
    },
    {
      "epoch": 1.84,
      "eval_loss": 0.24864718317985535,
      "eval_runtime": 2.5833,
      "eval_samples_per_second": 889.167,
      "eval_steps_per_second": 13.936,
      "step": 165000
    },
    {
      "epoch": 1.85,
      "learning_rate": 0.00014495573042043683,
      "loss": 0.2571,
      "step": 165500
    },
    {
      "epoch": 1.85,
      "learning_rate": 0.00014491250056010758,
      "loss": 0.2568,
      "step": 166000
    },
    {
      "epoch": 1.85,
      "eval_loss": 0.24337711930274963,
      "eval_runtime": 2.518,
      "eval_samples_per_second": 912.24,
      "eval_steps_per_second": 14.297,
      "step": 166000
    },
    {
      "epoch": 1.86,
      "learning_rate": 0.00014486909323148042,
      "loss": 0.257,
      "step": 166500
    },
    {
      "epoch": 1.86,
      "learning_rate": 0.00014482550855322943,
      "loss": 0.257,
      "step": 167000
    },
    {
      "epoch": 1.86,
      "eval_loss": 0.24700835347175598,
      "eval_runtime": 2.5407,
      "eval_samples_per_second": 904.072,
      "eval_steps_per_second": 14.169,
      "step": 167000
    },
    {
      "epoch": 1.87,
      "learning_rate": 0.00014478174664451338,
      "loss": 0.2565,
      "step": 167500
    },
    {
      "epoch": 1.87,
      "learning_rate": 0.0001447378076249757,
      "loss": 0.2559,
      "step": 168000
    },
    {
      "epoch": 1.87,
      "eval_loss": 0.24551132321357727,
      "eval_runtime": 2.5565,
      "eval_samples_per_second": 898.491,
      "eval_steps_per_second": 14.082,
      "step": 168000
    },
    {
      "epoch": 1.88,
      "learning_rate": 0.00014469369161474398,
      "loss": 0.2563,
      "step": 168500
    },
    {
      "epoch": 1.88,
      "learning_rate": 0.00014464939873442973,
      "loss": 0.2558,
      "step": 169000
    },
    {
      "epoch": 1.88,
      "eval_loss": 0.2444145828485489,
      "eval_runtime": 2.5541,
      "eval_samples_per_second": 899.348,
      "eval_steps_per_second": 14.095,
      "step": 169000
    },
    {
      "epoch": 1.89,
      "learning_rate": 0.00014460492910512793,
      "loss": 0.2558,
      "step": 169500
    },
    {
      "epoch": 1.9,
      "learning_rate": 0.00014456028284841693,
      "loss": 0.2558,
      "step": 170000
    },
    {
      "epoch": 1.9,
      "eval_loss": 0.24412675201892853,
      "eval_runtime": 2.5587,
      "eval_samples_per_second": 897.729,
      "eval_steps_per_second": 14.07,
      "step": 170000
    },
    {
      "epoch": 1.9,
      "learning_rate": 0.00014451546008635783,
      "loss": 0.2557,
      "step": 170500
    },
    {
      "epoch": 1.91,
      "learning_rate": 0.00014447046094149437,
      "loss": 0.2559,
      "step": 171000
    },
    {
      "epoch": 1.91,
      "eval_loss": 0.24562311172485352,
      "eval_runtime": 2.585,
      "eval_samples_per_second": 888.573,
      "eval_steps_per_second": 13.926,
      "step": 171000
    },
    {
      "epoch": 1.91,
      "learning_rate": 0.0001444252855368525,
      "loss": 0.2556,
      "step": 171500
    },
    {
      "epoch": 1.92,
      "learning_rate": 0.00014437993399594003,
      "loss": 0.2554,
      "step": 172000
    },
    {
      "epoch": 1.92,
      "eval_loss": 0.24504482746124268,
      "eval_runtime": 2.5424,
      "eval_samples_per_second": 903.478,
      "eval_steps_per_second": 14.16,
      "step": 172000
    },
    {
      "epoch": 1.92,
      "learning_rate": 0.00014433440644274635,
      "loss": 0.2552,
      "step": 172500
    },
    {
      "epoch": 1.93,
      "learning_rate": 0.0001442887030017421,
      "loss": 0.2552,
      "step": 173000
    },
    {
      "epoch": 1.93,
      "eval_loss": 0.24396482110023499,
      "eval_runtime": 2.5384,
      "eval_samples_per_second": 904.888,
      "eval_steps_per_second": 14.182,
      "step": 173000
    },
    {
      "epoch": 1.94,
      "learning_rate": 0.00014424282379787865,
      "loss": 0.255,
      "step": 173500
    },
    {
      "epoch": 1.94,
      "learning_rate": 0.00014419676895658807,
      "loss": 0.255,
      "step": 174000
    },
    {
      "epoch": 1.94,
      "eval_loss": 0.2440991997718811,
      "eval_runtime": 2.6023,
      "eval_samples_per_second": 882.671,
      "eval_steps_per_second": 13.834,
      "step": 174000
    },
    {
      "epoch": 1.95,
      "learning_rate": 0.00014415053860378254,
      "loss": 0.2551,
      "step": 174500
    },
    {
      "epoch": 1.95,
      "learning_rate": 0.000144104132865854,
      "loss": 0.2544,
      "step": 175000
    },
    {
      "epoch": 1.95,
      "eval_loss": 0.24473923444747925,
      "eval_runtime": 2.5664,
      "eval_samples_per_second": 895.019,
      "eval_steps_per_second": 14.027,
      "step": 175000
    },
    {
      "epoch": 1.96,
      "learning_rate": 0.00014405755186967404,
      "loss": 0.2542,
      "step": 175500
    },
    {
      "epoch": 1.96,
      "learning_rate": 0.0001440107957425933,
      "loss": 0.2544,
      "step": 176000
    },
    {
      "epoch": 1.96,
      "eval_loss": 0.24544279277324677,
      "eval_runtime": 2.5371,
      "eval_samples_per_second": 905.36,
      "eval_steps_per_second": 14.189,
      "step": 176000
    },
    {
      "epoch": 1.97,
      "learning_rate": 0.00014396386461244123,
      "loss": 0.2541,
      "step": 176500
    },
    {
      "epoch": 1.97,
      "learning_rate": 0.0001439167586075258,
      "loss": 0.2539,
      "step": 177000
    },
    {
      "epoch": 1.97,
      "eval_loss": 0.24276654422283173,
      "eval_runtime": 2.5385,
      "eval_samples_per_second": 904.868,
      "eval_steps_per_second": 14.182,
      "step": 177000
    },
    {
      "epoch": 1.98,
      "learning_rate": 0.00014386947785663293,
      "loss": 0.2538,
      "step": 177500
    },
    {
      "epoch": 1.99,
      "learning_rate": 0.0001438220224890265,
      "loss": 0.2537,
      "step": 178000
    },
    {
      "epoch": 1.99,
      "eval_loss": 0.24593636393547058,
      "eval_runtime": 2.5426,
      "eval_samples_per_second": 903.395,
      "eval_steps_per_second": 14.159,
      "step": 178000
    },
    {
      "epoch": 1.99,
      "learning_rate": 0.00014377439263444763,
      "loss": 0.2535,
      "step": 178500
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00014372658842311449,
      "loss": 0.2536,
      "step": 179000
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.2417013943195343,
      "eval_runtime": 2.5938,
      "eval_samples_per_second": 885.579,
      "eval_steps_per_second": 13.879,
      "step": 179000
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00014367860998572198,
      "loss": 0.2535,
      "step": 179500
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00014363045745344137,
      "loss": 0.2532,
      "step": 180000
    },
    {
      "epoch": 2.01,
      "eval_loss": 0.24318550527095795,
      "eval_runtime": 2.633,
      "eval_samples_per_second": 872.381,
      "eval_steps_per_second": 13.672,
      "step": 180000
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00014358213095791978,
      "loss": 0.2525,
      "step": 180500
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00014353363063128005,
      "loss": 0.2529,
      "step": 181000
    },
    {
      "epoch": 2.02,
      "eval_loss": 0.24333837628364563,
      "eval_runtime": 2.6257,
      "eval_samples_per_second": 874.816,
      "eval_steps_per_second": 13.711,
      "step": 181000
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00014348495660612015,
      "loss": 0.2526,
      "step": 181500
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0001434361090155131,
      "loss": 0.2523,
      "step": 182000
    },
    {
      "epoch": 2.03,
      "eval_loss": 0.2430492788553238,
      "eval_runtime": 2.7075,
      "eval_samples_per_second": 848.394,
      "eval_steps_per_second": 13.297,
      "step": 182000
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00014338708799300633,
      "loss": 0.2528,
      "step": 182500
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00014333789367262136,
      "loss": 0.252,
      "step": 183000
    },
    {
      "epoch": 2.04,
      "eval_loss": 0.2429981827735901,
      "eval_runtime": 2.6505,
      "eval_samples_per_second": 866.622,
      "eval_steps_per_second": 13.582,
      "step": 183000
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00014328852618885365,
      "loss": 0.252,
      "step": 183500
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00014323898567667202,
      "loss": 0.2517,
      "step": 184000
    },
    {
      "epoch": 2.05,
      "eval_loss": 0.24395941197872162,
      "eval_runtime": 2.6201,
      "eval_samples_per_second": 876.688,
      "eval_steps_per_second": 13.74,
      "step": 184000
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00014318927227151832,
      "loss": 0.252,
      "step": 184500
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00014313938610930712,
      "loss": 0.2518,
      "step": 185000
    },
    {
      "epoch": 2.06,
      "eval_loss": 0.23878329992294312,
      "eval_runtime": 2.6492,
      "eval_samples_per_second": 867.063,
      "eval_steps_per_second": 13.589,
      "step": 185000
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00014308932732642524,
      "loss": 0.2515,
      "step": 185500
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00014303909605973154,
      "loss": 0.2514,
      "step": 186000
    },
    {
      "epoch": 2.07,
      "eval_loss": 0.2405225932598114,
      "eval_runtime": 2.5624,
      "eval_samples_per_second": 896.419,
      "eval_steps_per_second": 14.049,
      "step": 186000
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00014298869244655632,
      "loss": 0.2514,
      "step": 186500
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0001429381166247012,
      "loss": 0.2509,
      "step": 187000
    },
    {
      "epoch": 2.09,
      "eval_loss": 0.2413066029548645,
      "eval_runtime": 2.6728,
      "eval_samples_per_second": 859.401,
      "eval_steps_per_second": 13.469,
      "step": 187000
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0001428873687324385,
      "loss": 0.251,
      "step": 187500
    },
    {
      "epoch": 2.1,
      "learning_rate": 0.00014283644890851103,
      "loss": 0.2513,
      "step": 188000
    },
    {
      "epoch": 2.1,
      "eval_loss": 0.24260124564170837,
      "eval_runtime": 2.6325,
      "eval_samples_per_second": 872.57,
      "eval_steps_per_second": 13.675,
      "step": 188000
    },
    {
      "epoch": 2.1,
      "learning_rate": 0.00014278535729213168,
      "loss": 0.251,
      "step": 188500
    },
    {
      "epoch": 2.11,
      "learning_rate": 0.00014273409402298291,
      "loss": 0.2508,
      "step": 189000
    },
    {
      "epoch": 2.11,
      "eval_loss": 0.24098709225654602,
      "eval_runtime": 2.7201,
      "eval_samples_per_second": 844.45,
      "eval_steps_per_second": 13.235,
      "step": 189000
    },
    {
      "epoch": 2.11,
      "learning_rate": 0.0001426826592412166,
      "loss": 0.2508,
      "step": 189500
    },
    {
      "epoch": 2.12,
      "learning_rate": 0.00014263105308745343,
      "loss": 0.251,
      "step": 190000
    },
    {
      "epoch": 2.12,
      "eval_loss": 0.2388133704662323,
      "eval_runtime": 2.7019,
      "eval_samples_per_second": 850.14,
      "eval_steps_per_second": 13.324,
      "step": 190000
    },
    {
      "epoch": 2.12,
      "learning_rate": 0.0001425792757027827,
      "loss": 0.2505,
      "step": 190500
    },
    {
      "epoch": 2.13,
      "learning_rate": 0.00014252732722876176,
      "loss": 0.2503,
      "step": 191000
    },
    {
      "epoch": 2.13,
      "eval_loss": 0.24010741710662842,
      "eval_runtime": 2.646,
      "eval_samples_per_second": 868.09,
      "eval_steps_per_second": 13.605,
      "step": 191000
    },
    {
      "epoch": 2.14,
      "learning_rate": 0.00014247520780741581,
      "loss": 0.2502,
      "step": 191500
    },
    {
      "epoch": 2.14,
      "learning_rate": 0.0001424229175812373,
      "loss": 0.25,
      "step": 192000
    },
    {
      "epoch": 2.14,
      "eval_loss": 0.24023786187171936,
      "eval_runtime": 2.599,
      "eval_samples_per_second": 883.799,
      "eval_steps_per_second": 13.851,
      "step": 192000
    },
    {
      "epoch": 2.15,
      "learning_rate": 0.00014237045669318574,
      "loss": 0.2497,
      "step": 192500
    },
    {
      "epoch": 2.15,
      "learning_rate": 0.00014231782528668717,
      "loss": 0.2496,
      "step": 193000
    },
    {
      "epoch": 2.15,
      "eval_loss": 0.24067841470241547,
      "eval_runtime": 2.6221,
      "eval_samples_per_second": 876.005,
      "eval_steps_per_second": 13.729,
      "step": 193000
    },
    {
      "epoch": 2.16,
      "learning_rate": 0.00014226502350563392,
      "loss": 0.2495,
      "step": 193500
    },
    {
      "epoch": 2.16,
      "learning_rate": 0.00014221205149438394,
      "loss": 0.2498,
      "step": 194000
    },
    {
      "epoch": 2.16,
      "eval_loss": 0.23946070671081543,
      "eval_runtime": 2.679,
      "eval_samples_per_second": 857.416,
      "eval_steps_per_second": 13.438,
      "step": 194000
    },
    {
      "epoch": 2.17,
      "learning_rate": 0.00014215890939776074,
      "loss": 0.2494,
      "step": 194500
    },
    {
      "epoch": 2.17,
      "learning_rate": 0.0001421055973610528,
      "loss": 0.2494,
      "step": 195000
    },
    {
      "epoch": 2.17,
      "eval_loss": 0.2386200875043869,
      "eval_runtime": 2.6792,
      "eval_samples_per_second": 857.337,
      "eval_steps_per_second": 13.437,
      "step": 195000
    },
    {
      "epoch": 2.18,
      "learning_rate": 0.00014205211553001317,
      "loss": 0.2493,
      "step": 195500
    },
    {
      "epoch": 2.19,
      "learning_rate": 0.00014199846405085913,
      "loss": 0.2488,
      "step": 196000
    },
    {
      "epoch": 2.19,
      "eval_loss": 0.2384871244430542,
      "eval_runtime": 2.7009,
      "eval_samples_per_second": 850.466,
      "eval_steps_per_second": 13.329,
      "step": 196000
    },
    {
      "epoch": 2.19,
      "learning_rate": 0.00014194464307027178,
      "loss": 0.2488,
      "step": 196500
    },
    {
      "epoch": 2.2,
      "learning_rate": 0.00014189065273539564,
      "loss": 0.249,
      "step": 197000
    },
    {
      "epoch": 2.2,
      "eval_loss": 0.23911140859127045,
      "eval_runtime": 2.6027,
      "eval_samples_per_second": 882.552,
      "eval_steps_per_second": 13.832,
      "step": 197000
    },
    {
      "epoch": 2.2,
      "learning_rate": 0.0001418364931938382,
      "loss": 0.2488,
      "step": 197500
    },
    {
      "epoch": 2.21,
      "learning_rate": 0.00014178216459366958,
      "loss": 0.2483,
      "step": 198000
    },
    {
      "epoch": 2.21,
      "eval_loss": 0.2387627363204956,
      "eval_runtime": 2.5383,
      "eval_samples_per_second": 904.933,
      "eval_steps_per_second": 14.183,
      "step": 198000
    },
    {
      "epoch": 2.21,
      "learning_rate": 0.0001417276670834221,
      "loss": 0.2486,
      "step": 198500
    },
    {
      "epoch": 2.22,
      "learning_rate": 0.00014167300081208988,
      "loss": 0.2486,
      "step": 199000
    },
    {
      "epoch": 2.22,
      "eval_loss": 0.23652148246765137,
      "eval_runtime": 2.5882,
      "eval_samples_per_second": 887.478,
      "eval_steps_per_second": 13.909,
      "step": 199000
    },
    {
      "epoch": 2.22,
      "learning_rate": 0.00014161816592912844,
      "loss": 0.2484,
      "step": 199500
    },
    {
      "epoch": 2.23,
      "learning_rate": 0.00014156316258445421,
      "loss": 0.2484,
      "step": 200000
    },
    {
      "epoch": 2.23,
      "eval_loss": 0.23854629695415497,
      "eval_runtime": 2.5755,
      "eval_samples_per_second": 891.852,
      "eval_steps_per_second": 13.978,
      "step": 200000
    },
    {
      "epoch": 2.24,
      "learning_rate": 0.00014150799092844428,
      "loss": 0.2482,
      "step": 200500
    },
    {
      "epoch": 2.24,
      "learning_rate": 0.00014145265111193583,
      "loss": 0.2485,
      "step": 201000
    },
    {
      "epoch": 2.24,
      "eval_loss": 0.2385808229446411,
      "eval_runtime": 2.6672,
      "eval_samples_per_second": 861.187,
      "eval_steps_per_second": 13.497,
      "step": 201000
    },
    {
      "epoch": 2.25,
      "learning_rate": 0.0001413971432862258,
      "loss": 0.2474,
      "step": 201500
    },
    {
      "epoch": 2.25,
      "learning_rate": 0.00014134146760307043,
      "loss": 0.2478,
      "step": 202000
    },
    {
      "epoch": 2.25,
      "eval_loss": 0.23785722255706787,
      "eval_runtime": 2.5981,
      "eval_samples_per_second": 884.094,
      "eval_steps_per_second": 13.856,
      "step": 202000
    },
    {
      "epoch": 2.26,
      "learning_rate": 0.000141285624214685,
      "loss": 0.2475,
      "step": 202500
    },
    {
      "epoch": 2.26,
      "learning_rate": 0.00014122961327374313,
      "loss": 0.2477,
      "step": 203000
    },
    {
      "epoch": 2.26,
      "eval_loss": 0.23697885870933533,
      "eval_runtime": 2.6667,
      "eval_samples_per_second": 861.349,
      "eval_steps_per_second": 13.5,
      "step": 203000
    },
    {
      "epoch": 2.27,
      "learning_rate": 0.0001411734349333766,
      "loss": 0.2472,
      "step": 203500
    },
    {
      "epoch": 2.28,
      "learning_rate": 0.0001411170893471749,
      "loss": 0.2475,
      "step": 204000
    },
    {
      "epoch": 2.28,
      "eval_loss": 0.23942548036575317,
      "eval_runtime": 2.5207,
      "eval_samples_per_second": 911.252,
      "eval_steps_per_second": 14.282,
      "step": 204000
    },
    {
      "epoch": 2.28,
      "learning_rate": 0.00014106057666918466,
      "loss": 0.2471,
      "step": 204500
    },
    {
      "epoch": 2.29,
      "learning_rate": 0.00014100389705390938,
      "loss": 0.2473,
      "step": 205000
    },
    {
      "epoch": 2.29,
      "eval_loss": 0.23825885355472565,
      "eval_runtime": 2.5296,
      "eval_samples_per_second": 908.049,
      "eval_steps_per_second": 14.232,
      "step": 205000
    },
    {
      "epoch": 2.29,
      "learning_rate": 0.00014094705065630903,
      "loss": 0.247,
      "step": 205500
    },
    {
      "epoch": 2.3,
      "learning_rate": 0.0001408900376317994,
      "loss": 0.247,
      "step": 206000
    },
    {
      "epoch": 2.3,
      "eval_loss": 0.2336592674255371,
      "eval_runtime": 2.6942,
      "eval_samples_per_second": 852.576,
      "eval_steps_per_second": 13.362,
      "step": 206000
    },
    {
      "epoch": 2.3,
      "learning_rate": 0.00014083285813625197,
      "loss": 0.2466,
      "step": 206500
    },
    {
      "epoch": 2.31,
      "learning_rate": 0.0001407755123259933,
      "loss": 0.2464,
      "step": 207000
    },
    {
      "epoch": 2.31,
      "eval_loss": 0.2362833321094513,
      "eval_runtime": 2.6517,
      "eval_samples_per_second": 866.23,
      "eval_steps_per_second": 13.576,
      "step": 207000
    },
    {
      "epoch": 2.31,
      "learning_rate": 0.00014071800035780465,
      "loss": 0.2465,
      "step": 207500
    },
    {
      "epoch": 2.32,
      "learning_rate": 0.00014066032238892152,
      "loss": 0.2484,
      "step": 208000
    },
    {
      "epoch": 2.32,
      "eval_loss": 0.23411661386489868,
      "eval_runtime": 2.6833,
      "eval_samples_per_second": 856.047,
      "eval_steps_per_second": 13.416,
      "step": 208000
    },
    {
      "epoch": 2.33,
      "learning_rate": 0.0001406024785770333,
      "loss": 0.2461,
      "step": 208500
    },
    {
      "epoch": 2.33,
      "learning_rate": 0.00014054446908028272,
      "loss": 0.2461,
      "step": 209000
    },
    {
      "epoch": 2.33,
      "eval_loss": 0.23610949516296387,
      "eval_runtime": 2.5904,
      "eval_samples_per_second": 886.74,
      "eval_steps_per_second": 13.898,
      "step": 209000
    },
    {
      "epoch": 2.34,
      "learning_rate": 0.0001404862940572656,
      "loss": 0.2457,
      "step": 209500
    },
    {
      "epoch": 2.34,
      "learning_rate": 0.00014042795366703018,
      "loss": 0.2457,
      "step": 210000
    },
    {
      "epoch": 2.34,
      "eval_loss": 0.236760213971138,
      "eval_runtime": 2.7246,
      "eval_samples_per_second": 843.05,
      "eval_steps_per_second": 13.213,
      "step": 210000
    },
    {
      "epoch": 2.35,
      "learning_rate": 0.00014036944806907685,
      "loss": 0.2457,
      "step": 210500
    },
    {
      "epoch": 2.35,
      "learning_rate": 0.0001403107774233577,
      "loss": 0.2464,
      "step": 211000
    },
    {
      "epoch": 2.35,
      "eval_loss": 0.23806537687778473,
      "eval_runtime": 2.6593,
      "eval_samples_per_second": 863.775,
      "eval_steps_per_second": 13.538,
      "step": 211000
    },
    {
      "epoch": 2.36,
      "learning_rate": 0.00014025194189027604,
      "loss": 0.2456,
      "step": 211500
    },
    {
      "epoch": 2.36,
      "learning_rate": 0.00014019294163068597,
      "loss": 0.2456,
      "step": 212000
    },
    {
      "epoch": 2.36,
      "eval_loss": 0.2328302264213562,
      "eval_runtime": 2.5601,
      "eval_samples_per_second": 897.245,
      "eval_steps_per_second": 14.062,
      "step": 212000
    },
    {
      "epoch": 2.37,
      "learning_rate": 0.00014013377680589196,
      "loss": 0.2452,
      "step": 212500
    },
    {
      "epoch": 2.38,
      "learning_rate": 0.00014007444757764835,
      "loss": 0.2457,
      "step": 213000
    },
    {
      "epoch": 2.38,
      "eval_loss": 0.236920565366745,
      "eval_runtime": 2.6122,
      "eval_samples_per_second": 879.347,
      "eval_steps_per_second": 13.782,
      "step": 213000
    },
    {
      "epoch": 2.38,
      "learning_rate": 0.000140014954108159,
      "loss": 0.2455,
      "step": 213500
    },
    {
      "epoch": 2.39,
      "learning_rate": 0.0001399552965600768,
      "loss": 0.2449,
      "step": 214000
    },
    {
      "epoch": 2.39,
      "eval_loss": 0.23769885301589966,
      "eval_runtime": 2.6992,
      "eval_samples_per_second": 850.998,
      "eval_steps_per_second": 13.337,
      "step": 214000
    },
    {
      "epoch": 2.39,
      "learning_rate": 0.00013989547509650314,
      "loss": 0.2447,
      "step": 214500
    },
    {
      "epoch": 2.4,
      "learning_rate": 0.0001398354898809877,
      "loss": 0.2447,
      "step": 215000
    },
    {
      "epoch": 2.4,
      "eval_loss": 0.23445510864257812,
      "eval_runtime": 2.6684,
      "eval_samples_per_second": 860.825,
      "eval_steps_per_second": 13.491,
      "step": 215000
    },
    {
      "epoch": 2.4,
      "learning_rate": 0.00013977534107752776,
      "loss": 0.2448,
      "step": 215500
    },
    {
      "epoch": 2.41,
      "learning_rate": 0.0001397150288505678,
      "loss": 0.2446,
      "step": 216000
    },
    {
      "epoch": 2.41,
      "eval_loss": 0.23636819422245026,
      "eval_runtime": 2.5563,
      "eval_samples_per_second": 898.582,
      "eval_steps_per_second": 14.083,
      "step": 216000
    },
    {
      "epoch": 2.41,
      "learning_rate": 0.0001396545533649992,
      "loss": 0.2448,
      "step": 216500
    },
    {
      "epoch": 2.42,
      "learning_rate": 0.00013959391478615959,
      "loss": 0.2445,
      "step": 217000
    },
    {
      "epoch": 2.42,
      "eval_loss": 0.23241940140724182,
      "eval_runtime": 2.6771,
      "eval_samples_per_second": 858.01,
      "eval_steps_per_second": 13.447,
      "step": 217000
    },
    {
      "epoch": 2.43,
      "learning_rate": 0.00013953311327983254,
      "loss": 0.2445,
      "step": 217500
    },
    {
      "epoch": 2.43,
      "learning_rate": 0.00013947214901224706,
      "loss": 0.2442,
      "step": 218000
    },
    {
      "epoch": 2.43,
      "eval_loss": 0.23456549644470215,
      "eval_runtime": 2.606,
      "eval_samples_per_second": 881.439,
      "eval_steps_per_second": 13.814,
      "step": 218000
    },
    {
      "epoch": 2.44,
      "learning_rate": 0.0001394110221500771,
      "loss": 0.2444,
      "step": 218500
    },
    {
      "epoch": 2.44,
      "learning_rate": 0.0001393497328604412,
      "loss": 0.2445,
      "step": 219000
    },
    {
      "epoch": 2.44,
      "eval_loss": 0.23404227197170258,
      "eval_runtime": 2.5738,
      "eval_samples_per_second": 892.457,
      "eval_steps_per_second": 13.987,
      "step": 219000
    },
    {
      "epoch": 2.45,
      "learning_rate": 0.00013928828131090193,
      "loss": 0.2442,
      "step": 219500
    },
    {
      "epoch": 2.45,
      "learning_rate": 0.00013922666766946545,
      "loss": 0.2435,
      "step": 220000
    },
    {
      "epoch": 2.45,
      "eval_loss": 0.2341037541627884,
      "eval_runtime": 2.6129,
      "eval_samples_per_second": 879.091,
      "eval_steps_per_second": 13.778,
      "step": 220000
    },
    {
      "epoch": 2.46,
      "learning_rate": 0.00013916489210458118,
      "loss": 0.2439,
      "step": 220500
    },
    {
      "epoch": 2.46,
      "learning_rate": 0.00013910295478514106,
      "loss": 0.2433,
      "step": 221000
    },
    {
      "epoch": 2.46,
      "eval_loss": 0.23220616579055786,
      "eval_runtime": 2.6144,
      "eval_samples_per_second": 878.61,
      "eval_steps_per_second": 13.77,
      "step": 221000
    },
    {
      "epoch": 2.47,
      "learning_rate": 0.0001390408558804794,
      "loss": 0.2436,
      "step": 221500
    },
    {
      "epoch": 2.48,
      "learning_rate": 0.0001389785955603722,
      "loss": 0.2435,
      "step": 222000
    },
    {
      "epoch": 2.48,
      "eval_loss": 0.23245947062969208,
      "eval_runtime": 2.5605,
      "eval_samples_per_second": 897.091,
      "eval_steps_per_second": 14.06,
      "step": 222000
    },
    {
      "epoch": 2.48,
      "learning_rate": 0.00013891617399503688,
      "loss": 0.2437,
      "step": 222500
    },
    {
      "epoch": 2.49,
      "learning_rate": 0.00013885359135513154,
      "loss": 0.2434,
      "step": 223000
    },
    {
      "epoch": 2.49,
      "eval_loss": 0.23385584354400635,
      "eval_runtime": 2.5799,
      "eval_samples_per_second": 890.337,
      "eval_steps_per_second": 13.954,
      "step": 223000
    },
    {
      "epoch": 2.49,
      "learning_rate": 0.00013879084781175476,
      "loss": 0.2432,
      "step": 223500
    },
    {
      "epoch": 2.5,
      "learning_rate": 0.000138727943536445,
      "loss": 0.243,
      "step": 224000
    },
    {
      "epoch": 2.5,
      "eval_loss": 0.2319779098033905,
      "eval_runtime": 2.5676,
      "eval_samples_per_second": 894.613,
      "eval_steps_per_second": 14.021,
      "step": 224000
    },
    {
      "epoch": 2.5,
      "learning_rate": 0.0001386648787011801,
      "loss": 0.2428,
      "step": 224500
    },
    {
      "epoch": 2.51,
      "learning_rate": 0.00013860165347837698,
      "loss": 0.2428,
      "step": 225000
    },
    {
      "epoch": 2.51,
      "eval_loss": 0.23255208134651184,
      "eval_runtime": 2.611,
      "eval_samples_per_second": 879.75,
      "eval_steps_per_second": 13.788,
      "step": 225000
    },
    {
      "epoch": 2.51,
      "learning_rate": 0.00013853826804089095,
      "loss": 0.2426,
      "step": 225500
    },
    {
      "epoch": 2.52,
      "learning_rate": 0.00013847472256201535,
      "loss": 0.243,
      "step": 226000
    },
    {
      "epoch": 2.52,
      "eval_loss": 0.23200708627700806,
      "eval_runtime": 2.619,
      "eval_samples_per_second": 877.053,
      "eval_steps_per_second": 13.746,
      "step": 226000
    },
    {
      "epoch": 2.53,
      "learning_rate": 0.00013841101721548112,
      "loss": 0.2419,
      "step": 226500
    },
    {
      "epoch": 2.53,
      "learning_rate": 0.00013834715217545625,
      "loss": 0.2428,
      "step": 227000
    },
    {
      "epoch": 2.53,
      "eval_loss": 0.23063500225543976,
      "eval_runtime": 2.5764,
      "eval_samples_per_second": 891.542,
      "eval_steps_per_second": 13.973,
      "step": 227000
    },
    {
      "epoch": 2.54,
      "learning_rate": 0.00013828312761654532,
      "loss": 0.2424,
      "step": 227500
    },
    {
      "epoch": 2.54,
      "learning_rate": 0.000138218943713789,
      "loss": 0.2423,
      "step": 228000
    },
    {
      "epoch": 2.54,
      "eval_loss": 0.23263856768608093,
      "eval_runtime": 2.6433,
      "eval_samples_per_second": 868.989,
      "eval_steps_per_second": 13.619,
      "step": 228000
    },
    {
      "epoch": 2.55,
      "learning_rate": 0.00013815460064266368,
      "loss": 0.2425,
      "step": 228500
    },
    {
      "epoch": 2.55,
      "learning_rate": 0.0001380900985790808,
      "loss": 0.2421,
      "step": 229000
    },
    {
      "epoch": 2.55,
      "eval_loss": 0.2327061891555786,
      "eval_runtime": 2.6197,
      "eval_samples_per_second": 876.814,
      "eval_steps_per_second": 13.742,
      "step": 229000
    },
    {
      "epoch": 2.56,
      "learning_rate": 0.0001380254376993866,
      "loss": 0.242,
      "step": 229500
    },
    {
      "epoch": 2.57,
      "learning_rate": 0.00013796061818036138,
      "loss": 0.2417,
      "step": 230000
    },
    {
      "epoch": 2.57,
      "eval_loss": 0.23025751113891602,
      "eval_runtime": 2.6113,
      "eval_samples_per_second": 879.626,
      "eval_steps_per_second": 13.786,
      "step": 230000
    },
    {
      "epoch": 2.57,
      "learning_rate": 0.00013789564019921931,
      "loss": 0.2418,
      "step": 230500
    },
    {
      "epoch": 2.58,
      "learning_rate": 0.00013783050393360768,
      "loss": 0.2418,
      "step": 231000
    },
    {
      "epoch": 2.58,
      "eval_loss": 0.23109453916549683,
      "eval_runtime": 2.644,
      "eval_samples_per_second": 868.746,
      "eval_steps_per_second": 13.616,
      "step": 231000
    },
    {
      "epoch": 2.58,
      "learning_rate": 0.00013776520956160655,
      "loss": 0.2417,
      "step": 231500
    },
    {
      "epoch": 2.59,
      "learning_rate": 0.0001376997572617282,
      "loss": 0.2414,
      "step": 232000
    },
    {
      "epoch": 2.59,
      "eval_loss": 0.23072299361228943,
      "eval_runtime": 2.7341,
      "eval_samples_per_second": 840.132,
      "eval_steps_per_second": 13.167,
      "step": 232000
    },
    {
      "epoch": 2.59,
      "learning_rate": 0.0001376341472129168,
      "loss": 0.2414,
      "step": 232500
    },
    {
      "epoch": 2.6,
      "learning_rate": 0.00013756837959454766,
      "loss": 0.2415,
      "step": 233000
    },
    {
      "epoch": 2.6,
      "eval_loss": 0.22945719957351685,
      "eval_runtime": 2.6629,
      "eval_samples_per_second": 862.583,
      "eval_steps_per_second": 13.519,
      "step": 233000
    },
    {
      "epoch": 2.6,
      "learning_rate": 0.00013750245458642692,
      "loss": 0.2411,
      "step": 233500
    },
    {
      "epoch": 2.61,
      "learning_rate": 0.0001374363723687911,
      "loss": 0.2411,
      "step": 234000
    },
    {
      "epoch": 2.61,
      "eval_loss": 0.22973643243312836,
      "eval_runtime": 2.6682,
      "eval_samples_per_second": 860.868,
      "eval_steps_per_second": 13.492,
      "step": 234000
    },
    {
      "epoch": 2.62,
      "learning_rate": 0.0001373701331223064,
      "loss": 0.2409,
      "step": 234500
    },
    {
      "epoch": 2.62,
      "learning_rate": 0.00013730373702806846,
      "loss": 0.241,
      "step": 235000
    },
    {
      "epoch": 2.62,
      "eval_loss": 0.23086461424827576,
      "eval_runtime": 2.643,
      "eval_samples_per_second": 869.097,
      "eval_steps_per_second": 13.621,
      "step": 235000
    },
    {
      "epoch": 2.63,
      "learning_rate": 0.0001372371842676016,
      "loss": 0.2408,
      "step": 235500
    },
    {
      "epoch": 2.63,
      "learning_rate": 0.00013717047502285855,
      "loss": 0.2411,
      "step": 236000
    },
    {
      "epoch": 2.63,
      "eval_loss": 0.22972989082336426,
      "eval_runtime": 2.6826,
      "eval_samples_per_second": 856.274,
      "eval_steps_per_second": 13.42,
      "step": 236000
    },
    {
      "epoch": 2.64,
      "learning_rate": 0.0001371036094762198,
      "loss": 0.2407,
      "step": 236500
    },
    {
      "epoch": 2.64,
      "learning_rate": 0.0001370365878104933,
      "loss": 0.241,
      "step": 237000
    },
    {
      "epoch": 2.64,
      "eval_loss": 0.2288552075624466,
      "eval_runtime": 2.7019,
      "eval_samples_per_second": 850.158,
      "eval_steps_per_second": 13.324,
      "step": 237000
    },
    {
      "epoch": 2.65,
      "learning_rate": 0.00013696941020891363,
      "loss": 0.2406,
      "step": 237500
    },
    {
      "epoch": 2.65,
      "learning_rate": 0.00013690207685514185,
      "loss": 0.2402,
      "step": 238000
    },
    {
      "epoch": 2.65,
      "eval_loss": 0.2302161604166031,
      "eval_runtime": 2.6823,
      "eval_samples_per_second": 856.355,
      "eval_steps_per_second": 13.421,
      "step": 238000
    },
    {
      "epoch": 2.66,
      "learning_rate": 0.0001368345879332647,
      "loss": 0.2406,
      "step": 238500
    },
    {
      "epoch": 2.67,
      "learning_rate": 0.0001367669436277944,
      "loss": 0.2403,
      "step": 239000
    },
    {
      "epoch": 2.67,
      "eval_loss": 0.2315613031387329,
      "eval_runtime": 2.609,
      "eval_samples_per_second": 880.43,
      "eval_steps_per_second": 13.799,
      "step": 239000
    },
    {
      "epoch": 2.67,
      "learning_rate": 0.00013669914412366783,
      "loss": 0.2404,
      "step": 239500
    },
    {
      "epoch": 2.68,
      "learning_rate": 0.0001366311896062463,
      "loss": 0.2403,
      "step": 240000
    },
    {
      "epoch": 2.68,
      "eval_loss": 0.2302522510290146,
      "eval_runtime": 2.7632,
      "eval_samples_per_second": 831.29,
      "eval_steps_per_second": 13.028,
      "step": 240000
    },
    {
      "epoch": 2.68,
      "learning_rate": 0.00013656308026131485,
      "loss": 0.2403,
      "step": 240500
    },
    {
      "epoch": 2.69,
      "learning_rate": 0.00013649481627508181,
      "loss": 0.2401,
      "step": 241000
    },
    {
      "epoch": 2.69,
      "eval_loss": 0.23056615889072418,
      "eval_runtime": 2.5909,
      "eval_samples_per_second": 886.57,
      "eval_steps_per_second": 13.895,
      "step": 241000
    },
    {
      "epoch": 2.69,
      "learning_rate": 0.00013642639783417832,
      "loss": 0.2401,
      "step": 241500
    },
    {
      "epoch": 2.7,
      "learning_rate": 0.0001363578251256578,
      "loss": 0.2394,
      "step": 242000
    },
    {
      "epoch": 2.7,
      "eval_loss": 0.23005209863185883,
      "eval_runtime": 2.6861,
      "eval_samples_per_second": 855.147,
      "eval_steps_per_second": 13.402,
      "step": 242000
    },
    {
      "epoch": 2.7,
      "learning_rate": 0.0001362890983369954,
      "loss": 0.2395,
      "step": 242500
    },
    {
      "epoch": 2.71,
      "learning_rate": 0.00013622021765608754,
      "loss": 0.2399,
      "step": 243000
    },
    {
      "epoch": 2.71,
      "eval_loss": 0.22968289256095886,
      "eval_runtime": 2.6547,
      "eval_samples_per_second": 865.265,
      "eval_steps_per_second": 13.561,
      "step": 243000
    },
    {
      "epoch": 2.72,
      "learning_rate": 0.00013615118327125136,
      "loss": 0.2395,
      "step": 243500
    },
    {
      "epoch": 2.72,
      "learning_rate": 0.00013608199537122425,
      "loss": 0.2389,
      "step": 244000
    },
    {
      "epoch": 2.72,
      "eval_loss": 0.23073868453502655,
      "eval_runtime": 2.6945,
      "eval_samples_per_second": 852.485,
      "eval_steps_per_second": 13.361,
      "step": 244000
    },
    {
      "epoch": 2.73,
      "learning_rate": 0.0001360126541451633,
      "loss": 0.2394,
      "step": 244500
    },
    {
      "epoch": 2.73,
      "learning_rate": 0.0001359431597826447,
      "loss": 0.239,
      "step": 245000
    },
    {
      "epoch": 2.73,
      "eval_loss": 0.22961042821407318,
      "eval_runtime": 2.5955,
      "eval_samples_per_second": 885.008,
      "eval_steps_per_second": 13.87,
      "step": 245000
    },
    {
      "epoch": 2.74,
      "learning_rate": 0.0001358735124736635,
      "loss": 0.24,
      "step": 245500
    },
    {
      "epoch": 2.74,
      "learning_rate": 0.0001358037124086327,
      "loss": 0.2393,
      "step": 246000
    },
    {
      "epoch": 2.74,
      "eval_loss": 0.23064970970153809,
      "eval_runtime": 2.6363,
      "eval_samples_per_second": 871.309,
      "eval_steps_per_second": 13.656,
      "step": 246000
    },
    {
      "epoch": 2.75,
      "learning_rate": 0.00013573375977838303,
      "loss": 0.2389,
      "step": 246500
    },
    {
      "epoch": 2.75,
      "learning_rate": 0.00013566365477416233,
      "loss": 0.2386,
      "step": 247000
    },
    {
      "epoch": 2.75,
      "eval_loss": 0.22737060487270355,
      "eval_runtime": 2.5775,
      "eval_samples_per_second": 891.16,
      "eval_steps_per_second": 13.967,
      "step": 247000
    },
    {
      "epoch": 2.76,
      "learning_rate": 0.00013559339758763495,
      "loss": 0.239,
      "step": 247500
    },
    {
      "epoch": 2.77,
      "learning_rate": 0.00013552298841088144,
      "loss": 0.2389,
      "step": 248000
    },
    {
      "epoch": 2.77,
      "eval_loss": 0.22698281705379486,
      "eval_runtime": 2.6869,
      "eval_samples_per_second": 854.901,
      "eval_steps_per_second": 13.399,
      "step": 248000
    },
    {
      "epoch": 2.77,
      "learning_rate": 0.00013545242743639774,
      "loss": 0.2387,
      "step": 248500
    },
    {
      "epoch": 2.78,
      "learning_rate": 0.00013538171485709486,
      "loss": 0.2382,
      "step": 249000
    },
    {
      "epoch": 2.78,
      "eval_loss": 0.22949343919754028,
      "eval_runtime": 2.6923,
      "eval_samples_per_second": 853.185,
      "eval_steps_per_second": 13.372,
      "step": 249000
    },
    {
      "epoch": 2.78,
      "learning_rate": 0.00013531085086629832,
      "loss": 0.2384,
      "step": 249500
    },
    {
      "epoch": 2.79,
      "learning_rate": 0.00013523983565774753,
      "loss": 0.2384,
      "step": 250000
    },
    {
      "epoch": 2.79,
      "eval_loss": 0.22683538496494293,
      "eval_runtime": 2.6574,
      "eval_samples_per_second": 864.374,
      "eval_steps_per_second": 13.547,
      "step": 250000
    },
    {
      "epoch": 2.79,
      "learning_rate": 0.0001351686694255954,
      "loss": 0.2385,
      "step": 250500
    },
    {
      "epoch": 2.8,
      "learning_rate": 0.00013509735236440766,
      "loss": 0.2379,
      "step": 251000
    },
    {
      "epoch": 2.8,
      "eval_loss": 0.2289341241121292,
      "eval_runtime": 2.5642,
      "eval_samples_per_second": 895.794,
      "eval_steps_per_second": 14.039,
      "step": 251000
    },
    {
      "epoch": 2.8,
      "learning_rate": 0.0001350258846691625,
      "loss": 0.238,
      "step": 251500
    },
    {
      "epoch": 2.81,
      "learning_rate": 0.00013495426653524972,
      "loss": 0.2381,
      "step": 252000
    },
    {
      "epoch": 2.81,
      "eval_loss": 0.22911319136619568,
      "eval_runtime": 2.6235,
      "eval_samples_per_second": 875.54,
      "eval_steps_per_second": 13.722,
      "step": 252000
    },
    {
      "epoch": 2.82,
      "learning_rate": 0.0001348824981584707,
      "loss": 0.2383,
      "step": 252500
    },
    {
      "epoch": 2.82,
      "learning_rate": 0.00013481057973503742,
      "loss": 0.2378,
      "step": 253000
    },
    {
      "epoch": 2.82,
      "eval_loss": 0.22786130011081696,
      "eval_runtime": 2.5806,
      "eval_samples_per_second": 890.1,
      "eval_steps_per_second": 13.95,
      "step": 253000
    },
    {
      "epoch": 2.83,
      "learning_rate": 0.00013473851146157204,
      "loss": 0.2382,
      "step": 253500
    },
    {
      "epoch": 2.83,
      "learning_rate": 0.00013466629353510651,
      "loss": 0.2376,
      "step": 254000
    },
    {
      "epoch": 2.83,
      "eval_loss": 0.2276422381401062,
      "eval_runtime": 2.6007,
      "eval_samples_per_second": 883.226,
      "eval_steps_per_second": 13.842,
      "step": 254000
    },
    {
      "epoch": 2.84,
      "learning_rate": 0.00013459392615308192,
      "loss": 0.2382,
      "step": 254500
    },
    {
      "epoch": 2.84,
      "learning_rate": 0.00013452140951334787,
      "loss": 0.2373,
      "step": 255000
    },
    {
      "epoch": 2.84,
      "eval_loss": 0.22781763970851898,
      "eval_runtime": 2.5831,
      "eval_samples_per_second": 889.242,
      "eval_steps_per_second": 13.937,
      "step": 255000
    },
    {
      "epoch": 2.85,
      "learning_rate": 0.00013444874381416208,
      "loss": 0.2371,
      "step": 255500
    },
    {
      "epoch": 2.86,
      "learning_rate": 0.00013437592925418985,
      "loss": 0.2374,
      "step": 256000
    },
    {
      "epoch": 2.86,
      "eval_loss": 0.22769607603549957,
      "eval_runtime": 2.6139,
      "eval_samples_per_second": 878.768,
      "eval_steps_per_second": 13.773,
      "step": 256000
    },
    {
      "epoch": 2.86,
      "learning_rate": 0.00013430296603250338,
      "loss": 0.2375,
      "step": 256500
    },
    {
      "epoch": 2.87,
      "learning_rate": 0.00013422985434858133,
      "loss": 0.2372,
      "step": 257000
    },
    {
      "epoch": 2.87,
      "eval_loss": 0.22923217713832855,
      "eval_runtime": 2.5056,
      "eval_samples_per_second": 916.749,
      "eval_steps_per_second": 14.368,
      "step": 257000
    },
    {
      "epoch": 2.87,
      "learning_rate": 0.00013415659440230824,
      "loss": 0.2375,
      "step": 257500
    },
    {
      "epoch": 2.88,
      "learning_rate": 0.00013408318639397405,
      "loss": 0.2364,
      "step": 258000
    },
    {
      "epoch": 2.88,
      "eval_loss": 0.2288464456796646,
      "eval_runtime": 2.5629,
      "eval_samples_per_second": 896.264,
      "eval_steps_per_second": 14.047,
      "step": 258000
    },
    {
      "epoch": 2.88,
      "learning_rate": 0.00013400963052427337,
      "loss": 0.237,
      "step": 258500
    },
    {
      "epoch": 2.89,
      "learning_rate": 0.00013393592699430525,
      "loss": 0.2371,
      "step": 259000
    },
    {
      "epoch": 2.89,
      "eval_loss": 0.22805295884609222,
      "eval_runtime": 2.5516,
      "eval_samples_per_second": 900.222,
      "eval_steps_per_second": 14.109,
      "step": 259000
    },
    {
      "epoch": 2.89,
      "learning_rate": 0.00013386207600557222,
      "loss": 0.2363,
      "step": 259500
    },
    {
      "epoch": 2.9,
      "learning_rate": 0.00013378807775998012,
      "loss": 0.237,
      "step": 260000
    },
    {
      "epoch": 2.9,
      "eval_loss": 0.22537867724895477,
      "eval_runtime": 2.54,
      "eval_samples_per_second": 904.347,
      "eval_steps_per_second": 14.173,
      "step": 260000
    },
    {
      "epoch": 2.91,
      "learning_rate": 0.0001337139324598373,
      "loss": 0.2369,
      "step": 260500
    },
    {
      "epoch": 2.91,
      "learning_rate": 0.00013363964030785422,
      "loss": 0.2366,
      "step": 261000
    },
    {
      "epoch": 2.91,
      "eval_loss": 0.22471630573272705,
      "eval_runtime": 2.4998,
      "eval_samples_per_second": 918.881,
      "eval_steps_per_second": 14.401,
      "step": 261000
    },
    {
      "epoch": 2.92,
      "learning_rate": 0.00013356520150714277,
      "loss": 0.2368,
      "step": 261500
    },
    {
      "epoch": 2.92,
      "learning_rate": 0.00013349061626121578,
      "loss": 0.2366,
      "step": 262000
    },
    {
      "epoch": 2.92,
      "eval_loss": 0.22358164191246033,
      "eval_runtime": 2.454,
      "eval_samples_per_second": 936.011,
      "eval_steps_per_second": 14.67,
      "step": 262000
    },
    {
      "epoch": 2.93,
      "learning_rate": 0.00013341588477398645,
      "loss": 0.2365,
      "step": 262500
    },
    {
      "epoch": 2.93,
      "learning_rate": 0.00013334100724976783,
      "loss": 0.2358,
      "step": 263000
    },
    {
      "epoch": 2.93,
      "eval_loss": 0.22610020637512207,
      "eval_runtime": 2.4695,
      "eval_samples_per_second": 930.133,
      "eval_steps_per_second": 14.578,
      "step": 263000
    },
    {
      "epoch": 2.94,
      "learning_rate": 0.00013326598389327223,
      "loss": 0.2368,
      "step": 263500
    },
    {
      "epoch": 2.94,
      "learning_rate": 0.0001331908149096106,
      "loss": 0.2361,
      "step": 264000
    },
    {
      "epoch": 2.94,
      "eval_loss": 0.22605933248996735,
      "eval_runtime": 2.4948,
      "eval_samples_per_second": 920.701,
      "eval_steps_per_second": 14.43,
      "step": 264000
    },
    {
      "epoch": 2.95,
      "learning_rate": 0.0001331155005042921,
      "loss": 0.2358,
      "step": 264500
    },
    {
      "epoch": 2.96,
      "learning_rate": 0.00013304004088322342,
      "loss": 0.2361,
      "step": 265000
    },
    {
      "epoch": 2.96,
      "eval_loss": 0.22790436446666718,
      "eval_runtime": 2.407,
      "eval_samples_per_second": 954.282,
      "eval_steps_per_second": 14.956,
      "step": 265000
    },
    {
      "epoch": 2.96,
      "learning_rate": 0.00013296443625270828,
      "loss": 0.2363,
      "step": 265500
    },
    {
      "epoch": 2.97,
      "learning_rate": 0.00013288868681944692,
      "loss": 0.236,
      "step": 266000
    },
    {
      "epoch": 2.97,
      "eval_loss": 0.22779372334480286,
      "eval_runtime": 2.5112,
      "eval_samples_per_second": 914.702,
      "eval_steps_per_second": 14.336,
      "step": 266000
    },
    {
      "epoch": 2.97,
      "learning_rate": 0.00013281279279053532,
      "loss": 0.2357,
      "step": 266500
    },
    {
      "epoch": 2.98,
      "learning_rate": 0.00013273675437346487,
      "loss": 0.2353,
      "step": 267000
    },
    {
      "epoch": 2.98,
      "eval_loss": 0.22502335906028748,
      "eval_runtime": 2.3705,
      "eval_samples_per_second": 968.999,
      "eval_steps_per_second": 15.187,
      "step": 267000
    },
    {
      "epoch": 2.98,
      "learning_rate": 0.00013266057177612172,
      "loss": 0.2356,
      "step": 267500
    },
    {
      "epoch": 2.99,
      "learning_rate": 0.00013258424520678618,
      "loss": 0.2357,
      "step": 268000
    },
    {
      "epoch": 2.99,
      "eval_loss": 0.2256162017583847,
      "eval_runtime": 15.618,
      "eval_samples_per_second": 147.074,
      "eval_steps_per_second": 2.305,
      "step": 268000
    },
    {
      "epoch": 2.99,
      "learning_rate": 0.00013250777487413217,
      "loss": 0.2355,
      "step": 268500
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00013243116098722663,
      "loss": 0.2357,
      "step": 269000
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.2276550680398941,
      "eval_runtime": 2.6194,
      "eval_samples_per_second": 876.914,
      "eval_steps_per_second": 13.744,
      "step": 269000
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00013235440375552906,
      "loss": 0.2354,
      "step": 269500
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00013227750338889077,
      "loss": 0.2354,
      "step": 270000
    },
    {
      "epoch": 3.01,
      "eval_loss": 0.22490766644477844,
      "eval_runtime": 2.524,
      "eval_samples_per_second": 910.07,
      "eval_steps_per_second": 14.263,
      "step": 270000
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0001322004600975544,
      "loss": 0.2351,
      "step": 270500
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00013212327409215343,
      "loss": 0.2349,
      "step": 271000
    },
    {
      "epoch": 3.02,
      "eval_loss": 0.2255261242389679,
      "eval_runtime": 2.5651,
      "eval_samples_per_second": 895.473,
      "eval_steps_per_second": 14.034,
      "step": 271000
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0001320459455837114,
      "loss": 0.2348,
      "step": 271500
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0001319684747836415,
      "loss": 0.2348,
      "step": 272000
    },
    {
      "epoch": 3.03,
      "eval_loss": 0.2229447066783905,
      "eval_runtime": 2.6131,
      "eval_samples_per_second": 879.025,
      "eval_steps_per_second": 13.777,
      "step": 272000
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00013189086190374595,
      "loss": 0.2348,
      "step": 272500
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0001318131071562154,
      "loss": 0.2345,
      "step": 273000
    },
    {
      "epoch": 3.04,
      "eval_loss": 0.22524061799049377,
      "eval_runtime": 2.5512,
      "eval_samples_per_second": 900.358,
      "eval_steps_per_second": 14.111,
      "step": 273000
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0001317352107536284,
      "loss": 0.2345,
      "step": 273500
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00013165717290895067,
      "loss": 0.2344,
      "step": 274000
    },
    {
      "epoch": 3.06,
      "eval_loss": 0.22475993633270264,
      "eval_runtime": 2.529,
      "eval_samples_per_second": 908.276,
      "eval_steps_per_second": 14.235,
      "step": 274000
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00013157899383553474,
      "loss": 0.2344,
      "step": 274500
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0001315006737471192,
      "loss": 0.2343,
      "step": 275000
    },
    {
      "epoch": 3.07,
      "eval_loss": 0.2223171889781952,
      "eval_runtime": 2.571,
      "eval_samples_per_second": 893.41,
      "eval_steps_per_second": 14.002,
      "step": 275000
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0001314222128578282,
      "loss": 0.2342,
      "step": 275500
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0001313436113821708,
      "loss": 0.2344,
      "step": 276000
    },
    {
      "epoch": 3.08,
      "eval_loss": 0.22429585456848145,
      "eval_runtime": 2.7431,
      "eval_samples_per_second": 837.378,
      "eval_steps_per_second": 13.124,
      "step": 276000
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0001312648695350404,
      "loss": 0.2344,
      "step": 276500
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00013118598753171425,
      "loss": 0.2343,
      "step": 277000
    },
    {
      "epoch": 3.09,
      "eval_loss": 0.22134767472743988,
      "eval_runtime": 2.5289,
      "eval_samples_per_second": 908.315,
      "eval_steps_per_second": 14.236,
      "step": 277000
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00013110696558785273,
      "loss": 0.2337,
      "step": 277500
    },
    {
      "epoch": 3.1,
      "learning_rate": 0.0001310278039194988,
      "loss": 0.2335,
      "step": 278000
    },
    {
      "epoch": 3.1,
      "eval_loss": 0.22428634762763977,
      "eval_runtime": 2.646,
      "eval_samples_per_second": 868.114,
      "eval_steps_per_second": 13.606,
      "step": 278000
    },
    {
      "epoch": 3.11,
      "learning_rate": 0.00013094850274307745,
      "loss": 0.234,
      "step": 278500
    },
    {
      "epoch": 3.11,
      "learning_rate": 0.00013086906227539506,
      "loss": 0.2336,
      "step": 279000
    },
    {
      "epoch": 3.11,
      "eval_loss": 0.22171786427497864,
      "eval_runtime": 4.1933,
      "eval_samples_per_second": 547.781,
      "eval_steps_per_second": 8.585,
      "step": 279000
    },
    {
      "epoch": 3.12,
      "learning_rate": 0.00013078948273363884,
      "loss": 0.2336,
      "step": 279500
    },
    {
      "epoch": 3.12,
      "learning_rate": 0.00013070976433537623,
      "loss": 0.2339,
      "step": 280000
    },
    {
      "epoch": 3.12,
      "eval_loss": 0.22290877997875214,
      "eval_runtime": 2.542,
      "eval_samples_per_second": 903.625,
      "eval_steps_per_second": 14.162,
      "step": 280000
    },
    {
      "epoch": 3.13,
      "learning_rate": 0.00013062990729855427,
      "loss": 0.2338,
      "step": 280500
    },
    {
      "epoch": 3.13,
      "learning_rate": 0.00013054991184149905,
      "loss": 0.2337,
      "step": 281000
    },
    {
      "epoch": 3.13,
      "eval_loss": 0.2214738130569458,
      "eval_runtime": 2.5831,
      "eval_samples_per_second": 889.227,
      "eval_steps_per_second": 13.937,
      "step": 281000
    },
    {
      "epoch": 3.14,
      "learning_rate": 0.00013046977818291508,
      "loss": 0.2337,
      "step": 281500
    },
    {
      "epoch": 3.15,
      "learning_rate": 0.00013038950654188476,
      "loss": 0.2334,
      "step": 282000
    },
    {
      "epoch": 3.15,
      "eval_loss": 0.2210046350955963,
      "eval_runtime": 2.5814,
      "eval_samples_per_second": 889.83,
      "eval_steps_per_second": 13.946,
      "step": 282000
    },
    {
      "epoch": 3.15,
      "learning_rate": 0.00013030909713786768,
      "loss": 0.2334,
      "step": 282500
    },
    {
      "epoch": 3.16,
      "learning_rate": 0.00013022855019070005,
      "loss": 0.2335,
      "step": 283000
    },
    {
      "epoch": 3.16,
      "eval_loss": 0.22165259718894958,
      "eval_runtime": 2.6243,
      "eval_samples_per_second": 875.295,
      "eval_steps_per_second": 13.718,
      "step": 283000
    },
    {
      "epoch": 3.16,
      "learning_rate": 0.00013014786592059418,
      "loss": 0.2331,
      "step": 283500
    },
    {
      "epoch": 3.17,
      "learning_rate": 0.0001300670445481378,
      "loss": 0.2328,
      "step": 284000
    },
    {
      "epoch": 3.17,
      "eval_loss": 0.2222437709569931,
      "eval_runtime": 2.5682,
      "eval_samples_per_second": 894.402,
      "eval_steps_per_second": 14.018,
      "step": 284000
    },
    {
      "epoch": 3.17,
      "learning_rate": 0.0001299860862942934,
      "loss": 0.2328,
      "step": 284500
    },
    {
      "epoch": 3.18,
      "learning_rate": 0.0001299049913803978,
      "loss": 0.2326,
      "step": 285000
    },
    {
      "epoch": 3.18,
      "eval_loss": 0.22183333337306976,
      "eval_runtime": 2.553,
      "eval_samples_per_second": 899.722,
      "eval_steps_per_second": 14.101,
      "step": 285000
    },
    {
      "epoch": 3.18,
      "learning_rate": 0.00012982376002816138,
      "loss": 0.2328,
      "step": 285500
    },
    {
      "epoch": 3.19,
      "learning_rate": 0.00012974239245966754,
      "loss": 0.2332,
      "step": 286000
    },
    {
      "epoch": 3.19,
      "eval_loss": 0.22312231361865997,
      "eval_runtime": 2.5852,
      "eval_samples_per_second": 888.513,
      "eval_steps_per_second": 13.925,
      "step": 286000
    },
    {
      "epoch": 3.2,
      "learning_rate": 0.00012966088889737216,
      "loss": 0.233,
      "step": 286500
    },
    {
      "epoch": 3.2,
      "learning_rate": 0.0001295792495641028,
      "loss": 0.2325,
      "step": 287000
    },
    {
      "epoch": 3.2,
      "eval_loss": 0.22235532104969025,
      "eval_runtime": 2.6014,
      "eval_samples_per_second": 882.973,
      "eval_steps_per_second": 13.838,
      "step": 287000
    },
    {
      "epoch": 3.21,
      "learning_rate": 0.0001294974746830583,
      "loss": 0.2325,
      "step": 287500
    },
    {
      "epoch": 3.21,
      "learning_rate": 0.00012941556447780813,
      "loss": 0.2326,
      "step": 288000
    },
    {
      "epoch": 3.21,
      "eval_loss": 0.22186285257339478,
      "eval_runtime": 2.5183,
      "eval_samples_per_second": 912.138,
      "eval_steps_per_second": 14.296,
      "step": 288000
    },
    {
      "epoch": 3.22,
      "learning_rate": 0.00012933351917229156,
      "loss": 0.2324,
      "step": 288500
    },
    {
      "epoch": 3.22,
      "learning_rate": 0.0001292513389908174,
      "loss": 0.2323,
      "step": 289000
    },
    {
      "epoch": 3.22,
      "eval_loss": 0.22211854159832,
      "eval_runtime": 2.5459,
      "eval_samples_per_second": 902.227,
      "eval_steps_per_second": 14.14,
      "step": 289000
    },
    {
      "epoch": 3.23,
      "learning_rate": 0.00012916902415806305,
      "loss": 0.2322,
      "step": 289500
    },
    {
      "epoch": 3.23,
      "learning_rate": 0.0001290865748990742,
      "loss": 0.2321,
      "step": 290000
    },
    {
      "epoch": 3.23,
      "eval_loss": 0.22493046522140503,
      "eval_runtime": 2.5418,
      "eval_samples_per_second": 903.676,
      "eval_steps_per_second": 14.163,
      "step": 290000
    },
    {
      "epoch": 3.24,
      "learning_rate": 0.00012900399143926395,
      "loss": 0.2322,
      "step": 290500
    },
    {
      "epoch": 3.25,
      "learning_rate": 0.00012892127400441228,
      "loss": 0.232,
      "step": 291000
    },
    {
      "epoch": 3.25,
      "eval_loss": 0.22251969575881958,
      "eval_runtime": 2.6362,
      "eval_samples_per_second": 871.331,
      "eval_steps_per_second": 13.656,
      "step": 291000
    },
    {
      "epoch": 3.25,
      "learning_rate": 0.00012883842282066557,
      "loss": 0.2322,
      "step": 291500
    },
    {
      "epoch": 3.26,
      "learning_rate": 0.00012875543811453576,
      "loss": 0.2321,
      "step": 292000
    },
    {
      "epoch": 3.26,
      "eval_loss": 0.22058303654193878,
      "eval_runtime": 2.601,
      "eval_samples_per_second": 883.125,
      "eval_steps_per_second": 13.841,
      "step": 292000
    },
    {
      "epoch": 3.26,
      "learning_rate": 0.00012867232011289984,
      "loss": 0.2315,
      "step": 292500
    },
    {
      "epoch": 3.27,
      "learning_rate": 0.0001285890690429993,
      "loss": 0.2318,
      "step": 293000
    },
    {
      "epoch": 3.27,
      "eval_loss": 0.22088617086410522,
      "eval_runtime": 2.572,
      "eval_samples_per_second": 893.074,
      "eval_steps_per_second": 13.997,
      "step": 293000
    },
    {
      "epoch": 3.27,
      "learning_rate": 0.00012850568513243934,
      "loss": 0.2318,
      "step": 293500
    },
    {
      "epoch": 3.28,
      "learning_rate": 0.00012842216860918846,
      "loss": 0.2318,
      "step": 294000
    },
    {
      "epoch": 3.28,
      "eval_loss": 0.2187003344297409,
      "eval_runtime": 2.5477,
      "eval_samples_per_second": 901.582,
      "eval_steps_per_second": 14.13,
      "step": 294000
    },
    {
      "epoch": 3.28,
      "learning_rate": 0.00012833851970157757,
      "loss": 0.2316,
      "step": 294500
    },
    {
      "epoch": 3.29,
      "learning_rate": 0.0001282547386382996,
      "loss": 0.2314,
      "step": 295000
    },
    {
      "epoch": 3.29,
      "eval_loss": 0.22022201120853424,
      "eval_runtime": 2.585,
      "eval_samples_per_second": 888.583,
      "eval_steps_per_second": 13.926,
      "step": 295000
    },
    {
      "epoch": 3.3,
      "learning_rate": 0.00012817082564840881,
      "loss": 0.2315,
      "step": 295500
    },
    {
      "epoch": 3.3,
      "learning_rate": 0.0001280867809613201,
      "loss": 0.2314,
      "step": 296000
    },
    {
      "epoch": 3.3,
      "eval_loss": 0.22136667370796204,
      "eval_runtime": 2.501,
      "eval_samples_per_second": 918.422,
      "eval_steps_per_second": 14.394,
      "step": 296000
    },
    {
      "epoch": 3.31,
      "learning_rate": 0.00012800260480680845,
      "loss": 0.2316,
      "step": 296500
    },
    {
      "epoch": 3.31,
      "learning_rate": 0.0001279182974150082,
      "loss": 0.2313,
      "step": 297000
    },
    {
      "epoch": 3.31,
      "eval_loss": 0.21788428723812103,
      "eval_runtime": 2.5879,
      "eval_samples_per_second": 887.603,
      "eval_steps_per_second": 13.911,
      "step": 297000
    },
    {
      "epoch": 3.32,
      "learning_rate": 0.00012783385901641258,
      "loss": 0.2312,
      "step": 297500
    },
    {
      "epoch": 3.32,
      "learning_rate": 0.00012774928984187297,
      "loss": 0.2309,
      "step": 298000
    },
    {
      "epoch": 3.32,
      "eval_loss": 0.22021803259849548,
      "eval_runtime": 2.7305,
      "eval_samples_per_second": 841.227,
      "eval_steps_per_second": 13.184,
      "step": 298000
    },
    {
      "epoch": 3.33,
      "learning_rate": 0.00012766459012259818,
      "loss": 0.2316,
      "step": 298500
    },
    {
      "epoch": 3.33,
      "learning_rate": 0.00012757976009015413,
      "loss": 0.2309,
      "step": 299000
    },
    {
      "epoch": 3.33,
      "eval_loss": 0.2220044583082199,
      "eval_runtime": 2.5329,
      "eval_samples_per_second": 906.883,
      "eval_steps_per_second": 14.213,
      "step": 299000
    },
    {
      "epoch": 3.34,
      "learning_rate": 0.00012749479997646275,
      "loss": 0.2307,
      "step": 299500
    },
    {
      "epoch": 3.35,
      "learning_rate": 0.0001274097100138019,
      "loss": 0.2313,
      "step": 300000
    },
    {
      "epoch": 3.35,
      "eval_loss": 0.2212502360343933,
      "eval_runtime": 2.5517,
      "eval_samples_per_second": 900.187,
      "eval_steps_per_second": 14.108,
      "step": 300000
    },
    {
      "epoch": 3.35,
      "learning_rate": 0.00012732449043480413,
      "loss": 0.2303,
      "step": 300500
    },
    {
      "epoch": 3.36,
      "learning_rate": 0.00012723914147245663,
      "loss": 0.2309,
      "step": 301000
    },
    {
      "epoch": 3.36,
      "eval_loss": 0.22217054665088654,
      "eval_runtime": 2.6319,
      "eval_samples_per_second": 872.746,
      "eval_steps_per_second": 13.678,
      "step": 301000
    },
    {
      "epoch": 3.36,
      "learning_rate": 0.00012715366336010016,
      "loss": 0.2309,
      "step": 301500
    },
    {
      "epoch": 3.37,
      "learning_rate": 0.00012706805633142863,
      "loss": 0.2306,
      "step": 302000
    },
    {
      "epoch": 3.37,
      "eval_loss": 0.2169163078069687,
      "eval_runtime": 2.56,
      "eval_samples_per_second": 897.252,
      "eval_steps_per_second": 14.062,
      "step": 302000
    },
    {
      "epoch": 3.37,
      "learning_rate": 0.00012698232062048837,
      "loss": 0.2306,
      "step": 302500
    },
    {
      "epoch": 3.38,
      "learning_rate": 0.00012689645646167755,
      "loss": 0.2302,
      "step": 303000
    },
    {
      "epoch": 3.38,
      "eval_loss": 0.21775402128696442,
      "eval_runtime": 2.5371,
      "eval_samples_per_second": 905.368,
      "eval_steps_per_second": 14.189,
      "step": 303000
    },
    {
      "epoch": 3.38,
      "learning_rate": 0.0001268104640897455,
      "loss": 0.2303,
      "step": 303500
    },
    {
      "epoch": 3.39,
      "learning_rate": 0.00012672434373979207,
      "loss": 0.2299,
      "step": 304000
    },
    {
      "epoch": 3.39,
      "eval_loss": 0.22127319872379303,
      "eval_runtime": 2.5832,
      "eval_samples_per_second": 889.194,
      "eval_steps_per_second": 13.936,
      "step": 304000
    },
    {
      "epoch": 3.4,
      "learning_rate": 0.00012663809564726706,
      "loss": 0.2302,
      "step": 304500
    },
    {
      "epoch": 3.4,
      "learning_rate": 0.00012655172004796936,
      "loss": 0.2304,
      "step": 305000
    },
    {
      "epoch": 3.4,
      "eval_loss": 0.21901114284992218,
      "eval_runtime": 2.5923,
      "eval_samples_per_second": 886.081,
      "eval_steps_per_second": 13.887,
      "step": 305000
    },
    {
      "epoch": 3.41,
      "learning_rate": 0.00012646521717804668,
      "loss": 0.2301,
      "step": 305500
    },
    {
      "epoch": 3.41,
      "learning_rate": 0.00012637858727399448,
      "loss": 0.2301,
      "step": 306000
    },
    {
      "epoch": 3.41,
      "eval_loss": 0.21952645480632782,
      "eval_runtime": 2.5862,
      "eval_samples_per_second": 888.187,
      "eval_steps_per_second": 13.92,
      "step": 306000
    },
    {
      "epoch": 3.42,
      "learning_rate": 0.00012629183057265563,
      "loss": 0.2301,
      "step": 306500
    },
    {
      "epoch": 3.42,
      "learning_rate": 0.00012620494731121966,
      "loss": 0.2297,
      "step": 307000
    },
    {
      "epoch": 3.42,
      "eval_loss": 0.2185007482767105,
      "eval_runtime": 2.5465,
      "eval_samples_per_second": 902.023,
      "eval_steps_per_second": 14.137,
      "step": 307000
    },
    {
      "epoch": 3.43,
      "learning_rate": 0.00012611793772722204,
      "loss": 0.2299,
      "step": 307500
    },
    {
      "epoch": 3.44,
      "learning_rate": 0.00012603080205854372,
      "loss": 0.2297,
      "step": 308000
    },
    {
      "epoch": 3.44,
      "eval_loss": 0.21959775686264038,
      "eval_runtime": 2.6732,
      "eval_samples_per_second": 859.262,
      "eval_steps_per_second": 13.467,
      "step": 308000
    },
    {
      "epoch": 3.44,
      "learning_rate": 0.00012594354054341024,
      "loss": 0.2299,
      "step": 308500
    },
    {
      "epoch": 3.45,
      "learning_rate": 0.00012585615342039126,
      "loss": 0.2299,
      "step": 309000
    },
    {
      "epoch": 3.45,
      "eval_loss": 0.21925170719623566,
      "eval_runtime": 2.6993,
      "eval_samples_per_second": 850.968,
      "eval_steps_per_second": 13.337,
      "step": 309000
    },
    {
      "epoch": 3.45,
      "learning_rate": 0.00012576864092839985,
      "loss": 0.2295,
      "step": 309500
    },
    {
      "epoch": 3.46,
      "learning_rate": 0.0001256810033066918,
      "loss": 0.2293,
      "step": 310000
    },
    {
      "epoch": 3.46,
      "eval_loss": 0.21877124905586243,
      "eval_runtime": 2.5234,
      "eval_samples_per_second": 910.263,
      "eval_steps_per_second": 14.266,
      "step": 310000
    },
    {
      "epoch": 3.46,
      "learning_rate": 0.00012559324079486505,
      "loss": 0.2299,
      "step": 310500
    },
    {
      "epoch": 3.47,
      "learning_rate": 0.0001255053536328589,
      "loss": 0.2294,
      "step": 311000
    },
    {
      "epoch": 3.47,
      "eval_loss": 0.2199290245771408,
      "eval_runtime": 2.5325,
      "eval_samples_per_second": 907.016,
      "eval_steps_per_second": 14.215,
      "step": 311000
    },
    {
      "epoch": 3.47,
      "learning_rate": 0.0001254173420609536,
      "loss": 0.2293,
      "step": 311500
    },
    {
      "epoch": 3.48,
      "learning_rate": 0.0001253292063197693,
      "loss": 0.2294,
      "step": 312000
    },
    {
      "epoch": 3.48,
      "eval_loss": 0.2200849950313568,
      "eval_runtime": 2.5639,
      "eval_samples_per_second": 895.889,
      "eval_steps_per_second": 14.041,
      "step": 312000
    },
    {
      "epoch": 3.49,
      "learning_rate": 0.00012524094665026584,
      "loss": 0.229,
      "step": 312500
    },
    {
      "epoch": 3.49,
      "learning_rate": 0.0001251525632937418,
      "loss": 0.2292,
      "step": 313000
    },
    {
      "epoch": 3.49,
      "eval_loss": 0.21636785566806793,
      "eval_runtime": 2.5639,
      "eval_samples_per_second": 895.902,
      "eval_steps_per_second": 14.041,
      "step": 313000
    },
    {
      "epoch": 3.5,
      "learning_rate": 0.0001250640564918338,
      "loss": 0.229,
      "step": 313500
    },
    {
      "epoch": 3.5,
      "learning_rate": 0.00012497542648651615,
      "loss": 0.229,
      "step": 314000
    },
    {
      "epoch": 3.5,
      "eval_loss": 0.21922779083251953,
      "eval_runtime": 2.5515,
      "eval_samples_per_second": 900.238,
      "eval_steps_per_second": 14.109,
      "step": 314000
    },
    {
      "epoch": 3.51,
      "learning_rate": 0.00012488667352009985,
      "loss": 0.229,
      "step": 314500
    },
    {
      "epoch": 3.51,
      "learning_rate": 0.00012479779783523216,
      "loss": 0.2287,
      "step": 315000
    },
    {
      "epoch": 3.51,
      "eval_loss": 0.21868407726287842,
      "eval_runtime": 2.5974,
      "eval_samples_per_second": 884.334,
      "eval_steps_per_second": 13.86,
      "step": 315000
    },
    {
      "epoch": 3.52,
      "learning_rate": 0.00012470879967489579,
      "loss": 0.2291,
      "step": 315500
    },
    {
      "epoch": 3.52,
      "learning_rate": 0.00012461967928240828,
      "loss": 0.2291,
      "step": 316000
    },
    {
      "epoch": 3.52,
      "eval_loss": 0.22106607258319855,
      "eval_runtime": 2.4653,
      "eval_samples_per_second": 931.733,
      "eval_steps_per_second": 14.603,
      "step": 316000
    },
    {
      "epoch": 3.53,
      "learning_rate": 0.00012453043690142143,
      "loss": 0.2285,
      "step": 316500
    },
    {
      "epoch": 3.54,
      "learning_rate": 0.00012444107277592047,
      "loss": 0.2285,
      "step": 317000
    },
    {
      "epoch": 3.54,
      "eval_loss": 0.21759365499019623,
      "eval_runtime": 2.6088,
      "eval_samples_per_second": 880.492,
      "eval_steps_per_second": 13.8,
      "step": 317000
    },
    {
      "epoch": 3.54,
      "learning_rate": 0.00012435158715022352,
      "loss": 0.2289,
      "step": 317500
    },
    {
      "epoch": 3.55,
      "learning_rate": 0.0001242619802689809,
      "loss": 0.2283,
      "step": 318000
    },
    {
      "epoch": 3.55,
      "eval_loss": 0.21786309778690338,
      "eval_runtime": 2.5783,
      "eval_samples_per_second": 890.904,
      "eval_steps_per_second": 13.963,
      "step": 318000
    },
    {
      "epoch": 3.55,
      "learning_rate": 0.00012417225237717434,
      "loss": 0.2287,
      "step": 318500
    },
    {
      "epoch": 3.56,
      "learning_rate": 0.00012408240372011647,
      "loss": 0.2282,
      "step": 319000
    },
    {
      "epoch": 3.56,
      "eval_loss": 0.2200136035680771,
      "eval_runtime": 2.6445,
      "eval_samples_per_second": 868.601,
      "eval_steps_per_second": 13.613,
      "step": 319000
    },
    {
      "epoch": 3.56,
      "learning_rate": 0.00012399243454345012,
      "loss": 0.2286,
      "step": 319500
    },
    {
      "epoch": 3.57,
      "learning_rate": 0.0001239023450931476,
      "loss": 0.2281,
      "step": 320000
    },
    {
      "epoch": 3.57,
      "eval_loss": 0.21773354709148407,
      "eval_runtime": 2.5181,
      "eval_samples_per_second": 912.179,
      "eval_steps_per_second": 14.296,
      "step": 320000
    },
    {
      "epoch": 3.57,
      "learning_rate": 0.00012381213561550995,
      "loss": 0.228,
      "step": 320500
    },
    {
      "epoch": 3.58,
      "learning_rate": 0.00012372180635716656,
      "loss": 0.2279,
      "step": 321000
    },
    {
      "epoch": 3.58,
      "eval_loss": 0.21474304795265198,
      "eval_runtime": 2.495,
      "eval_samples_per_second": 920.639,
      "eval_steps_per_second": 14.429,
      "step": 321000
    },
    {
      "epoch": 3.59,
      "learning_rate": 0.00012363135756507406,
      "loss": 0.2283,
      "step": 321500
    },
    {
      "epoch": 3.59,
      "learning_rate": 0.00012354078948651604,
      "loss": 0.228,
      "step": 322000
    },
    {
      "epoch": 3.59,
      "eval_loss": 0.21709482371807098,
      "eval_runtime": 2.5983,
      "eval_samples_per_second": 884.027,
      "eval_steps_per_second": 13.855,
      "step": 322000
    },
    {
      "epoch": 3.6,
      "learning_rate": 0.00012345010236910217,
      "loss": 0.2278,
      "step": 322500
    },
    {
      "epoch": 3.6,
      "learning_rate": 0.00012335929646076758,
      "loss": 0.2284,
      "step": 323000
    },
    {
      "epoch": 3.6,
      "eval_loss": 0.21743525564670563,
      "eval_runtime": 2.5879,
      "eval_samples_per_second": 887.578,
      "eval_steps_per_second": 13.911,
      "step": 323000
    },
    {
      "epoch": 3.61,
      "learning_rate": 0.0001232683720097721,
      "loss": 0.2278,
      "step": 323500
    },
    {
      "epoch": 3.61,
      "learning_rate": 0.00012317732926469976,
      "loss": 0.2282,
      "step": 324000
    },
    {
      "epoch": 3.61,
      "eval_loss": 0.2163660079240799,
      "eval_runtime": 2.509,
      "eval_samples_per_second": 915.491,
      "eval_steps_per_second": 14.348,
      "step": 324000
    },
    {
      "epoch": 3.62,
      "learning_rate": 0.00012308616847445794,
      "loss": 0.2278,
      "step": 324500
    },
    {
      "epoch": 3.62,
      "learning_rate": 0.00012299488988827675,
      "loss": 0.2278,
      "step": 325000
    },
    {
      "epoch": 3.62,
      "eval_loss": 0.21636153757572174,
      "eval_runtime": 2.5942,
      "eval_samples_per_second": 885.428,
      "eval_steps_per_second": 13.877,
      "step": 325000
    },
    {
      "epoch": 3.63,
      "learning_rate": 0.00012290349375570836,
      "loss": 0.2275,
      "step": 325500
    },
    {
      "epoch": 3.64,
      "learning_rate": 0.0001228119803266263,
      "loss": 0.2278,
      "step": 326000
    },
    {
      "epoch": 3.64,
      "eval_loss": 0.21803778409957886,
      "eval_runtime": 2.5627,
      "eval_samples_per_second": 896.316,
      "eval_steps_per_second": 14.048,
      "step": 326000
    },
    {
      "epoch": 3.64,
      "learning_rate": 0.0001227203498512248,
      "loss": 0.2276,
      "step": 326500
    },
    {
      "epoch": 3.65,
      "learning_rate": 0.0001226286025800181,
      "loss": 0.2276,
      "step": 327000
    },
    {
      "epoch": 3.65,
      "eval_loss": 0.21890942752361298,
      "eval_runtime": 2.5935,
      "eval_samples_per_second": 885.689,
      "eval_steps_per_second": 13.881,
      "step": 327000
    },
    {
      "epoch": 3.65,
      "learning_rate": 0.00012253673876383967,
      "loss": 0.2271,
      "step": 327500
    },
    {
      "epoch": 3.66,
      "learning_rate": 0.00012244475865384177,
      "loss": 0.2275,
      "step": 328000
    },
    {
      "epoch": 3.66,
      "eval_loss": 0.21888494491577148,
      "eval_runtime": 2.6221,
      "eval_samples_per_second": 876.02,
      "eval_steps_per_second": 13.73,
      "step": 328000
    },
    {
      "epoch": 3.66,
      "learning_rate": 0.00012235266250149444,
      "loss": 0.2273,
      "step": 328500
    },
    {
      "epoch": 3.67,
      "learning_rate": 0.00012226045055858505,
      "loss": 0.228,
      "step": 329000
    },
    {
      "epoch": 3.67,
      "eval_loss": 0.2192571461200714,
      "eval_runtime": 2.5089,
      "eval_samples_per_second": 915.552,
      "eval_steps_per_second": 14.349,
      "step": 329000
    },
    {
      "epoch": 3.67,
      "learning_rate": 0.00012216812307721755,
      "loss": 0.227,
      "step": 329500
    },
    {
      "epoch": 3.68,
      "learning_rate": 0.00012207568030981174,
      "loss": 0.2269,
      "step": 330000
    },
    {
      "epoch": 3.68,
      "eval_loss": 0.21729490160942078,
      "eval_runtime": 2.5927,
      "eval_samples_per_second": 885.965,
      "eval_steps_per_second": 13.885,
      "step": 330000
    },
    {
      "epoch": 3.69,
      "learning_rate": 0.00012198312250910265,
      "loss": 0.2271,
      "step": 330500
    },
    {
      "epoch": 3.69,
      "learning_rate": 0.00012189044992813972,
      "loss": 0.2271,
      "step": 331000
    },
    {
      "epoch": 3.69,
      "eval_loss": 0.21838127076625824,
      "eval_runtime": 2.5783,
      "eval_samples_per_second": 890.899,
      "eval_steps_per_second": 13.963,
      "step": 331000
    },
    {
      "epoch": 3.7,
      "learning_rate": 0.00012179766282028625,
      "loss": 0.2267,
      "step": 331500
    },
    {
      "epoch": 3.7,
      "learning_rate": 0.0001217047614392187,
      "loss": 0.2265,
      "step": 332000
    },
    {
      "epoch": 3.7,
      "eval_loss": 0.2152900993824005,
      "eval_runtime": 2.5995,
      "eval_samples_per_second": 883.629,
      "eval_steps_per_second": 13.849,
      "step": 332000
    },
    {
      "epoch": 3.71,
      "learning_rate": 0.00012161174603892584,
      "loss": 0.2271,
      "step": 332500
    },
    {
      "epoch": 3.71,
      "learning_rate": 0.00012151861687370828,
      "loss": 0.2269,
      "step": 333000
    },
    {
      "epoch": 3.71,
      "eval_loss": 0.21577809751033783,
      "eval_runtime": 2.6271,
      "eval_samples_per_second": 874.351,
      "eval_steps_per_second": 13.703,
      "step": 333000
    },
    {
      "epoch": 3.72,
      "learning_rate": 0.00012142537419817753,
      "loss": 0.2273,
      "step": 333500
    },
    {
      "epoch": 3.73,
      "learning_rate": 0.00012133201826725558,
      "loss": 0.2267,
      "step": 334000
    },
    {
      "epoch": 3.73,
      "eval_loss": 0.21679390966892242,
      "eval_runtime": 3.7147,
      "eval_samples_per_second": 618.35,
      "eval_steps_per_second": 9.691,
      "step": 334000
    },
    {
      "epoch": 3.73,
      "learning_rate": 0.00012123854933617394,
      "loss": 0.2264,
      "step": 334500
    },
    {
      "epoch": 3.74,
      "learning_rate": 0.0001211449676604731,
      "loss": 0.2269,
      "step": 335000
    },
    {
      "epoch": 3.74,
      "eval_loss": 0.21456079185009003,
      "eval_runtime": 2.5549,
      "eval_samples_per_second": 899.059,
      "eval_steps_per_second": 14.091,
      "step": 335000
    },
    {
      "epoch": 3.74,
      "learning_rate": 0.0001210512734960018,
      "loss": 0.2266,
      "step": 335500
    },
    {
      "epoch": 3.75,
      "learning_rate": 0.00012095746709891632,
      "loss": 0.2257,
      "step": 336000
    },
    {
      "epoch": 3.75,
      "eval_loss": 0.21742412447929382,
      "eval_runtime": 2.5117,
      "eval_samples_per_second": 914.52,
      "eval_steps_per_second": 14.333,
      "step": 336000
    },
    {
      "epoch": 3.75,
      "learning_rate": 0.00012086354872567969,
      "loss": 0.2269,
      "step": 336500
    },
    {
      "epoch": 3.76,
      "learning_rate": 0.00012076951863306127,
      "loss": 0.2263,
      "step": 337000
    },
    {
      "epoch": 3.76,
      "eval_loss": 0.21796566247940063,
      "eval_runtime": 2.5462,
      "eval_samples_per_second": 902.138,
      "eval_steps_per_second": 14.139,
      "step": 337000
    },
    {
      "epoch": 3.76,
      "learning_rate": 0.00012067537707813568,
      "loss": 0.226,
      "step": 337500
    },
    {
      "epoch": 3.77,
      "learning_rate": 0.0001205811243182823,
      "loss": 0.2263,
      "step": 338000
    },
    {
      "epoch": 3.77,
      "eval_loss": 0.21637655794620514,
      "eval_runtime": 2.5145,
      "eval_samples_per_second": 913.508,
      "eval_steps_per_second": 14.317,
      "step": 338000
    },
    {
      "epoch": 3.78,
      "learning_rate": 0.00012048676061118467,
      "loss": 0.226,
      "step": 338500
    },
    {
      "epoch": 3.78,
      "learning_rate": 0.00012039228621482949,
      "loss": 0.2262,
      "step": 339000
    },
    {
      "epoch": 3.78,
      "eval_loss": 0.21473273634910583,
      "eval_runtime": 2.5773,
      "eval_samples_per_second": 891.235,
      "eval_steps_per_second": 13.968,
      "step": 339000
    },
    {
      "epoch": 3.79,
      "learning_rate": 0.0001202977013875062,
      "loss": 0.2257,
      "step": 339500
    },
    {
      "epoch": 3.79,
      "learning_rate": 0.00012020300638780604,
      "loss": 0.2259,
      "step": 340000
    },
    {
      "epoch": 3.79,
      "eval_loss": 0.21654388308525085,
      "eval_runtime": 2.5215,
      "eval_samples_per_second": 910.974,
      "eval_steps_per_second": 14.277,
      "step": 340000
    },
    {
      "epoch": 3.8,
      "learning_rate": 0.0001201082014746216,
      "loss": 0.226,
      "step": 340500
    },
    {
      "epoch": 3.8,
      "learning_rate": 0.00012001328690714582,
      "loss": 0.2263,
      "step": 341000
    },
    {
      "epoch": 3.8,
      "eval_loss": 0.2129027545452118,
      "eval_runtime": 2.51,
      "eval_samples_per_second": 915.143,
      "eval_steps_per_second": 14.343,
      "step": 341000
    },
    {
      "epoch": 3.81,
      "learning_rate": 0.00011991826294487155,
      "loss": 0.2259,
      "step": 341500
    },
    {
      "epoch": 3.81,
      "learning_rate": 0.00011982312984759068,
      "loss": 0.2259,
      "step": 342000
    },
    {
      "epoch": 3.81,
      "eval_loss": 0.21433599293231964,
      "eval_runtime": 2.5873,
      "eval_samples_per_second": 887.783,
      "eval_steps_per_second": 13.914,
      "step": 342000
    },
    {
      "epoch": 3.82,
      "learning_rate": 0.00011972788787539345,
      "loss": 0.2254,
      "step": 342500
    },
    {
      "epoch": 3.83,
      "learning_rate": 0.00011963253728866778,
      "loss": 0.2254,
      "step": 343000
    },
    {
      "epoch": 3.83,
      "eval_loss": 0.21325699985027313,
      "eval_runtime": 2.551,
      "eval_samples_per_second": 900.437,
      "eval_steps_per_second": 14.112,
      "step": 343000
    },
    {
      "epoch": 3.83,
      "learning_rate": 0.00011953707834809848,
      "loss": 0.2257,
      "step": 343500
    },
    {
      "epoch": 3.84,
      "learning_rate": 0.00011944151131466675,
      "loss": 0.2259,
      "step": 344000
    },
    {
      "epoch": 3.84,
      "eval_loss": 0.2161191999912262,
      "eval_runtime": 2.6489,
      "eval_samples_per_second": 867.163,
      "eval_steps_per_second": 13.591,
      "step": 344000
    },
    {
      "epoch": 3.84,
      "learning_rate": 0.00011934583644964913,
      "loss": 0.2255,
      "step": 344500
    },
    {
      "epoch": 3.85,
      "learning_rate": 0.00011925005401461709,
      "loss": 0.2256,
      "step": 345000
    },
    {
      "epoch": 3.85,
      "eval_loss": 0.21498842537403107,
      "eval_runtime": 2.5418,
      "eval_samples_per_second": 903.695,
      "eval_steps_per_second": 14.163,
      "step": 345000
    },
    {
      "epoch": 3.85,
      "learning_rate": 0.00011915416427143613,
      "loss": 0.225,
      "step": 345500
    },
    {
      "epoch": 3.86,
      "learning_rate": 0.00011905816748226513,
      "loss": 0.2257,
      "step": 346000
    },
    {
      "epoch": 3.86,
      "eval_loss": 0.21445859968662262,
      "eval_runtime": 2.5612,
      "eval_samples_per_second": 896.853,
      "eval_steps_per_second": 14.056,
      "step": 346000
    },
    {
      "epoch": 3.86,
      "learning_rate": 0.00011896206390955567,
      "loss": 0.225,
      "step": 346500
    },
    {
      "epoch": 3.87,
      "learning_rate": 0.00011886585381605125,
      "loss": 0.2251,
      "step": 347000
    },
    {
      "epoch": 3.87,
      "eval_loss": 0.2150752693414688,
      "eval_runtime": 2.605,
      "eval_samples_per_second": 881.779,
      "eval_steps_per_second": 13.82,
      "step": 347000
    },
    {
      "epoch": 3.88,
      "learning_rate": 0.0001187695374647866,
      "loss": 0.2251,
      "step": 347500
    },
    {
      "epoch": 3.88,
      "learning_rate": 0.00011867311511908693,
      "loss": 0.2249,
      "step": 348000
    },
    {
      "epoch": 3.88,
      "eval_loss": 0.212222158908844,
      "eval_runtime": 2.552,
      "eval_samples_per_second": 900.074,
      "eval_steps_per_second": 14.107,
      "step": 348000
    },
    {
      "epoch": 3.89,
      "learning_rate": 0.00011857658704256721,
      "loss": 0.2249,
      "step": 348500
    },
    {
      "epoch": 3.89,
      "learning_rate": 0.00011847995349913162,
      "loss": 0.2252,
      "step": 349000
    },
    {
      "epoch": 3.89,
      "eval_loss": 0.21469330787658691,
      "eval_runtime": 2.6867,
      "eval_samples_per_second": 854.946,
      "eval_steps_per_second": 13.399,
      "step": 349000
    },
    {
      "epoch": 3.9,
      "learning_rate": 0.00011838321475297247,
      "loss": 0.2243,
      "step": 349500
    },
    {
      "epoch": 3.9,
      "learning_rate": 0.00011828637106856989,
      "loss": 0.2248,
      "step": 350000
    },
    {
      "epoch": 3.9,
      "eval_loss": 0.21383698284626007,
      "eval_runtime": 2.5687,
      "eval_samples_per_second": 894.235,
      "eval_steps_per_second": 14.015,
      "step": 350000
    },
    {
      "epoch": 3.91,
      "learning_rate": 0.00011818942271069073,
      "loss": 0.2252,
      "step": 350500
    },
    {
      "epoch": 3.91,
      "learning_rate": 0.00011809236994438816,
      "loss": 0.2245,
      "step": 351000
    },
    {
      "epoch": 3.91,
      "eval_loss": 0.21465420722961426,
      "eval_runtime": 2.6241,
      "eval_samples_per_second": 875.34,
      "eval_steps_per_second": 13.719,
      "step": 351000
    },
    {
      "epoch": 3.92,
      "learning_rate": 0.0001179952130350007,
      "loss": 0.2245,
      "step": 351500
    },
    {
      "epoch": 3.93,
      "learning_rate": 0.00011789795224815164,
      "loss": 0.2246,
      "step": 352000
    },
    {
      "epoch": 3.93,
      "eval_loss": 0.21523068845272064,
      "eval_runtime": 2.5097,
      "eval_samples_per_second": 915.238,
      "eval_steps_per_second": 14.344,
      "step": 352000
    },
    {
      "epoch": 3.93,
      "learning_rate": 0.00011780058784974831,
      "loss": 0.2247,
      "step": 352500
    },
    {
      "epoch": 3.94,
      "learning_rate": 0.00011770312010598116,
      "loss": 0.2243,
      "step": 353000
    },
    {
      "epoch": 3.94,
      "eval_loss": 0.21631552278995514,
      "eval_runtime": 2.5552,
      "eval_samples_per_second": 898.94,
      "eval_steps_per_second": 14.089,
      "step": 353000
    },
    {
      "epoch": 3.94,
      "learning_rate": 0.00011760554928332333,
      "loss": 0.2242,
      "step": 353500
    },
    {
      "epoch": 3.95,
      "learning_rate": 0.00011750787564852973,
      "loss": 0.2245,
      "step": 354000
    },
    {
      "epoch": 3.95,
      "eval_loss": 0.21260108053684235,
      "eval_runtime": 2.5458,
      "eval_samples_per_second": 902.259,
      "eval_steps_per_second": 14.141,
      "step": 354000
    },
    {
      "epoch": 3.95,
      "learning_rate": 0.00011741009946863639,
      "loss": 0.2242,
      "step": 354500
    },
    {
      "epoch": 3.96,
      "learning_rate": 0.00011731222101095955,
      "loss": 0.2239,
      "step": 355000
    },
    {
      "epoch": 3.96,
      "eval_loss": 0.2130313217639923,
      "eval_runtime": 2.5436,
      "eval_samples_per_second": 903.054,
      "eval_steps_per_second": 14.153,
      "step": 355000
    },
    {
      "epoch": 3.96,
      "learning_rate": 0.00011721424054309525,
      "loss": 0.2236,
      "step": 355500
    },
    {
      "epoch": 3.97,
      "learning_rate": 0.00011711615833291833,
      "loss": 0.2239,
      "step": 356000
    },
    {
      "epoch": 3.97,
      "eval_loss": 0.21391764283180237,
      "eval_runtime": 2.5894,
      "eval_samples_per_second": 887.095,
      "eval_steps_per_second": 13.903,
      "step": 356000
    },
    {
      "epoch": 3.98,
      "learning_rate": 0.00011701797464858178,
      "loss": 0.2243,
      "step": 356500
    },
    {
      "epoch": 3.98,
      "learning_rate": 0.0001169196897585161,
      "loss": 0.2241,
      "step": 357000
    },
    {
      "epoch": 3.98,
      "eval_loss": 0.2131282538175583,
      "eval_runtime": 2.6054,
      "eval_samples_per_second": 881.617,
      "eval_steps_per_second": 13.817,
      "step": 357000
    },
    {
      "epoch": 3.99,
      "learning_rate": 0.00011682130393142838,
      "loss": 0.2236,
      "step": 357500
    },
    {
      "epoch": 3.99,
      "learning_rate": 0.00011672281743630175,
      "loss": 0.2239,
      "step": 358000
    },
    {
      "epoch": 3.99,
      "eval_loss": 0.21421436965465546,
      "eval_runtime": 2.6143,
      "eval_samples_per_second": 878.642,
      "eval_steps_per_second": 13.771,
      "step": 358000
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00011662423054239445,
      "loss": 0.2254,
      "step": 358500
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0001165255435192394,
      "loss": 0.2237,
      "step": 359000
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.21473369002342224,
      "eval_runtime": 2.5965,
      "eval_samples_per_second": 884.643,
      "eval_steps_per_second": 13.865,
      "step": 359000
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00011642675663664308,
      "loss": 0.2234,
      "step": 359500
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00011632787016468506,
      "loss": 0.2232,
      "step": 360000
    },
    {
      "epoch": 4.02,
      "eval_loss": 0.21288304030895233,
      "eval_runtime": 2.644,
      "eval_samples_per_second": 868.758,
      "eval_steps_per_second": 13.616,
      "step": 360000
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00011622888437371719,
      "loss": 0.2232,
      "step": 360500
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0001161297995343628,
      "loss": 0.2234,
      "step": 361000
    },
    {
      "epoch": 4.03,
      "eval_loss": 0.21240203082561493,
      "eval_runtime": 2.5699,
      "eval_samples_per_second": 893.801,
      "eval_steps_per_second": 14.008,
      "step": 361000
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00011603061591751615,
      "loss": 0.2238,
      "step": 361500
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00011593133379434138,
      "loss": 0.2238,
      "step": 362000
    },
    {
      "epoch": 4.04,
      "eval_loss": 0.2136167287826538,
      "eval_runtime": 2.4904,
      "eval_samples_per_second": 922.34,
      "eval_steps_per_second": 14.455,
      "step": 362000
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00011583195343627207,
      "loss": 0.2234,
      "step": 362500
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00011573247511501028,
      "loss": 0.2235,
      "step": 363000
    },
    {
      "epoch": 4.05,
      "eval_loss": 0.2138611227273941,
      "eval_runtime": 2.4582,
      "eval_samples_per_second": 934.423,
      "eval_steps_per_second": 14.645,
      "step": 363000
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00011563289910252599,
      "loss": 0.2229,
      "step": 363500
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00011553322567105619,
      "loss": 0.2229,
      "step": 364000
    },
    {
      "epoch": 4.06,
      "eval_loss": 0.21355026960372925,
      "eval_runtime": 2.5444,
      "eval_samples_per_second": 902.768,
      "eval_steps_per_second": 14.149,
      "step": 364000
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00011543345509310421,
      "loss": 0.223,
      "step": 364500
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00011533358764143905,
      "loss": 0.2225,
      "step": 365000
    },
    {
      "epoch": 4.07,
      "eval_loss": 0.21412673592567444,
      "eval_runtime": 2.6104,
      "eval_samples_per_second": 879.928,
      "eval_steps_per_second": 13.791,
      "step": 365000
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00011523362358909449,
      "loss": 0.223,
      "step": 365500
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00011513356320936841,
      "loss": 0.2231,
      "step": 366000
    },
    {
      "epoch": 4.08,
      "eval_loss": 0.21194761991500854,
      "eval_runtime": 2.4955,
      "eval_samples_per_second": 920.443,
      "eval_steps_per_second": 14.426,
      "step": 366000
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00011503340677582213,
      "loss": 0.223,
      "step": 366500
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00011493315456227943,
      "loss": 0.2231,
      "step": 367000
    },
    {
      "epoch": 4.09,
      "eval_loss": 0.21338239312171936,
      "eval_runtime": 2.5254,
      "eval_samples_per_second": 909.575,
      "eval_steps_per_second": 14.255,
      "step": 367000
    },
    {
      "epoch": 4.1,
      "learning_rate": 0.00011483280684282611,
      "loss": 0.2226,
      "step": 367500
    },
    {
      "epoch": 4.1,
      "learning_rate": 0.00011473236389180894,
      "loss": 0.2228,
      "step": 368000
    },
    {
      "epoch": 4.1,
      "eval_loss": 0.21372437477111816,
      "eval_runtime": 2.6109,
      "eval_samples_per_second": 879.784,
      "eval_steps_per_second": 13.789,
      "step": 368000
    },
    {
      "epoch": 4.11,
      "learning_rate": 0.00011463182598383516,
      "loss": 0.2226,
      "step": 368500
    },
    {
      "epoch": 4.12,
      "learning_rate": 0.00011453119339377154,
      "loss": 0.2226,
      "step": 369000
    },
    {
      "epoch": 4.12,
      "eval_loss": 0.2124049812555313,
      "eval_runtime": 2.4867,
      "eval_samples_per_second": 923.72,
      "eval_steps_per_second": 14.477,
      "step": 369000
    },
    {
      "epoch": 4.12,
      "learning_rate": 0.00011443046639674375,
      "loss": 0.2224,
      "step": 369500
    },
    {
      "epoch": 4.13,
      "learning_rate": 0.00011432964526813558,
      "loss": 0.2229,
      "step": 370000
    },
    {
      "epoch": 4.13,
      "eval_loss": 0.21308059990406036,
      "eval_runtime": 2.6387,
      "eval_samples_per_second": 870.49,
      "eval_steps_per_second": 13.643,
      "step": 370000
    },
    {
      "epoch": 4.13,
      "learning_rate": 0.00011422873028358807,
      "loss": 0.2223,
      "step": 370500
    },
    {
      "epoch": 4.14,
      "learning_rate": 0.00011412772171899904,
      "loss": 0.2224,
      "step": 371000
    },
    {
      "epoch": 4.14,
      "eval_loss": 0.212064728140831,
      "eval_runtime": 2.5187,
      "eval_samples_per_second": 911.981,
      "eval_steps_per_second": 14.293,
      "step": 371000
    },
    {
      "epoch": 4.14,
      "learning_rate": 0.00011402661985052197,
      "loss": 0.222,
      "step": 371500
    },
    {
      "epoch": 4.15,
      "learning_rate": 0.00011392542495456556,
      "loss": 0.222,
      "step": 372000
    },
    {
      "epoch": 4.15,
      "eval_loss": 0.21111257374286652,
      "eval_runtime": 2.5956,
      "eval_samples_per_second": 884.946,
      "eval_steps_per_second": 13.869,
      "step": 372000
    },
    {
      "epoch": 4.15,
      "learning_rate": 0.00011382413730779273,
      "loss": 0.222,
      "step": 372500
    },
    {
      "epoch": 4.16,
      "learning_rate": 0.00011372275718712006,
      "loss": 0.2223,
      "step": 373000
    },
    {
      "epoch": 4.16,
      "eval_loss": 0.21289856731891632,
      "eval_runtime": 2.6026,
      "eval_samples_per_second": 882.577,
      "eval_steps_per_second": 13.832,
      "step": 373000
    },
    {
      "epoch": 4.17,
      "learning_rate": 0.00011362128486971696,
      "loss": 0.2224,
      "step": 373500
    },
    {
      "epoch": 4.17,
      "learning_rate": 0.00011351972063300484,
      "loss": 0.2222,
      "step": 374000
    },
    {
      "epoch": 4.17,
      "eval_loss": 0.2087879329919815,
      "eval_runtime": 2.5302,
      "eval_samples_per_second": 907.82,
      "eval_steps_per_second": 14.228,
      "step": 374000
    },
    {
      "epoch": 4.18,
      "learning_rate": 0.0001134180647546565,
      "loss": 0.2218,
      "step": 374500
    },
    {
      "epoch": 4.18,
      "learning_rate": 0.00011331631751259515,
      "loss": 0.222,
      "step": 375000
    },
    {
      "epoch": 4.18,
      "eval_loss": 0.21286322176456451,
      "eval_runtime": 2.6262,
      "eval_samples_per_second": 874.641,
      "eval_steps_per_second": 13.708,
      "step": 375000
    },
    {
      "epoch": 4.19,
      "learning_rate": 0.00011321447918499391,
      "loss": 0.2218,
      "step": 375500
    },
    {
      "epoch": 4.19,
      "learning_rate": 0.00011311255005027487,
      "loss": 0.222,
      "step": 376000
    },
    {
      "epoch": 4.19,
      "eval_loss": 0.210297092795372,
      "eval_runtime": 2.5937,
      "eval_samples_per_second": 885.619,
      "eval_steps_per_second": 13.88,
      "step": 376000
    },
    {
      "epoch": 4.2,
      "learning_rate": 0.00011301053038710837,
      "loss": 0.2218,
      "step": 376500
    },
    {
      "epoch": 4.2,
      "learning_rate": 0.00011290842047441232,
      "loss": 0.222,
      "step": 377000
    },
    {
      "epoch": 4.2,
      "eval_loss": 0.21230094134807587,
      "eval_runtime": 2.5876,
      "eval_samples_per_second": 887.679,
      "eval_steps_per_second": 13.912,
      "step": 377000
    },
    {
      "epoch": 4.21,
      "learning_rate": 0.0001128062205913513,
      "loss": 0.2217,
      "step": 377500
    },
    {
      "epoch": 4.22,
      "learning_rate": 0.00011270393101733585,
      "loss": 0.222,
      "step": 378000
    },
    {
      "epoch": 4.22,
      "eval_loss": 0.2095794677734375,
      "eval_runtime": 2.6193,
      "eval_samples_per_second": 876.967,
      "eval_steps_per_second": 13.744,
      "step": 378000
    },
    {
      "epoch": 4.22,
      "learning_rate": 0.00011260155203202183,
      "loss": 0.2215,
      "step": 378500
    },
    {
      "epoch": 4.23,
      "learning_rate": 0.00011249908391530946,
      "loss": 0.2214,
      "step": 379000
    },
    {
      "epoch": 4.23,
      "eval_loss": 0.21008515357971191,
      "eval_runtime": 2.6054,
      "eval_samples_per_second": 881.646,
      "eval_steps_per_second": 13.818,
      "step": 379000
    },
    {
      "epoch": 4.23,
      "learning_rate": 0.0001123965269473426,
      "loss": 0.2217,
      "step": 379500
    },
    {
      "epoch": 4.24,
      "learning_rate": 0.00011229388140850814,
      "loss": 0.2215,
      "step": 380000
    },
    {
      "epoch": 4.24,
      "eval_loss": 0.21324031054973602,
      "eval_runtime": 2.6426,
      "eval_samples_per_second": 869.205,
      "eval_steps_per_second": 13.623,
      "step": 380000
    },
    {
      "epoch": 4.24,
      "learning_rate": 0.00011219114757943505,
      "loss": 0.2215,
      "step": 380500
    },
    {
      "epoch": 4.25,
      "learning_rate": 0.00011208832574099368,
      "loss": 0.2215,
      "step": 381000
    },
    {
      "epoch": 4.25,
      "eval_loss": 0.20935191214084625,
      "eval_runtime": 2.5516,
      "eval_samples_per_second": 900.235,
      "eval_steps_per_second": 14.109,
      "step": 381000
    },
    {
      "epoch": 4.25,
      "learning_rate": 0.00011198541617429504,
      "loss": 0.2213,
      "step": 381500
    },
    {
      "epoch": 4.26,
      "learning_rate": 0.00011188241916068993,
      "loss": 0.2211,
      "step": 382000
    },
    {
      "epoch": 4.26,
      "eval_loss": 0.21030767261981964,
      "eval_runtime": 2.5649,
      "eval_samples_per_second": 895.555,
      "eval_steps_per_second": 14.036,
      "step": 382000
    },
    {
      "epoch": 4.27,
      "learning_rate": 0.00011177933498176826,
      "loss": 0.221,
      "step": 382500
    },
    {
      "epoch": 4.27,
      "learning_rate": 0.00011167616391935826,
      "loss": 0.221,
      "step": 383000
    },
    {
      "epoch": 4.27,
      "eval_loss": 0.21177974343299866,
      "eval_runtime": 2.5071,
      "eval_samples_per_second": 916.184,
      "eval_steps_per_second": 14.359,
      "step": 383000
    },
    {
      "epoch": 4.28,
      "learning_rate": 0.00011157290625552563,
      "loss": 0.2206,
      "step": 383500
    },
    {
      "epoch": 4.28,
      "learning_rate": 0.00011146956227257293,
      "loss": 0.2209,
      "step": 384000
    },
    {
      "epoch": 4.28,
      "eval_loss": 0.2104860246181488,
      "eval_runtime": 2.5213,
      "eval_samples_per_second": 911.034,
      "eval_steps_per_second": 14.278,
      "step": 384000
    },
    {
      "epoch": 4.29,
      "learning_rate": 0.0001113661322530386,
      "loss": 0.2204,
      "step": 384500
    },
    {
      "epoch": 4.29,
      "learning_rate": 0.00011126261647969645,
      "loss": 0.221,
      "step": 385000
    },
    {
      "epoch": 4.29,
      "eval_loss": 0.20968575775623322,
      "eval_runtime": 2.69,
      "eval_samples_per_second": 853.888,
      "eval_steps_per_second": 13.383,
      "step": 385000
    },
    {
      "epoch": 4.3,
      "learning_rate": 0.00011115901523555457,
      "loss": 0.2206,
      "step": 385500
    },
    {
      "epoch": 4.31,
      "learning_rate": 0.00011105532880385487,
      "loss": 0.2207,
      "step": 386000
    },
    {
      "epoch": 4.31,
      "eval_loss": 0.21161584556102753,
      "eval_runtime": 2.6386,
      "eval_samples_per_second": 870.537,
      "eval_steps_per_second": 13.644,
      "step": 386000
    },
    {
      "epoch": 4.31,
      "learning_rate": 0.00011095155746807206,
      "loss": 0.2211,
      "step": 386500
    },
    {
      "epoch": 4.32,
      "learning_rate": 0.00011084770151191299,
      "loss": 0.2207,
      "step": 387000
    },
    {
      "epoch": 4.32,
      "eval_loss": 0.20934854447841644,
      "eval_runtime": 2.6571,
      "eval_samples_per_second": 864.463,
      "eval_steps_per_second": 13.548,
      "step": 387000
    },
    {
      "epoch": 4.32,
      "learning_rate": 0.00011074376121931591,
      "loss": 0.2204,
      "step": 387500
    },
    {
      "epoch": 4.33,
      "learning_rate": 0.00011063973687444962,
      "loss": 0.2205,
      "step": 388000
    },
    {
      "epoch": 4.33,
      "eval_loss": 0.20918506383895874,
      "eval_runtime": 2.6259,
      "eval_samples_per_second": 874.731,
      "eval_steps_per_second": 13.709,
      "step": 388000
    },
    {
      "epoch": 4.33,
      "learning_rate": 0.00011053562876171268,
      "loss": 0.2204,
      "step": 388500
    },
    {
      "epoch": 4.34,
      "learning_rate": 0.00011043143716573272,
      "loss": 0.2204,
      "step": 389000
    },
    {
      "epoch": 4.34,
      "eval_loss": 0.2104416787624359,
      "eval_runtime": 2.543,
      "eval_samples_per_second": 903.26,
      "eval_steps_per_second": 14.156,
      "step": 389000
    },
    {
      "epoch": 4.34,
      "learning_rate": 0.00011032716237136557,
      "loss": 0.2206,
      "step": 389500
    },
    {
      "epoch": 4.35,
      "learning_rate": 0.00011022280466369448,
      "loss": 0.2204,
      "step": 390000
    },
    {
      "epoch": 4.35,
      "eval_loss": 0.20674540102481842,
      "eval_runtime": 3.9957,
      "eval_samples_per_second": 574.865,
      "eval_steps_per_second": 9.01,
      "step": 390000
    },
    {
      "epoch": 4.36,
      "learning_rate": 0.00011011836432802956,
      "loss": 0.2203,
      "step": 390500
    },
    {
      "epoch": 4.36,
      "learning_rate": 0.00011001384164990662,
      "loss": 0.2204,
      "step": 391000
    },
    {
      "epoch": 4.36,
      "eval_loss": 0.20982348918914795,
      "eval_runtime": 2.6878,
      "eval_samples_per_second": 854.612,
      "eval_steps_per_second": 13.394,
      "step": 391000
    },
    {
      "epoch": 4.37,
      "learning_rate": 0.00010990923691508666,
      "loss": 0.2206,
      "step": 391500
    },
    {
      "epoch": 4.37,
      "learning_rate": 0.00010980455040955506,
      "loss": 0.2201,
      "step": 392000
    },
    {
      "epoch": 4.37,
      "eval_loss": 0.20970028638839722,
      "eval_runtime": 2.66,
      "eval_samples_per_second": 863.53,
      "eval_steps_per_second": 13.534,
      "step": 392000
    },
    {
      "epoch": 4.38,
      "learning_rate": 0.00010969978241952076,
      "loss": 0.2199,
      "step": 392500
    },
    {
      "epoch": 4.38,
      "learning_rate": 0.00010959493323141538,
      "loss": 0.2198,
      "step": 393000
    },
    {
      "epoch": 4.38,
      "eval_loss": 0.2100575715303421,
      "eval_runtime": 2.6711,
      "eval_samples_per_second": 859.949,
      "eval_steps_per_second": 13.478,
      "step": 393000
    },
    {
      "epoch": 4.39,
      "learning_rate": 0.00010949000313189264,
      "loss": 0.2201,
      "step": 393500
    },
    {
      "epoch": 4.39,
      "learning_rate": 0.00010938499240782739,
      "loss": 0.2197,
      "step": 394000
    },
    {
      "epoch": 4.39,
      "eval_loss": 0.21025221049785614,
      "eval_runtime": 2.6759,
      "eval_samples_per_second": 858.394,
      "eval_steps_per_second": 13.453,
      "step": 394000
    },
    {
      "epoch": 4.4,
      "learning_rate": 0.00010927990134631496,
      "loss": 0.2201,
      "step": 394500
    },
    {
      "epoch": 4.41,
      "learning_rate": 0.00010917473023467032,
      "loss": 0.2197,
      "step": 395000
    },
    {
      "epoch": 4.41,
      "eval_loss": 0.21049723029136658,
      "eval_runtime": 2.4803,
      "eval_samples_per_second": 926.101,
      "eval_steps_per_second": 14.514,
      "step": 395000
    },
    {
      "epoch": 4.41,
      "learning_rate": 0.00010906947936042724,
      "loss": 0.2192,
      "step": 395500
    },
    {
      "epoch": 4.42,
      "learning_rate": 0.00010896414901133761,
      "loss": 0.2196,
      "step": 396000
    },
    {
      "epoch": 4.42,
      "eval_loss": 0.2069980949163437,
      "eval_runtime": 2.5039,
      "eval_samples_per_second": 917.356,
      "eval_steps_per_second": 14.377,
      "step": 396000
    },
    {
      "epoch": 4.42,
      "learning_rate": 0.00010885873947537058,
      "loss": 0.2197,
      "step": 396500
    },
    {
      "epoch": 4.43,
      "learning_rate": 0.00010875325104071177,
      "loss": 0.22,
      "step": 397000
    },
    {
      "epoch": 4.43,
      "eval_loss": 0.21147653460502625,
      "eval_runtime": 2.5638,
      "eval_samples_per_second": 895.951,
      "eval_steps_per_second": 14.042,
      "step": 397000
    },
    {
      "epoch": 4.43,
      "learning_rate": 0.00010864768399576257,
      "loss": 0.2196,
      "step": 397500
    },
    {
      "epoch": 4.44,
      "learning_rate": 0.00010854203862913927,
      "loss": 0.2194,
      "step": 398000
    },
    {
      "epoch": 4.44,
      "eval_loss": 0.21261727809906006,
      "eval_runtime": 2.567,
      "eval_samples_per_second": 894.808,
      "eval_steps_per_second": 14.024,
      "step": 398000
    },
    {
      "epoch": 4.44,
      "learning_rate": 0.00010843631522967218,
      "loss": 0.2193,
      "step": 398500
    },
    {
      "epoch": 4.45,
      "learning_rate": 0.00010833051408640509,
      "loss": 0.2193,
      "step": 399000
    },
    {
      "epoch": 4.45,
      "eval_loss": 0.2084684520959854,
      "eval_runtime": 2.5323,
      "eval_samples_per_second": 907.094,
      "eval_steps_per_second": 14.217,
      "step": 399000
    },
    {
      "epoch": 4.46,
      "learning_rate": 0.0001082246354885943,
      "loss": 0.2187,
      "step": 399500
    },
    {
      "epoch": 4.46,
      "learning_rate": 0.00010811867972570786,
      "loss": 0.2195,
      "step": 400000
    },
    {
      "epoch": 4.46,
      "eval_loss": 0.20896346867084503,
      "eval_runtime": 2.5912,
      "eval_samples_per_second": 886.47,
      "eval_steps_per_second": 13.893,
      "step": 400000
    },
    {
      "epoch": 4.47,
      "learning_rate": 0.00010801264708742474,
      "loss": 0.219,
      "step": 400500
    },
    {
      "epoch": 4.47,
      "learning_rate": 0.00010790653786363416,
      "loss": 0.6229,
      "step": 401000
    },
    {
      "epoch": 4.47,
      "eval_loss": 0.6781117916107178,
      "eval_runtime": 2.5271,
      "eval_samples_per_second": 908.933,
      "eval_steps_per_second": 14.245,
      "step": 401000
    },
    {
      "epoch": 4.48,
      "learning_rate": 0.00010780035234443463,
      "loss": 0.6782,
      "step": 401500
    },
    {
      "epoch": 4.48,
      "learning_rate": 0.00010769409082013337,
      "loss": 0.6775,
      "step": 402000
    },
    {
      "epoch": 4.48,
      "eval_loss": 0.6771849989891052,
      "eval_runtime": 2.5891,
      "eval_samples_per_second": 887.183,
      "eval_steps_per_second": 13.904,
      "step": 402000
    },
    {
      "epoch": 4.49,
      "learning_rate": 0.00010758775358124532,
      "loss": 0.6774,
      "step": 402500
    },
    {
      "epoch": 4.49,
      "learning_rate": 0.00010748134091849238,
      "loss": 0.6773,
      "step": 403000
    },
    {
      "epoch": 4.49,
      "eval_loss": 0.6773160099983215,
      "eval_runtime": 2.6017,
      "eval_samples_per_second": 882.892,
      "eval_steps_per_second": 13.837,
      "step": 403000
    },
    {
      "epoch": 4.5,
      "learning_rate": 0.00010737485312280277,
      "loss": 0.6773,
      "step": 403500
    },
    {
      "epoch": 4.51,
      "learning_rate": 0.00010726829048531,
      "loss": 0.6773,
      "step": 404000
    },
    {
      "epoch": 4.51,
      "eval_loss": 0.6771775484085083,
      "eval_runtime": 2.5177,
      "eval_samples_per_second": 912.323,
      "eval_steps_per_second": 14.298,
      "step": 404000
    },
    {
      "epoch": 4.51,
      "learning_rate": 0.00010716165329735229,
      "loss": 0.6773,
      "step": 404500
    },
    {
      "epoch": 4.52,
      "learning_rate": 0.00010705494185047165,
      "loss": 0.6773,
      "step": 405000
    },
    {
      "epoch": 4.52,
      "eval_loss": 0.6771337389945984,
      "eval_runtime": 2.4415,
      "eval_samples_per_second": 940.818,
      "eval_steps_per_second": 14.745,
      "step": 405000
    },
    {
      "epoch": 4.52,
      "learning_rate": 0.00010694815643641308,
      "loss": 0.6773,
      "step": 405500
    },
    {
      "epoch": 4.53,
      "learning_rate": 0.0001068412973471238,
      "loss": 0.6774,
      "step": 406000
    },
    {
      "epoch": 4.53,
      "eval_loss": 0.6766112446784973,
      "eval_runtime": 2.5325,
      "eval_samples_per_second": 906.994,
      "eval_steps_per_second": 14.215,
      "step": 406000
    },
    {
      "epoch": 4.53,
      "learning_rate": 0.00010673436487475252,
      "loss": 0.6771,
      "step": 406500
    },
    {
      "epoch": 4.54,
      "learning_rate": 0.00010662735931164853,
      "loss": 0.6774,
      "step": 407000
    },
    {
      "epoch": 4.54,
      "eval_loss": 0.6774923801422119,
      "eval_runtime": 2.5078,
      "eval_samples_per_second": 915.958,
      "eval_steps_per_second": 14.355,
      "step": 407000
    },
    {
      "epoch": 4.54,
      "learning_rate": 0.00010652028095036092,
      "loss": 0.6773,
      "step": 407500
    },
    {
      "epoch": 4.55,
      "learning_rate": 0.0001064131300836379,
      "loss": 0.6774,
      "step": 408000
    },
    {
      "epoch": 4.55,
      "eval_loss": 0.6766784191131592,
      "eval_runtime": 2.6482,
      "eval_samples_per_second": 867.379,
      "eval_steps_per_second": 13.594,
      "step": 408000
    },
    {
      "epoch": 4.56,
      "learning_rate": 0.0001063059070044258,
      "loss": 0.6772,
      "step": 408500
    },
    {
      "epoch": 4.56,
      "learning_rate": 0.0001061986120058684,
      "loss": 0.6713,
      "step": 409000
    },
    {
      "epoch": 4.56,
      "eval_loss": 0.6385542154312134,
      "eval_runtime": 2.5784,
      "eval_samples_per_second": 890.853,
      "eval_steps_per_second": 13.962,
      "step": 409000
    },
    {
      "epoch": 4.57,
      "learning_rate": 0.00010609124538130623,
      "loss": 0.4505,
      "step": 409500
    },
    {
      "epoch": 4.57,
      "learning_rate": 0.00010598380742427543,
      "loss": 0.2349,
      "step": 410000
    },
    {
      "epoch": 4.57,
      "eval_loss": 0.22073650360107422,
      "eval_runtime": 2.623,
      "eval_samples_per_second": 875.727,
      "eval_steps_per_second": 13.725,
      "step": 410000
    },
    {
      "epoch": 4.58,
      "learning_rate": 0.00010587629842850737,
      "loss": 0.2291,
      "step": 410500
    },
    {
      "epoch": 4.58,
      "learning_rate": 0.00010576871868792746,
      "loss": 0.2266,
      "step": 411000
    },
    {
      "epoch": 4.58,
      "eval_loss": 0.21682733297348022,
      "eval_runtime": 2.6588,
      "eval_samples_per_second": 863.931,
      "eval_steps_per_second": 13.54,
      "step": 411000
    },
    {
      "epoch": 4.59,
      "learning_rate": 0.00010566106849665463,
      "loss": 0.2249,
      "step": 411500
    },
    {
      "epoch": 4.59,
      "learning_rate": 0.0001055533481490004,
      "loss": 0.2234,
      "step": 412000
    },
    {
      "epoch": 4.59,
      "eval_loss": 0.2134321630001068,
      "eval_runtime": 2.6016,
      "eval_samples_per_second": 882.919,
      "eval_steps_per_second": 13.838,
      "step": 412000
    },
    {
      "epoch": 4.6,
      "learning_rate": 0.00010544555793946805,
      "loss": 0.2232,
      "step": 412500
    },
    {
      "epoch": 4.61,
      "learning_rate": 0.000105337698162752,
      "loss": 0.222,
      "step": 413000
    },
    {
      "epoch": 4.61,
      "eval_loss": 0.21415403485298157,
      "eval_runtime": 2.6959,
      "eval_samples_per_second": 852.048,
      "eval_steps_per_second": 13.354,
      "step": 413000
    },
    {
      "epoch": 4.61,
      "learning_rate": 0.00010522976911373667,
      "loss": 0.2219,
      "step": 413500
    },
    {
      "epoch": 4.62,
      "learning_rate": 0.00010512177108749594,
      "loss": 0.2217,
      "step": 414000
    },
    {
      "epoch": 4.62,
      "eval_loss": 0.21083350479602814,
      "eval_runtime": 2.6458,
      "eval_samples_per_second": 868.155,
      "eval_steps_per_second": 13.606,
      "step": 414000
    },
    {
      "epoch": 4.62,
      "learning_rate": 0.00010501370437929234,
      "loss": 0.2211,
      "step": 414500
    },
    {
      "epoch": 4.63,
      "learning_rate": 0.00010490556928457616,
      "loss": 0.2215,
      "step": 415000
    },
    {
      "epoch": 4.63,
      "eval_loss": 0.21060839295387268,
      "eval_runtime": 2.64,
      "eval_samples_per_second": 870.073,
      "eval_steps_per_second": 13.636,
      "step": 415000
    },
    {
      "epoch": 4.63,
      "learning_rate": 0.00010479736609898454,
      "loss": 0.2206,
      "step": 415500
    },
    {
      "epoch": 4.64,
      "learning_rate": 0.00010468909511834088,
      "loss": 0.2209,
      "step": 416000
    },
    {
      "epoch": 4.64,
      "eval_loss": 0.20775271952152252,
      "eval_runtime": 2.6856,
      "eval_samples_per_second": 855.309,
      "eval_steps_per_second": 13.405,
      "step": 416000
    },
    {
      "epoch": 4.65,
      "learning_rate": 0.00010458075663865392,
      "loss": 0.2205,
      "step": 416500
    },
    {
      "epoch": 4.65,
      "learning_rate": 0.00010447235095611692,
      "loss": 0.2201,
      "step": 417000
    },
    {
      "epoch": 4.65,
      "eval_loss": 0.20872974395751953,
      "eval_runtime": 2.6782,
      "eval_samples_per_second": 857.663,
      "eval_steps_per_second": 13.442,
      "step": 417000
    },
    {
      "epoch": 4.66,
      "learning_rate": 0.0001043638783671069,
      "loss": 0.2206,
      "step": 417500
    },
    {
      "epoch": 4.66,
      "learning_rate": 0.00010425533916818376,
      "loss": 0.2199,
      "step": 418000
    },
    {
      "epoch": 4.66,
      "eval_loss": 0.20818735659122467,
      "eval_runtime": 2.6703,
      "eval_samples_per_second": 860.216,
      "eval_steps_per_second": 13.482,
      "step": 418000
    },
    {
      "epoch": 4.67,
      "learning_rate": 0.0001041467336560895,
      "loss": 0.22,
      "step": 418500
    },
    {
      "epoch": 4.67,
      "learning_rate": 0.00010403806212774747,
      "loss": 0.2197,
      "step": 419000
    },
    {
      "epoch": 4.67,
      "eval_loss": 0.20917142927646637,
      "eval_runtime": 2.6142,
      "eval_samples_per_second": 878.667,
      "eval_steps_per_second": 13.771,
      "step": 419000
    },
    {
      "epoch": 4.68,
      "learning_rate": 0.00010392932488026147,
      "loss": 0.2194,
      "step": 419500
    },
    {
      "epoch": 4.68,
      "learning_rate": 0.000103820522210915,
      "loss": 0.2196,
      "step": 420000
    },
    {
      "epoch": 4.68,
      "eval_loss": 0.20849910378456116,
      "eval_runtime": 2.6737,
      "eval_samples_per_second": 859.121,
      "eval_steps_per_second": 13.465,
      "step": 420000
    },
    {
      "epoch": 4.69,
      "learning_rate": 0.00010371165441717041,
      "loss": 0.2194,
      "step": 420500
    },
    {
      "epoch": 4.7,
      "learning_rate": 0.00010360272179666802,
      "loss": 0.2192,
      "step": 421000
    },
    {
      "epoch": 4.7,
      "eval_loss": 0.2099054604768753,
      "eval_runtime": 2.6463,
      "eval_samples_per_second": 868.003,
      "eval_steps_per_second": 13.604,
      "step": 421000
    },
    {
      "epoch": 4.7,
      "learning_rate": 0.00010349372464722555,
      "loss": 0.2192,
      "step": 421500
    },
    {
      "epoch": 4.71,
      "learning_rate": 0.00010338466326683697,
      "loss": 0.2192,
      "step": 422000
    },
    {
      "epoch": 4.71,
      "eval_loss": 0.20990021526813507,
      "eval_runtime": 2.699,
      "eval_samples_per_second": 851.063,
      "eval_steps_per_second": 13.338,
      "step": 422000
    },
    {
      "epoch": 4.71,
      "learning_rate": 0.00010327553795367197,
      "loss": 0.2191,
      "step": 422500
    },
    {
      "epoch": 4.72,
      "learning_rate": 0.00010316634900607497,
      "loss": 0.2187,
      "step": 423000
    },
    {
      "epoch": 4.72,
      "eval_loss": 0.21041807532310486,
      "eval_runtime": 2.6813,
      "eval_samples_per_second": 856.689,
      "eval_steps_per_second": 13.427,
      "step": 423000
    },
    {
      "epoch": 4.72,
      "learning_rate": 0.0001030570967225644,
      "loss": 0.219,
      "step": 423500
    },
    {
      "epoch": 4.73,
      "learning_rate": 0.00010294778140183182,
      "loss": 0.2188,
      "step": 424000
    },
    {
      "epoch": 4.73,
      "eval_loss": 0.20911771059036255,
      "eval_runtime": 2.7061,
      "eval_samples_per_second": 848.811,
      "eval_steps_per_second": 13.303,
      "step": 424000
    },
    {
      "epoch": 4.73,
      "learning_rate": 0.00010283840334274117,
      "loss": 0.2189,
      "step": 424500
    },
    {
      "epoch": 4.74,
      "learning_rate": 0.00010272896284432785,
      "loss": 0.2187,
      "step": 425000
    },
    {
      "epoch": 4.74,
      "eval_loss": 0.20740678906440735,
      "eval_runtime": 2.6921,
      "eval_samples_per_second": 853.226,
      "eval_steps_per_second": 13.372,
      "step": 425000
    },
    {
      "epoch": 4.75,
      "learning_rate": 0.00010261946020579799,
      "loss": 0.2188,
      "step": 425500
    },
    {
      "epoch": 4.75,
      "learning_rate": 0.00010250989572652766,
      "loss": 0.2182,
      "step": 426000
    },
    {
      "epoch": 4.75,
      "eval_loss": 0.20986562967300415,
      "eval_runtime": 2.6422,
      "eval_samples_per_second": 869.349,
      "eval_steps_per_second": 13.625,
      "step": 426000
    },
    {
      "epoch": 4.76,
      "learning_rate": 0.00010240026970606198,
      "loss": 0.2189,
      "step": 426500
    },
    {
      "epoch": 4.76,
      "learning_rate": 0.00010229058244411427,
      "loss": 0.2179,
      "step": 427000
    },
    {
      "epoch": 4.76,
      "eval_loss": 0.20823934674263,
      "eval_runtime": 2.6772,
      "eval_samples_per_second": 857.994,
      "eval_steps_per_second": 13.447,
      "step": 427000
    },
    {
      "epoch": 4.77,
      "learning_rate": 0.0001021808342405653,
      "loss": 0.2183,
      "step": 427500
    },
    {
      "epoch": 4.77,
      "learning_rate": 0.00010207102539546251,
      "loss": 0.2183,
      "step": 428000
    },
    {
      "epoch": 4.77,
      "eval_loss": 0.2062913179397583,
      "eval_runtime": 2.6582,
      "eval_samples_per_second": 864.108,
      "eval_steps_per_second": 13.543,
      "step": 428000
    },
    {
      "epoch": 4.78,
      "learning_rate": 0.00010196115620901904,
      "loss": 0.2184,
      "step": 428500
    },
    {
      "epoch": 4.78,
      "learning_rate": 0.00010185122698161311,
      "loss": 0.2178,
      "step": 429000
    },
    {
      "epoch": 4.78,
      "eval_loss": 0.2081955075263977,
      "eval_runtime": 2.6196,
      "eval_samples_per_second": 876.837,
      "eval_steps_per_second": 13.742,
      "step": 429000
    },
    {
      "epoch": 4.79,
      "learning_rate": 0.00010174123801378698,
      "loss": 0.2178,
      "step": 429500
    },
    {
      "epoch": 4.8,
      "learning_rate": 0.00010163118960624632,
      "loss": 0.2178,
      "step": 430000
    },
    {
      "epoch": 4.8,
      "eval_loss": 0.2074689269065857,
      "eval_runtime": 2.661,
      "eval_samples_per_second": 863.204,
      "eval_steps_per_second": 13.529,
      "step": 430000
    },
    {
      "epoch": 4.8,
      "learning_rate": 0.00010152108205985925,
      "loss": 0.2179,
      "step": 430500
    },
    {
      "epoch": 4.81,
      "learning_rate": 0.00010141091567565561,
      "loss": 0.2176,
      "step": 431000
    },
    {
      "epoch": 4.81,
      "eval_loss": 0.20468705892562866,
      "eval_runtime": 2.535,
      "eval_samples_per_second": 906.124,
      "eval_steps_per_second": 14.201,
      "step": 431000
    },
    {
      "epoch": 4.81,
      "learning_rate": 0.00010130069075482611,
      "loss": 0.2174,
      "step": 431500
    },
    {
      "epoch": 4.82,
      "learning_rate": 0.00010119040759872142,
      "loss": 0.2178,
      "step": 432000
    },
    {
      "epoch": 4.82,
      "eval_loss": 0.20778746902942657,
      "eval_runtime": 2.5027,
      "eval_samples_per_second": 917.795,
      "eval_steps_per_second": 14.384,
      "step": 432000
    },
    {
      "epoch": 4.82,
      "learning_rate": 0.00010108006650885151,
      "loss": 0.2171,
      "step": 432500
    },
    {
      "epoch": 4.83,
      "learning_rate": 0.00010096966778688472,
      "loss": 0.2175,
      "step": 433000
    },
    {
      "epoch": 4.83,
      "eval_loss": 0.2072180211544037,
      "eval_runtime": 2.6325,
      "eval_samples_per_second": 872.544,
      "eval_steps_per_second": 13.675,
      "step": 433000
    },
    {
      "epoch": 4.83,
      "learning_rate": 0.00010085921173464691,
      "loss": 0.2175,
      "step": 433500
    },
    {
      "epoch": 4.84,
      "learning_rate": 0.00010074869865412074,
      "loss": 0.217,
      "step": 434000
    },
    {
      "epoch": 4.84,
      "eval_loss": 0.20796707272529602,
      "eval_runtime": 2.5383,
      "eval_samples_per_second": 904.951,
      "eval_steps_per_second": 14.183,
      "step": 434000
    },
    {
      "epoch": 4.85,
      "learning_rate": 0.00010063812884744475,
      "loss": 0.217,
      "step": 434500
    },
    {
      "epoch": 4.85,
      "learning_rate": 0.00010052750261691254,
      "loss": 0.2172,
      "step": 435000
    },
    {
      "epoch": 4.85,
      "eval_loss": 0.2071193903684616,
      "eval_runtime": 2.5047,
      "eval_samples_per_second": 917.082,
      "eval_steps_per_second": 14.373,
      "step": 435000
    },
    {
      "epoch": 4.86,
      "learning_rate": 0.00010041682026497199,
      "loss": 0.2165,
      "step": 435500
    },
    {
      "epoch": 4.86,
      "learning_rate": 0.0001003060820942245,
      "loss": 0.2168,
      "step": 436000
    },
    {
      "epoch": 4.86,
      "eval_loss": 0.205715611577034,
      "eval_runtime": 2.6955,
      "eval_samples_per_second": 852.165,
      "eval_steps_per_second": 13.356,
      "step": 436000
    },
    {
      "epoch": 4.87,
      "learning_rate": 0.00010019528840742392,
      "loss": 0.2172,
      "step": 436500
    },
    {
      "epoch": 4.87,
      "learning_rate": 0.00010008443950747599,
      "loss": 0.2166,
      "step": 437000
    },
    {
      "epoch": 4.87,
      "eval_loss": 0.20823287963867188,
      "eval_runtime": 2.581,
      "eval_samples_per_second": 889.967,
      "eval_steps_per_second": 13.948,
      "step": 437000
    },
    {
      "epoch": 4.88,
      "learning_rate": 9.997353569743736e-05,
      "loss": 0.2168,
      "step": 437500
    },
    {
      "epoch": 4.88,
      "learning_rate": 9.986257728051483e-05,
      "loss": 0.2173,
      "step": 438000
    },
    {
      "epoch": 4.88,
      "eval_loss": 0.2068750560283661,
      "eval_runtime": 2.5735,
      "eval_samples_per_second": 892.546,
      "eval_steps_per_second": 13.989,
      "step": 438000
    },
    {
      "epoch": 4.89,
      "learning_rate": 9.975156456006448e-05,
      "loss": 0.2164,
      "step": 438500
    },
    {
      "epoch": 4.9,
      "learning_rate": 9.964049783959082e-05,
      "loss": 0.2168,
      "step": 439000
    },
    {
      "epoch": 4.9,
      "eval_loss": 0.2076042890548706,
      "eval_runtime": 2.5663,
      "eval_samples_per_second": 895.048,
      "eval_steps_per_second": 14.028,
      "step": 439000
    },
    {
      "epoch": 4.9,
      "learning_rate": 9.952937742274605e-05,
      "loss": 0.2166,
      "step": 439500
    },
    {
      "epoch": 4.91,
      "learning_rate": 9.94182036133291e-05,
      "loss": 0.2167,
      "step": 440000
    },
    {
      "epoch": 4.91,
      "eval_loss": 0.20752529799938202,
      "eval_runtime": 2.6206,
      "eval_samples_per_second": 876.507,
      "eval_steps_per_second": 13.737,
      "step": 440000
    },
    {
      "epoch": 4.91,
      "learning_rate": 9.930697671528499e-05,
      "loss": 0.2165,
      "step": 440500
    },
    {
      "epoch": 4.92,
      "learning_rate": 9.919569703270376e-05,
      "loss": 0.2166,
      "step": 441000
    },
    {
      "epoch": 4.92,
      "eval_loss": 0.20720918476581573,
      "eval_runtime": 2.6091,
      "eval_samples_per_second": 880.39,
      "eval_steps_per_second": 13.798,
      "step": 441000
    },
    {
      "epoch": 4.92,
      "learning_rate": 9.908436486981984e-05,
      "loss": 0.2164,
      "step": 441500
    },
    {
      "epoch": 4.93,
      "learning_rate": 9.89729805310111e-05,
      "loss": 0.217,
      "step": 442000
    },
    {
      "epoch": 4.93,
      "eval_loss": 0.20782898366451263,
      "eval_runtime": 2.5759,
      "eval_samples_per_second": 891.724,
      "eval_steps_per_second": 13.976,
      "step": 442000
    },
    {
      "epoch": 4.94,
      "learning_rate": 9.886154432079803e-05,
      "loss": 0.2166,
      "step": 442500
    },
    {
      "epoch": 4.94,
      "learning_rate": 9.875005654384307e-05,
      "loss": 0.2163,
      "step": 443000
    },
    {
      "epoch": 4.94,
      "eval_loss": 0.20723912119865417,
      "eval_runtime": 2.6842,
      "eval_samples_per_second": 855.749,
      "eval_steps_per_second": 13.412,
      "step": 443000
    },
    {
      "epoch": 4.95,
      "learning_rate": 9.863851750494944e-05,
      "loss": 0.2168,
      "step": 443500
    },
    {
      "epoch": 4.95,
      "learning_rate": 9.852692750906071e-05,
      "loss": 0.2165,
      "step": 444000
    },
    {
      "epoch": 4.95,
      "eval_loss": 0.20486100018024445,
      "eval_runtime": 2.6649,
      "eval_samples_per_second": 861.956,
      "eval_steps_per_second": 13.509,
      "step": 444000
    },
    {
      "epoch": 4.96,
      "learning_rate": 9.841528686125961e-05,
      "loss": 0.2158,
      "step": 444500
    },
    {
      "epoch": 4.96,
      "learning_rate": 9.830359586676737e-05,
      "loss": 0.2157,
      "step": 445000
    },
    {
      "epoch": 4.96,
      "eval_loss": 0.20740051567554474,
      "eval_runtime": 2.6655,
      "eval_samples_per_second": 861.742,
      "eval_steps_per_second": 13.506,
      "step": 445000
    },
    {
      "epoch": 4.97,
      "learning_rate": 9.819185483094299e-05,
      "loss": 0.2154,
      "step": 445500
    },
    {
      "epoch": 4.97,
      "learning_rate": 9.808006405928215e-05,
      "loss": 0.216,
      "step": 446000
    },
    {
      "epoch": 4.97,
      "eval_loss": 0.20571541786193848,
      "eval_runtime": 2.5424,
      "eval_samples_per_second": 903.461,
      "eval_steps_per_second": 14.16,
      "step": 446000
    },
    {
      "epoch": 4.98,
      "learning_rate": 9.796822385741657e-05,
      "loss": 0.216,
      "step": 446500
    },
    {
      "epoch": 4.99,
      "learning_rate": 9.785633453111306e-05,
      "loss": 0.2159,
      "step": 447000
    },
    {
      "epoch": 4.99,
      "eval_loss": 0.20638670027256012,
      "eval_runtime": 2.5508,
      "eval_samples_per_second": 900.496,
      "eval_steps_per_second": 14.113,
      "step": 447000
    },
    {
      "epoch": 4.99,
      "learning_rate": 9.774439638627277e-05,
      "loss": 0.2156,
      "step": 447500
    },
    {
      "epoch": 5.0,
      "learning_rate": 9.763240972893037e-05,
      "loss": 0.216,
      "step": 448000
    },
    {
      "epoch": 5.0,
      "eval_loss": 0.20472365617752075,
      "eval_runtime": 2.6318,
      "eval_samples_per_second": 872.783,
      "eval_steps_per_second": 13.679,
      "step": 448000
    },
    {
      "epoch": 5.0,
      "learning_rate": 9.752037486525302e-05,
      "loss": 0.2157,
      "step": 448500
    },
    {
      "epoch": 5.01,
      "learning_rate": 9.740829210153984e-05,
      "loss": 0.2153,
      "step": 449000
    },
    {
      "epoch": 5.01,
      "eval_loss": 0.2034875452518463,
      "eval_runtime": 2.6776,
      "eval_samples_per_second": 857.846,
      "eval_steps_per_second": 13.445,
      "step": 449000
    },
    {
      "epoch": 5.01,
      "learning_rate": 9.729616174422077e-05,
      "loss": 0.2153,
      "step": 449500
    },
    {
      "epoch": 5.02,
      "learning_rate": 9.718398409985593e-05,
      "loss": 0.2157,
      "step": 450000
    },
    {
      "epoch": 5.02,
      "eval_loss": 0.20429137349128723,
      "eval_runtime": 2.6747,
      "eval_samples_per_second": 858.786,
      "eval_steps_per_second": 13.459,
      "step": 450000
    },
    {
      "epoch": 5.02,
      "learning_rate": 9.707175947513475e-05,
      "loss": 0.215,
      "step": 450500
    },
    {
      "epoch": 5.03,
      "learning_rate": 9.695948817687504e-05,
      "loss": 0.2153,
      "step": 451000
    },
    {
      "epoch": 5.03,
      "eval_loss": 0.20547254383563995,
      "eval_runtime": 2.5571,
      "eval_samples_per_second": 898.273,
      "eval_steps_per_second": 14.078,
      "step": 451000
    },
    {
      "epoch": 5.04,
      "learning_rate": 9.684717051202227e-05,
      "loss": 0.2154,
      "step": 451500
    },
    {
      "epoch": 5.04,
      "learning_rate": 9.673480678764858e-05,
      "loss": 0.2151,
      "step": 452000
    },
    {
      "epoch": 5.04,
      "eval_loss": 0.204800084233284,
      "eval_runtime": 2.6202,
      "eval_samples_per_second": 876.642,
      "eval_steps_per_second": 13.739,
      "step": 452000
    },
    {
      "epoch": 5.05,
      "learning_rate": 9.662239731095222e-05,
      "loss": 0.2153,
      "step": 452500
    },
    {
      "epoch": 5.05,
      "learning_rate": 9.650994238925626e-05,
      "loss": 0.2149,
      "step": 453000
    },
    {
      "epoch": 5.05,
      "eval_loss": 0.20622511208057404,
      "eval_runtime": 2.6307,
      "eval_samples_per_second": 873.167,
      "eval_steps_per_second": 13.685,
      "step": 453000
    },
    {
      "epoch": 5.06,
      "learning_rate": 9.63974423300083e-05,
      "loss": 0.2148,
      "step": 453500
    },
    {
      "epoch": 5.06,
      "learning_rate": 9.628489744077911e-05,
      "loss": 0.2151,
      "step": 454000
    },
    {
      "epoch": 5.06,
      "eval_loss": 0.20128118991851807,
      "eval_runtime": 2.6992,
      "eval_samples_per_second": 851.002,
      "eval_steps_per_second": 13.337,
      "step": 454000
    },
    {
      "epoch": 5.07,
      "learning_rate": 9.617230802926214e-05,
      "loss": 0.2149,
      "step": 454500
    },
    {
      "epoch": 5.07,
      "learning_rate": 9.60596744032726e-05,
      "loss": 0.2151,
      "step": 455000
    },
    {
      "epoch": 5.07,
      "eval_loss": 0.2051314264535904,
      "eval_runtime": 2.5516,
      "eval_samples_per_second": 900.232,
      "eval_steps_per_second": 14.109,
      "step": 455000
    },
    {
      "epoch": 5.08,
      "learning_rate": 9.594699687074648e-05,
      "loss": 0.2147,
      "step": 455500
    },
    {
      "epoch": 5.09,
      "learning_rate": 9.583427573973982e-05,
      "loss": 0.2151,
      "step": 456000
    },
    {
      "epoch": 5.09,
      "eval_loss": 0.2050078958272934,
      "eval_runtime": 2.6417,
      "eval_samples_per_second": 869.501,
      "eval_steps_per_second": 13.627,
      "step": 456000
    },
    {
      "epoch": 5.09,
      "learning_rate": 9.57215113184279e-05,
      "loss": 0.2144,
      "step": 456500
    },
    {
      "epoch": 5.1,
      "learning_rate": 9.560870391510441e-05,
      "loss": 0.2149,
      "step": 457000
    },
    {
      "epoch": 5.1,
      "eval_loss": 0.20550428330898285,
      "eval_runtime": 2.6145,
      "eval_samples_per_second": 878.571,
      "eval_steps_per_second": 13.77,
      "step": 457000
    },
    {
      "epoch": 5.1,
      "learning_rate": 9.549585383818041e-05,
      "loss": 0.2152,
      "step": 457500
    },
    {
      "epoch": 5.11,
      "learning_rate": 9.538296139618371e-05,
      "loss": 0.2147,
      "step": 458000
    },
    {
      "epoch": 5.11,
      "eval_loss": 0.20618882775306702,
      "eval_runtime": 2.5432,
      "eval_samples_per_second": 903.188,
      "eval_steps_per_second": 14.155,
      "step": 458000
    },
    {
      "epoch": 5.11,
      "learning_rate": 9.527002689775799e-05,
      "loss": 0.2147,
      "step": 458500
    },
    {
      "epoch": 5.12,
      "learning_rate": 9.515705065166178e-05,
      "loss": 0.2143,
      "step": 459000
    },
    {
      "epoch": 5.12,
      "eval_loss": 0.20430730283260345,
      "eval_runtime": 2.5628,
      "eval_samples_per_second": 896.27,
      "eval_steps_per_second": 14.047,
      "step": 459000
    },
    {
      "epoch": 5.12,
      "learning_rate": 9.504403296676786e-05,
      "loss": 0.2148,
      "step": 459500
    },
    {
      "epoch": 5.13,
      "learning_rate": 9.493097415206228e-05,
      "loss": 0.2142,
      "step": 460000
    },
    {
      "epoch": 5.13,
      "eval_loss": 0.20442381501197815,
      "eval_runtime": 2.5555,
      "eval_samples_per_second": 898.857,
      "eval_steps_per_second": 14.087,
      "step": 460000
    },
    {
      "epoch": 5.14,
      "learning_rate": 9.481787451664349e-05,
      "loss": 0.2139,
      "step": 460500
    },
    {
      "epoch": 5.14,
      "learning_rate": 9.47047343697216e-05,
      "loss": 0.2141,
      "step": 461000
    },
    {
      "epoch": 5.14,
      "eval_loss": 0.2020626962184906,
      "eval_runtime": 2.9504,
      "eval_samples_per_second": 778.533,
      "eval_steps_per_second": 12.202,
      "step": 461000
    },
    {
      "epoch": 5.15,
      "learning_rate": 9.459155402061744e-05,
      "loss": 0.2143,
      "step": 461500
    },
    {
      "epoch": 5.15,
      "learning_rate": 9.447833377876176e-05,
      "loss": 0.2146,
      "step": 462000
    },
    {
      "epoch": 5.15,
      "eval_loss": 0.20509694516658783,
      "eval_runtime": 2.6371,
      "eval_samples_per_second": 871.021,
      "eval_steps_per_second": 13.651,
      "step": 462000
    },
    {
      "epoch": 5.16,
      "learning_rate": 9.436507395369439e-05,
      "loss": 0.2146,
      "step": 462500
    },
    {
      "epoch": 5.16,
      "learning_rate": 9.425177485506336e-05,
      "loss": 0.2143,
      "step": 463000
    },
    {
      "epoch": 5.16,
      "eval_loss": 0.20591478049755096,
      "eval_runtime": 2.5305,
      "eval_samples_per_second": 907.719,
      "eval_steps_per_second": 14.226,
      "step": 463000
    },
    {
      "epoch": 5.17,
      "learning_rate": 9.413843679262408e-05,
      "loss": 0.2143,
      "step": 463500
    },
    {
      "epoch": 5.17,
      "learning_rate": 9.402506007623848e-05,
      "loss": 0.2152,
      "step": 464000
    },
    {
      "epoch": 5.17,
      "eval_loss": 0.20412901043891907,
      "eval_runtime": 2.4892,
      "eval_samples_per_second": 922.805,
      "eval_steps_per_second": 14.463,
      "step": 464000
    },
    {
      "epoch": 5.18,
      "learning_rate": 9.391164501587417e-05,
      "loss": 0.2141,
      "step": 464500
    },
    {
      "epoch": 5.19,
      "learning_rate": 9.379819192160362e-05,
      "loss": 0.214,
      "step": 465000
    },
    {
      "epoch": 5.19,
      "eval_loss": 0.2041676938533783,
      "eval_runtime": 6.2441,
      "eval_samples_per_second": 367.868,
      "eval_steps_per_second": 5.765,
      "step": 465000
    },
    {
      "epoch": 5.19,
      "learning_rate": 9.368470110360323e-05,
      "loss": 0.2141,
      "step": 465500
    },
    {
      "epoch": 5.2,
      "learning_rate": 9.357117287215258e-05,
      "loss": 0.2138,
      "step": 466000
    },
    {
      "epoch": 5.2,
      "eval_loss": 0.20280323922634125,
      "eval_runtime": 2.5119,
      "eval_samples_per_second": 914.43,
      "eval_steps_per_second": 14.332,
      "step": 466000
    },
    {
      "epoch": 5.2,
      "learning_rate": 9.345760753763347e-05,
      "loss": 0.2139,
      "step": 466500
    },
    {
      "epoch": 5.21,
      "learning_rate": 9.334400541052928e-05,
      "loss": 0.2133,
      "step": 467000
    },
    {
      "epoch": 5.21,
      "eval_loss": 0.20342190563678741,
      "eval_runtime": 2.466,
      "eval_samples_per_second": 931.458,
      "eval_steps_per_second": 14.598,
      "step": 467000
    },
    {
      "epoch": 5.21,
      "learning_rate": 9.323036680142382e-05,
      "loss": 0.2138,
      "step": 467500
    },
    {
      "epoch": 5.22,
      "learning_rate": 9.311669202100073e-05,
      "loss": 0.2136,
      "step": 468000
    },
    {
      "epoch": 5.22,
      "eval_loss": 0.20149648189544678,
      "eval_runtime": 2.5403,
      "eval_samples_per_second": 904.23,
      "eval_steps_per_second": 14.172,
      "step": 468000
    },
    {
      "epoch": 5.23,
      "learning_rate": 9.300298138004249e-05,
      "loss": 0.2137,
      "step": 468500
    },
    {
      "epoch": 5.23,
      "learning_rate": 9.288923518942968e-05,
      "loss": 0.2135,
      "step": 469000
    },
    {
      "epoch": 5.23,
      "eval_loss": 0.2041381597518921,
      "eval_runtime": 3.0567,
      "eval_samples_per_second": 751.455,
      "eval_steps_per_second": 11.777,
      "step": 469000
    },
    {
      "epoch": 5.24,
      "learning_rate": 9.277545376014005e-05,
      "loss": 0.2134,
      "step": 469500
    },
    {
      "epoch": 5.24,
      "learning_rate": 9.26616374032477e-05,
      "loss": 0.2135,
      "step": 470000
    },
    {
      "epoch": 5.24,
      "eval_loss": 0.2072334885597229,
      "eval_runtime": 2.5117,
      "eval_samples_per_second": 914.527,
      "eval_steps_per_second": 14.333,
      "step": 470000
    },
    {
      "epoch": 5.25,
      "learning_rate": 9.254778642992213e-05,
      "loss": 0.2133,
      "step": 470500
    },
    {
      "epoch": 5.25,
      "learning_rate": 9.243390115142761e-05,
      "loss": 0.2132,
      "step": 471000
    },
    {
      "epoch": 5.25,
      "eval_loss": 0.20281285047531128,
      "eval_runtime": 2.5255,
      "eval_samples_per_second": 909.541,
      "eval_steps_per_second": 14.255,
      "step": 471000
    },
    {
      "epoch": 5.26,
      "learning_rate": 9.231998187912211e-05,
      "loss": 0.2135,
      "step": 471500
    },
    {
      "epoch": 5.26,
      "learning_rate": 9.220602892445661e-05,
      "loss": 0.2131,
      "step": 472000
    },
    {
      "epoch": 5.26,
      "eval_loss": 0.20354430377483368,
      "eval_runtime": 2.6683,
      "eval_samples_per_second": 860.841,
      "eval_steps_per_second": 13.492,
      "step": 472000
    },
    {
      "epoch": 5.27,
      "learning_rate": 9.209204259897412e-05,
      "loss": 0.2135,
      "step": 472500
    },
    {
      "epoch": 5.28,
      "learning_rate": 9.197802321430889e-05,
      "loss": 0.2131,
      "step": 473000
    },
    {
      "epoch": 5.28,
      "eval_loss": 0.20278283953666687,
      "eval_runtime": 2.6368,
      "eval_samples_per_second": 871.12,
      "eval_steps_per_second": 13.653,
      "step": 473000
    },
    {
      "epoch": 5.28,
      "learning_rate": 9.186397108218558e-05,
      "loss": 0.2131,
      "step": 473500
    },
    {
      "epoch": 5.29,
      "learning_rate": 9.174988651441833e-05,
      "loss": 0.2133,
      "step": 474000
    },
    {
      "epoch": 5.29,
      "eval_loss": 0.20307588577270508,
      "eval_runtime": 2.5584,
      "eval_samples_per_second": 897.827,
      "eval_steps_per_second": 14.071,
      "step": 474000
    },
    {
      "epoch": 5.29,
      "learning_rate": 9.163576982291006e-05,
      "loss": 0.2128,
      "step": 474500
    },
    {
      "epoch": 5.3,
      "learning_rate": 9.152162131965137e-05,
      "loss": 0.213,
      "step": 475000
    },
    {
      "epoch": 5.3,
      "eval_loss": 0.2002372145652771,
      "eval_runtime": 2.521,
      "eval_samples_per_second": 911.138,
      "eval_steps_per_second": 14.28,
      "step": 475000
    },
    {
      "epoch": 5.3,
      "learning_rate": 9.140744131671994e-05,
      "loss": 0.2133,
      "step": 475500
    },
    {
      "epoch": 5.31,
      "learning_rate": 9.129323012627956e-05,
      "loss": 0.2129,
      "step": 476000
    },
    {
      "epoch": 5.31,
      "eval_loss": 0.2028408944606781,
      "eval_runtime": 2.5037,
      "eval_samples_per_second": 917.448,
      "eval_steps_per_second": 14.379,
      "step": 476000
    },
    {
      "epoch": 5.31,
      "learning_rate": 9.117898806057925e-05,
      "loss": 0.2129,
      "step": 476500
    },
    {
      "epoch": 5.32,
      "learning_rate": 9.106471543195244e-05,
      "loss": 0.2131,
      "step": 477000
    },
    {
      "epoch": 5.32,
      "eval_loss": 0.20160850882530212,
      "eval_runtime": 2.4426,
      "eval_samples_per_second": 940.394,
      "eval_steps_per_second": 14.738,
      "step": 477000
    },
    {
      "epoch": 5.33,
      "learning_rate": 9.095041255281616e-05,
      "loss": 0.2124,
      "step": 477500
    },
    {
      "epoch": 5.33,
      "learning_rate": 9.08360797356701e-05,
      "loss": 0.2129,
      "step": 478000
    },
    {
      "epoch": 5.33,
      "eval_loss": 0.2036493867635727,
      "eval_runtime": 2.4816,
      "eval_samples_per_second": 925.609,
      "eval_steps_per_second": 14.507,
      "step": 478000
    },
    {
      "epoch": 5.34,
      "learning_rate": 9.07217172930958e-05,
      "loss": 0.2126,
      "step": 478500
    },
    {
      "epoch": 5.34,
      "learning_rate": 9.060732553775582e-05,
      "loss": 0.2127,
      "step": 479000
    },
    {
      "epoch": 5.34,
      "eval_loss": 0.19979223608970642,
      "eval_runtime": 2.5343,
      "eval_samples_per_second": 906.38,
      "eval_steps_per_second": 14.205,
      "step": 479000
    },
    {
      "epoch": 5.35,
      "learning_rate": 9.049290478239287e-05,
      "loss": 0.2126,
      "step": 479500
    },
    {
      "epoch": 5.35,
      "learning_rate": 9.037845533982892e-05,
      "loss": 0.2129,
      "step": 480000
    },
    {
      "epoch": 5.35,
      "eval_loss": 0.20060274004936218,
      "eval_runtime": 2.5428,
      "eval_samples_per_second": 903.348,
      "eval_steps_per_second": 14.158,
      "step": 480000
    },
    {
      "epoch": 5.36,
      "learning_rate": 9.02639775229644e-05,
      "loss": 0.2124,
      "step": 480500
    },
    {
      "epoch": 5.36,
      "learning_rate": 9.014947164477721e-05,
      "loss": 0.2126,
      "step": 481000
    },
    {
      "epoch": 5.36,
      "eval_loss": 0.20033051073551178,
      "eval_runtime": 2.4935,
      "eval_samples_per_second": 921.209,
      "eval_steps_per_second": 14.438,
      "step": 481000
    },
    {
      "epoch": 5.37,
      "learning_rate": 9.003493801832213e-05,
      "loss": 0.212,
      "step": 481500
    },
    {
      "epoch": 5.38,
      "learning_rate": 8.992037695672967e-05,
      "loss": 0.2127,
      "step": 482000
    },
    {
      "epoch": 5.38,
      "eval_loss": 0.2025318741798401,
      "eval_runtime": 2.5419,
      "eval_samples_per_second": 903.65,
      "eval_steps_per_second": 14.163,
      "step": 482000
    },
    {
      "epoch": 5.38,
      "learning_rate": 8.980578877320544e-05,
      "loss": 0.2128,
      "step": 482500
    },
    {
      "epoch": 5.39,
      "learning_rate": 8.969117378102912e-05,
      "loss": 0.2122,
      "step": 483000
    },
    {
      "epoch": 5.39,
      "eval_loss": 0.2024153620004654,
      "eval_runtime": 2.5406,
      "eval_samples_per_second": 904.131,
      "eval_steps_per_second": 14.17,
      "step": 483000
    },
    {
      "epoch": 5.39,
      "learning_rate": 8.957653229355374e-05,
      "loss": 0.212,
      "step": 483500
    },
    {
      "epoch": 5.4,
      "learning_rate": 8.946186462420478e-05,
      "loss": 0.2123,
      "step": 484000
    },
    {
      "epoch": 5.4,
      "eval_loss": 0.20227807760238647,
      "eval_runtime": 2.6148,
      "eval_samples_per_second": 878.471,
      "eval_steps_per_second": 13.768,
      "step": 484000
    },
    {
      "epoch": 5.4,
      "learning_rate": 8.934717108647922e-05,
      "loss": 0.2122,
      "step": 484500
    },
    {
      "epoch": 5.41,
      "learning_rate": 8.923245199394482e-05,
      "loss": 0.2123,
      "step": 485000
    },
    {
      "epoch": 5.41,
      "eval_loss": 0.2004445344209671,
      "eval_runtime": 2.6169,
      "eval_samples_per_second": 877.747,
      "eval_steps_per_second": 13.757,
      "step": 485000
    },
    {
      "epoch": 5.41,
      "learning_rate": 8.911770766023921e-05,
      "loss": 0.212,
      "step": 485500
    },
    {
      "epoch": 5.42,
      "learning_rate": 8.900293839906903e-05,
      "loss": 0.2113,
      "step": 486000
    },
    {
      "epoch": 5.42,
      "eval_loss": 0.20107805728912354,
      "eval_runtime": 2.6497,
      "eval_samples_per_second": 866.906,
      "eval_steps_per_second": 13.587,
      "step": 486000
    },
    {
      "epoch": 5.43,
      "learning_rate": 8.888814452420903e-05,
      "loss": 0.2238,
      "step": 486500
    },
    {
      "epoch": 5.43,
      "learning_rate": 8.87733263495013e-05,
      "loss": 0.2125,
      "step": 487000
    },
    {
      "epoch": 5.43,
      "eval_loss": 0.20239748060703278,
      "eval_runtime": 2.6391,
      "eval_samples_per_second": 870.377,
      "eval_steps_per_second": 13.641,
      "step": 487000
    },
    {
      "epoch": 5.44,
      "learning_rate": 8.865848418885434e-05,
      "loss": 0.2116,
      "step": 487500
    },
    {
      "epoch": 5.44,
      "learning_rate": 8.85436183562422e-05,
      "loss": 0.2117,
      "step": 488000
    },
    {
      "epoch": 5.44,
      "eval_loss": 0.20209769904613495,
      "eval_runtime": 2.6167,
      "eval_samples_per_second": 877.814,
      "eval_steps_per_second": 13.758,
      "step": 488000
    },
    {
      "epoch": 5.45,
      "learning_rate": 8.842872916570374e-05,
      "loss": 0.2117,
      "step": 488500
    },
    {
      "epoch": 5.45,
      "learning_rate": 8.83138169313416e-05,
      "loss": 0.2117,
      "step": 489000
    },
    {
      "epoch": 5.45,
      "eval_loss": 0.20107334852218628,
      "eval_runtime": 2.6263,
      "eval_samples_per_second": 874.613,
      "eval_steps_per_second": 13.707,
      "step": 489000
    },
    {
      "epoch": 5.46,
      "learning_rate": 8.819888196732144e-05,
      "loss": 0.2116,
      "step": 489500
    },
    {
      "epoch": 5.46,
      "learning_rate": 8.808392458787103e-05,
      "loss": 0.2116,
      "step": 490000
    },
    {
      "epoch": 5.46,
      "eval_loss": 0.2020297795534134,
      "eval_runtime": 2.6641,
      "eval_samples_per_second": 862.22,
      "eval_steps_per_second": 13.513,
      "step": 490000
    },
    {
      "epoch": 5.47,
      "learning_rate": 8.796894510727945e-05,
      "loss": 0.2116,
      "step": 490500
    },
    {
      "epoch": 5.48,
      "learning_rate": 8.78539438398963e-05,
      "loss": 0.2112,
      "step": 491000
    },
    {
      "epoch": 5.48,
      "eval_loss": 0.20125530660152435,
      "eval_runtime": 2.6454,
      "eval_samples_per_second": 868.302,
      "eval_steps_per_second": 13.609,
      "step": 491000
    },
    {
      "epoch": 5.48,
      "learning_rate": 8.773892110013058e-05,
      "loss": 0.2116,
      "step": 491500
    },
    {
      "epoch": 5.49,
      "learning_rate": 8.762387720245008e-05,
      "loss": 0.2114,
      "step": 492000
    },
    {
      "epoch": 5.49,
      "eval_loss": 0.2021346092224121,
      "eval_runtime": 2.6349,
      "eval_samples_per_second": 871.772,
      "eval_steps_per_second": 13.663,
      "step": 492000
    },
    {
      "epoch": 5.49,
      "learning_rate": 8.750881246138043e-05,
      "loss": 0.2115,
      "step": 492500
    },
    {
      "epoch": 5.5,
      "learning_rate": 8.73937271915042e-05,
      "loss": 0.2115,
      "step": 493000
    },
    {
      "epoch": 5.5,
      "eval_loss": 0.20063868165016174,
      "eval_runtime": 2.6134,
      "eval_samples_per_second": 878.933,
      "eval_steps_per_second": 13.775,
      "step": 493000
    },
    {
      "epoch": 5.5,
      "learning_rate": 8.727862170746019e-05,
      "loss": 0.2112,
      "step": 493500
    },
    {
      "epoch": 5.51,
      "learning_rate": 8.716349632394235e-05,
      "loss": 0.2113,
      "step": 494000
    },
    {
      "epoch": 5.51,
      "eval_loss": 0.2013174444437027,
      "eval_runtime": 2.5993,
      "eval_samples_per_second": 883.689,
      "eval_steps_per_second": 13.85,
      "step": 494000
    },
    {
      "epoch": 5.52,
      "learning_rate": 8.70483513556991e-05,
      "loss": 0.2111,
      "step": 494500
    },
    {
      "epoch": 5.52,
      "learning_rate": 8.69331871175324e-05,
      "loss": 0.2111,
      "step": 495000
    },
    {
      "epoch": 5.52,
      "eval_loss": 0.20253530144691467,
      "eval_runtime": 2.655,
      "eval_samples_per_second": 865.174,
      "eval_steps_per_second": 13.56,
      "step": 495000
    },
    {
      "epoch": 5.53,
      "learning_rate": 8.681800392429684e-05,
      "loss": 0.2111,
      "step": 495500
    },
    {
      "epoch": 5.53,
      "learning_rate": 8.67028020908989e-05,
      "loss": 0.2112,
      "step": 496000
    },
    {
      "epoch": 5.53,
      "eval_loss": 0.2021927386522293,
      "eval_runtime": 2.5884,
      "eval_samples_per_second": 887.415,
      "eval_steps_per_second": 13.908,
      "step": 496000
    },
    {
      "epoch": 5.54,
      "learning_rate": 8.658758193229601e-05,
      "loss": 0.211,
      "step": 496500
    },
    {
      "epoch": 5.54,
      "learning_rate": 8.647234376349565e-05,
      "loss": 0.2107,
      "step": 497000
    },
    {
      "epoch": 5.54,
      "eval_loss": 0.19982150197029114,
      "eval_runtime": 2.5228,
      "eval_samples_per_second": 910.512,
      "eval_steps_per_second": 14.27,
      "step": 497000
    },
    {
      "epoch": 5.55,
      "learning_rate": 8.635708789955458e-05,
      "loss": 0.2108,
      "step": 497500
    },
    {
      "epoch": 5.55,
      "learning_rate": 8.624181465557794e-05,
      "loss": 0.2109,
      "step": 498000
    },
    {
      "epoch": 5.55,
      "eval_loss": 0.2010236382484436,
      "eval_runtime": 2.4638,
      "eval_samples_per_second": 932.313,
      "eval_steps_per_second": 14.612,
      "step": 498000
    },
    {
      "epoch": 5.56,
      "learning_rate": 8.612652434671837e-05,
      "loss": 0.2108,
      "step": 498500
    },
    {
      "epoch": 5.57,
      "learning_rate": 8.601121728817519e-05,
      "loss": 0.211,
      "step": 499000
    },
    {
      "epoch": 5.57,
      "eval_loss": 0.19817927479743958,
      "eval_runtime": 2.5254,
      "eval_samples_per_second": 909.547,
      "eval_steps_per_second": 14.255,
      "step": 499000
    },
    {
      "epoch": 5.57,
      "learning_rate": 8.589589379519346e-05,
      "loss": 0.2109,
      "step": 499500
    },
    {
      "epoch": 5.58,
      "learning_rate": 8.578055418306327e-05,
      "loss": 0.2107,
      "step": 500000
    },
    {
      "epoch": 5.58,
      "eval_loss": 0.199160635471344,
      "eval_runtime": 2.6076,
      "eval_samples_per_second": 880.873,
      "eval_steps_per_second": 13.806,
      "step": 500000
    },
    {
      "epoch": 5.58,
      "learning_rate": 8.566519876711864e-05,
      "loss": 0.2109,
      "step": 500500
    },
    {
      "epoch": 5.59,
      "learning_rate": 8.55498278627369e-05,
      "loss": 0.2107,
      "step": 501000
    },
    {
      "epoch": 5.59,
      "eval_loss": 0.1986326277256012,
      "eval_runtime": 2.5229,
      "eval_samples_per_second": 910.451,
      "eval_steps_per_second": 14.269,
      "step": 501000
    },
    {
      "epoch": 5.59,
      "learning_rate": 8.543444178533773e-05,
      "loss": 0.2101,
      "step": 501500
    },
    {
      "epoch": 5.6,
      "learning_rate": 8.531904085038221e-05,
      "loss": 0.2104,
      "step": 502000
    },
    {
      "epoch": 5.6,
      "eval_loss": 0.1985243856906891,
      "eval_runtime": 2.533,
      "eval_samples_per_second": 906.836,
      "eval_steps_per_second": 14.212,
      "step": 502000
    },
    {
      "epoch": 5.6,
      "learning_rate": 8.520362537337214e-05,
      "loss": 0.2104,
      "step": 502500
    },
    {
      "epoch": 5.61,
      "learning_rate": 8.508819566984897e-05,
      "loss": 0.2105,
      "step": 503000
    },
    {
      "epoch": 5.61,
      "eval_loss": 0.19903981685638428,
      "eval_runtime": 2.5776,
      "eval_samples_per_second": 891.131,
      "eval_steps_per_second": 13.966,
      "step": 503000
    },
    {
      "epoch": 5.62,
      "learning_rate": 8.497275205539314e-05,
      "loss": 0.2103,
      "step": 503500
    },
    {
      "epoch": 5.62,
      "learning_rate": 8.485729484562307e-05,
      "loss": 0.2104,
      "step": 504000
    },
    {
      "epoch": 5.62,
      "eval_loss": 0.20253504812717438,
      "eval_runtime": 2.542,
      "eval_samples_per_second": 903.632,
      "eval_steps_per_second": 14.162,
      "step": 504000
    },
    {
      "epoch": 5.63,
      "learning_rate": 8.474182435619437e-05,
      "loss": 0.21,
      "step": 504500
    },
    {
      "epoch": 5.63,
      "learning_rate": 8.462634090279895e-05,
      "loss": 0.2099,
      "step": 505000
    },
    {
      "epoch": 5.63,
      "eval_loss": 0.20055775344371796,
      "eval_runtime": 2.5318,
      "eval_samples_per_second": 907.265,
      "eval_steps_per_second": 14.219,
      "step": 505000
    },
    {
      "epoch": 5.64,
      "learning_rate": 8.451084480116415e-05,
      "loss": 0.2099,
      "step": 505500
    },
    {
      "epoch": 5.64,
      "learning_rate": 8.439533636705194e-05,
      "loss": 0.2102,
      "step": 506000
    },
    {
      "epoch": 5.64,
      "eval_loss": 0.1988956183195114,
      "eval_runtime": 2.5177,
      "eval_samples_per_second": 912.354,
      "eval_steps_per_second": 14.299,
      "step": 506000
    },
    {
      "epoch": 5.65,
      "learning_rate": 8.427981591625791e-05,
      "loss": 0.2097,
      "step": 506500
    },
    {
      "epoch": 5.65,
      "learning_rate": 8.416428376461061e-05,
      "loss": 0.2101,
      "step": 507000
    },
    {
      "epoch": 5.65,
      "eval_loss": 0.20103515684604645,
      "eval_runtime": 2.4839,
      "eval_samples_per_second": 924.771,
      "eval_steps_per_second": 14.494,
      "step": 507000
    },
    {
      "epoch": 5.66,
      "learning_rate": 8.404874022797049e-05,
      "loss": 0.21,
      "step": 507500
    },
    {
      "epoch": 5.67,
      "learning_rate": 8.393318562222916e-05,
      "loss": 0.2102,
      "step": 508000
    },
    {
      "epoch": 5.67,
      "eval_loss": 0.19926241040229797,
      "eval_runtime": 2.6513,
      "eval_samples_per_second": 866.384,
      "eval_steps_per_second": 13.578,
      "step": 508000
    },
    {
      "epoch": 5.67,
      "learning_rate": 8.381762026330858e-05,
      "loss": 0.2101,
      "step": 508500
    },
    {
      "epoch": 5.68,
      "learning_rate": 8.370204446715997e-05,
      "loss": 0.2098,
      "step": 509000
    },
    {
      "epoch": 5.68,
      "eval_loss": 0.2018594741821289,
      "eval_runtime": 2.6012,
      "eval_samples_per_second": 883.058,
      "eval_steps_per_second": 13.84,
      "step": 509000
    },
    {
      "epoch": 5.68,
      "learning_rate": 8.358645854976311e-05,
      "loss": 0.2101,
      "step": 509500
    },
    {
      "epoch": 5.69,
      "learning_rate": 8.347086282712556e-05,
      "loss": 0.2097,
      "step": 510000
    },
    {
      "epoch": 5.69,
      "eval_loss": 0.19868962466716766,
      "eval_runtime": 2.618,
      "eval_samples_per_second": 877.377,
      "eval_steps_per_second": 13.751,
      "step": 510000
    },
    {
      "epoch": 5.69,
      "learning_rate": 8.335525761528157e-05,
      "loss": 0.2097,
      "step": 510500
    },
    {
      "epoch": 5.7,
      "learning_rate": 8.323964323029136e-05,
      "loss": 0.2097,
      "step": 511000
    },
    {
      "epoch": 5.7,
      "eval_loss": 0.19855724275112152,
      "eval_runtime": 2.5751,
      "eval_samples_per_second": 892.001,
      "eval_steps_per_second": 13.98,
      "step": 511000
    },
    {
      "epoch": 5.7,
      "learning_rate": 8.312401998824027e-05,
      "loss": 0.2094,
      "step": 511500
    },
    {
      "epoch": 5.71,
      "learning_rate": 8.300838820523784e-05,
      "loss": 0.2098,
      "step": 512000
    },
    {
      "epoch": 5.71,
      "eval_loss": 0.20061442255973816,
      "eval_runtime": 2.5891,
      "eval_samples_per_second": 887.184,
      "eval_steps_per_second": 13.904,
      "step": 512000
    },
    {
      "epoch": 5.72,
      "learning_rate": 8.289274819741691e-05,
      "loss": 0.2094,
      "step": 512500
    },
    {
      "epoch": 5.72,
      "learning_rate": 8.277710028093289e-05,
      "loss": 0.2095,
      "step": 513000
    },
    {
      "epoch": 5.72,
      "eval_loss": 0.19978360831737518,
      "eval_runtime": 2.5112,
      "eval_samples_per_second": 914.704,
      "eval_steps_per_second": 14.336,
      "step": 513000
    },
    {
      "epoch": 5.73,
      "learning_rate": 8.266144477196274e-05,
      "loss": 0.2097,
      "step": 513500
    },
    {
      "epoch": 5.73,
      "learning_rate": 8.254578198670421e-05,
      "loss": 0.2096,
      "step": 514000
    },
    {
      "epoch": 5.73,
      "eval_loss": 0.19981496036052704,
      "eval_runtime": 2.6312,
      "eval_samples_per_second": 872.987,
      "eval_steps_per_second": 13.682,
      "step": 514000
    },
    {
      "epoch": 5.74,
      "learning_rate": 8.243011224137492e-05,
      "loss": 0.2091,
      "step": 514500
    },
    {
      "epoch": 5.74,
      "learning_rate": 8.231443585221157e-05,
      "loss": 0.2094,
      "step": 515000
    },
    {
      "epoch": 5.74,
      "eval_loss": 0.19933128356933594,
      "eval_runtime": 2.5527,
      "eval_samples_per_second": 899.838,
      "eval_steps_per_second": 14.103,
      "step": 515000
    },
    {
      "epoch": 5.75,
      "learning_rate": 8.219875313546898e-05,
      "loss": 0.2097,
      "step": 515500
    },
    {
      "epoch": 5.75,
      "learning_rate": 8.208306440741926e-05,
      "loss": 0.209,
      "step": 516000
    },
    {
      "epoch": 5.75,
      "eval_loss": 0.19918790459632874,
      "eval_runtime": 2.5243,
      "eval_samples_per_second": 909.957,
      "eval_steps_per_second": 14.261,
      "step": 516000
    },
    {
      "epoch": 5.76,
      "learning_rate": 8.196736998435101e-05,
      "loss": 0.2091,
      "step": 516500
    },
    {
      "epoch": 5.77,
      "learning_rate": 8.185167018256834e-05,
      "loss": 0.2092,
      "step": 517000
    },
    {
      "epoch": 5.77,
      "eval_loss": 0.19863919913768768,
      "eval_runtime": 2.6312,
      "eval_samples_per_second": 872.998,
      "eval_steps_per_second": 13.682,
      "step": 517000
    },
    {
      "epoch": 5.77,
      "learning_rate": 8.173596531839011e-05,
      "loss": 0.2089,
      "step": 517500
    },
    {
      "epoch": 5.78,
      "learning_rate": 8.162025570814896e-05,
      "loss": 0.2088,
      "step": 518000
    },
    {
      "epoch": 5.78,
      "eval_loss": 0.19913460314273834,
      "eval_runtime": 2.6329,
      "eval_samples_per_second": 872.416,
      "eval_steps_per_second": 13.673,
      "step": 518000
    },
    {
      "epoch": 5.78,
      "learning_rate": 8.150454166819059e-05,
      "loss": 0.2087,
      "step": 518500
    },
    {
      "epoch": 5.79,
      "learning_rate": 8.138882351487275e-05,
      "loss": 0.2086,
      "step": 519000
    },
    {
      "epoch": 5.79,
      "eval_loss": 0.20065449178218842,
      "eval_runtime": 2.6728,
      "eval_samples_per_second": 859.388,
      "eval_steps_per_second": 13.469,
      "step": 519000
    },
    {
      "epoch": 5.79,
      "learning_rate": 8.127310156456445e-05,
      "loss": 0.2088,
      "step": 519500
    },
    {
      "epoch": 5.8,
      "learning_rate": 8.115737613364511e-05,
      "loss": 0.2091,
      "step": 520000
    },
    {
      "epoch": 5.8,
      "eval_loss": 0.19827573001384735,
      "eval_runtime": 2.5025,
      "eval_samples_per_second": 917.889,
      "eval_steps_per_second": 14.386,
      "step": 520000
    },
    {
      "epoch": 5.81,
      "learning_rate": 8.104164753850357e-05,
      "loss": 0.2086,
      "step": 520500
    },
    {
      "epoch": 5.81,
      "learning_rate": 8.092591609553747e-05,
      "loss": 0.2084,
      "step": 521000
    },
    {
      "epoch": 5.81,
      "eval_loss": 0.19897180795669556,
      "eval_runtime": 2.6826,
      "eval_samples_per_second": 856.274,
      "eval_steps_per_second": 13.42,
      "step": 521000
    },
    {
      "epoch": 5.82,
      "learning_rate": 8.081018212115208e-05,
      "loss": 0.2088,
      "step": 521500
    },
    {
      "epoch": 5.82,
      "learning_rate": 8.069444593175975e-05,
      "loss": 0.2087,
      "step": 522000
    },
    {
      "epoch": 5.82,
      "eval_loss": 0.19715061783790588,
      "eval_runtime": 2.574,
      "eval_samples_per_second": 892.373,
      "eval_steps_per_second": 13.986,
      "step": 522000
    },
    {
      "epoch": 5.83,
      "learning_rate": 8.057870784377874e-05,
      "loss": 0.2091,
      "step": 522500
    },
    {
      "epoch": 5.83,
      "learning_rate": 8.046296817363259e-05,
      "loss": 0.2084,
      "step": 523000
    },
    {
      "epoch": 5.83,
      "eval_loss": 0.19694332778453827,
      "eval_runtime": 2.6562,
      "eval_samples_per_second": 864.754,
      "eval_steps_per_second": 13.553,
      "step": 523000
    },
    {
      "epoch": 5.84,
      "learning_rate": 8.034722723774913e-05,
      "loss": 0.2087,
      "step": 523500
    },
    {
      "epoch": 5.84,
      "learning_rate": 8.023148535255965e-05,
      "loss": 0.2085,
      "step": 524000
    },
    {
      "epoch": 5.84,
      "eval_loss": 0.19816145300865173,
      "eval_runtime": 2.6212,
      "eval_samples_per_second": 876.329,
      "eval_steps_per_second": 13.734,
      "step": 524000
    },
    {
      "epoch": 5.85,
      "learning_rate": 8.011574283449807e-05,
      "loss": 0.2083,
      "step": 524500
    },
    {
      "epoch": 5.86,
      "learning_rate": 7.999999999999999e-05,
      "loss": 0.2084,
      "step": 525000
    },
    {
      "epoch": 5.86,
      "eval_loss": 0.1987900733947754,
      "eval_runtime": 2.6058,
      "eval_samples_per_second": 881.485,
      "eval_steps_per_second": 13.815,
      "step": 525000
    },
    {
      "epoch": 5.86,
      "learning_rate": 7.98842571655019e-05,
      "loss": 0.2084,
      "step": 525500
    },
    {
      "epoch": 5.87,
      "learning_rate": 7.976851464744033e-05,
      "loss": 0.2085,
      "step": 526000
    },
    {
      "epoch": 5.87,
      "eval_loss": 0.1988108605146408,
      "eval_runtime": 2.5389,
      "eval_samples_per_second": 904.735,
      "eval_steps_per_second": 14.18,
      "step": 526000
    },
    {
      "epoch": 5.87,
      "learning_rate": 7.965277276225087e-05,
      "loss": 0.208,
      "step": 526500
    },
    {
      "epoch": 5.88,
      "learning_rate": 7.953703182636741e-05,
      "loss": 0.208,
      "step": 527000
    },
    {
      "epoch": 5.88,
      "eval_loss": 0.19850178062915802,
      "eval_runtime": 2.6174,
      "eval_samples_per_second": 877.592,
      "eval_steps_per_second": 13.754,
      "step": 527000
    },
    {
      "epoch": 5.88,
      "learning_rate": 7.942129215622125e-05,
      "loss": 0.2086,
      "step": 527500
    },
    {
      "epoch": 5.89,
      "learning_rate": 7.930555406824026e-05,
      "loss": 0.208,
      "step": 528000
    },
    {
      "epoch": 5.89,
      "eval_loss": 0.19715136289596558,
      "eval_runtime": 2.6066,
      "eval_samples_per_second": 881.233,
      "eval_steps_per_second": 13.811,
      "step": 528000
    },
    {
      "epoch": 5.89,
      "learning_rate": 7.91898178788479e-05,
      "loss": 0.2078,
      "step": 528500
    },
    {
      "epoch": 5.9,
      "learning_rate": 7.907408390446254e-05,
      "loss": 0.2085,
      "step": 529000
    },
    {
      "epoch": 5.9,
      "eval_loss": 0.1981622874736786,
      "eval_runtime": 2.6077,
      "eval_samples_per_second": 880.843,
      "eval_steps_per_second": 13.805,
      "step": 529000
    },
    {
      "epoch": 5.91,
      "learning_rate": 7.895835246149643e-05,
      "loss": 0.2077,
      "step": 529500
    },
    {
      "epoch": 5.91,
      "learning_rate": 7.884262386635489e-05,
      "loss": 0.2081,
      "step": 530000
    },
    {
      "epoch": 5.91,
      "eval_loss": 0.19613459706306458,
      "eval_runtime": 2.6351,
      "eval_samples_per_second": 871.683,
      "eval_steps_per_second": 13.662,
      "step": 530000
    },
    {
      "epoch": 5.92,
      "learning_rate": 7.872689843543554e-05,
      "loss": 0.208,
      "step": 530500
    },
    {
      "epoch": 5.92,
      "learning_rate": 7.861117648512725e-05,
      "loss": 0.2078,
      "step": 531000
    },
    {
      "epoch": 5.92,
      "eval_loss": 0.19699302315711975,
      "eval_runtime": 2.6704,
      "eval_samples_per_second": 860.181,
      "eval_steps_per_second": 13.481,
      "step": 531000
    },
    {
      "epoch": 5.93,
      "learning_rate": 7.849545833180941e-05,
      "loss": 0.2078,
      "step": 531500
    },
    {
      "epoch": 5.93,
      "learning_rate": 7.837974429185103e-05,
      "loss": 0.2084,
      "step": 532000
    },
    {
      "epoch": 5.93,
      "eval_loss": 0.19729405641555786,
      "eval_runtime": 2.4788,
      "eval_samples_per_second": 926.65,
      "eval_steps_per_second": 14.523,
      "step": 532000
    },
    {
      "epoch": 5.94,
      "learning_rate": 7.82640346816099e-05,
      "loss": 0.2078,
      "step": 532500
    },
    {
      "epoch": 5.94,
      "learning_rate": 7.814832981743164e-05,
      "loss": 0.2075,
      "step": 533000
    },
    {
      "epoch": 5.94,
      "eval_loss": 0.1986730396747589,
      "eval_runtime": 2.5108,
      "eval_samples_per_second": 914.858,
      "eval_steps_per_second": 14.338,
      "step": 533000
    },
    {
      "epoch": 5.95,
      "learning_rate": 7.803263001564899e-05,
      "loss": 0.2079,
      "step": 533500
    },
    {
      "epoch": 5.96,
      "learning_rate": 7.791693559258072e-05,
      "loss": 0.2073,
      "step": 534000
    },
    {
      "epoch": 5.96,
      "eval_loss": 0.19755910336971283,
      "eval_runtime": 2.4538,
      "eval_samples_per_second": 936.101,
      "eval_steps_per_second": 14.671,
      "step": 534000
    },
    {
      "epoch": 5.96,
      "learning_rate": 7.780124686453101e-05,
      "loss": 0.208,
      "step": 534500
    },
    {
      "epoch": 5.97,
      "learning_rate": 7.768556414778842e-05,
      "loss": 0.2076,
      "step": 535000
    },
    {
      "epoch": 5.97,
      "eval_loss": 0.19874924421310425,
      "eval_runtime": 2.4792,
      "eval_samples_per_second": 926.525,
      "eval_steps_per_second": 14.521,
      "step": 535000
    },
    {
      "epoch": 5.97,
      "learning_rate": 7.756988775862508e-05,
      "loss": 0.2076,
      "step": 535500
    },
    {
      "epoch": 5.98,
      "learning_rate": 7.74542180132958e-05,
      "loss": 0.2076,
      "step": 536000
    },
    {
      "epoch": 5.98,
      "eval_loss": 0.1981775462627411,
      "eval_runtime": 2.5707,
      "eval_samples_per_second": 893.538,
      "eval_steps_per_second": 14.004,
      "step": 536000
    },
    {
      "epoch": 5.98,
      "learning_rate": 7.733855522803725e-05,
      "loss": 0.2075,
      "step": 536500
    },
    {
      "epoch": 5.99,
      "learning_rate": 7.72228997190671e-05,
      "loss": 0.2074,
      "step": 537000
    },
    {
      "epoch": 5.99,
      "eval_loss": 0.19704852998256683,
      "eval_runtime": 2.4868,
      "eval_samples_per_second": 923.663,
      "eval_steps_per_second": 14.476,
      "step": 537000
    },
    {
      "epoch": 5.99,
      "learning_rate": 7.710725180258306e-05,
      "loss": 0.2077,
      "step": 537500
    },
    {
      "epoch": 6.0,
      "learning_rate": 7.699161179476217e-05,
      "loss": 0.2073,
      "step": 538000
    },
    {
      "epoch": 6.0,
      "eval_loss": 0.19944582879543304,
      "eval_runtime": 2.6063,
      "eval_samples_per_second": 881.323,
      "eval_steps_per_second": 13.813,
      "step": 538000
    },
    {
      "epoch": 6.01,
      "learning_rate": 7.687598001175972e-05,
      "loss": 0.2075,
      "step": 538500
    },
    {
      "epoch": 6.01,
      "learning_rate": 7.676035676970863e-05,
      "loss": 0.207,
      "step": 539000
    },
    {
      "epoch": 6.01,
      "eval_loss": 0.19753174483776093,
      "eval_runtime": 2.6631,
      "eval_samples_per_second": 862.522,
      "eval_steps_per_second": 13.518,
      "step": 539000
    },
    {
      "epoch": 6.02,
      "learning_rate": 7.664474238471844e-05,
      "loss": 0.2071,
      "step": 539500
    },
    {
      "epoch": 6.02,
      "learning_rate": 7.652913717287443e-05,
      "loss": 0.2069,
      "step": 540000
    },
    {
      "epoch": 6.02,
      "eval_loss": 0.19679389894008636,
      "eval_runtime": 2.5175,
      "eval_samples_per_second": 912.43,
      "eval_steps_per_second": 14.3,
      "step": 540000
    },
    {
      "epoch": 6.03,
      "learning_rate": 7.641354145023687e-05,
      "loss": 0.2072,
      "step": 540500
    },
    {
      "epoch": 6.03,
      "learning_rate": 7.629795553284005e-05,
      "loss": 0.2071,
      "step": 541000
    },
    {
      "epoch": 6.03,
      "eval_loss": 0.19721636176109314,
      "eval_runtime": 2.791,
      "eval_samples_per_second": 822.988,
      "eval_steps_per_second": 12.898,
      "step": 541000
    },
    {
      "epoch": 6.04,
      "learning_rate": 7.61823797366914e-05,
      "loss": 0.2071,
      "step": 541500
    },
    {
      "epoch": 6.04,
      "learning_rate": 7.606681437777081e-05,
      "loss": 0.2067,
      "step": 542000
    },
    {
      "epoch": 6.04,
      "eval_loss": 0.19431854784488678,
      "eval_runtime": 2.5424,
      "eval_samples_per_second": 903.464,
      "eval_steps_per_second": 14.16,
      "step": 542000
    },
    {
      "epoch": 6.05,
      "learning_rate": 7.595125977202952e-05,
      "loss": 0.207,
      "step": 542500
    },
    {
      "epoch": 6.06,
      "learning_rate": 7.583571623538939e-05,
      "loss": 0.2073,
      "step": 543000
    },
    {
      "epoch": 6.06,
      "eval_loss": 0.1961280107498169,
      "eval_runtime": 2.5047,
      "eval_samples_per_second": 917.065,
      "eval_steps_per_second": 14.373,
      "step": 543000
    },
    {
      "epoch": 6.06,
      "learning_rate": 7.572018408374208e-05,
      "loss": 0.2069,
      "step": 543500
    },
    {
      "epoch": 6.07,
      "learning_rate": 7.560466363294806e-05,
      "loss": 0.2068,
      "step": 544000
    },
    {
      "epoch": 6.07,
      "eval_loss": 0.19669890403747559,
      "eval_runtime": 2.4907,
      "eval_samples_per_second": 922.238,
      "eval_steps_per_second": 14.454,
      "step": 544000
    },
    {
      "epoch": 6.07,
      "learning_rate": 7.548915519883582e-05,
      "loss": 0.2066,
      "step": 544500
    },
    {
      "epoch": 6.08,
      "learning_rate": 7.537365909720104e-05,
      "loss": 0.2065,
      "step": 545000
    },
    {
      "epoch": 6.08,
      "eval_loss": 0.1973814219236374,
      "eval_runtime": 2.6437,
      "eval_samples_per_second": 868.843,
      "eval_steps_per_second": 13.617,
      "step": 545000
    },
    {
      "epoch": 6.08,
      "learning_rate": 7.525817564380562e-05,
      "loss": 0.2068,
      "step": 545500
    },
    {
      "epoch": 6.09,
      "learning_rate": 7.514270515437691e-05,
      "loss": 0.2062,
      "step": 546000
    },
    {
      "epoch": 6.09,
      "eval_loss": 0.19645407795906067,
      "eval_runtime": 2.6493,
      "eval_samples_per_second": 867.035,
      "eval_steps_per_second": 13.589,
      "step": 546000
    },
    {
      "epoch": 6.1,
      "learning_rate": 7.502724794460685e-05,
      "loss": 0.2066,
      "step": 546500
    },
    {
      "epoch": 6.1,
      "learning_rate": 7.491180433015101e-05,
      "loss": 0.2062,
      "step": 547000
    },
    {
      "epoch": 6.1,
      "eval_loss": 0.19736704230308533,
      "eval_runtime": 2.5133,
      "eval_samples_per_second": 913.953,
      "eval_steps_per_second": 14.324,
      "step": 547000
    },
    {
      "epoch": 6.11,
      "learning_rate": 7.479637462662786e-05,
      "loss": 0.2065,
      "step": 547500
    },
    {
      "epoch": 6.11,
      "learning_rate": 7.468095914961777e-05,
      "loss": 0.2063,
      "step": 548000
    },
    {
      "epoch": 6.11,
      "eval_loss": 0.19768109917640686,
      "eval_runtime": 2.5796,
      "eval_samples_per_second": 890.452,
      "eval_steps_per_second": 13.956,
      "step": 548000
    },
    {
      "epoch": 6.12,
      "learning_rate": 7.456555821466225e-05,
      "loss": 0.2063,
      "step": 548500
    },
    {
      "epoch": 6.12,
      "learning_rate": 7.445017213726307e-05,
      "loss": 0.2068,
      "step": 549000
    },
    {
      "epoch": 6.12,
      "eval_loss": 0.19542047381401062,
      "eval_runtime": 2.4958,
      "eval_samples_per_second": 920.364,
      "eval_steps_per_second": 14.425,
      "step": 549000
    },
    {
      "epoch": 6.13,
      "learning_rate": 7.433480123288138e-05,
      "loss": 0.2065,
      "step": 549500
    },
    {
      "epoch": 6.13,
      "learning_rate": 7.421944581693674e-05,
      "loss": 0.2062,
      "step": 550000
    },
    {
      "epoch": 6.13,
      "eval_loss": 0.19721828401088715,
      "eval_runtime": 2.5551,
      "eval_samples_per_second": 898.971,
      "eval_steps_per_second": 14.089,
      "step": 550000
    },
    {
      "epoch": 6.14,
      "learning_rate": 7.410410620480651e-05,
      "loss": 0.2065,
      "step": 550500
    },
    {
      "epoch": 6.15,
      "learning_rate": 7.39887827118248e-05,
      "loss": 0.2063,
      "step": 551000
    },
    {
      "epoch": 6.15,
      "eval_loss": 0.1966421753168106,
      "eval_runtime": 2.4465,
      "eval_samples_per_second": 938.899,
      "eval_steps_per_second": 14.715,
      "step": 551000
    },
    {
      "epoch": 6.15,
      "learning_rate": 7.38734756532816e-05,
      "loss": 0.2062,
      "step": 551500
    },
    {
      "epoch": 6.16,
      "learning_rate": 7.375818534442207e-05,
      "loss": 0.2063,
      "step": 552000
    },
    {
      "epoch": 6.16,
      "eval_loss": 0.19533967971801758,
      "eval_runtime": 2.4902,
      "eval_samples_per_second": 922.429,
      "eval_steps_per_second": 14.457,
      "step": 552000
    },
    {
      "epoch": 6.16,
      "learning_rate": 7.364291210044542e-05,
      "loss": 0.2058,
      "step": 552500
    },
    {
      "epoch": 6.17,
      "learning_rate": 7.352765623650435e-05,
      "loss": 0.2061,
      "step": 553000
    },
    {
      "epoch": 6.17,
      "eval_loss": 0.1968429684638977,
      "eval_runtime": 2.5279,
      "eval_samples_per_second": 908.659,
      "eval_steps_per_second": 14.241,
      "step": 553000
    },
    {
      "epoch": 6.17,
      "learning_rate": 7.341241806770399e-05,
      "loss": 0.2064,
      "step": 553500
    },
    {
      "epoch": 6.18,
      "learning_rate": 7.329719790910108e-05,
      "loss": 0.2056,
      "step": 554000
    },
    {
      "epoch": 6.18,
      "eval_loss": 0.19725964963436127,
      "eval_runtime": 2.5134,
      "eval_samples_per_second": 913.905,
      "eval_steps_per_second": 14.323,
      "step": 554000
    },
    {
      "epoch": 6.18,
      "learning_rate": 7.318199607570318e-05,
      "loss": 0.2057,
      "step": 554500
    },
    {
      "epoch": 6.19,
      "learning_rate": 7.30668128824676e-05,
      "loss": 0.2061,
      "step": 555000
    },
    {
      "epoch": 6.19,
      "eval_loss": 0.19744634628295898,
      "eval_runtime": 2.5342,
      "eval_samples_per_second": 906.409,
      "eval_steps_per_second": 14.206,
      "step": 555000
    },
    {
      "epoch": 6.2,
      "learning_rate": 7.295164864430088e-05,
      "loss": 0.2056,
      "step": 555500
    },
    {
      "epoch": 6.2,
      "learning_rate": 7.283650367605764e-05,
      "loss": 0.2062,
      "step": 556000
    },
    {
      "epoch": 6.2,
      "eval_loss": 0.19634658098220825,
      "eval_runtime": 2.4766,
      "eval_samples_per_second": 927.495,
      "eval_steps_per_second": 14.536,
      "step": 556000
    },
    {
      "epoch": 6.21,
      "learning_rate": 7.272137829253983e-05,
      "loss": 0.2059,
      "step": 556500
    },
    {
      "epoch": 6.21,
      "learning_rate": 7.260627280849581e-05,
      "loss": 0.2061,
      "step": 557000
    },
    {
      "epoch": 6.21,
      "eval_loss": 0.19559474289417267,
      "eval_runtime": 2.4504,
      "eval_samples_per_second": 937.415,
      "eval_steps_per_second": 14.692,
      "step": 557000
    },
    {
      "epoch": 6.22,
      "learning_rate": 7.249118753861958e-05,
      "loss": 0.2056,
      "step": 557500
    },
    {
      "epoch": 6.22,
      "learning_rate": 7.23761227975499e-05,
      "loss": 0.2055,
      "step": 558000
    },
    {
      "epoch": 6.22,
      "eval_loss": 0.1950286328792572,
      "eval_runtime": 2.5356,
      "eval_samples_per_second": 905.898,
      "eval_steps_per_second": 14.198,
      "step": 558000
    },
    {
      "epoch": 6.23,
      "learning_rate": 7.22610788998694e-05,
      "loss": 0.2056,
      "step": 558500
    },
    {
      "epoch": 6.23,
      "learning_rate": 7.21460561601037e-05,
      "loss": 0.2055,
      "step": 559000
    },
    {
      "epoch": 6.23,
      "eval_loss": 0.19409525394439697,
      "eval_runtime": 2.5975,
      "eval_samples_per_second": 884.304,
      "eval_steps_per_second": 13.859,
      "step": 559000
    },
    {
      "epoch": 6.24,
      "learning_rate": 7.203105489272053e-05,
      "loss": 0.2055,
      "step": 559500
    },
    {
      "epoch": 6.25,
      "learning_rate": 7.191607541212897e-05,
      "loss": 0.2057,
      "step": 560000
    },
    {
      "epoch": 6.25,
      "eval_loss": 0.19587305188179016,
      "eval_runtime": 2.6038,
      "eval_samples_per_second": 882.169,
      "eval_steps_per_second": 13.826,
      "step": 560000
    },
    {
      "epoch": 6.25,
      "learning_rate": 7.180111803267856e-05,
      "loss": 0.2054,
      "step": 560500
    },
    {
      "epoch": 6.26,
      "learning_rate": 7.168618306865838e-05,
      "loss": 0.2051,
      "step": 561000
    },
    {
      "epoch": 6.26,
      "eval_loss": 0.195304736495018,
      "eval_runtime": 2.4332,
      "eval_samples_per_second": 944.014,
      "eval_steps_per_second": 14.795,
      "step": 561000
    },
    {
      "epoch": 6.26,
      "learning_rate": 7.157127083429626e-05,
      "loss": 0.2053,
      "step": 561500
    },
    {
      "epoch": 6.27,
      "learning_rate": 7.145638164375779e-05,
      "loss": 0.205,
      "step": 562000
    },
    {
      "epoch": 6.27,
      "eval_loss": 0.1958540827035904,
      "eval_runtime": 2.6408,
      "eval_samples_per_second": 869.814,
      "eval_steps_per_second": 13.632,
      "step": 562000
    },
    {
      "epoch": 6.27,
      "learning_rate": 7.134151581114565e-05,
      "loss": 0.2053,
      "step": 562500
    },
    {
      "epoch": 6.28,
      "learning_rate": 7.122667365049869e-05,
      "loss": 0.2052,
      "step": 563000
    },
    {
      "epoch": 6.28,
      "eval_loss": 0.19526307284832,
      "eval_runtime": 2.6193,
      "eval_samples_per_second": 876.966,
      "eval_steps_per_second": 13.744,
      "step": 563000
    },
    {
      "epoch": 6.28,
      "learning_rate": 7.111185547579099e-05,
      "loss": 0.205,
      "step": 563500
    },
    {
      "epoch": 6.29,
      "learning_rate": 7.099706160093098e-05,
      "loss": 0.2051,
      "step": 564000
    },
    {
      "epoch": 6.29,
      "eval_loss": 0.1962643265724182,
      "eval_runtime": 2.4959,
      "eval_samples_per_second": 920.299,
      "eval_steps_per_second": 14.423,
      "step": 564000
    },
    {
      "epoch": 6.3,
      "learning_rate": 7.08822923397608e-05,
      "loss": 0.2054,
      "step": 564500
    },
    {
      "epoch": 6.3,
      "learning_rate": 7.076754800605516e-05,
      "loss": 0.2053,
      "step": 565000
    },
    {
      "epoch": 6.3,
      "eval_loss": 0.19500210881233215,
      "eval_runtime": 2.6355,
      "eval_samples_per_second": 871.546,
      "eval_steps_per_second": 13.659,
      "step": 565000
    },
    {
      "epoch": 6.31,
      "learning_rate": 7.065282891352078e-05,
      "loss": 0.2049,
      "step": 565500
    },
    {
      "epoch": 6.31,
      "learning_rate": 7.053813537579523e-05,
      "loss": 0.2052,
      "step": 566000
    },
    {
      "epoch": 6.31,
      "eval_loss": 0.1964665800333023,
      "eval_runtime": 2.6178,
      "eval_samples_per_second": 877.444,
      "eval_steps_per_second": 13.752,
      "step": 566000
    },
    {
      "epoch": 6.32,
      "learning_rate": 7.042346770644624e-05,
      "loss": 0.2046,
      "step": 566500
    },
    {
      "epoch": 6.32,
      "learning_rate": 7.030882621897088e-05,
      "loss": 0.2046,
      "step": 567000
    },
    {
      "epoch": 6.32,
      "eval_loss": 0.19378143548965454,
      "eval_runtime": 2.6471,
      "eval_samples_per_second": 867.729,
      "eval_steps_per_second": 13.6,
      "step": 567000
    },
    {
      "epoch": 6.33,
      "learning_rate": 7.019421122679455e-05,
      "loss": 0.2052,
      "step": 567500
    },
    {
      "epoch": 6.33,
      "learning_rate": 7.00796230432703e-05,
      "loss": 0.2045,
      "step": 568000
    },
    {
      "epoch": 6.33,
      "eval_loss": 0.1938391774892807,
      "eval_runtime": 2.5793,
      "eval_samples_per_second": 890.552,
      "eval_steps_per_second": 13.957,
      "step": 568000
    },
    {
      "epoch": 6.34,
      "learning_rate": 6.996506198167789e-05,
      "loss": 0.2046,
      "step": 568500
    },
    {
      "epoch": 6.35,
      "learning_rate": 6.985052835522279e-05,
      "loss": 0.2045,
      "step": 569000
    },
    {
      "epoch": 6.35,
      "eval_loss": 0.19408397376537323,
      "eval_runtime": 2.5021,
      "eval_samples_per_second": 918.029,
      "eval_steps_per_second": 14.388,
      "step": 569000
    },
    {
      "epoch": 6.35,
      "learning_rate": 6.973602247703561e-05,
      "loss": 0.2047,
      "step": 569500
    },
    {
      "epoch": 6.36,
      "learning_rate": 6.962154466017105e-05,
      "loss": 0.2047,
      "step": 570000
    },
    {
      "epoch": 6.36,
      "eval_loss": 0.19305509328842163,
      "eval_runtime": 2.5881,
      "eval_samples_per_second": 887.536,
      "eval_steps_per_second": 13.91,
      "step": 570000
    },
    {
      "epoch": 6.01,
      "learning_rate": 6.950709521760712e-05,
      "loss": 0.2048,
      "step": 570500
    },
    {
      "epoch": 6.01,
      "learning_rate": 6.939267446224418e-05,
      "loss": 0.2046,
      "step": 571000
    },
    {
      "epoch": 6.01,
      "eval_loss": 0.19425606727600098,
      "eval_runtime": 2.5589,
      "eval_samples_per_second": 897.645,
      "eval_steps_per_second": 14.068,
      "step": 571000
    },
    {
      "epoch": 6.02,
      "learning_rate": 6.927828270690422e-05,
      "loss": 0.2047,
      "step": 571500
    },
    {
      "epoch": 6.02,
      "learning_rate": 6.91639202643299e-05,
      "loss": 0.2042,
      "step": 572000
    },
    {
      "epoch": 6.02,
      "eval_loss": 0.19538278877735138,
      "eval_runtime": 2.5656,
      "eval_samples_per_second": 895.308,
      "eval_steps_per_second": 14.032,
      "step": 572000
    },
    {
      "epoch": 6.03,
      "learning_rate": 6.904958744718383e-05,
      "loss": 0.2046,
      "step": 572500
    },
    {
      "epoch": 6.03,
      "learning_rate": 6.893528456804756e-05,
      "loss": 0.2042,
      "step": 573000
    },
    {
      "epoch": 6.03,
      "eval_loss": 0.19409753382205963,
      "eval_runtime": 2.5486,
      "eval_samples_per_second": 901.27,
      "eval_steps_per_second": 14.125,
      "step": 573000
    },
    {
      "epoch": 6.04,
      "learning_rate": 6.882101193942075e-05,
      "loss": 0.2042,
      "step": 573500
    },
    {
      "epoch": 6.04,
      "learning_rate": 6.870676987372044e-05,
      "loss": 0.2041,
      "step": 574000
    },
    {
      "epoch": 6.04,
      "eval_loss": 0.1939525604248047,
      "eval_runtime": 2.5437,
      "eval_samples_per_second": 903.03,
      "eval_steps_per_second": 14.153,
      "step": 574000
    },
    {
      "epoch": 6.05,
      "learning_rate": 6.859255868328003e-05,
      "loss": 0.2039,
      "step": 574500
    },
    {
      "epoch": 6.06,
      "learning_rate": 6.847837868034861e-05,
      "loss": 0.2042,
      "step": 575000
    },
    {
      "epoch": 6.06,
      "eval_loss": 0.1951504349708557,
      "eval_runtime": 2.5798,
      "eval_samples_per_second": 890.377,
      "eval_steps_per_second": 13.955,
      "step": 575000
    },
    {
      "epoch": 6.06,
      "learning_rate": 6.836423017708996e-05,
      "loss": 0.2038,
      "step": 575500
    },
    {
      "epoch": 6.07,
      "learning_rate": 6.825011348558167e-05,
      "loss": 0.204,
      "step": 576000
    },
    {
      "epoch": 6.07,
      "eval_loss": 0.19505272805690765,
      "eval_runtime": 2.5475,
      "eval_samples_per_second": 901.67,
      "eval_steps_per_second": 14.132,
      "step": 576000
    },
    {
      "epoch": 6.07,
      "learning_rate": 6.813602891781443e-05,
      "loss": 0.2039,
      "step": 576500
    },
    {
      "epoch": 6.08,
      "learning_rate": 6.802197678569109e-05,
      "loss": 0.2038,
      "step": 577000
    },
    {
      "epoch": 6.08,
      "eval_loss": 0.19440634548664093,
      "eval_runtime": 2.537,
      "eval_samples_per_second": 905.395,
      "eval_steps_per_second": 14.19,
      "step": 577000
    },
    {
      "epoch": 6.08,
      "learning_rate": 6.790795740102589e-05,
      "loss": 0.2038,
      "step": 577500
    },
    {
      "epoch": 6.09,
      "learning_rate": 6.779397107554339e-05,
      "loss": 0.2038,
      "step": 578000
    },
    {
      "epoch": 6.09,
      "eval_loss": 0.19268804788589478,
      "eval_runtime": 2.5143,
      "eval_samples_per_second": 913.559,
      "eval_steps_per_second": 14.318,
      "step": 578000
    },
    {
      "epoch": 6.09,
      "learning_rate": 6.768001812087789e-05,
      "loss": 0.2038,
      "step": 578500
    },
    {
      "epoch": 6.1,
      "learning_rate": 6.756609884857239e-05,
      "loss": 0.2037,
      "step": 579000
    },
    {
      "epoch": 6.1,
      "eval_loss": 0.19323283433914185,
      "eval_runtime": 2.5075,
      "eval_samples_per_second": 916.04,
      "eval_steps_per_second": 14.357,
      "step": 579000
    },
    {
      "epoch": 6.11,
      "learning_rate": 6.745221357007786e-05,
      "loss": 0.2037,
      "step": 579500
    },
    {
      "epoch": 6.11,
      "learning_rate": 6.733836259675233e-05,
      "loss": 0.2036,
      "step": 580000
    },
    {
      "epoch": 6.11,
      "eval_loss": 0.19253070652484894,
      "eval_runtime": 2.4507,
      "eval_samples_per_second": 937.277,
      "eval_steps_per_second": 14.69,
      "step": 580000
    },
    {
      "epoch": 6.12,
      "learning_rate": 6.722454623985994e-05,
      "loss": 0.2036,
      "step": 580500
    },
    {
      "epoch": 6.12,
      "learning_rate": 6.71107648105703e-05,
      "loss": 0.2038,
      "step": 581000
    },
    {
      "epoch": 6.12,
      "eval_loss": 0.1942874640226364,
      "eval_runtime": 2.6604,
      "eval_samples_per_second": 863.404,
      "eval_steps_per_second": 13.532,
      "step": 581000
    },
    {
      "epoch": 6.13,
      "learning_rate": 6.69970186199575e-05,
      "loss": 0.204,
      "step": 581500
    },
    {
      "epoch": 6.13,
      "learning_rate": 6.688330797899925e-05,
      "loss": 0.2036,
      "step": 582000
    },
    {
      "epoch": 6.13,
      "eval_loss": 0.1936686784029007,
      "eval_runtime": 2.5783,
      "eval_samples_per_second": 890.887,
      "eval_steps_per_second": 13.963,
      "step": 582000
    },
    {
      "epoch": 6.14,
      "learning_rate": 6.676963319857618e-05,
      "loss": 0.2038,
      "step": 582500
    },
    {
      "epoch": 6.14,
      "learning_rate": 6.665599458947072e-05,
      "loss": 0.2035,
      "step": 583000
    },
    {
      "epoch": 6.14,
      "eval_loss": 0.1947861611843109,
      "eval_runtime": 2.5506,
      "eval_samples_per_second": 900.557,
      "eval_steps_per_second": 14.114,
      "step": 583000
    },
    {
      "epoch": 6.15,
      "learning_rate": 6.654239246236651e-05,
      "loss": 0.2037,
      "step": 583500
    },
    {
      "epoch": 6.16,
      "learning_rate": 6.642882712784742e-05,
      "loss": 0.2033,
      "step": 584000
    },
    {
      "epoch": 6.16,
      "eval_loss": 0.1927429735660553,
      "eval_runtime": 2.5138,
      "eval_samples_per_second": 913.772,
      "eval_steps_per_second": 14.321,
      "step": 584000
    },
    {
      "epoch": 6.16,
      "learning_rate": 6.631529889639679e-05,
      "loss": 0.2032,
      "step": 584500
    },
    {
      "epoch": 6.17,
      "learning_rate": 6.620180807839639e-05,
      "loss": 0.2033,
      "step": 585000
    },
    {
      "epoch": 6.17,
      "eval_loss": 0.19153708219528198,
      "eval_runtime": 2.5442,
      "eval_samples_per_second": 902.822,
      "eval_steps_per_second": 14.15,
      "step": 585000
    },
    {
      "epoch": 6.17,
      "learning_rate": 6.608835498412583e-05,
      "loss": 0.2034,
      "step": 585500
    },
    {
      "epoch": 6.18,
      "learning_rate": 6.597493992376152e-05,
      "loss": 0.2031,
      "step": 586000
    },
    {
      "epoch": 6.18,
      "eval_loss": 0.1941889226436615,
      "eval_runtime": 2.5658,
      "eval_samples_per_second": 895.227,
      "eval_steps_per_second": 14.031,
      "step": 586000
    },
    {
      "epoch": 6.18,
      "learning_rate": 6.586156320737592e-05,
      "loss": 0.2032,
      "step": 586500
    },
    {
      "epoch": 6.19,
      "learning_rate": 6.574822514493664e-05,
      "loss": 0.2029,
      "step": 587000
    },
    {
      "epoch": 6.19,
      "eval_loss": 0.19126850366592407,
      "eval_runtime": 2.5788,
      "eval_samples_per_second": 890.72,
      "eval_steps_per_second": 13.96,
      "step": 587000
    },
    {
      "epoch": 6.2,
      "learning_rate": 6.56349260463056e-05,
      "loss": 0.2026,
      "step": 587500
    },
    {
      "epoch": 6.2,
      "learning_rate": 6.552166622123824e-05,
      "loss": 0.2031,
      "step": 588000
    },
    {
      "epoch": 6.2,
      "eval_loss": 0.19138653576374054,
      "eval_runtime": 2.5292,
      "eval_samples_per_second": 908.193,
      "eval_steps_per_second": 14.234,
      "step": 588000
    },
    {
      "epoch": 6.21,
      "learning_rate": 6.540844597938256e-05,
      "loss": 0.203,
      "step": 588500
    },
    {
      "epoch": 6.21,
      "learning_rate": 6.52952656302784e-05,
      "loss": 0.203,
      "step": 589000
    },
    {
      "epoch": 6.21,
      "eval_loss": 0.1942255049943924,
      "eval_runtime": 2.5063,
      "eval_samples_per_second": 916.493,
      "eval_steps_per_second": 14.364,
      "step": 589000
    },
    {
      "epoch": 6.22,
      "learning_rate": 6.518212548335651e-05,
      "loss": 0.2027,
      "step": 589500
    },
    {
      "epoch": 6.22,
      "learning_rate": 6.506902584793773e-05,
      "loss": 0.2032,
      "step": 590000
    },
    {
      "epoch": 6.22,
      "eval_loss": 0.1933428943157196,
      "eval_runtime": 2.5059,
      "eval_samples_per_second": 916.62,
      "eval_steps_per_second": 14.366,
      "step": 590000
    },
    {
      "epoch": 6.23,
      "learning_rate": 6.495596703323214e-05,
      "loss": 0.203,
      "step": 590500
    },
    {
      "epoch": 6.23,
      "learning_rate": 6.484294934833822e-05,
      "loss": 0.203,
      "step": 591000
    },
    {
      "epoch": 6.23,
      "eval_loss": 0.19498485326766968,
      "eval_runtime": 2.4612,
      "eval_samples_per_second": 933.27,
      "eval_steps_per_second": 14.627,
      "step": 591000
    },
    {
      "epoch": 6.24,
      "learning_rate": 6.472997310224204e-05,
      "loss": 0.2028,
      "step": 591500
    },
    {
      "epoch": 6.25,
      "learning_rate": 6.461703860381628e-05,
      "loss": 0.2029,
      "step": 592000
    },
    {
      "epoch": 6.25,
      "eval_loss": 0.19485318660736084,
      "eval_runtime": 2.5316,
      "eval_samples_per_second": 907.333,
      "eval_steps_per_second": 14.22,
      "step": 592000
    },
    {
      "epoch": 6.25,
      "learning_rate": 6.450414616181959e-05,
      "loss": 0.2027,
      "step": 592500
    },
    {
      "epoch": 6.26,
      "learning_rate": 6.439129608489559e-05,
      "loss": 0.2023,
      "step": 593000
    },
    {
      "epoch": 6.26,
      "eval_loss": 0.19170017540454865,
      "eval_runtime": 2.5465,
      "eval_samples_per_second": 902.036,
      "eval_steps_per_second": 14.137,
      "step": 593000
    },
    {
      "epoch": 6.26,
      "learning_rate": 6.427848868157208e-05,
      "loss": 0.2021,
      "step": 593500
    },
    {
      "epoch": 6.27,
      "learning_rate": 6.41657242602602e-05,
      "loss": 0.2024,
      "step": 594000
    },
    {
      "epoch": 6.27,
      "eval_loss": 0.19145548343658447,
      "eval_runtime": 2.5956,
      "eval_samples_per_second": 884.955,
      "eval_steps_per_second": 13.87,
      "step": 594000
    },
    {
      "epoch": 6.27,
      "learning_rate": 6.405300312925353e-05,
      "loss": 0.2026,
      "step": 594500
    },
    {
      "epoch": 6.28,
      "learning_rate": 6.39403255967274e-05,
      "loss": 0.2021,
      "step": 595000
    },
    {
      "epoch": 6.28,
      "eval_loss": 0.19004245102405548,
      "eval_runtime": 2.516,
      "eval_samples_per_second": 912.956,
      "eval_steps_per_second": 14.308,
      "step": 595000
    },
    {
      "epoch": 6.28,
      "learning_rate": 6.382769197073783e-05,
      "loss": 0.2023,
      "step": 595500
    },
    {
      "epoch": 6.29,
      "learning_rate": 6.371510255922088e-05,
      "loss": 0.2024,
      "step": 596000
    },
    {
      "epoch": 6.29,
      "eval_loss": 0.19270072877407074,
      "eval_runtime": 2.4967,
      "eval_samples_per_second": 920.021,
      "eval_steps_per_second": 14.419,
      "step": 596000
    },
    {
      "epoch": 6.3,
      "learning_rate": 6.360255766999172e-05,
      "loss": 0.2025,
      "step": 596500
    },
    {
      "epoch": 6.3,
      "learning_rate": 6.349005761074372e-05,
      "loss": 0.2027,
      "step": 597000
    },
    {
      "epoch": 6.3,
      "eval_loss": 0.18951117992401123,
      "eval_runtime": 2.5319,
      "eval_samples_per_second": 907.212,
      "eval_steps_per_second": 14.218,
      "step": 597000
    },
    {
      "epoch": 6.31,
      "learning_rate": 6.33776026890478e-05,
      "loss": 0.2019,
      "step": 597500
    },
    {
      "epoch": 6.31,
      "learning_rate": 6.326519321235139e-05,
      "loss": 0.2025,
      "step": 598000
    },
    {
      "epoch": 6.31,
      "eval_loss": 0.19144007563591003,
      "eval_runtime": 2.4829,
      "eval_samples_per_second": 925.137,
      "eval_steps_per_second": 14.499,
      "step": 598000
    },
    {
      "epoch": 6.32,
      "learning_rate": 6.315282948797776e-05,
      "loss": 0.2022,
      "step": 598500
    },
    {
      "epoch": 6.32,
      "learning_rate": 6.304051182312496e-05,
      "loss": 0.2021,
      "step": 599000
    },
    {
      "epoch": 6.32,
      "eval_loss": 0.18959102034568787,
      "eval_runtime": 2.5681,
      "eval_samples_per_second": 894.45,
      "eval_steps_per_second": 14.018,
      "step": 599000
    },
    {
      "epoch": 6.33,
      "learning_rate": 6.292824052486525e-05,
      "loss": 0.2023,
      "step": 599500
    },
    {
      "epoch": 6.33,
      "learning_rate": 6.281601590014407e-05,
      "loss": 0.2018,
      "step": 600000
    },
    {
      "epoch": 6.33,
      "eval_loss": 0.1930903047323227,
      "eval_runtime": 2.4861,
      "eval_samples_per_second": 923.926,
      "eval_steps_per_second": 14.48,
      "step": 600000
    },
    {
      "epoch": 6.34,
      "learning_rate": 6.270383825577923e-05,
      "loss": 0.2017,
      "step": 600500
    },
    {
      "epoch": 6.35,
      "learning_rate": 6.259170789846017e-05,
      "loss": 0.2019,
      "step": 601000
    },
    {
      "epoch": 6.35,
      "eval_loss": 0.1939515918493271,
      "eval_runtime": 2.5332,
      "eval_samples_per_second": 906.761,
      "eval_steps_per_second": 14.211,
      "step": 601000
    },
    {
      "epoch": 6.35,
      "learning_rate": 6.247962513474697e-05,
      "loss": 0.2021,
      "step": 601500
    },
    {
      "epoch": 6.36,
      "learning_rate": 6.236759027106965e-05,
      "loss": 0.2019,
      "step": 602000
    },
    {
      "epoch": 6.36,
      "eval_loss": 0.192647323012352,
      "eval_runtime": 2.5565,
      "eval_samples_per_second": 898.498,
      "eval_steps_per_second": 14.082,
      "step": 602000
    },
    {
      "epoch": 6.36,
      "learning_rate": 6.225560361372722e-05,
      "loss": 0.202,
      "step": 602500
    },
    {
      "epoch": 6.37,
      "learning_rate": 6.214366546888694e-05,
      "loss": 0.2014,
      "step": 603000
    },
    {
      "epoch": 6.37,
      "eval_loss": 0.19305016100406647,
      "eval_runtime": 2.5415,
      "eval_samples_per_second": 903.802,
      "eval_steps_per_second": 14.165,
      "step": 603000
    },
    {
      "epoch": 6.37,
      "learning_rate": 6.203177614258345e-05,
      "loss": 0.2016,
      "step": 603500
    },
    {
      "epoch": 6.38,
      "learning_rate": 6.191993594071785e-05,
      "loss": 0.2021,
      "step": 604000
    },
    {
      "epoch": 6.38,
      "eval_loss": 0.19003823399543762,
      "eval_runtime": 2.6362,
      "eval_samples_per_second": 871.343,
      "eval_steps_per_second": 13.656,
      "step": 604000
    },
    {
      "epoch": 6.38,
      "learning_rate": 6.180814516905701e-05,
      "loss": 0.2019,
      "step": 604500
    },
    {
      "epoch": 6.39,
      "learning_rate": 6.169640413323262e-05,
      "loss": 0.2016,
      "step": 605000
    },
    {
      "epoch": 6.39,
      "eval_loss": 0.19246533513069153,
      "eval_runtime": 2.5718,
      "eval_samples_per_second": 893.139,
      "eval_steps_per_second": 13.998,
      "step": 605000
    },
    {
      "epoch": 6.4,
      "learning_rate": 6.158471313874041e-05,
      "loss": 0.2017,
      "step": 605500
    },
    {
      "epoch": 6.4,
      "learning_rate": 6.147307249093929e-05,
      "loss": 0.2016,
      "step": 606000
    },
    {
      "epoch": 6.4,
      "eval_loss": 0.19287154078483582,
      "eval_runtime": 2.4757,
      "eval_samples_per_second": 927.823,
      "eval_steps_per_second": 14.541,
      "step": 606000
    },
    {
      "epoch": 6.41,
      "learning_rate": 6.136148249505053e-05,
      "loss": 0.2016,
      "step": 606500
    },
    {
      "epoch": 6.41,
      "learning_rate": 6.124994345615693e-05,
      "loss": 0.2018,
      "step": 607000
    },
    {
      "epoch": 6.41,
      "eval_loss": 0.19156165421009064,
      "eval_runtime": 2.5771,
      "eval_samples_per_second": 891.317,
      "eval_steps_per_second": 13.969,
      "step": 607000
    },
    {
      "epoch": 6.42,
      "learning_rate": 6.113845567920194e-05,
      "loss": 0.201,
      "step": 607500
    },
    {
      "epoch": 6.42,
      "learning_rate": 6.102701946898891e-05,
      "loss": 0.2013,
      "step": 608000
    },
    {
      "epoch": 6.42,
      "eval_loss": 0.19082282483577728,
      "eval_runtime": 2.6476,
      "eval_samples_per_second": 867.578,
      "eval_steps_per_second": 13.597,
      "step": 608000
    },
    {
      "epoch": 6.43,
      "learning_rate": 6.0915635130180154e-05,
      "loss": 0.2013,
      "step": 608500
    },
    {
      "epoch": 6.43,
      "learning_rate": 6.0804302967296225e-05,
      "loss": 0.2019,
      "step": 609000
    },
    {
      "epoch": 6.43,
      "eval_loss": 0.19137336313724518,
      "eval_runtime": 2.6058,
      "eval_samples_per_second": 881.501,
      "eval_steps_per_second": 13.815,
      "step": 609000
    },
    {
      "epoch": 6.44,
      "learning_rate": 6.0693023284715e-05,
      "loss": 0.2014,
      "step": 609500
    },
    {
      "epoch": 6.45,
      "learning_rate": 6.058179638667089e-05,
      "loss": 0.201,
      "step": 610000
    },
    {
      "epoch": 6.45,
      "eval_loss": 0.19182415306568146,
      "eval_runtime": 2.5256,
      "eval_samples_per_second": 909.494,
      "eval_steps_per_second": 14.254,
      "step": 610000
    },
    {
      "epoch": 6.45,
      "learning_rate": 6.047062257725395e-05,
      "loss": 0.2011,
      "step": 610500
    },
    {
      "epoch": 6.46,
      "learning_rate": 6.035950216040917e-05,
      "loss": 0.2014,
      "step": 611000
    },
    {
      "epoch": 6.46,
      "eval_loss": 0.19371135532855988,
      "eval_runtime": 2.5383,
      "eval_samples_per_second": 904.933,
      "eval_steps_per_second": 14.183,
      "step": 611000
    },
    {
      "epoch": 6.46,
      "learning_rate": 6.0248435439935516e-05,
      "loss": 0.2018,
      "step": 611500
    },
    {
      "epoch": 6.47,
      "learning_rate": 6.0137422719485145e-05,
      "loss": 0.2009,
      "step": 612000
    },
    {
      "epoch": 6.47,
      "eval_loss": 0.18979497253894806,
      "eval_runtime": 2.61,
      "eval_samples_per_second": 880.076,
      "eval_steps_per_second": 13.793,
      "step": 612000
    },
    {
      "epoch": 6.47,
      "learning_rate": 6.0026464302562636e-05,
      "loss": 0.2009,
      "step": 612500
    },
    {
      "epoch": 6.48,
      "learning_rate": 5.991556049252401e-05,
      "loss": 0.2012,
      "step": 613000
    },
    {
      "epoch": 6.48,
      "eval_loss": 0.19082209467887878,
      "eval_runtime": 2.5296,
      "eval_samples_per_second": 908.057,
      "eval_steps_per_second": 14.232,
      "step": 613000
    },
    {
      "epoch": 6.49,
      "learning_rate": 5.980471159257609e-05,
      "loss": 0.2015,
      "step": 613500
    },
    {
      "epoch": 6.49,
      "learning_rate": 5.969391790577551e-05,
      "loss": 0.201,
      "step": 614000
    },
    {
      "epoch": 6.49,
      "eval_loss": 0.19177234172821045,
      "eval_runtime": 2.4578,
      "eval_samples_per_second": 934.56,
      "eval_steps_per_second": 14.647,
      "step": 614000
    },
    {
      "epoch": 6.5,
      "learning_rate": 5.958317973502798e-05,
      "loss": 0.2009,
      "step": 614500
    },
    {
      "epoch": 6.5,
      "learning_rate": 5.947249738308747e-05,
      "loss": 0.2009,
      "step": 615000
    },
    {
      "epoch": 6.5,
      "eval_loss": 0.18852779269218445,
      "eval_runtime": 2.5225,
      "eval_samples_per_second": 910.6,
      "eval_steps_per_second": 14.271,
      "step": 615000
    },
    {
      "epoch": 6.51,
      "learning_rate": 5.9361871152555254e-05,
      "loss": 0.2008,
      "step": 615500
    },
    {
      "epoch": 6.51,
      "learning_rate": 5.925130134587924e-05,
      "loss": 0.2007,
      "step": 616000
    },
    {
      "epoch": 6.51,
      "eval_loss": 0.18964450061321259,
      "eval_runtime": 2.6115,
      "eval_samples_per_second": 879.56,
      "eval_steps_per_second": 13.785,
      "step": 616000
    },
    {
      "epoch": 6.52,
      "learning_rate": 5.914078826535307e-05,
      "loss": 0.2009,
      "step": 616500
    },
    {
      "epoch": 6.52,
      "learning_rate": 5.903033221311528e-05,
      "loss": 0.2005,
      "step": 617000
    },
    {
      "epoch": 6.52,
      "eval_loss": 0.19303825497627258,
      "eval_runtime": 2.4819,
      "eval_samples_per_second": 925.494,
      "eval_steps_per_second": 14.505,
      "step": 617000
    },
    {
      "epoch": 6.53,
      "learning_rate": 5.891993349114847e-05,
      "loss": 0.201,
      "step": 617500
    },
    {
      "epoch": 6.54,
      "learning_rate": 5.880959240127858e-05,
      "loss": 0.2008,
      "step": 618000
    },
    {
      "epoch": 6.54,
      "eval_loss": 0.1906815767288208,
      "eval_runtime": 2.5316,
      "eval_samples_per_second": 907.334,
      "eval_steps_per_second": 14.22,
      "step": 618000
    },
    {
      "epoch": 6.54,
      "learning_rate": 5.86993092451739e-05,
      "loss": 0.2005,
      "step": 618500
    },
    {
      "epoch": 6.55,
      "learning_rate": 5.858908432434438e-05,
      "loss": 0.2002,
      "step": 619000
    },
    {
      "epoch": 6.55,
      "eval_loss": 0.18906356394290924,
      "eval_runtime": 2.5512,
      "eval_samples_per_second": 900.346,
      "eval_steps_per_second": 14.111,
      "step": 619000
    },
    {
      "epoch": 6.55,
      "learning_rate": 5.847891794014074e-05,
      "loss": 0.2007,
      "step": 619500
    },
    {
      "epoch": 6.56,
      "learning_rate": 5.8368810393753684e-05,
      "loss": 0.2003,
      "step": 620000
    },
    {
      "epoch": 6.56,
      "eval_loss": 0.19198235869407654,
      "eval_runtime": 2.5115,
      "eval_samples_per_second": 914.603,
      "eval_steps_per_second": 14.334,
      "step": 620000
    },
    {
      "epoch": 6.56,
      "learning_rate": 5.8258761986213015e-05,
      "loss": 0.2004,
      "step": 620500
    },
    {
      "epoch": 6.57,
      "learning_rate": 5.814877301838688e-05,
      "loss": 0.2002,
      "step": 621000
    },
    {
      "epoch": 6.57,
      "eval_loss": 0.19093768298625946,
      "eval_runtime": 2.5116,
      "eval_samples_per_second": 914.543,
      "eval_steps_per_second": 14.333,
      "step": 621000
    },
    {
      "epoch": 6.57,
      "learning_rate": 5.803884379098094e-05,
      "loss": 0.2005,
      "step": 621500
    },
    {
      "epoch": 6.58,
      "learning_rate": 5.7928974604537494e-05,
      "loss": 0.2003,
      "step": 622000
    },
    {
      "epoch": 6.58,
      "eval_loss": 0.19122566282749176,
      "eval_runtime": 2.5785,
      "eval_samples_per_second": 890.823,
      "eval_steps_per_second": 13.962,
      "step": 622000
    },
    {
      "epoch": 6.59,
      "learning_rate": 5.781916575943469e-05,
      "loss": 0.2,
      "step": 622500
    },
    {
      "epoch": 6.59,
      "learning_rate": 5.770941755588573e-05,
      "loss": 0.2005,
      "step": 623000
    },
    {
      "epoch": 6.59,
      "eval_loss": 0.19022968411445618,
      "eval_runtime": 2.5145,
      "eval_samples_per_second": 913.487,
      "eval_steps_per_second": 14.317,
      "step": 623000
    },
    {
      "epoch": 6.6,
      "learning_rate": 5.7599730293938e-05,
      "loss": 0.2005,
      "step": 623500
    },
    {
      "epoch": 6.6,
      "learning_rate": 5.749010427347233e-05,
      "loss": 0.1998,
      "step": 624000
    },
    {
      "epoch": 6.6,
      "eval_loss": 0.18952861428260803,
      "eval_runtime": 2.6789,
      "eval_samples_per_second": 857.447,
      "eval_steps_per_second": 13.438,
      "step": 624000
    },
    {
      "epoch": 6.61,
      "learning_rate": 5.738053979420199e-05,
      "loss": 0.2004,
      "step": 624500
    },
    {
      "epoch": 6.61,
      "learning_rate": 5.7271037155672156e-05,
      "loss": 0.2,
      "step": 625000
    },
    {
      "epoch": 6.61,
      "eval_loss": 0.18912938237190247,
      "eval_runtime": 2.5074,
      "eval_samples_per_second": 916.101,
      "eval_steps_per_second": 14.358,
      "step": 625000
    },
    {
      "epoch": 6.62,
      "learning_rate": 5.716159665725883e-05,
      "loss": 0.1998,
      "step": 625500
    },
    {
      "epoch": 6.62,
      "learning_rate": 5.7052218598168154e-05,
      "loss": 0.1996,
      "step": 626000
    },
    {
      "epoch": 6.62,
      "eval_loss": 0.19050002098083496,
      "eval_runtime": 2.5363,
      "eval_samples_per_second": 905.665,
      "eval_steps_per_second": 14.194,
      "step": 626000
    },
    {
      "epoch": 6.63,
      "learning_rate": 5.69429032774356e-05,
      "loss": 0.1999,
      "step": 626500
    },
    {
      "epoch": 6.64,
      "learning_rate": 5.6833650993925016e-05,
      "loss": 0.1998,
      "step": 627000
    },
    {
      "epoch": 6.64,
      "eval_loss": 0.19091765582561493,
      "eval_runtime": 2.6002,
      "eval_samples_per_second": 883.405,
      "eval_steps_per_second": 13.845,
      "step": 627000
    },
    {
      "epoch": 6.64,
      "learning_rate": 5.6724462046328025e-05,
      "loss": 0.2,
      "step": 627500
    },
    {
      "epoch": 6.65,
      "learning_rate": 5.661533673316303e-05,
      "loss": 0.2,
      "step": 628000
    },
    {
      "epoch": 6.65,
      "eval_loss": 0.19069619476795197,
      "eval_runtime": 2.549,
      "eval_samples_per_second": 901.121,
      "eval_steps_per_second": 14.123,
      "step": 628000
    },
    {
      "epoch": 6.65,
      "learning_rate": 5.6506275352774447e-05,
      "loss": 0.1998,
      "step": 628500
    },
    {
      "epoch": 6.66,
      "learning_rate": 5.639727820333198e-05,
      "loss": 0.2001,
      "step": 629000
    },
    {
      "epoch": 6.66,
      "eval_loss": 0.19251221418380737,
      "eval_runtime": 2.5685,
      "eval_samples_per_second": 894.303,
      "eval_steps_per_second": 14.016,
      "step": 629000
    },
    {
      "epoch": 6.66,
      "learning_rate": 5.62883455828296e-05,
      "loss": 0.1996,
      "step": 629500
    },
    {
      "epoch": 6.67,
      "learning_rate": 5.617947778908498e-05,
      "loss": 0.2,
      "step": 630000
    },
    {
      "epoch": 6.67,
      "eval_loss": 0.19011762738227844,
      "eval_runtime": 2.6047,
      "eval_samples_per_second": 881.86,
      "eval_steps_per_second": 13.821,
      "step": 630000
    },
    {
      "epoch": 6.67,
      "learning_rate": 5.60706751197385e-05,
      "loss": 0.1996,
      "step": 630500
    },
    {
      "epoch": 6.68,
      "learning_rate": 5.596193787225254e-05,
      "loss": 0.1998,
      "step": 631000
    },
    {
      "epoch": 6.68,
      "eval_loss": 0.19056497514247894,
      "eval_runtime": 2.5606,
      "eval_samples_per_second": 897.047,
      "eval_steps_per_second": 14.059,
      "step": 631000
    },
    {
      "epoch": 6.69,
      "learning_rate": 5.585326634391049e-05,
      "loss": 0.1993,
      "step": 631500
    },
    {
      "epoch": 6.69,
      "learning_rate": 5.574466083181624e-05,
      "loss": 0.1993,
      "step": 632000
    },
    {
      "epoch": 6.69,
      "eval_loss": 0.18938562273979187,
      "eval_runtime": 2.5436,
      "eval_samples_per_second": 903.047,
      "eval_steps_per_second": 14.153,
      "step": 632000
    },
    {
      "epoch": 6.7,
      "learning_rate": 5.563612163289308e-05,
      "loss": 0.1995,
      "step": 632500
    },
    {
      "epoch": 6.7,
      "learning_rate": 5.552764904388305e-05,
      "loss": 0.1992,
      "step": 633000
    },
    {
      "epoch": 6.7,
      "eval_loss": 0.18923692405223846,
      "eval_runtime": 2.5198,
      "eval_samples_per_second": 911.583,
      "eval_steps_per_second": 14.287,
      "step": 633000
    },
    {
      "epoch": 6.71,
      "learning_rate": 5.541924336134609e-05,
      "loss": 0.1998,
      "step": 633500
    },
    {
      "epoch": 6.71,
      "learning_rate": 5.5310904881659116e-05,
      "loss": 0.1991,
      "step": 634000
    },
    {
      "epoch": 6.71,
      "eval_loss": 0.18939541280269623,
      "eval_runtime": 2.5614,
      "eval_samples_per_second": 896.775,
      "eval_steps_per_second": 14.055,
      "step": 634000
    },
    {
      "epoch": 6.72,
      "learning_rate": 5.5202633901015464e-05,
      "loss": 0.1989,
      "step": 634500
    },
    {
      "epoch": 6.72,
      "learning_rate": 5.5094430715423835e-05,
      "loss": 0.1989,
      "step": 635000
    },
    {
      "epoch": 6.72,
      "eval_loss": 0.18934237957000732,
      "eval_runtime": 2.4647,
      "eval_samples_per_second": 931.977,
      "eval_steps_per_second": 14.607,
      "step": 635000
    },
    {
      "epoch": 6.73,
      "learning_rate": 5.4986295620707626e-05,
      "loss": 0.1993,
      "step": 635500
    },
    {
      "epoch": 6.74,
      "learning_rate": 5.487822891250406e-05,
      "loss": 0.1991,
      "step": 636000
    },
    {
      "epoch": 6.74,
      "eval_loss": 0.18926899135112762,
      "eval_runtime": 2.5769,
      "eval_samples_per_second": 891.373,
      "eval_steps_per_second": 13.97,
      "step": 636000
    },
    {
      "epoch": 6.74,
      "learning_rate": 5.477023088626334e-05,
      "loss": 0.1994,
      "step": 636500
    },
    {
      "epoch": 6.75,
      "learning_rate": 5.4662301837247985e-05,
      "loss": 0.1988,
      "step": 637000
    },
    {
      "epoch": 6.75,
      "eval_loss": 0.18590334057807922,
      "eval_runtime": 2.6094,
      "eval_samples_per_second": 880.292,
      "eval_steps_per_second": 13.796,
      "step": 637000
    },
    {
      "epoch": 6.75,
      "learning_rate": 5.45544420605319e-05,
      "loss": 0.1991,
      "step": 637500
    },
    {
      "epoch": 6.76,
      "learning_rate": 5.4446651850999604e-05,
      "loss": 0.1987,
      "step": 638000
    },
    {
      "epoch": 6.76,
      "eval_loss": 0.1874515414237976,
      "eval_runtime": 2.6122,
      "eval_samples_per_second": 879.338,
      "eval_steps_per_second": 13.782,
      "step": 638000
    },
    {
      "epoch": 6.76,
      "learning_rate": 5.433893150334538e-05,
      "loss": 0.1991,
      "step": 638500
    },
    {
      "epoch": 6.77,
      "learning_rate": 5.4231281312072544e-05,
      "loss": 0.1989,
      "step": 639000
    },
    {
      "epoch": 6.77,
      "eval_loss": 0.1883026510477066,
      "eval_runtime": 2.5837,
      "eval_samples_per_second": 889.018,
      "eval_steps_per_second": 13.933,
      "step": 639000
    },
    {
      "epoch": 6.78,
      "learning_rate": 5.4123701571492636e-05,
      "loss": 0.1985,
      "step": 639500
    },
    {
      "epoch": 6.78,
      "learning_rate": 5.401619257572453e-05,
      "loss": 0.1987,
      "step": 640000
    },
    {
      "epoch": 6.78,
      "eval_loss": 0.18854811787605286,
      "eval_runtime": 2.5902,
      "eval_samples_per_second": 886.813,
      "eval_steps_per_second": 13.899,
      "step": 640000
    },
    {
      "epoch": 6.79,
      "learning_rate": 5.390875461869379e-05,
      "loss": 0.1994,
      "step": 640500
    },
    {
      "epoch": 6.79,
      "learning_rate": 5.3801387994131576e-05,
      "loss": 0.1988,
      "step": 641000
    },
    {
      "epoch": 6.79,
      "eval_loss": 0.1891511082649231,
      "eval_runtime": 2.5529,
      "eval_samples_per_second": 899.746,
      "eval_steps_per_second": 14.101,
      "step": 641000
    },
    {
      "epoch": 6.8,
      "learning_rate": 5.36940929955742e-05,
      "loss": 0.1984,
      "step": 641500
    },
    {
      "epoch": 6.8,
      "learning_rate": 5.358686991636209e-05,
      "loss": 0.1986,
      "step": 642000
    },
    {
      "epoch": 6.8,
      "eval_loss": 0.1878819614648819,
      "eval_runtime": 2.5682,
      "eval_samples_per_second": 894.392,
      "eval_steps_per_second": 14.017,
      "step": 642000
    },
    {
      "epoch": 6.81,
      "learning_rate": 5.347971904963904e-05,
      "loss": 0.1984,
      "step": 642500
    },
    {
      "epoch": 6.81,
      "learning_rate": 5.3372640688351476e-05,
      "loss": 0.1988,
      "step": 643000
    },
    {
      "epoch": 6.81,
      "eval_loss": 0.18957975506782532,
      "eval_runtime": 2.5593,
      "eval_samples_per_second": 897.526,
      "eval_steps_per_second": 14.067,
      "step": 643000
    },
    {
      "epoch": 6.82,
      "learning_rate": 5.326563512524748e-05,
      "loss": 0.1984,
      "step": 643500
    },
    {
      "epoch": 6.83,
      "learning_rate": 5.315870265287618e-05,
      "loss": 0.1985,
      "step": 644000
    },
    {
      "epoch": 6.83,
      "eval_loss": 0.18896985054016113,
      "eval_runtime": 2.5541,
      "eval_samples_per_second": 899.348,
      "eval_steps_per_second": 14.095,
      "step": 644000
    },
    {
      "epoch": 6.83,
      "learning_rate": 5.3051843563586914e-05,
      "loss": 0.1985,
      "step": 644500
    },
    {
      "epoch": 6.84,
      "learning_rate": 5.294505814952835e-05,
      "loss": 0.1981,
      "step": 645000
    },
    {
      "epoch": 6.84,
      "eval_loss": 0.18807749450206757,
      "eval_runtime": 2.6155,
      "eval_samples_per_second": 878.24,
      "eval_steps_per_second": 13.764,
      "step": 645000
    },
    {
      "epoch": 6.84,
      "learning_rate": 5.28383467026477e-05,
      "loss": 0.1983,
      "step": 645500
    },
    {
      "epoch": 6.85,
      "learning_rate": 5.2731709514689995e-05,
      "loss": 0.1982,
      "step": 646000
    },
    {
      "epoch": 6.85,
      "eval_loss": 0.18697316944599152,
      "eval_runtime": 2.5109,
      "eval_samples_per_second": 914.815,
      "eval_steps_per_second": 14.338,
      "step": 646000
    },
    {
      "epoch": 6.85,
      "learning_rate": 5.262514687719722e-05,
      "loss": 0.1983,
      "step": 646500
    },
    {
      "epoch": 6.86,
      "learning_rate": 5.25186590815076e-05,
      "loss": 0.1985,
      "step": 647000
    },
    {
      "epoch": 6.86,
      "eval_loss": 0.1897994428873062,
      "eval_runtime": 2.4914,
      "eval_samples_per_second": 921.982,
      "eval_steps_per_second": 14.45,
      "step": 647000
    },
    {
      "epoch": 6.86,
      "learning_rate": 5.24122464187547e-05,
      "loss": 0.1981,
      "step": 647500
    },
    {
      "epoch": 6.87,
      "learning_rate": 5.2305909179866635e-05,
      "loss": 0.1982,
      "step": 648000
    },
    {
      "epoch": 6.87,
      "eval_loss": 0.18750005960464478,
      "eval_runtime": 2.5048,
      "eval_samples_per_second": 917.025,
      "eval_steps_per_second": 14.372,
      "step": 648000
    },
    {
      "epoch": 6.88,
      "learning_rate": 5.219964765556536e-05,
      "loss": 0.1982,
      "step": 648500
    },
    {
      "epoch": 6.88,
      "learning_rate": 5.209346213636584e-05,
      "loss": 0.198,
      "step": 649000
    },
    {
      "epoch": 6.88,
      "eval_loss": 0.18683280050754547,
      "eval_runtime": 2.5207,
      "eval_samples_per_second": 911.266,
      "eval_steps_per_second": 14.282,
      "step": 649000
    },
    {
      "epoch": 6.89,
      "learning_rate": 5.1987352912575244e-05,
      "loss": 0.1982,
      "step": 649500
    },
    {
      "epoch": 6.89,
      "learning_rate": 5.188132027429215e-05,
      "loss": 0.198,
      "step": 650000
    },
    {
      "epoch": 6.89,
      "eval_loss": 0.18715369701385498,
      "eval_runtime": 2.4622,
      "eval_samples_per_second": 932.918,
      "eval_steps_per_second": 14.621,
      "step": 650000
    },
    {
      "epoch": 6.9,
      "learning_rate": 5.177536451140569e-05,
      "loss": 0.1982,
      "step": 650500
    },
    {
      "epoch": 6.9,
      "learning_rate": 5.166948591359489e-05,
      "loss": 0.1983,
      "step": 651000
    },
    {
      "epoch": 6.9,
      "eval_loss": 0.18645188212394714,
      "eval_runtime": 2.4913,
      "eval_samples_per_second": 922.024,
      "eval_steps_per_second": 14.451,
      "step": 651000
    },
    {
      "epoch": 6.91,
      "learning_rate": 5.1563684770327804e-05,
      "loss": 0.1983,
      "step": 651500
    },
    {
      "epoch": 6.91,
      "learning_rate": 5.145796137086076e-05,
      "loss": 0.1981,
      "step": 652000
    },
    {
      "epoch": 6.91,
      "eval_loss": 0.1894109547138214,
      "eval_runtime": 2.5469,
      "eval_samples_per_second": 901.873,
      "eval_steps_per_second": 14.135,
      "step": 652000
    },
    {
      "epoch": 6.92,
      "learning_rate": 5.135231600423742e-05,
      "loss": 0.1979,
      "step": 652500
    },
    {
      "epoch": 6.93,
      "learning_rate": 5.124674895928823e-05,
      "loss": 0.1976,
      "step": 653000
    },
    {
      "epoch": 6.93,
      "eval_loss": 0.18761011958122253,
      "eval_runtime": 2.6483,
      "eval_samples_per_second": 867.333,
      "eval_steps_per_second": 13.593,
      "step": 653000
    },
    {
      "epoch": 6.93,
      "learning_rate": 5.114126052462943e-05,
      "loss": 0.1978,
      "step": 653500
    },
    {
      "epoch": 6.94,
      "learning_rate": 5.103585098866237e-05,
      "loss": 0.1976,
      "step": 654000
    },
    {
      "epoch": 6.94,
      "eval_loss": 0.1896512508392334,
      "eval_runtime": 2.6053,
      "eval_samples_per_second": 881.651,
      "eval_steps_per_second": 13.818,
      "step": 654000
    },
    {
      "epoch": 6.94,
      "learning_rate": 5.093052063957276e-05,
      "loss": 0.1978,
      "step": 654500
    },
    {
      "epoch": 6.95,
      "learning_rate": 5.082526976532968e-05,
      "loss": 0.1982,
      "step": 655000
    },
    {
      "epoch": 6.95,
      "eval_loss": 0.18722392618656158,
      "eval_runtime": 2.5063,
      "eval_samples_per_second": 916.492,
      "eval_steps_per_second": 14.364,
      "step": 655000
    },
    {
      "epoch": 6.95,
      "learning_rate": 5.072009865368501e-05,
      "loss": 0.1977,
      "step": 655500
    },
    {
      "epoch": 6.96,
      "learning_rate": 5.061500759217261e-05,
      "loss": 0.1976,
      "step": 656000
    },
    {
      "epoch": 6.96,
      "eval_loss": 0.1882035732269287,
      "eval_runtime": 2.625,
      "eval_samples_per_second": 875.059,
      "eval_steps_per_second": 13.714,
      "step": 656000
    },
    {
      "epoch": 6.96,
      "learning_rate": 5.050999686810735e-05,
      "loss": 0.1977,
      "step": 656500
    },
    {
      "epoch": 6.97,
      "learning_rate": 5.04050667685846e-05,
      "loss": 0.1979,
      "step": 657000
    },
    {
      "epoch": 6.97,
      "eval_loss": 0.18653704226016998,
      "eval_runtime": 2.582,
      "eval_samples_per_second": 889.617,
      "eval_steps_per_second": 13.943,
      "step": 657000
    },
    {
      "epoch": 6.98,
      "learning_rate": 5.0300217580479244e-05,
      "loss": 0.1979,
      "step": 657500
    },
    {
      "epoch": 6.98,
      "learning_rate": 5.01954495904449e-05,
      "loss": 0.1973,
      "step": 658000
    },
    {
      "epoch": 6.98,
      "eval_loss": 0.1866321712732315,
      "eval_runtime": 2.5422,
      "eval_samples_per_second": 903.55,
      "eval_steps_per_second": 14.161,
      "step": 658000
    },
    {
      "epoch": 6.99,
      "learning_rate": 5.0090763084913336e-05,
      "loss": 0.1975,
      "step": 658500
    },
    {
      "epoch": 6.99,
      "learning_rate": 4.998615835009339e-05,
      "loss": 0.197,
      "step": 659000
    },
    {
      "epoch": 6.99,
      "eval_loss": 0.18548274040222168,
      "eval_runtime": 2.5558,
      "eval_samples_per_second": 898.724,
      "eval_steps_per_second": 14.085,
      "step": 659000
    },
    {
      "epoch": 7.0,
      "learning_rate": 4.988163567197043e-05,
      "loss": 0.1975,
      "step": 659500
    },
    {
      "epoch": 7.0,
      "learning_rate": 4.97771953363055e-05,
      "loss": 0.1977,
      "step": 660000
    },
    {
      "epoch": 7.0,
      "eval_loss": 0.1880265176296234,
      "eval_runtime": 2.6377,
      "eval_samples_per_second": 870.836,
      "eval_steps_per_second": 13.648,
      "step": 660000
    },
    {
      "epoch": 7.01,
      "learning_rate": 4.967283762863444e-05,
      "loss": 0.1974,
      "step": 660500
    },
    {
      "epoch": 7.01,
      "learning_rate": 4.956856283426728e-05,
      "loss": 0.1972,
      "step": 661000
    },
    {
      "epoch": 7.01,
      "eval_loss": 0.18715223670005798,
      "eval_runtime": 2.6628,
      "eval_samples_per_second": 862.627,
      "eval_steps_per_second": 13.52,
      "step": 661000
    },
    {
      "epoch": 7.02,
      "learning_rate": 4.946437123828732e-05,
      "loss": 0.1971,
      "step": 661500
    },
    {
      "epoch": 7.03,
      "learning_rate": 4.936026312555037e-05,
      "loss": 0.1967,
      "step": 662000
    },
    {
      "epoch": 7.03,
      "eval_loss": 0.18600988388061523,
      "eval_runtime": 2.614,
      "eval_samples_per_second": 878.714,
      "eval_steps_per_second": 13.772,
      "step": 662000
    },
    {
      "epoch": 7.03,
      "learning_rate": 4.925623878068408e-05,
      "loss": 0.1972,
      "step": 662500
    },
    {
      "epoch": 7.04,
      "learning_rate": 4.915229848808698e-05,
      "loss": 0.1973,
      "step": 663000
    },
    {
      "epoch": 7.04,
      "eval_loss": 0.1860048770904541,
      "eval_runtime": 2.673,
      "eval_samples_per_second": 859.331,
      "eval_steps_per_second": 13.468,
      "step": 663000
    },
    {
      "epoch": 7.04,
      "learning_rate": 4.904844253192795e-05,
      "loss": 0.1965,
      "step": 663500
    },
    {
      "epoch": 7.05,
      "learning_rate": 4.8944671196145136e-05,
      "loss": 0.1967,
      "step": 664000
    },
    {
      "epoch": 7.05,
      "eval_loss": 0.18750624358654022,
      "eval_runtime": 2.7099,
      "eval_samples_per_second": 847.62,
      "eval_steps_per_second": 13.284,
      "step": 664000
    },
    {
      "epoch": 7.05,
      "learning_rate": 4.884098476444539e-05,
      "loss": 0.1968,
      "step": 664500
    },
    {
      "epoch": 7.06,
      "learning_rate": 4.8737383520303546e-05,
      "loss": 0.1965,
      "step": 665000
    },
    {
      "epoch": 7.06,
      "eval_loss": 0.18656140565872192,
      "eval_runtime": 2.728,
      "eval_samples_per_second": 842.009,
      "eval_steps_per_second": 13.196,
      "step": 665000
    },
    {
      "epoch": 7.07,
      "learning_rate": 4.8633867746961356e-05,
      "loss": 0.1968,
      "step": 665500
    },
    {
      "epoch": 7.07,
      "learning_rate": 4.853043772742709e-05,
      "loss": 0.1969,
      "step": 666000
    },
    {
      "epoch": 7.07,
      "eval_loss": 0.18721851706504822,
      "eval_runtime": 2.6334,
      "eval_samples_per_second": 872.245,
      "eval_steps_per_second": 13.67,
      "step": 666000
    },
    {
      "epoch": 7.08,
      "learning_rate": 4.8427093744474364e-05,
      "loss": 0.1971,
      "step": 666500
    },
    {
      "epoch": 7.08,
      "learning_rate": 4.832383608064172e-05,
      "loss": 0.1968,
      "step": 667000
    },
    {
      "epoch": 7.08,
      "eval_loss": 0.18801318109035492,
      "eval_runtime": 2.7526,
      "eval_samples_per_second": 834.477,
      "eval_steps_per_second": 13.078,
      "step": 667000
    },
    {
      "epoch": 7.09,
      "learning_rate": 4.822066501823172e-05,
      "loss": 0.1967,
      "step": 667500
    },
    {
      "epoch": 7.09,
      "learning_rate": 4.811758083931005e-05,
      "loss": 0.1967,
      "step": 668000
    },
    {
      "epoch": 7.09,
      "eval_loss": 0.18405194580554962,
      "eval_runtime": 2.6183,
      "eval_samples_per_second": 877.299,
      "eval_steps_per_second": 13.75,
      "step": 668000
    },
    {
      "epoch": 7.1,
      "learning_rate": 4.8014583825704976e-05,
      "loss": 0.1965,
      "step": 668500
    },
    {
      "epoch": 7.1,
      "learning_rate": 4.791167425900632e-05,
      "loss": 0.1968,
      "step": 669000
    },
    {
      "epoch": 7.1,
      "eval_loss": 0.18537673354148865,
      "eval_runtime": 2.6514,
      "eval_samples_per_second": 866.335,
      "eval_steps_per_second": 13.578,
      "step": 669000
    },
    {
      "epoch": 7.11,
      "learning_rate": 4.780885242056493e-05,
      "loss": 0.1963,
      "step": 669500
    },
    {
      "epoch": 7.12,
      "learning_rate": 4.770611859149185e-05,
      "loss": 0.1963,
      "step": 670000
    },
    {
      "epoch": 7.12,
      "eval_loss": 0.187167689204216,
      "eval_runtime": 2.6705,
      "eval_samples_per_second": 860.127,
      "eval_steps_per_second": 13.48,
      "step": 670000
    },
    {
      "epoch": 7.12,
      "learning_rate": 4.7603473052657374e-05,
      "loss": 0.1966,
      "step": 670500
    },
    {
      "epoch": 7.13,
      "learning_rate": 4.7500916084690564e-05,
      "loss": 0.1967,
      "step": 671000
    },
    {
      "epoch": 7.13,
      "eval_loss": 0.186045840382576,
      "eval_runtime": 2.6321,
      "eval_samples_per_second": 872.688,
      "eval_steps_per_second": 13.677,
      "step": 671000
    },
    {
      "epoch": 7.13,
      "learning_rate": 4.7398447967978165e-05,
      "loss": 0.1963,
      "step": 671500
    },
    {
      "epoch": 7.14,
      "learning_rate": 4.729606898266411e-05,
      "loss": 0.1969,
      "step": 672000
    },
    {
      "epoch": 7.14,
      "eval_loss": 0.1882571130990982,
      "eval_runtime": 2.6187,
      "eval_samples_per_second": 877.14,
      "eval_steps_per_second": 13.747,
      "step": 672000
    },
    {
      "epoch": 7.14,
      "learning_rate": 4.71937794086487e-05,
      "loss": 0.1963,
      "step": 672500
    },
    {
      "epoch": 7.15,
      "learning_rate": 4.709157952558768e-05,
      "loss": 0.1963,
      "step": 673000
    },
    {
      "epoch": 7.15,
      "eval_loss": 0.18748030066490173,
      "eval_runtime": 2.6428,
      "eval_samples_per_second": 869.147,
      "eval_steps_per_second": 13.622,
      "step": 673000
    },
    {
      "epoch": 7.15,
      "learning_rate": 4.698946961289163e-05,
      "loss": 0.1962,
      "step": 673500
    },
    {
      "epoch": 7.16,
      "learning_rate": 4.688744994972514e-05,
      "loss": 0.1962,
      "step": 674000
    },
    {
      "epoch": 7.16,
      "eval_loss": 0.18707983195781708,
      "eval_runtime": 2.661,
      "eval_samples_per_second": 863.195,
      "eval_steps_per_second": 13.529,
      "step": 674000
    },
    {
      "epoch": 7.17,
      "learning_rate": 4.6785520815006085e-05,
      "loss": 0.1965,
      "step": 674500
    },
    {
      "epoch": 7.17,
      "learning_rate": 4.668368248740485e-05,
      "loss": 0.196,
      "step": 675000
    },
    {
      "epoch": 7.17,
      "eval_loss": 0.18474501371383667,
      "eval_runtime": 2.6847,
      "eval_samples_per_second": 855.597,
      "eval_steps_per_second": 13.409,
      "step": 675000
    },
    {
      "epoch": 7.18,
      "learning_rate": 4.658193524534351e-05,
      "loss": 0.1959,
      "step": 675500
    },
    {
      "epoch": 7.18,
      "learning_rate": 4.6480279366995116e-05,
      "loss": 0.1959,
      "step": 676000
    },
    {
      "epoch": 7.18,
      "eval_loss": 0.1878414750099182,
      "eval_runtime": 2.6912,
      "eval_samples_per_second": 853.536,
      "eval_steps_per_second": 13.377,
      "step": 676000
    },
    {
      "epoch": 7.19,
      "learning_rate": 4.637871513028303e-05,
      "loss": 0.196,
      "step": 676500
    },
    {
      "epoch": 7.19,
      "learning_rate": 4.6277242812879914e-05,
      "loss": 0.1959,
      "step": 677000
    },
    {
      "epoch": 7.19,
      "eval_loss": 0.18550430238246918,
      "eval_runtime": 2.7131,
      "eval_samples_per_second": 846.623,
      "eval_steps_per_second": 13.269,
      "step": 677000
    },
    {
      "epoch": 7.2,
      "learning_rate": 4.617586269220728e-05,
      "loss": 0.1959,
      "step": 677500
    },
    {
      "epoch": 7.2,
      "learning_rate": 4.607457504543447e-05,
      "loss": 0.1962,
      "step": 678000
    },
    {
      "epoch": 7.2,
      "eval_loss": 0.1865391880273819,
      "eval_runtime": 2.6562,
      "eval_samples_per_second": 864.771,
      "eval_steps_per_second": 13.553,
      "step": 678000
    },
    {
      "epoch": 7.21,
      "learning_rate": 4.597338014947801e-05,
      "loss": 0.1956,
      "step": 678500
    },
    {
      "epoch": 7.22,
      "learning_rate": 4.5872278281000955e-05,
      "loss": 0.1954,
      "step": 679000
    },
    {
      "epoch": 7.22,
      "eval_loss": 0.18739037215709686,
      "eval_runtime": 2.5862,
      "eval_samples_per_second": 888.174,
      "eval_steps_per_second": 13.92,
      "step": 679000
    },
    {
      "epoch": 7.22,
      "learning_rate": 4.577126971641189e-05,
      "loss": 0.1959,
      "step": 679500
    },
    {
      "epoch": 7.23,
      "learning_rate": 4.567035473186444e-05,
      "loss": 0.1955,
      "step": 680000
    },
    {
      "epoch": 7.23,
      "eval_loss": 0.18440315127372742,
      "eval_runtime": 2.6652,
      "eval_samples_per_second": 861.845,
      "eval_steps_per_second": 13.507,
      "step": 680000
    },
    {
      "epoch": 7.23,
      "learning_rate": 4.556953360325625e-05,
      "loss": 0.1953,
      "step": 680500
    },
    {
      "epoch": 7.24,
      "learning_rate": 4.546880660622845e-05,
      "loss": 0.1956,
      "step": 681000
    },
    {
      "epoch": 7.24,
      "eval_loss": 0.1863844096660614,
      "eval_runtime": 2.65,
      "eval_samples_per_second": 866.784,
      "eval_steps_per_second": 13.585,
      "step": 681000
    },
    {
      "epoch": 7.24,
      "learning_rate": 4.5368174016164844e-05,
      "loss": 0.1959,
      "step": 681500
    },
    {
      "epoch": 7.25,
      "learning_rate": 4.5267636108191036e-05,
      "loss": 0.1954,
      "step": 682000
    },
    {
      "epoch": 7.25,
      "eval_loss": 0.18577434122562408,
      "eval_runtime": 2.6836,
      "eval_samples_per_second": 855.93,
      "eval_steps_per_second": 13.415,
      "step": 682000
    },
    {
      "epoch": 7.25,
      "learning_rate": 4.5167193157173913e-05,
      "loss": 0.1954,
      "step": 682500
    },
    {
      "epoch": 7.26,
      "learning_rate": 4.5066845437720555e-05,
      "loss": 0.1956,
      "step": 683000
    },
    {
      "epoch": 7.26,
      "eval_loss": 0.1866350919008255,
      "eval_runtime": 2.6163,
      "eval_samples_per_second": 877.947,
      "eval_steps_per_second": 13.76,
      "step": 683000
    },
    {
      "epoch": 7.27,
      "learning_rate": 4.4966593224177866e-05,
      "loss": 0.1954,
      "step": 683500
    },
    {
      "epoch": 7.27,
      "learning_rate": 4.4866436790631564e-05,
      "loss": 0.1956,
      "step": 684000
    },
    {
      "epoch": 7.27,
      "eval_loss": 0.1872914433479309,
      "eval_runtime": 2.7089,
      "eval_samples_per_second": 847.935,
      "eval_steps_per_second": 13.289,
      "step": 684000
    },
    {
      "epoch": 7.28,
      "learning_rate": 4.476637641090551e-05,
      "loss": 0.1951,
      "step": 684500
    },
    {
      "epoch": 7.28,
      "learning_rate": 4.4666412358560955e-05,
      "loss": 0.195,
      "step": 685000
    },
    {
      "epoch": 7.28,
      "eval_loss": 0.1856185644865036,
      "eval_runtime": 2.6121,
      "eval_samples_per_second": 879.375,
      "eval_steps_per_second": 13.782,
      "step": 685000
    },
    {
      "epoch": 7.29,
      "learning_rate": 4.456654490689578e-05,
      "loss": 0.1952,
      "step": 685500
    },
    {
      "epoch": 7.29,
      "learning_rate": 4.4466774328943796e-05,
      "loss": 0.1953,
      "step": 686000
    },
    {
      "epoch": 7.29,
      "eval_loss": 0.18476100265979767,
      "eval_runtime": 2.6153,
      "eval_samples_per_second": 878.305,
      "eval_steps_per_second": 13.765,
      "step": 686000
    },
    {
      "epoch": 7.3,
      "learning_rate": 4.4367100897474e-05,
      "loss": 0.1955,
      "step": 686500
    },
    {
      "epoch": 7.3,
      "learning_rate": 4.426752488498972e-05,
      "loss": 0.1952,
      "step": 687000
    },
    {
      "epoch": 7.3,
      "eval_loss": 0.18481390178203583,
      "eval_runtime": 2.6488,
      "eval_samples_per_second": 867.183,
      "eval_steps_per_second": 13.591,
      "step": 687000
    },
    {
      "epoch": 7.31,
      "learning_rate": 4.4168046563727945e-05,
      "loss": 0.1953,
      "step": 687500
    },
    {
      "epoch": 7.32,
      "learning_rate": 4.406866620565862e-05,
      "loss": 0.1948,
      "step": 688000
    },
    {
      "epoch": 7.32,
      "eval_loss": 0.185079887509346,
      "eval_runtime": 2.6355,
      "eval_samples_per_second": 871.569,
      "eval_steps_per_second": 13.66,
      "step": 688000
    },
    {
      "epoch": 7.32,
      "learning_rate": 4.396938408248383e-05,
      "loss": 0.1952,
      "step": 688500
    },
    {
      "epoch": 7.33,
      "learning_rate": 4.3870200465637164e-05,
      "loss": 0.1951,
      "step": 689000
    },
    {
      "epoch": 7.33,
      "eval_loss": 0.1854194700717926,
      "eval_runtime": 2.6418,
      "eval_samples_per_second": 869.478,
      "eval_steps_per_second": 13.627,
      "step": 689000
    },
    {
      "epoch": 7.33,
      "learning_rate": 4.377111562628282e-05,
      "loss": 0.1948,
      "step": 689500
    },
    {
      "epoch": 7.34,
      "learning_rate": 4.3672129835314955e-05,
      "loss": 0.1951,
      "step": 690000
    },
    {
      "epoch": 7.34,
      "eval_loss": 0.18585672974586487,
      "eval_runtime": 2.7189,
      "eval_samples_per_second": 844.818,
      "eval_steps_per_second": 13.241,
      "step": 690000
    },
    {
      "epoch": 7.34,
      "learning_rate": 4.3573243363356916e-05,
      "loss": 0.1948,
      "step": 690500
    },
    {
      "epoch": 7.35,
      "learning_rate": 4.347445648076057e-05,
      "loss": 0.1953,
      "step": 691000
    },
    {
      "epoch": 7.35,
      "eval_loss": 0.18430623412132263,
      "eval_runtime": 2.6745,
      "eval_samples_per_second": 858.843,
      "eval_steps_per_second": 13.46,
      "step": 691000
    },
    {
      "epoch": 7.36,
      "learning_rate": 4.337576945760554e-05,
      "loss": 0.1949,
      "step": 691500
    },
    {
      "epoch": 7.36,
      "learning_rate": 4.327718256369826e-05,
      "loss": 0.1947,
      "step": 692000
    },
    {
      "epoch": 7.36,
      "eval_loss": 0.18453127145767212,
      "eval_runtime": 2.6094,
      "eval_samples_per_second": 880.283,
      "eval_steps_per_second": 13.796,
      "step": 692000
    },
    {
      "epoch": 7.37,
      "learning_rate": 4.317869606857162e-05,
      "loss": 0.1949,
      "step": 692500
    },
    {
      "epoch": 7.37,
      "learning_rate": 4.3080310241483885e-05,
      "loss": 0.1947,
      "step": 693000
    },
    {
      "epoch": 7.37,
      "eval_loss": 0.18398995697498322,
      "eval_runtime": 2.7028,
      "eval_samples_per_second": 849.851,
      "eval_steps_per_second": 13.319,
      "step": 693000
    },
    {
      "epoch": 7.38,
      "learning_rate": 4.298202535141818e-05,
      "loss": 0.1944,
      "step": 693500
    },
    {
      "epoch": 7.38,
      "learning_rate": 4.2883841667081675e-05,
      "loss": 0.1945,
      "step": 694000
    },
    {
      "epoch": 7.38,
      "eval_loss": 0.1815757304430008,
      "eval_runtime": 2.6223,
      "eval_samples_per_second": 875.945,
      "eval_steps_per_second": 13.728,
      "step": 694000
    },
    {
      "epoch": 7.39,
      "learning_rate": 4.2785759456904745e-05,
      "loss": 0.1943,
      "step": 694500
    },
    {
      "epoch": 7.39,
      "learning_rate": 4.268777898904044e-05,
      "loss": 0.1945,
      "step": 695000
    },
    {
      "epoch": 7.39,
      "eval_loss": 0.1843911111354828,
      "eval_runtime": 2.6472,
      "eval_samples_per_second": 867.72,
      "eval_steps_per_second": 13.599,
      "step": 695000
    },
    {
      "epoch": 7.4,
      "learning_rate": 4.2589900531363606e-05,
      "loss": 0.1947,
      "step": 695500
    },
    {
      "epoch": 7.41,
      "learning_rate": 4.2492124351470214e-05,
      "loss": 0.1946,
      "step": 696000
    },
    {
      "epoch": 7.41,
      "eval_loss": 0.18493103981018066,
      "eval_runtime": 2.6337,
      "eval_samples_per_second": 872.173,
      "eval_steps_per_second": 13.669,
      "step": 696000
    },
    {
      "epoch": 7.41,
      "learning_rate": 4.239445071667666e-05,
      "loss": 0.1946,
      "step": 696500
    },
    {
      "epoch": 7.42,
      "learning_rate": 4.2296879894018835e-05,
      "loss": 0.1944,
      "step": 697000
    },
    {
      "epoch": 7.42,
      "eval_loss": 0.18418000638484955,
      "eval_runtime": 2.5716,
      "eval_samples_per_second": 893.203,
      "eval_steps_per_second": 13.999,
      "step": 697000
    },
    {
      "epoch": 7.42,
      "learning_rate": 4.219941215025171e-05,
      "loss": 0.1943,
      "step": 697500
    },
    {
      "epoch": 7.43,
      "learning_rate": 4.210204775184834e-05,
      "loss": 0.1948,
      "step": 698000
    },
    {
      "epoch": 7.43,
      "eval_loss": 0.18401235342025757,
      "eval_runtime": 2.6433,
      "eval_samples_per_second": 868.996,
      "eval_steps_per_second": 13.619,
      "step": 698000
    },
    {
      "epoch": 7.43,
      "learning_rate": 4.2004786964999304e-05,
      "loss": 0.1944,
      "step": 698500
    },
    {
      "epoch": 7.44,
      "learning_rate": 4.190763005561186e-05,
      "loss": 0.1943,
      "step": 699000
    },
    {
      "epoch": 7.44,
      "eval_loss": 0.1846538931131363,
      "eval_runtime": 2.6279,
      "eval_samples_per_second": 874.077,
      "eval_steps_per_second": 13.699,
      "step": 699000
    },
    {
      "epoch": 7.44,
      "learning_rate": 4.1810577289309266e-05,
      "loss": 0.1941,
      "step": 699500
    },
    {
      "epoch": 7.45,
      "learning_rate": 4.171362893143013e-05,
      "loss": 0.1944,
      "step": 700000
    },
    {
      "epoch": 7.45,
      "eval_loss": 0.1839645951986313,
      "eval_runtime": 2.6536,
      "eval_samples_per_second": 865.616,
      "eval_steps_per_second": 13.566,
      "step": 700000
    },
    {
      "epoch": 7.46,
      "learning_rate": 4.1616785247027506e-05,
      "loss": 0.1942,
      "step": 700500
    },
    {
      "epoch": 7.46,
      "learning_rate": 4.1520046500868384e-05,
      "loss": 0.1941,
      "step": 701000
    },
    {
      "epoch": 7.46,
      "eval_loss": 0.18309521675109863,
      "eval_runtime": 2.6732,
      "eval_samples_per_second": 859.262,
      "eval_steps_per_second": 13.467,
      "step": 701000
    },
    {
      "epoch": 7.47,
      "learning_rate": 4.1423412957432775e-05,
      "loss": 0.1943,
      "step": 701500
    },
    {
      "epoch": 7.47,
      "learning_rate": 4.1326884880913074e-05,
      "loss": 0.1939,
      "step": 702000
    },
    {
      "epoch": 7.47,
      "eval_loss": 0.18348637223243713,
      "eval_runtime": 2.6604,
      "eval_samples_per_second": 863.415,
      "eval_steps_per_second": 13.532,
      "step": 702000
    },
    {
      "epoch": 7.48,
      "learning_rate": 4.123046253521341e-05,
      "loss": 0.1939,
      "step": 702500
    },
    {
      "epoch": 7.48,
      "learning_rate": 4.1134146183948724e-05,
      "loss": 0.1942,
      "step": 703000
    },
    {
      "epoch": 7.48,
      "eval_loss": 0.18514606356620789,
      "eval_runtime": 2.6612,
      "eval_samples_per_second": 863.158,
      "eval_steps_per_second": 13.528,
      "step": 703000
    },
    {
      "epoch": 7.49,
      "learning_rate": 4.1037936090444315e-05,
      "loss": 0.1945,
      "step": 703500
    },
    {
      "epoch": 7.49,
      "learning_rate": 4.0941832517734885e-05,
      "loss": 0.1938,
      "step": 704000
    },
    {
      "epoch": 7.49,
      "eval_loss": 0.18320082128047943,
      "eval_runtime": 2.5921,
      "eval_samples_per_second": 886.166,
      "eval_steps_per_second": 13.889,
      "step": 704000
    },
    {
      "epoch": 7.5,
      "learning_rate": 4.084583572856388e-05,
      "loss": 0.1937,
      "step": 704500
    },
    {
      "epoch": 7.51,
      "learning_rate": 4.0749945985382915e-05,
      "loss": 0.1938,
      "step": 705000
    },
    {
      "epoch": 7.51,
      "eval_loss": 0.181605264544487,
      "eval_runtime": 2.7114,
      "eval_samples_per_second": 847.159,
      "eval_steps_per_second": 13.277,
      "step": 705000
    },
    {
      "epoch": 7.51,
      "learning_rate": 4.065416355035087e-05,
      "loss": 0.1941,
      "step": 705500
    },
    {
      "epoch": 7.52,
      "learning_rate": 4.0558488685333235e-05,
      "loss": 0.1938,
      "step": 706000
    },
    {
      "epoch": 7.52,
      "eval_loss": 0.18272297084331512,
      "eval_runtime": 2.6623,
      "eval_samples_per_second": 862.792,
      "eval_steps_per_second": 13.522,
      "step": 706000
    },
    {
      "epoch": 7.52,
      "learning_rate": 4.04629216519015e-05,
      "loss": 0.1937,
      "step": 706500
    },
    {
      "epoch": 7.53,
      "learning_rate": 4.036746271133223e-05,
      "loss": 0.1936,
      "step": 707000
    },
    {
      "epoch": 7.53,
      "eval_loss": 0.1840648055076599,
      "eval_runtime": 2.6725,
      "eval_samples_per_second": 859.485,
      "eval_steps_per_second": 13.47,
      "step": 707000
    },
    {
      "epoch": 7.53,
      "learning_rate": 4.0272112124606546e-05,
      "loss": 0.1936,
      "step": 707500
    },
    {
      "epoch": 7.54,
      "learning_rate": 4.0176870152409324e-05,
      "loss": 0.1937,
      "step": 708000
    },
    {
      "epoch": 7.54,
      "eval_loss": 0.1836749017238617,
      "eval_runtime": 2.7112,
      "eval_samples_per_second": 847.228,
      "eval_steps_per_second": 13.278,
      "step": 708000
    },
    {
      "epoch": 7.54,
      "learning_rate": 4.008173705512842e-05,
      "loss": 0.1938,
      "step": 708500
    },
    {
      "epoch": 7.55,
      "learning_rate": 3.998671309285417e-05,
      "loss": 0.1937,
      "step": 709000
    },
    {
      "epoch": 7.55,
      "eval_loss": 0.18317854404449463,
      "eval_runtime": 2.6768,
      "eval_samples_per_second": 858.105,
      "eval_steps_per_second": 13.449,
      "step": 709000
    },
    {
      "epoch": 7.56,
      "learning_rate": 3.989179852537839e-05,
      "loss": 0.1937,
      "step": 709500
    },
    {
      "epoch": 7.56,
      "learning_rate": 3.979699361219395e-05,
      "loss": 0.1938,
      "step": 710000
    },
    {
      "epoch": 7.56,
      "eval_loss": 0.18444040417671204,
      "eval_runtime": 2.6261,
      "eval_samples_per_second": 874.67,
      "eval_steps_per_second": 13.708,
      "step": 710000
    },
    {
      "epoch": 7.57,
      "learning_rate": 3.9702298612493816e-05,
      "loss": 0.1938,
      "step": 710500
    },
    {
      "epoch": 7.57,
      "learning_rate": 3.960771378517049e-05,
      "loss": 0.1935,
      "step": 711000
    },
    {
      "epoch": 7.57,
      "eval_loss": 0.183274045586586,
      "eval_runtime": 2.6269,
      "eval_samples_per_second": 874.423,
      "eval_steps_per_second": 13.705,
      "step": 711000
    },
    {
      "epoch": 7.58,
      "learning_rate": 3.951323938881533e-05,
      "loss": 0.1936,
      "step": 711500
    },
    {
      "epoch": 7.58,
      "learning_rate": 3.941887568171766e-05,
      "loss": 0.1934,
      "step": 712000
    },
    {
      "epoch": 7.58,
      "eval_loss": 0.18676131963729858,
      "eval_runtime": 2.7142,
      "eval_samples_per_second": 846.278,
      "eval_steps_per_second": 13.263,
      "step": 712000
    },
    {
      "epoch": 7.59,
      "learning_rate": 3.9324622921864323e-05,
      "loss": 0.1933,
      "step": 712500
    },
    {
      "epoch": 7.59,
      "learning_rate": 3.923048136693873e-05,
      "loss": 0.1933,
      "step": 713000
    },
    {
      "epoch": 7.59,
      "eval_loss": 0.18421515822410583,
      "eval_runtime": 2.6188,
      "eval_samples_per_second": 877.135,
      "eval_steps_per_second": 13.747,
      "step": 713000
    },
    {
      "epoch": 7.6,
      "learning_rate": 3.913645127432028e-05,
      "loss": 0.193,
      "step": 713500
    },
    {
      "epoch": 7.61,
      "learning_rate": 3.904253290108369e-05,
      "loss": 0.193,
      "step": 714000
    },
    {
      "epoch": 7.61,
      "eval_loss": 0.18448708951473236,
      "eval_runtime": 2.6531,
      "eval_samples_per_second": 865.779,
      "eval_steps_per_second": 13.569,
      "step": 714000
    },
    {
      "epoch": 7.61,
      "learning_rate": 3.8948726503998176e-05,
      "loss": 0.1931,
      "step": 714500
    },
    {
      "epoch": 7.62,
      "learning_rate": 3.885503233952689e-05,
      "loss": 0.1937,
      "step": 715000
    },
    {
      "epoch": 7.62,
      "eval_loss": 0.18292050063610077,
      "eval_runtime": 2.5788,
      "eval_samples_per_second": 890.71,
      "eval_steps_per_second": 13.96,
      "step": 715000
    },
    {
      "epoch": 7.62,
      "learning_rate": 3.876145066382606e-05,
      "loss": 0.1931,
      "step": 715500
    },
    {
      "epoch": 7.63,
      "learning_rate": 3.86679817327444e-05,
      "loss": 0.1931,
      "step": 716000
    },
    {
      "epoch": 7.63,
      "eval_loss": 0.18449412286281586,
      "eval_runtime": 2.6103,
      "eval_samples_per_second": 879.974,
      "eval_steps_per_second": 13.792,
      "step": 716000
    },
    {
      "epoch": 7.63,
      "learning_rate": 3.857462580182245e-05,
      "loss": 0.1931,
      "step": 716500
    },
    {
      "epoch": 7.64,
      "learning_rate": 3.848138312629171e-05,
      "loss": 0.193,
      "step": 717000
    },
    {
      "epoch": 7.64,
      "eval_loss": 0.18389998376369476,
      "eval_runtime": 2.6596,
      "eval_samples_per_second": 863.655,
      "eval_steps_per_second": 13.536,
      "step": 717000
    },
    {
      "epoch": 7.65,
      "learning_rate": 3.838825396107415e-05,
      "loss": 0.1926,
      "step": 717500
    },
    {
      "epoch": 7.65,
      "learning_rate": 3.8295238560781317e-05,
      "loss": 0.1927,
      "step": 718000
    },
    {
      "epoch": 7.65,
      "eval_loss": 0.18192943930625916,
      "eval_runtime": 2.6184,
      "eval_samples_per_second": 877.259,
      "eval_steps_per_second": 13.749,
      "step": 718000
    },
    {
      "epoch": 7.66,
      "learning_rate": 3.820233717971374e-05,
      "loss": 0.193,
      "step": 718500
    },
    {
      "epoch": 7.66,
      "learning_rate": 3.810955007186029e-05,
      "loss": 0.1928,
      "step": 719000
    },
    {
      "epoch": 7.66,
      "eval_loss": 0.1814703792333603,
      "eval_runtime": 2.627,
      "eval_samples_per_second": 874.39,
      "eval_steps_per_second": 13.704,
      "step": 719000
    },
    {
      "epoch": 7.67,
      "learning_rate": 3.801687749089737e-05,
      "loss": 0.1928,
      "step": 719500
    },
    {
      "epoch": 7.67,
      "learning_rate": 3.792431969018824e-05,
      "loss": 0.1927,
      "step": 720000
    },
    {
      "epoch": 7.67,
      "eval_loss": 0.18235580623149872,
      "eval_runtime": 2.6689,
      "eval_samples_per_second": 860.641,
      "eval_steps_per_second": 13.488,
      "step": 720000
    },
    {
      "epoch": 7.68,
      "learning_rate": 3.783187692278245e-05,
      "loss": 0.1924,
      "step": 720500
    },
    {
      "epoch": 7.68,
      "learning_rate": 3.7739549441414945e-05,
      "loss": 0.1927,
      "step": 721000
    },
    {
      "epoch": 7.68,
      "eval_loss": 0.18421824276447296,
      "eval_runtime": 2.699,
      "eval_samples_per_second": 851.051,
      "eval_steps_per_second": 13.338,
      "step": 721000
    },
    {
      "epoch": 7.69,
      "learning_rate": 3.764733749850558e-05,
      "loss": 0.1928,
      "step": 721500
    },
    {
      "epoch": 7.7,
      "learning_rate": 3.755524134615825e-05,
      "loss": 0.1928,
      "step": 722000
    },
    {
      "epoch": 7.7,
      "eval_loss": 0.1823606938123703,
      "eval_runtime": 2.6367,
      "eval_samples_per_second": 871.176,
      "eval_steps_per_second": 13.654,
      "step": 722000
    },
    {
      "epoch": 7.7,
      "learning_rate": 3.746326123616032e-05,
      "loss": 0.1927,
      "step": 722500
    },
    {
      "epoch": 7.71,
      "learning_rate": 3.7371397419981925e-05,
      "loss": 0.1924,
      "step": 723000
    },
    {
      "epoch": 7.71,
      "eval_loss": 0.18170827627182007,
      "eval_runtime": 2.6026,
      "eval_samples_per_second": 882.563,
      "eval_steps_per_second": 13.832,
      "step": 723000
    },
    {
      "epoch": 7.71,
      "learning_rate": 3.7279650148775196e-05,
      "loss": 0.1927,
      "step": 723500
    },
    {
      "epoch": 7.72,
      "learning_rate": 3.7188019673373706e-05,
      "loss": 0.1923,
      "step": 724000
    },
    {
      "epoch": 7.72,
      "eval_loss": 0.18194718658924103,
      "eval_runtime": 2.729,
      "eval_samples_per_second": 841.686,
      "eval_steps_per_second": 13.191,
      "step": 724000
    },
    {
      "epoch": 7.72,
      "learning_rate": 3.709650624429166e-05,
      "loss": 0.1921,
      "step": 724500
    },
    {
      "epoch": 7.73,
      "learning_rate": 3.700511011172325e-05,
      "loss": 0.1924,
      "step": 725000
    },
    {
      "epoch": 7.73,
      "eval_loss": 0.18209348618984222,
      "eval_runtime": 2.6667,
      "eval_samples_per_second": 861.365,
      "eval_steps_per_second": 13.5,
      "step": 725000
    },
    {
      "epoch": 7.73,
      "learning_rate": 3.691383152554207e-05,
      "loss": 0.1926,
      "step": 725500
    },
    {
      "epoch": 7.74,
      "learning_rate": 3.682267073530023e-05,
      "loss": 0.1922,
      "step": 726000
    },
    {
      "epoch": 7.74,
      "eval_loss": 0.18260575830936432,
      "eval_runtime": 2.7378,
      "eval_samples_per_second": 838.986,
      "eval_steps_per_second": 13.149,
      "step": 726000
    },
    {
      "epoch": 7.75,
      "learning_rate": 3.67316279902279e-05,
      "loss": 0.1921,
      "step": 726500
    },
    {
      "epoch": 7.75,
      "learning_rate": 3.664070353923245e-05,
      "loss": 0.1923,
      "step": 727000
    },
    {
      "epoch": 7.75,
      "eval_loss": 0.18323123455047607,
      "eval_runtime": 2.6275,
      "eval_samples_per_second": 874.227,
      "eval_steps_per_second": 13.701,
      "step": 727000
    },
    {
      "epoch": 7.76,
      "learning_rate": 3.654989763089782e-05,
      "loss": 0.1921,
      "step": 727500
    },
    {
      "epoch": 7.76,
      "learning_rate": 3.645921051348396e-05,
      "loss": 0.1926,
      "step": 728000
    },
    {
      "epoch": 7.76,
      "eval_loss": 0.18396110832691193,
      "eval_runtime": 2.6742,
      "eval_samples_per_second": 858.943,
      "eval_steps_per_second": 13.462,
      "step": 728000
    },
    {
      "epoch": 7.77,
      "learning_rate": 3.6368642434925924e-05,
      "loss": 0.192,
      "step": 728500
    },
    {
      "epoch": 7.77,
      "learning_rate": 3.627819364283345e-05,
      "loss": 0.1921,
      "step": 729000
    },
    {
      "epoch": 7.77,
      "eval_loss": 0.18145744502544403,
      "eval_runtime": 2.6278,
      "eval_samples_per_second": 874.116,
      "eval_steps_per_second": 13.7,
      "step": 729000
    },
    {
      "epoch": 7.78,
      "learning_rate": 3.6187864384490035e-05,
      "loss": 0.1916,
      "step": 729500
    },
    {
      "epoch": 7.78,
      "learning_rate": 3.6097654906852405e-05,
      "loss": 0.1917,
      "step": 730000
    },
    {
      "epoch": 7.78,
      "eval_loss": 0.1821279525756836,
      "eval_runtime": 2.6606,
      "eval_samples_per_second": 863.347,
      "eval_steps_per_second": 13.531,
      "step": 730000
    },
    {
      "epoch": 7.79,
      "learning_rate": 3.600756545654988e-05,
      "loss": 0.1924,
      "step": 730500
    },
    {
      "epoch": 7.8,
      "learning_rate": 3.591759627988353e-05,
      "loss": 0.1921,
      "step": 731000
    },
    {
      "epoch": 7.8,
      "eval_loss": 0.1826609969139099,
      "eval_runtime": 2.5577,
      "eval_samples_per_second": 898.086,
      "eval_steps_per_second": 14.075,
      "step": 731000
    },
    {
      "epoch": 7.8,
      "learning_rate": 3.582774762282568e-05,
      "loss": 0.1926,
      "step": 731500
    },
    {
      "epoch": 7.81,
      "learning_rate": 3.573801973101913e-05,
      "loss": 0.1919,
      "step": 732000
    },
    {
      "epoch": 7.81,
      "eval_loss": 0.1833614856004715,
      "eval_runtime": 2.5691,
      "eval_samples_per_second": 894.085,
      "eval_steps_per_second": 14.013,
      "step": 732000
    },
    {
      "epoch": 7.81,
      "learning_rate": 3.564841284977646e-05,
      "loss": 0.1918,
      "step": 732500
    },
    {
      "epoch": 7.82,
      "learning_rate": 3.5558927224079534e-05,
      "loss": 0.1922,
      "step": 733000
    },
    {
      "epoch": 7.82,
      "eval_loss": 0.18190455436706543,
      "eval_runtime": 2.6712,
      "eval_samples_per_second": 859.911,
      "eval_steps_per_second": 13.477,
      "step": 733000
    },
    {
      "epoch": 7.82,
      "learning_rate": 3.546956309857859e-05,
      "loss": 0.1918,
      "step": 733500
    },
    {
      "epoch": 7.83,
      "learning_rate": 3.5380320717591716e-05,
      "loss": 0.1918,
      "step": 734000
    },
    {
      "epoch": 7.83,
      "eval_loss": 0.18287116289138794,
      "eval_runtime": 2.6384,
      "eval_samples_per_second": 870.604,
      "eval_steps_per_second": 13.645,
      "step": 734000
    },
    {
      "epoch": 7.83,
      "learning_rate": 3.5291200325104234e-05,
      "loss": 0.1916,
      "step": 734500
    },
    {
      "epoch": 7.84,
      "learning_rate": 3.5202202164767836e-05,
      "loss": 0.1916,
      "step": 735000
    },
    {
      "epoch": 7.84,
      "eval_loss": 0.18400417268276215,
      "eval_runtime": 2.6593,
      "eval_samples_per_second": 863.766,
      "eval_steps_per_second": 13.537,
      "step": 735000
    },
    {
      "epoch": 7.85,
      "learning_rate": 3.511332647990014e-05,
      "loss": 0.1918,
      "step": 735500
    },
    {
      "epoch": 7.85,
      "learning_rate": 3.5024573513483864e-05,
      "loss": 0.1918,
      "step": 736000
    },
    {
      "epoch": 7.85,
      "eval_loss": 0.17965777218341827,
      "eval_runtime": 2.6148,
      "eval_samples_per_second": 878.457,
      "eval_steps_per_second": 13.768,
      "step": 736000
    },
    {
      "epoch": 7.86,
      "learning_rate": 3.493594350816619e-05,
      "loss": 0.1916,
      "step": 736500
    },
    {
      "epoch": 7.86,
      "learning_rate": 3.484743670625822e-05,
      "loss": 0.1913,
      "step": 737000
    },
    {
      "epoch": 7.86,
      "eval_loss": 0.18245890736579895,
      "eval_runtime": 2.6872,
      "eval_samples_per_second": 854.791,
      "eval_steps_per_second": 13.397,
      "step": 737000
    },
    {
      "epoch": 7.87,
      "learning_rate": 3.4759053349734126e-05,
      "loss": 0.1916,
      "step": 737500
    },
    {
      "epoch": 7.87,
      "learning_rate": 3.467079368023068e-05,
      "loss": 0.1913,
      "step": 738000
    },
    {
      "epoch": 7.87,
      "eval_loss": 0.18023112416267395,
      "eval_runtime": 2.6258,
      "eval_samples_per_second": 874.787,
      "eval_steps_per_second": 13.71,
      "step": 738000
    },
    {
      "epoch": 7.88,
      "learning_rate": 3.458265793904642e-05,
      "loss": 0.1912,
      "step": 738500
    },
    {
      "epoch": 7.88,
      "learning_rate": 3.449464636714107e-05,
      "loss": 0.1911,
      "step": 739000
    },
    {
      "epoch": 7.88,
      "eval_loss": 0.18116475641727448,
      "eval_runtime": 2.5092,
      "eval_samples_per_second": 915.413,
      "eval_steps_per_second": 14.347,
      "step": 739000
    },
    {
      "epoch": 7.89,
      "learning_rate": 3.4406759205134966e-05,
      "loss": 0.191,
      "step": 739500
    },
    {
      "epoch": 7.9,
      "learning_rate": 3.431899669330819e-05,
      "loss": 0.1914,
      "step": 740000
    },
    {
      "epoch": 7.9,
      "eval_loss": 0.18056176602840424,
      "eval_runtime": 2.7172,
      "eval_samples_per_second": 845.342,
      "eval_steps_per_second": 13.249,
      "step": 740000
    },
    {
      "epoch": 7.9,
      "learning_rate": 3.4231359071600156e-05,
      "loss": 0.1915,
      "step": 740500
    },
    {
      "epoch": 7.91,
      "learning_rate": 3.4143846579608744e-05,
      "loss": 0.1915,
      "step": 741000
    },
    {
      "epoch": 7.91,
      "eval_loss": 0.18201898038387299,
      "eval_runtime": 2.6375,
      "eval_samples_per_second": 870.9,
      "eval_steps_per_second": 13.649,
      "step": 741000
    },
    {
      "epoch": 7.91,
      "learning_rate": 3.405645945658976e-05,
      "loss": 0.191,
      "step": 741500
    },
    {
      "epoch": 7.92,
      "learning_rate": 3.396919794145629e-05,
      "loss": 0.1915,
      "step": 742000
    },
    {
      "epoch": 7.92,
      "eval_loss": 0.1818230152130127,
      "eval_runtime": 2.6616,
      "eval_samples_per_second": 863.023,
      "eval_steps_per_second": 13.526,
      "step": 742000
    },
    {
      "epoch": 7.92,
      "learning_rate": 3.3882062272777936e-05,
      "loss": 0.191,
      "step": 742500
    },
    {
      "epoch": 7.93,
      "learning_rate": 3.3795052688780345e-05,
      "loss": 0.191,
      "step": 743000
    },
    {
      "epoch": 7.93,
      "eval_loss": 0.18135210871696472,
      "eval_runtime": 2.7035,
      "eval_samples_per_second": 849.632,
      "eval_steps_per_second": 13.316,
      "step": 743000
    },
    {
      "epoch": 7.94,
      "learning_rate": 3.370816942734438e-05,
      "loss": 0.1912,
      "step": 743500
    },
    {
      "epoch": 7.94,
      "learning_rate": 3.362141272600552e-05,
      "loss": 0.1907,
      "step": 744000
    },
    {
      "epoch": 7.94,
      "eval_loss": 0.18246738612651825,
      "eval_runtime": 2.6098,
      "eval_samples_per_second": 880.139,
      "eval_steps_per_second": 13.794,
      "step": 744000
    },
    {
      "epoch": 7.95,
      "learning_rate": 3.3534782821953325e-05,
      "loss": 0.1907,
      "step": 744500
    },
    {
      "epoch": 7.95,
      "learning_rate": 3.3448279952030615e-05,
      "loss": 0.191,
      "step": 745000
    },
    {
      "epoch": 7.95,
      "eval_loss": 0.1824055016040802,
      "eval_runtime": 2.604,
      "eval_samples_per_second": 882.092,
      "eval_steps_per_second": 13.825,
      "step": 745000
    },
    {
      "epoch": 7.96,
      "learning_rate": 3.336190435273295e-05,
      "loss": 0.1911,
      "step": 745500
    },
    {
      "epoch": 7.96,
      "learning_rate": 3.327565626020793e-05,
      "loss": 0.1909,
      "step": 746000
    },
    {
      "epoch": 7.96,
      "eval_loss": 0.18200530111789703,
      "eval_runtime": 2.6615,
      "eval_samples_per_second": 863.04,
      "eval_steps_per_second": 13.526,
      "step": 746000
    },
    {
      "epoch": 7.97,
      "learning_rate": 3.31895359102545e-05,
      "loss": 0.191,
      "step": 746500
    },
    {
      "epoch": 7.97,
      "learning_rate": 3.3103543538322455e-05,
      "loss": 0.1908,
      "step": 747000
    },
    {
      "epoch": 7.97,
      "eval_loss": 0.17920434474945068,
      "eval_runtime": 2.6801,
      "eval_samples_per_second": 857.049,
      "eval_steps_per_second": 13.432,
      "step": 747000
    },
    {
      "epoch": 7.98,
      "learning_rate": 3.3017679379511645e-05,
      "loss": 0.191,
      "step": 747500
    },
    {
      "epoch": 7.99,
      "learning_rate": 3.293194366857137e-05,
      "loss": 0.1905,
      "step": 748000
    },
    {
      "epoch": 7.99,
      "eval_loss": 0.18376828730106354,
      "eval_runtime": 2.6586,
      "eval_samples_per_second": 863.997,
      "eval_steps_per_second": 13.541,
      "step": 748000
    },
    {
      "epoch": 7.99,
      "learning_rate": 3.2846336639899845e-05,
      "loss": 0.1904,
      "step": 748500
    },
    {
      "epoch": 8.0,
      "learning_rate": 3.276085852754336e-05,
      "loss": 0.1907,
      "step": 749000
    },
    {
      "epoch": 8.0,
      "eval_loss": 0.18018439412117004,
      "eval_runtime": 2.6652,
      "eval_samples_per_second": 861.854,
      "eval_steps_per_second": 13.508,
      "step": 749000
    },
    {
      "epoch": 8.0,
      "learning_rate": 3.267550956519586e-05,
      "loss": 0.1904,
      "step": 749500
    },
    {
      "epoch": 8.01,
      "learning_rate": 3.259028998619814e-05,
      "loss": 0.1904,
      "step": 750000
    },
    {
      "epoch": 8.01,
      "eval_loss": 0.1812291294336319,
      "eval_runtime": 2.6153,
      "eval_samples_per_second": 878.279,
      "eval_steps_per_second": 13.765,
      "step": 750000
    },
    {
      "epoch": 8.01,
      "learning_rate": 3.2505200023537225e-05,
      "loss": 0.1906,
      "step": 750500
    },
    {
      "epoch": 8.02,
      "learning_rate": 3.2420239909845894e-05,
      "loss": 0.1903,
      "step": 751000
    },
    {
      "epoch": 8.02,
      "eval_loss": 0.18331938982009888,
      "eval_runtime": 2.6413,
      "eval_samples_per_second": 869.633,
      "eval_steps_per_second": 13.629,
      "step": 751000
    },
    {
      "epoch": 8.02,
      "learning_rate": 3.233540987740179e-05,
      "loss": 0.1904,
      "step": 751500
    },
    {
      "epoch": 8.03,
      "learning_rate": 3.2250710158127045e-05,
      "loss": 0.1903,
      "step": 752000
    },
    {
      "epoch": 8.03,
      "eval_loss": 0.18277068436145782,
      "eval_runtime": 2.5885,
      "eval_samples_per_second": 887.393,
      "eval_steps_per_second": 13.908,
      "step": 752000
    },
    {
      "epoch": 8.04,
      "learning_rate": 3.216614098358741e-05,
      "loss": 0.1902,
      "step": 752500
    },
    {
      "epoch": 8.04,
      "learning_rate": 3.2081702584991786e-05,
      "loss": 0.1905,
      "step": 753000
    },
    {
      "epoch": 8.04,
      "eval_loss": 0.18270383775234222,
      "eval_runtime": 2.6624,
      "eval_samples_per_second": 862.74,
      "eval_steps_per_second": 13.521,
      "step": 753000
    },
    {
      "epoch": 8.05,
      "learning_rate": 3.1997395193191565e-05,
      "loss": 0.1901,
      "step": 753500
    },
    {
      "epoch": 8.05,
      "learning_rate": 3.191321903867988e-05,
      "loss": 0.1903,
      "step": 754000
    },
    {
      "epoch": 8.05,
      "eval_loss": 0.1816052347421646,
      "eval_runtime": 2.6428,
      "eval_samples_per_second": 869.149,
      "eval_steps_per_second": 13.622,
      "step": 754000
    },
    {
      "epoch": 8.06,
      "learning_rate": 3.18291743515912e-05,
      "loss": 0.1903,
      "step": 754500
    },
    {
      "epoch": 8.06,
      "learning_rate": 3.174526136170039e-05,
      "loss": 0.1905,
      "step": 755000
    },
    {
      "epoch": 8.06,
      "eval_loss": 0.1785338819026947,
      "eval_runtime": 2.6371,
      "eval_samples_per_second": 871.041,
      "eval_steps_per_second": 13.651,
      "step": 755000
    },
    {
      "epoch": 8.07,
      "learning_rate": 3.1661480298422433e-05,
      "loss": 0.1905,
      "step": 755500
    },
    {
      "epoch": 8.07,
      "learning_rate": 3.157783139081155e-05,
      "loss": 0.19,
      "step": 756000
    },
    {
      "epoch": 8.07,
      "eval_loss": 0.18119803071022034,
      "eval_runtime": 2.5992,
      "eval_samples_per_second": 883.749,
      "eval_steps_per_second": 13.851,
      "step": 756000
    },
    {
      "epoch": 8.08,
      "learning_rate": 3.149431486756063e-05,
      "loss": 0.19,
      "step": 756500
    },
    {
      "epoch": 8.09,
      "learning_rate": 3.141093095700072e-05,
      "loss": 0.1899,
      "step": 757000
    },
    {
      "epoch": 8.09,
      "eval_loss": 0.18124920129776,
      "eval_runtime": 2.6586,
      "eval_samples_per_second": 863.984,
      "eval_steps_per_second": 13.541,
      "step": 757000
    },
    {
      "epoch": 8.09,
      "learning_rate": 3.132767988710016e-05,
      "loss": 0.1898,
      "step": 757500
    },
    {
      "epoch": 8.1,
      "learning_rate": 3.1244561885464244e-05,
      "loss": 0.1902,
      "step": 758000
    },
    {
      "epoch": 8.1,
      "eval_loss": 0.1819985955953598,
      "eval_runtime": 2.6588,
      "eval_samples_per_second": 863.933,
      "eval_steps_per_second": 13.54,
      "step": 758000
    },
    {
      "epoch": 8.1,
      "learning_rate": 3.116157717933443e-05,
      "loss": 0.1902,
      "step": 758500
    },
    {
      "epoch": 8.11,
      "learning_rate": 3.107872599558769e-05,
      "loss": 0.19,
      "step": 759000
    },
    {
      "epoch": 8.11,
      "eval_loss": 0.18176889419555664,
      "eval_runtime": 2.6674,
      "eval_samples_per_second": 861.13,
      "eval_steps_per_second": 13.496,
      "step": 759000
    },
    {
      "epoch": 8.11,
      "learning_rate": 3.0996008560736083e-05,
      "loss": 0.1901,
      "step": 759500
    },
    {
      "epoch": 8.12,
      "learning_rate": 3.0913425100925795e-05,
      "loss": 0.19,
      "step": 760000
    },
    {
      "epoch": 8.12,
      "eval_loss": 0.17839893698692322,
      "eval_runtime": 2.7183,
      "eval_samples_per_second": 845.029,
      "eval_steps_per_second": 13.244,
      "step": 760000
    },
    {
      "epoch": 8.12,
      "learning_rate": 3.083097584193693e-05,
      "loss": 0.1896,
      "step": 760500
    },
    {
      "epoch": 8.13,
      "learning_rate": 3.0748661009182616e-05,
      "loss": 0.1898,
      "step": 761000
    },
    {
      "epoch": 8.13,
      "eval_loss": 0.17937108874320984,
      "eval_runtime": 2.6571,
      "eval_samples_per_second": 864.488,
      "eval_steps_per_second": 13.549,
      "step": 761000
    },
    {
      "epoch": 8.14,
      "learning_rate": 3.066648082770845e-05,
      "loss": 0.19,
      "step": 761500
    },
    {
      "epoch": 8.14,
      "learning_rate": 3.0584435522191896e-05,
      "loss": 0.1893,
      "step": 762000
    },
    {
      "epoch": 8.14,
      "eval_loss": 0.18026237189769745,
      "eval_runtime": 2.6859,
      "eval_samples_per_second": 855.204,
      "eval_steps_per_second": 13.403,
      "step": 762000
    },
    {
      "epoch": 8.15,
      "learning_rate": 3.0502525316941673e-05,
      "loss": 0.1896,
      "step": 762500
    },
    {
      "epoch": 8.15,
      "learning_rate": 3.0420750435897183e-05,
      "loss": 0.1895,
      "step": 763000
    },
    {
      "epoch": 8.15,
      "eval_loss": 0.18032604455947876,
      "eval_runtime": 2.6799,
      "eval_samples_per_second": 857.123,
      "eval_steps_per_second": 13.433,
      "step": 763000
    },
    {
      "epoch": 8.16,
      "learning_rate": 3.0339111102627846e-05,
      "loss": 0.1897,
      "step": 763500
    },
    {
      "epoch": 8.16,
      "learning_rate": 3.025760754033246e-05,
      "loss": 0.1892,
      "step": 764000
    },
    {
      "epoch": 8.16,
      "eval_loss": 0.18092449009418488,
      "eval_runtime": 2.6398,
      "eval_samples_per_second": 870.157,
      "eval_steps_per_second": 13.638,
      "step": 764000
    },
    {
      "epoch": 8.17,
      "learning_rate": 3.017623997183864e-05,
      "loss": 0.1897,
      "step": 764500
    },
    {
      "epoch": 8.17,
      "learning_rate": 3.0095008619602206e-05,
      "loss": 0.1894,
      "step": 765000
    },
    {
      "epoch": 8.17,
      "eval_loss": 0.17851532995700836,
      "eval_runtime": 2.6678,
      "eval_samples_per_second": 861.0,
      "eval_steps_per_second": 13.494,
      "step": 765000
    },
    {
      "epoch": 8.18,
      "learning_rate": 3.0013913705706587e-05,
      "loss": 0.1892,
      "step": 765500
    },
    {
      "epoch": 8.19,
      "learning_rate": 2.993295545186223e-05,
      "loss": 0.19,
      "step": 766000
    },
    {
      "epoch": 8.19,
      "eval_loss": 0.18047097325325012,
      "eval_runtime": 2.6997,
      "eval_samples_per_second": 850.838,
      "eval_steps_per_second": 13.335,
      "step": 766000
    },
    {
      "epoch": 8.19,
      "learning_rate": 2.9852134079405817e-05,
      "loss": 0.1897,
      "step": 766500
    },
    {
      "epoch": 8.2,
      "learning_rate": 2.977144980929996e-05,
      "loss": 0.1895,
      "step": 767000
    },
    {
      "epoch": 8.2,
      "eval_loss": 0.1800650954246521,
      "eval_runtime": 2.6124,
      "eval_samples_per_second": 879.254,
      "eval_steps_per_second": 13.78,
      "step": 767000
    },
    {
      "epoch": 8.2,
      "learning_rate": 2.969090286213233e-05,
      "loss": 0.1896,
      "step": 767500
    },
    {
      "epoch": 8.21,
      "learning_rate": 2.961049345811523e-05,
      "loss": 0.1891,
      "step": 768000
    },
    {
      "epoch": 8.21,
      "eval_loss": 0.18048371374607086,
      "eval_runtime": 2.7369,
      "eval_samples_per_second": 839.265,
      "eval_steps_per_second": 13.153,
      "step": 768000
    },
    {
      "epoch": 8.21,
      "learning_rate": 2.9530221817084937e-05,
      "loss": 0.1894,
      "step": 768500
    },
    {
      "epoch": 8.22,
      "learning_rate": 2.945008815850097e-05,
      "loss": 0.1895,
      "step": 769000
    },
    {
      "epoch": 8.22,
      "eval_loss": 0.17877112329006195,
      "eval_runtime": 2.6697,
      "eval_samples_per_second": 860.384,
      "eval_steps_per_second": 13.484,
      "step": 769000
    },
    {
      "epoch": 8.22,
      "learning_rate": 2.9370092701445748e-05,
      "loss": 0.1891,
      "step": 769500
    },
    {
      "epoch": 8.23,
      "learning_rate": 2.929023566462377e-05,
      "loss": 0.1894,
      "step": 770000
    },
    {
      "epoch": 8.23,
      "eval_loss": 0.1808948516845703,
      "eval_runtime": 2.6258,
      "eval_samples_per_second": 874.77,
      "eval_steps_per_second": 13.71,
      "step": 770000
    },
    {
      "epoch": 8.24,
      "learning_rate": 2.921051726636114e-05,
      "loss": 0.1892,
      "step": 770500
    },
    {
      "epoch": 8.24,
      "learning_rate": 2.9130937724604947e-05,
      "loss": 0.1895,
      "step": 771000
    },
    {
      "epoch": 8.24,
      "eval_loss": 0.17894434928894043,
      "eval_runtime": 2.6506,
      "eval_samples_per_second": 866.601,
      "eval_steps_per_second": 13.582,
      "step": 771000
    },
    {
      "epoch": 8.25,
      "learning_rate": 2.9051497256922545e-05,
      "loss": 0.1891,
      "step": 771500
    },
    {
      "epoch": 8.25,
      "learning_rate": 2.8972196080501208e-05,
      "loss": 0.1888,
      "step": 772000
    },
    {
      "epoch": 8.25,
      "eval_loss": 0.17959585785865784,
      "eval_runtime": 2.688,
      "eval_samples_per_second": 854.55,
      "eval_steps_per_second": 13.393,
      "step": 772000
    },
    {
      "epoch": 8.26,
      "learning_rate": 2.8893034412147268e-05,
      "loss": 0.189,
      "step": 772500
    },
    {
      "epoch": 8.26,
      "learning_rate": 2.8814012468285748e-05,
      "loss": 0.189,
      "step": 773000
    },
    {
      "epoch": 8.26,
      "eval_loss": 0.1775985211133957,
      "eval_runtime": 2.6568,
      "eval_samples_per_second": 864.565,
      "eval_steps_per_second": 13.55,
      "step": 773000
    },
    {
      "epoch": 8.27,
      "learning_rate": 2.8735130464959604e-05,
      "loss": 0.1888,
      "step": 773500
    },
    {
      "epoch": 8.28,
      "learning_rate": 2.865638861782922e-05,
      "loss": 0.1889,
      "step": 774000
    },
    {
      "epoch": 8.28,
      "eval_loss": 0.18201559782028198,
      "eval_runtime": 2.6162,
      "eval_samples_per_second": 877.995,
      "eval_steps_per_second": 13.76,
      "step": 774000
    },
    {
      "epoch": 8.28,
      "learning_rate": 2.8577787142171804e-05,
      "loss": 0.1888,
      "step": 774500
    },
    {
      "epoch": 8.29,
      "learning_rate": 2.849932625288079e-05,
      "loss": 0.1895,
      "step": 775000
    },
    {
      "epoch": 8.29,
      "eval_loss": 0.18033553659915924,
      "eval_runtime": 2.6434,
      "eval_samples_per_second": 868.969,
      "eval_steps_per_second": 13.619,
      "step": 775000
    },
    {
      "epoch": 8.29,
      "learning_rate": 2.8421006164465254e-05,
      "loss": 0.1891,
      "step": 775500
    },
    {
      "epoch": 8.3,
      "learning_rate": 2.8342827091049336e-05,
      "loss": 0.1888,
      "step": 776000
    },
    {
      "epoch": 8.3,
      "eval_loss": 0.17612183094024658,
      "eval_runtime": 2.6473,
      "eval_samples_per_second": 867.687,
      "eval_steps_per_second": 13.599,
      "step": 776000
    },
    {
      "epoch": 8.3,
      "learning_rate": 2.8264789246371605e-05,
      "loss": 0.189,
      "step": 776500
    },
    {
      "epoch": 8.31,
      "learning_rate": 2.8186892843784587e-05,
      "loss": 0.1889,
      "step": 777000
    },
    {
      "epoch": 8.31,
      "eval_loss": 0.17975667119026184,
      "eval_runtime": 2.6953,
      "eval_samples_per_second": 852.233,
      "eval_steps_per_second": 13.357,
      "step": 777000
    },
    {
      "epoch": 8.31,
      "learning_rate": 2.810913809625404e-05,
      "loss": 0.1884,
      "step": 777500
    },
    {
      "epoch": 8.32,
      "learning_rate": 2.803152521635851e-05,
      "loss": 0.1885,
      "step": 778000
    },
    {
      "epoch": 8.32,
      "eval_loss": 0.17638467252254486,
      "eval_runtime": 2.5904,
      "eval_samples_per_second": 886.746,
      "eval_steps_per_second": 13.898,
      "step": 778000
    },
    {
      "epoch": 8.33,
      "learning_rate": 2.795405441628862e-05,
      "loss": 0.1883,
      "step": 778500
    },
    {
      "epoch": 8.33,
      "learning_rate": 2.7876725907846578e-05,
      "loss": 0.1891,
      "step": 779000
    },
    {
      "epoch": 8.33,
      "eval_loss": 0.18025876581668854,
      "eval_runtime": 2.6761,
      "eval_samples_per_second": 858.354,
      "eval_steps_per_second": 13.453,
      "step": 779000
    },
    {
      "epoch": 8.34,
      "learning_rate": 2.7799539902445596e-05,
      "loss": 0.1883,
      "step": 779500
    },
    {
      "epoch": 8.34,
      "learning_rate": 2.7722496611109243e-05,
      "loss": 0.1888,
      "step": 780000
    },
    {
      "epoch": 8.34,
      "eval_loss": 0.1802951842546463,
      "eval_runtime": 2.7322,
      "eval_samples_per_second": 840.721,
      "eval_steps_per_second": 13.176,
      "step": 780000
    },
    {
      "epoch": 8.35,
      "learning_rate": 2.7645596244470935e-05,
      "loss": 0.1885,
      "step": 780500
    },
    {
      "epoch": 8.35,
      "learning_rate": 2.7568839012773365e-05,
      "loss": 0.1884,
      "step": 781000
    },
    {
      "epoch": 8.35,
      "eval_loss": 0.1805545538663864,
      "eval_runtime": 2.6452,
      "eval_samples_per_second": 868.354,
      "eval_steps_per_second": 13.609,
      "step": 781000
    },
    {
      "epoch": 8.36,
      "learning_rate": 2.7492225125867825e-05,
      "loss": 0.1889,
      "step": 781500
    },
    {
      "epoch": 8.36,
      "learning_rate": 2.7415754793213826e-05,
      "loss": 0.1886,
      "step": 782000
    },
    {
      "epoch": 8.36,
      "eval_loss": 0.1767302304506302,
      "eval_runtime": 2.6579,
      "eval_samples_per_second": 864.226,
      "eval_steps_per_second": 13.545,
      "step": 782000
    },
    {
      "epoch": 8.37,
      "learning_rate": 2.7339428223878283e-05,
      "loss": 0.1883,
      "step": 782500
    },
    {
      "epoch": 8.38,
      "learning_rate": 2.7263245626535116e-05,
      "loss": 0.1884,
      "step": 783000
    },
    {
      "epoch": 8.38,
      "eval_loss": 0.1802656203508377,
      "eval_runtime": 2.7131,
      "eval_samples_per_second": 846.621,
      "eval_steps_per_second": 13.269,
      "step": 783000
    },
    {
      "epoch": 8.38,
      "learning_rate": 2.7187207209464687e-05,
      "loss": 0.1883,
      "step": 783500
    },
    {
      "epoch": 8.39,
      "learning_rate": 2.7111313180553077e-05,
      "loss": 0.1882,
      "step": 784000
    },
    {
      "epoch": 8.39,
      "eval_loss": 0.17976997792720795,
      "eval_runtime": 2.668,
      "eval_samples_per_second": 860.932,
      "eval_steps_per_second": 13.493,
      "step": 784000
    },
    {
      "epoch": 8.39,
      "learning_rate": 2.703556374729169e-05,
      "loss": 0.1885,
      "step": 784500
    },
    {
      "epoch": 8.4,
      "learning_rate": 2.6959959116776587e-05,
      "loss": 0.188,
      "step": 785000
    },
    {
      "epoch": 8.4,
      "eval_loss": 0.1783231794834137,
      "eval_runtime": 2.6459,
      "eval_samples_per_second": 868.123,
      "eval_steps_per_second": 13.606,
      "step": 785000
    },
    {
      "epoch": 8.4,
      "learning_rate": 2.68844994957079e-05,
      "loss": 0.1881,
      "step": 785500
    },
    {
      "epoch": 8.41,
      "learning_rate": 2.6809185090389406e-05,
      "loss": 0.1884,
      "step": 786000
    },
    {
      "epoch": 8.41,
      "eval_loss": 0.18017184734344482,
      "eval_runtime": 2.6671,
      "eval_samples_per_second": 861.233,
      "eval_steps_per_second": 13.498,
      "step": 786000
    },
    {
      "epoch": 8.41,
      "learning_rate": 2.6734016106727777e-05,
      "loss": 0.1881,
      "step": 786500
    },
    {
      "epoch": 8.42,
      "learning_rate": 2.6658992750232167e-05,
      "loss": 0.188,
      "step": 787000
    },
    {
      "epoch": 8.42,
      "eval_loss": 0.17710144817829132,
      "eval_runtime": 2.6667,
      "eval_samples_per_second": 861.367,
      "eval_steps_per_second": 13.5,
      "step": 787000
    },
    {
      "epoch": 8.43,
      "learning_rate": 2.6584115226013553e-05,
      "loss": 0.1883,
      "step": 787500
    },
    {
      "epoch": 8.43,
      "learning_rate": 2.6509383738784218e-05,
      "loss": 0.188,
      "step": 788000
    },
    {
      "epoch": 8.43,
      "eval_loss": 0.1786525398492813,
      "eval_runtime": 2.6579,
      "eval_samples_per_second": 864.232,
      "eval_steps_per_second": 13.545,
      "step": 788000
    },
    {
      "epoch": 8.44,
      "learning_rate": 2.6434798492857228e-05,
      "loss": 0.1881,
      "step": 788500
    },
    {
      "epoch": 8.44,
      "learning_rate": 2.6360359692145757e-05,
      "loss": 0.1882,
      "step": 789000
    },
    {
      "epoch": 8.44,
      "eval_loss": 0.17897970974445343,
      "eval_runtime": 2.6253,
      "eval_samples_per_second": 874.933,
      "eval_steps_per_second": 13.712,
      "step": 789000
    },
    {
      "epoch": 8.45,
      "learning_rate": 2.6286067540162677e-05,
      "loss": 0.1882,
      "step": 789500
    },
    {
      "epoch": 8.45,
      "learning_rate": 2.6211922240019883e-05,
      "loss": 0.1883,
      "step": 790000
    },
    {
      "epoch": 8.45,
      "eval_loss": 0.17872017621994019,
      "eval_runtime": 2.5868,
      "eval_samples_per_second": 887.972,
      "eval_steps_per_second": 13.917,
      "step": 790000
    },
    {
      "epoch": 8.46,
      "learning_rate": 2.6137923994427768e-05,
      "loss": 0.1881,
      "step": 790500
    },
    {
      "epoch": 8.46,
      "learning_rate": 2.6064073005694758e-05,
      "loss": 0.1876,
      "step": 791000
    },
    {
      "epoch": 8.46,
      "eval_loss": 0.17775288224220276,
      "eval_runtime": 2.6373,
      "eval_samples_per_second": 870.95,
      "eval_steps_per_second": 13.65,
      "step": 791000
    },
    {
      "epoch": 8.47,
      "learning_rate": 2.5990369475726598e-05,
      "loss": 0.1878,
      "step": 791500
    },
    {
      "epoch": 8.48,
      "learning_rate": 2.591681360602595e-05,
      "loss": 0.1876,
      "step": 792000
    },
    {
      "epoch": 8.48,
      "eval_loss": 0.1769527643918991,
      "eval_runtime": 2.7236,
      "eval_samples_per_second": 843.383,
      "eval_steps_per_second": 13.218,
      "step": 792000
    },
    {
      "epoch": 8.48,
      "learning_rate": 2.5843405597691748e-05,
      "loss": 0.188,
      "step": 792500
    },
    {
      "epoch": 8.49,
      "learning_rate": 2.577014565141866e-05,
      "loss": 0.1872,
      "step": 793000
    },
    {
      "epoch": 8.49,
      "eval_loss": 0.1795121282339096,
      "eval_runtime": 2.6103,
      "eval_samples_per_second": 879.963,
      "eval_steps_per_second": 13.791,
      "step": 793000
    },
    {
      "epoch": 8.49,
      "learning_rate": 2.569703396749661e-05,
      "loss": 0.1875,
      "step": 793500
    },
    {
      "epoch": 8.5,
      "learning_rate": 2.562407074581014e-05,
      "loss": 0.188,
      "step": 794000
    },
    {
      "epoch": 8.5,
      "eval_loss": 0.17704832553863525,
      "eval_runtime": 2.6547,
      "eval_samples_per_second": 865.258,
      "eval_steps_per_second": 13.561,
      "step": 794000
    },
    {
      "epoch": 8.5,
      "learning_rate": 2.5551256185837897e-05,
      "loss": 0.1878,
      "step": 794500
    },
    {
      "epoch": 8.51,
      "learning_rate": 2.5478590486652137e-05,
      "loss": 0.1879,
      "step": 795000
    },
    {
      "epoch": 8.51,
      "eval_loss": 0.1775512397289276,
      "eval_runtime": 2.704,
      "eval_samples_per_second": 849.487,
      "eval_steps_per_second": 13.314,
      "step": 795000
    },
    {
      "epoch": 8.51,
      "learning_rate": 2.5406073846918076e-05,
      "loss": 0.1873,
      "step": 795500
    },
    {
      "epoch": 8.52,
      "learning_rate": 2.533370646489347e-05,
      "loss": 0.1872,
      "step": 796000
    },
    {
      "epoch": 8.52,
      "eval_loss": 0.17661112546920776,
      "eval_runtime": 2.6375,
      "eval_samples_per_second": 870.909,
      "eval_steps_per_second": 13.649,
      "step": 796000
    },
    {
      "epoch": 8.53,
      "learning_rate": 2.526148853842796e-05,
      "loss": 0.1874,
      "step": 796500
    },
    {
      "epoch": 8.53,
      "learning_rate": 2.5189420264962586e-05,
      "loss": 0.1875,
      "step": 797000
    },
    {
      "epoch": 8.53,
      "eval_loss": 0.17768479883670807,
      "eval_runtime": 2.6644,
      "eval_samples_per_second": 862.092,
      "eval_steps_per_second": 13.511,
      "step": 797000
    },
    {
      "epoch": 8.54,
      "learning_rate": 2.5117501841529297e-05,
      "loss": 0.1871,
      "step": 797500
    },
    {
      "epoch": 8.54,
      "learning_rate": 2.504573346475026e-05,
      "loss": 0.1874,
      "step": 798000
    },
    {
      "epoch": 8.54,
      "eval_loss": 0.17816244065761566,
      "eval_runtime": 2.6304,
      "eval_samples_per_second": 873.254,
      "eval_steps_per_second": 13.686,
      "step": 798000
    },
    {
      "epoch": 8.55,
      "learning_rate": 2.497411533083753e-05,
      "loss": 0.1874,
      "step": 798500
    },
    {
      "epoch": 8.55,
      "learning_rate": 2.4902647635592324e-05,
      "loss": 0.187,
      "step": 799000
    },
    {
      "epoch": 8.55,
      "eval_loss": 0.17790637910366058,
      "eval_runtime": 2.6765,
      "eval_samples_per_second": 858.201,
      "eval_steps_per_second": 13.45,
      "step": 799000
    },
    {
      "epoch": 8.56,
      "learning_rate": 2.483133057440458e-05,
      "loss": 0.1871,
      "step": 799500
    },
    {
      "epoch": 8.57,
      "learning_rate": 2.476016434225246e-05,
      "loss": 0.1872,
      "step": 800000
    },
    {
      "epoch": 8.57,
      "eval_loss": 0.17626047134399414,
      "eval_runtime": 2.6916,
      "eval_samples_per_second": 853.408,
      "eval_steps_per_second": 13.375,
      "step": 800000
    },
    {
      "epoch": 8.57,
      "learning_rate": 2.4689149133701672e-05,
      "loss": 0.1867,
      "step": 800500
    },
    {
      "epoch": 8.58,
      "learning_rate": 2.461828514290513e-05,
      "loss": 0.1869,
      "step": 801000
    },
    {
      "epoch": 8.58,
      "eval_loss": 0.17704518139362335,
      "eval_runtime": 2.7376,
      "eval_samples_per_second": 839.048,
      "eval_steps_per_second": 13.15,
      "step": 801000
    },
    {
      "epoch": 8.58,
      "learning_rate": 2.4547572563602267e-05,
      "loss": 0.1872,
      "step": 801500
    },
    {
      "epoch": 8.59,
      "learning_rate": 2.447701158911855e-05,
      "loss": 0.1868,
      "step": 802000
    },
    {
      "epoch": 8.59,
      "eval_loss": 0.17794357240200043,
      "eval_runtime": 2.6487,
      "eval_samples_per_second": 867.219,
      "eval_steps_per_second": 13.592,
      "step": 802000
    },
    {
      "epoch": 8.59,
      "learning_rate": 2.4406602412365027e-05,
      "loss": 0.187,
      "step": 802500
    },
    {
      "epoch": 8.6,
      "learning_rate": 2.4336345225837658e-05,
      "loss": 0.1872,
      "step": 803000
    },
    {
      "epoch": 8.6,
      "eval_loss": 0.1776154637336731,
      "eval_runtime": 2.7206,
      "eval_samples_per_second": 844.313,
      "eval_steps_per_second": 13.233,
      "step": 803000
    },
    {
      "epoch": 8.6,
      "learning_rate": 2.4266240221616956e-05,
      "loss": 0.1873,
      "step": 803500
    },
    {
      "epoch": 8.61,
      "learning_rate": 2.4196287591367296e-05,
      "loss": 0.1868,
      "step": 804000
    },
    {
      "epoch": 8.61,
      "eval_loss": 0.176628977060318,
      "eval_runtime": 2.6337,
      "eval_samples_per_second": 872.144,
      "eval_steps_per_second": 13.669,
      "step": 804000
    },
    {
      "epoch": 8.62,
      "learning_rate": 2.412648752633649e-05,
      "loss": 0.1869,
      "step": 804500
    },
    {
      "epoch": 8.62,
      "learning_rate": 2.405684021735527e-05,
      "loss": 0.1866,
      "step": 805000
    },
    {
      "epoch": 8.62,
      "eval_loss": 0.1774420291185379,
      "eval_runtime": 2.6036,
      "eval_samples_per_second": 882.223,
      "eval_steps_per_second": 13.827,
      "step": 805000
    },
    {
      "epoch": 8.63,
      "learning_rate": 2.39873458548367e-05,
      "loss": 0.1871,
      "step": 805500
    },
    {
      "epoch": 8.63,
      "learning_rate": 2.3918004628775736e-05,
      "loss": 0.1871,
      "step": 806000
    },
    {
      "epoch": 8.63,
      "eval_loss": 0.1766408383846283,
      "eval_runtime": 2.6553,
      "eval_samples_per_second": 865.055,
      "eval_steps_per_second": 13.558,
      "step": 806000
    },
    {
      "epoch": 8.64,
      "learning_rate": 2.3848816728748643e-05,
      "loss": 0.187,
      "step": 806500
    },
    {
      "epoch": 8.64,
      "learning_rate": 2.3779782343912463e-05,
      "loss": 0.1871,
      "step": 807000
    },
    {
      "epoch": 8.64,
      "eval_loss": 0.17578239738941193,
      "eval_runtime": 2.6261,
      "eval_samples_per_second": 874.674,
      "eval_steps_per_second": 13.708,
      "step": 807000
    },
    {
      "epoch": 8.65,
      "learning_rate": 2.3710901663004604e-05,
      "loss": 0.1867,
      "step": 807500
    },
    {
      "epoch": 8.65,
      "learning_rate": 2.364217487434221e-05,
      "loss": 0.1867,
      "step": 808000
    },
    {
      "epoch": 8.65,
      "eval_loss": 0.17686133086681366,
      "eval_runtime": 2.5799,
      "eval_samples_per_second": 890.356,
      "eval_steps_per_second": 13.954,
      "step": 808000
    },
    {
      "epoch": 8.66,
      "learning_rate": 2.3573602165821668e-05,
      "loss": 0.187,
      "step": 808500
    },
    {
      "epoch": 8.67,
      "learning_rate": 2.3505183724918196e-05,
      "loss": 0.1867,
      "step": 809000
    },
    {
      "epoch": 8.67,
      "eval_loss": 0.1787070780992508,
      "eval_runtime": 2.6677,
      "eval_samples_per_second": 861.032,
      "eval_steps_per_second": 13.495,
      "step": 809000
    },
    {
      "epoch": 8.67,
      "learning_rate": 2.3436919738685132e-05,
      "loss": 0.1864,
      "step": 809500
    },
    {
      "epoch": 8.68,
      "learning_rate": 2.3368810393753687e-05,
      "loss": 0.1866,
      "step": 810000
    },
    {
      "epoch": 8.68,
      "eval_loss": 0.17782823741436005,
      "eval_runtime": 2.6203,
      "eval_samples_per_second": 876.614,
      "eval_steps_per_second": 13.739,
      "step": 810000
    },
    {
      "epoch": 8.68,
      "learning_rate": 2.3300855876332162e-05,
      "loss": 0.1868,
      "step": 810500
    },
    {
      "epoch": 8.69,
      "learning_rate": 2.32330563722056e-05,
      "loss": 0.1864,
      "step": 811000
    },
    {
      "epoch": 8.69,
      "eval_loss": 0.1779273897409439,
      "eval_runtime": 2.7339,
      "eval_samples_per_second": 840.196,
      "eval_steps_per_second": 13.168,
      "step": 811000
    },
    {
      "epoch": 8.69,
      "learning_rate": 2.316541206673529e-05,
      "loss": 0.1865,
      "step": 811500
    },
    {
      "epoch": 8.7,
      "learning_rate": 2.309792314485815e-05,
      "loss": 0.1866,
      "step": 812000
    },
    {
      "epoch": 8.7,
      "eval_loss": 0.17686782777309418,
      "eval_runtime": 2.8998,
      "eval_samples_per_second": 792.124,
      "eval_steps_per_second": 12.415,
      "step": 812000
    },
    {
      "epoch": 8.7,
      "learning_rate": 2.3030589791086353e-05,
      "loss": 0.1865,
      "step": 812500
    },
    {
      "epoch": 8.71,
      "learning_rate": 2.2963412189506695e-05,
      "loss": 0.1869,
      "step": 813000
    },
    {
      "epoch": 8.71,
      "eval_loss": 0.1769571304321289,
      "eval_runtime": 2.6694,
      "eval_samples_per_second": 860.483,
      "eval_steps_per_second": 13.486,
      "step": 813000
    },
    {
      "epoch": 8.72,
      "learning_rate": 2.2896390523780156e-05,
      "loss": 0.1865,
      "step": 813500
    },
    {
      "epoch": 8.72,
      "learning_rate": 2.282952497714145e-05,
      "loss": 0.186,
      "step": 814000
    },
    {
      "epoch": 8.72,
      "eval_loss": 0.17855176329612732,
      "eval_runtime": 2.6874,
      "eval_samples_per_second": 854.735,
      "eval_steps_per_second": 13.396,
      "step": 814000
    },
    {
      "epoch": 8.73,
      "learning_rate": 2.2762815732398387e-05,
      "loss": 0.1862,
      "step": 814500
    },
    {
      "epoch": 8.73,
      "learning_rate": 2.2696262971931538e-05,
      "loss": 0.1863,
      "step": 815000
    },
    {
      "epoch": 8.73,
      "eval_loss": 0.17720898985862732,
      "eval_runtime": 2.6031,
      "eval_samples_per_second": 882.4,
      "eval_steps_per_second": 13.83,
      "step": 815000
    },
    {
      "epoch": 8.74,
      "learning_rate": 2.2629866877693577e-05,
      "loss": 0.1865,
      "step": 815500
    },
    {
      "epoch": 8.74,
      "learning_rate": 2.2563627631208887e-05,
      "loss": 0.1869,
      "step": 816000
    },
    {
      "epoch": 8.74,
      "eval_loss": 0.17861302196979523,
      "eval_runtime": 2.6611,
      "eval_samples_per_second": 863.188,
      "eval_steps_per_second": 13.528,
      "step": 816000
    },
    {
      "epoch": 8.75,
      "learning_rate": 2.2497545413573065e-05,
      "loss": 0.1863,
      "step": 816500
    },
    {
      "epoch": 8.75,
      "learning_rate": 2.2431620405452336e-05,
      "loss": 0.1859,
      "step": 817000
    },
    {
      "epoch": 8.75,
      "eval_loss": 0.17606213688850403,
      "eval_runtime": 2.6726,
      "eval_samples_per_second": 859.474,
      "eval_steps_per_second": 13.47,
      "step": 817000
    },
    {
      "epoch": 8.76,
      "learning_rate": 2.23658527870832e-05,
      "loss": 0.1863,
      "step": 817500
    },
    {
      "epoch": 8.77,
      "learning_rate": 2.230024273827179e-05,
      "loss": 0.1862,
      "step": 818000
    },
    {
      "epoch": 8.77,
      "eval_loss": 0.17465642094612122,
      "eval_runtime": 2.6831,
      "eval_samples_per_second": 856.102,
      "eval_steps_per_second": 13.417,
      "step": 818000
    },
    {
      "epoch": 8.77,
      "learning_rate": 2.223479043839345e-05,
      "loss": 0.1867,
      "step": 818500
    },
    {
      "epoch": 8.78,
      "learning_rate": 2.216949606639231e-05,
      "loss": 0.1863,
      "step": 819000
    },
    {
      "epoch": 8.78,
      "eval_loss": 0.17773117125034332,
      "eval_runtime": 2.6871,
      "eval_samples_per_second": 854.84,
      "eval_steps_per_second": 13.398,
      "step": 819000
    },
    {
      "epoch": 8.78,
      "learning_rate": 2.2104359800780665e-05,
      "loss": 0.1859,
      "step": 819500
    },
    {
      "epoch": 8.79,
      "learning_rate": 2.2039381819638596e-05,
      "loss": 0.186,
      "step": 820000
    },
    {
      "epoch": 8.79,
      "eval_loss": 0.1770503968000412,
      "eval_runtime": 2.6357,
      "eval_samples_per_second": 871.505,
      "eval_steps_per_second": 13.659,
      "step": 820000
    },
    {
      "epoch": 8.79,
      "learning_rate": 2.1974562300613417e-05,
      "loss": 0.186,
      "step": 820500
    },
    {
      "epoch": 8.8,
      "learning_rate": 2.1909901420919184e-05,
      "loss": 0.1856,
      "step": 821000
    },
    {
      "epoch": 8.8,
      "eval_loss": 0.17747129499912262,
      "eval_runtime": 2.664,
      "eval_samples_per_second": 862.246,
      "eval_steps_per_second": 13.514,
      "step": 821000
    },
    {
      "epoch": 8.8,
      "learning_rate": 2.1845399357336326e-05,
      "loss": 0.186,
      "step": 821500
    },
    {
      "epoch": 8.81,
      "learning_rate": 2.1781056286210997e-05,
      "loss": 0.186,
      "step": 822000
    },
    {
      "epoch": 8.81,
      "eval_loss": 0.1773909628391266,
      "eval_runtime": 2.5828,
      "eval_samples_per_second": 889.354,
      "eval_steps_per_second": 13.939,
      "step": 822000
    },
    {
      "epoch": 8.82,
      "learning_rate": 2.1716872383454674e-05,
      "loss": 0.1861,
      "step": 822500
    },
    {
      "epoch": 8.82,
      "learning_rate": 2.1652847824543744e-05,
      "loss": 0.1856,
      "step": 823000
    },
    {
      "epoch": 8.82,
      "eval_loss": 0.1759449690580368,
      "eval_runtime": 2.6867,
      "eval_samples_per_second": 854.948,
      "eval_steps_per_second": 13.399,
      "step": 823000
    },
    {
      "epoch": 8.83,
      "learning_rate": 2.1588982784518853e-05,
      "loss": 0.1862,
      "step": 823500
    },
    {
      "epoch": 8.83,
      "learning_rate": 2.1525277437984636e-05,
      "loss": 0.1857,
      "step": 824000
    },
    {
      "epoch": 8.83,
      "eval_loss": 0.1774652898311615,
      "eval_runtime": 2.6123,
      "eval_samples_per_second": 879.304,
      "eval_steps_per_second": 13.781,
      "step": 824000
    },
    {
      "epoch": 8.84,
      "learning_rate": 2.1461731959109053e-05,
      "loss": 0.186,
      "step": 824500
    },
    {
      "epoch": 8.84,
      "learning_rate": 2.1398346521623e-05,
      "loss": 0.1857,
      "step": 825000
    },
    {
      "epoch": 8.84,
      "eval_loss": 0.17699038982391357,
      "eval_runtime": 2.654,
      "eval_samples_per_second": 865.476,
      "eval_steps_per_second": 13.564,
      "step": 825000
    },
    {
      "epoch": 8.85,
      "learning_rate": 2.1335121298819867e-05,
      "loss": 0.1859,
      "step": 825500
    },
    {
      "epoch": 8.86,
      "learning_rate": 2.1272056463554978e-05,
      "loss": 0.1862,
      "step": 826000
    },
    {
      "epoch": 8.86,
      "eval_loss": 0.17693667113780975,
      "eval_runtime": 2.6428,
      "eval_samples_per_second": 869.15,
      "eval_steps_per_second": 13.622,
      "step": 826000
    },
    {
      "epoch": 8.86,
      "learning_rate": 2.1209152188245214e-05,
      "loss": 0.1858,
      "step": 826500
    },
    {
      "epoch": 8.87,
      "learning_rate": 2.114640864486845e-05,
      "loss": 0.1857,
      "step": 827000
    },
    {
      "epoch": 8.87,
      "eval_loss": 0.1788521409034729,
      "eval_runtime": 2.6742,
      "eval_samples_per_second": 858.952,
      "eval_steps_per_second": 13.462,
      "step": 827000
    },
    {
      "epoch": 8.87,
      "learning_rate": 2.1083826004963102e-05,
      "loss": 0.1859,
      "step": 827500
    },
    {
      "epoch": 8.88,
      "learning_rate": 2.1021404439627775e-05,
      "loss": 0.1855,
      "step": 828000
    },
    {
      "epoch": 8.88,
      "eval_loss": 0.17763476073741913,
      "eval_runtime": 2.5581,
      "eval_samples_per_second": 897.942,
      "eval_steps_per_second": 14.073,
      "step": 828000
    },
    {
      "epoch": 8.88,
      "learning_rate": 2.09591441195206e-05,
      "loss": 0.1856,
      "step": 828500
    },
    {
      "epoch": 8.89,
      "learning_rate": 2.089704521485896e-05,
      "loss": 0.1858,
      "step": 829000
    },
    {
      "epoch": 8.89,
      "eval_loss": 0.17711400985717773,
      "eval_runtime": 2.6039,
      "eval_samples_per_second": 882.145,
      "eval_steps_per_second": 13.826,
      "step": 829000
    },
    {
      "epoch": 8.89,
      "learning_rate": 2.083510789541883e-05,
      "loss": 0.1852,
      "step": 829500
    },
    {
      "epoch": 8.9,
      "learning_rate": 2.0773332330534513e-05,
      "loss": 0.1857,
      "step": 830000
    },
    {
      "epoch": 8.9,
      "eval_loss": 0.17438167333602905,
      "eval_runtime": 2.6569,
      "eval_samples_per_second": 864.555,
      "eval_steps_per_second": 13.55,
      "step": 830000
    },
    {
      "epoch": 8.91,
      "learning_rate": 2.0711718689098057e-05,
      "loss": 0.1858,
      "step": 830500
    },
    {
      "epoch": 8.91,
      "learning_rate": 2.0650267139558772e-05,
      "loss": 0.1854,
      "step": 831000
    },
    {
      "epoch": 8.91,
      "eval_loss": 0.17501012980937958,
      "eval_runtime": 2.6667,
      "eval_samples_per_second": 861.35,
      "eval_steps_per_second": 13.5,
      "step": 831000
    },
    {
      "epoch": 8.92,
      "learning_rate": 2.058897784992289e-05,
      "loss": 0.1855,
      "step": 831500
    },
    {
      "epoch": 8.92,
      "learning_rate": 2.052785098775293e-05,
      "loss": 0.1855,
      "step": 832000
    },
    {
      "epoch": 8.92,
      "eval_loss": 0.17379425466060638,
      "eval_runtime": 2.675,
      "eval_samples_per_second": 858.689,
      "eval_steps_per_second": 13.458,
      "step": 832000
    },
    {
      "epoch": 8.93,
      "learning_rate": 2.0466886720167436e-05,
      "loss": 0.1847,
      "step": 832500
    },
    {
      "epoch": 8.93,
      "learning_rate": 2.04060852138404e-05,
      "loss": 0.1854,
      "step": 833000
    },
    {
      "epoch": 8.93,
      "eval_loss": 0.1764400452375412,
      "eval_runtime": 2.6031,
      "eval_samples_per_second": 882.425,
      "eval_steps_per_second": 13.83,
      "step": 833000
    },
    {
      "epoch": 8.94,
      "learning_rate": 2.0345446635000783e-05,
      "loss": 0.1856,
      "step": 833500
    },
    {
      "epoch": 8.94,
      "learning_rate": 2.028497114943219e-05,
      "loss": 0.1851,
      "step": 834000
    },
    {
      "epoch": 8.94,
      "eval_loss": 0.17593778669834137,
      "eval_runtime": 2.5824,
      "eval_samples_per_second": 889.497,
      "eval_steps_per_second": 13.941,
      "step": 834000
    },
    {
      "epoch": 8.95,
      "learning_rate": 2.022465892247223e-05,
      "loss": 0.1855,
      "step": 834500
    },
    {
      "epoch": 8.96,
      "learning_rate": 2.0164510119012263e-05,
      "loss": 0.1849,
      "step": 835000
    },
    {
      "epoch": 8.96,
      "eval_loss": 0.1772100031375885,
      "eval_runtime": 2.6877,
      "eval_samples_per_second": 854.619,
      "eval_steps_per_second": 13.394,
      "step": 835000
    },
    {
      "epoch": 8.96,
      "learning_rate": 2.0104524903496834e-05,
      "loss": 0.1852,
      "step": 835500
    },
    {
      "epoch": 8.97,
      "learning_rate": 2.0044703439923217e-05,
      "loss": 0.1854,
      "step": 836000
    },
    {
      "epoch": 8.97,
      "eval_loss": 0.17744192481040955,
      "eval_runtime": 2.6114,
      "eval_samples_per_second": 879.619,
      "eval_steps_per_second": 13.786,
      "step": 836000
    },
    {
      "epoch": 8.97,
      "learning_rate": 1.998504589184101e-05,
      "loss": 0.1851,
      "step": 836500
    },
    {
      "epoch": 8.98,
      "learning_rate": 1.9925552422351654e-05,
      "loss": 0.1849,
      "step": 837000
    },
    {
      "epoch": 8.98,
      "eval_loss": 0.1755765676498413,
      "eval_runtime": 2.6326,
      "eval_samples_per_second": 872.529,
      "eval_steps_per_second": 13.675,
      "step": 837000
    },
    {
      "epoch": 8.98,
      "learning_rate": 1.9866223194108028e-05,
      "loss": 0.1851,
      "step": 837500
    },
    {
      "epoch": 8.99,
      "learning_rate": 1.9807058369314016e-05,
      "loss": 0.1845,
      "step": 838000
    },
    {
      "epoch": 8.99,
      "eval_loss": 0.17676672339439392,
      "eval_runtime": 2.6846,
      "eval_samples_per_second": 855.61,
      "eval_steps_per_second": 13.41,
      "step": 838000
    },
    {
      "epoch": 8.99,
      "learning_rate": 1.9748058109723953e-05,
      "loss": 0.1852,
      "step": 838500
    },
    {
      "epoch": 9.0,
      "learning_rate": 1.968922257664231e-05,
      "loss": 0.1853,
      "step": 839000
    },
    {
      "epoch": 9.0,
      "eval_loss": 0.17678546905517578,
      "eval_runtime": 2.6872,
      "eval_samples_per_second": 854.778,
      "eval_steps_per_second": 13.397,
      "step": 839000
    },
    {
      "epoch": 9.01,
      "learning_rate": 1.9630551930923155e-05,
      "loss": 0.1851,
      "step": 839500
    },
    {
      "epoch": 9.01,
      "learning_rate": 1.9572046332969825e-05,
      "loss": 0.1848,
      "step": 840000
    },
    {
      "epoch": 9.01,
      "eval_loss": 0.1751183122396469,
      "eval_runtime": 2.6372,
      "eval_samples_per_second": 870.984,
      "eval_steps_per_second": 13.651,
      "step": 840000
    },
    {
      "epoch": 9.02,
      "learning_rate": 1.95137059427344e-05,
      "loss": 0.1851,
      "step": 840500
    },
    {
      "epoch": 9.02,
      "learning_rate": 1.945553091971727e-05,
      "loss": 0.1852,
      "step": 841000
    },
    {
      "epoch": 9.02,
      "eval_loss": 0.17605267465114594,
      "eval_runtime": 2.6415,
      "eval_samples_per_second": 869.593,
      "eval_steps_per_second": 13.629,
      "step": 841000
    },
    {
      "epoch": 9.03,
      "learning_rate": 1.93975214229667e-05,
      "loss": 0.185,
      "step": 841500
    },
    {
      "epoch": 9.03,
      "learning_rate": 1.933967761107847e-05,
      "loss": 0.1846,
      "step": 842000
    },
    {
      "epoch": 9.03,
      "eval_loss": 0.17334681749343872,
      "eval_runtime": 2.679,
      "eval_samples_per_second": 857.418,
      "eval_steps_per_second": 13.438,
      "step": 842000
    },
    {
      "epoch": 9.04,
      "learning_rate": 1.928199964219533e-05,
      "loss": 0.1851,
      "step": 842500
    },
    {
      "epoch": 9.04,
      "learning_rate": 1.9224487674006694e-05,
      "loss": 0.1848,
      "step": 843000
    },
    {
      "epoch": 9.04,
      "eval_loss": 0.17702366411685944,
      "eval_runtime": 2.7205,
      "eval_samples_per_second": 844.336,
      "eval_steps_per_second": 13.233,
      "step": 843000
    },
    {
      "epoch": 9.05,
      "learning_rate": 1.9167141863748015e-05,
      "loss": 0.1848,
      "step": 843500
    },
    {
      "epoch": 9.06,
      "learning_rate": 1.9109962368200602e-05,
      "loss": 0.1846,
      "step": 844000
    },
    {
      "epoch": 9.06,
      "eval_loss": 0.17539054155349731,
      "eval_runtime": 2.6244,
      "eval_samples_per_second": 875.248,
      "eval_steps_per_second": 13.717,
      "step": 844000
    },
    {
      "epoch": 9.06,
      "learning_rate": 1.9052949343690977e-05,
      "loss": 0.1848,
      "step": 844500
    },
    {
      "epoch": 9.07,
      "learning_rate": 1.8996102946090586e-05,
      "loss": 0.1846,
      "step": 845000
    },
    {
      "epoch": 9.07,
      "eval_loss": 0.17318959534168243,
      "eval_runtime": 2.7201,
      "eval_samples_per_second": 844.445,
      "eval_steps_per_second": 13.235,
      "step": 845000
    },
    {
      "epoch": 9.07,
      "learning_rate": 1.8939423330815345e-05,
      "loss": 0.1847,
      "step": 845500
    },
    {
      "epoch": 9.08,
      "learning_rate": 1.888291065282509e-05,
      "loss": 0.1847,
      "step": 846000
    },
    {
      "epoch": 9.08,
      "eval_loss": 0.17560191452503204,
      "eval_runtime": 2.7101,
      "eval_samples_per_second": 847.565,
      "eval_steps_per_second": 13.284,
      "step": 846000
    },
    {
      "epoch": 9.08,
      "learning_rate": 1.882656506662338e-05,
      "loss": 0.1846,
      "step": 846500
    },
    {
      "epoch": 9.09,
      "learning_rate": 1.8770386726256865e-05,
      "loss": 0.1844,
      "step": 847000
    },
    {
      "epoch": 9.09,
      "eval_loss": 0.17197825014591217,
      "eval_runtime": 2.7201,
      "eval_samples_per_second": 844.459,
      "eval_steps_per_second": 13.235,
      "step": 847000
    },
    {
      "epoch": 9.09,
      "learning_rate": 1.8714375785315006e-05,
      "loss": 0.1845,
      "step": 847500
    },
    {
      "epoch": 9.1,
      "learning_rate": 1.8658532396929565e-05,
      "loss": 0.184,
      "step": 848000
    },
    {
      "epoch": 9.1,
      "eval_loss": 0.17653484642505646,
      "eval_runtime": 2.6616,
      "eval_samples_per_second": 863.031,
      "eval_steps_per_second": 13.526,
      "step": 848000
    },
    {
      "epoch": 9.11,
      "learning_rate": 1.8602856713774208e-05,
      "loss": 0.1843,
      "step": 848500
    },
    {
      "epoch": 9.11,
      "learning_rate": 1.8547348888064178e-05,
      "loss": 0.1848,
      "step": 849000
    },
    {
      "epoch": 9.11,
      "eval_loss": 0.1734277456998825,
      "eval_runtime": 2.6737,
      "eval_samples_per_second": 859.112,
      "eval_steps_per_second": 13.465,
      "step": 849000
    },
    {
      "epoch": 9.12,
      "learning_rate": 1.8492009071555703e-05,
      "loss": 0.1846,
      "step": 849500
    },
    {
      "epoch": 9.12,
      "learning_rate": 1.8436837415545772e-05,
      "loss": 0.1848,
      "step": 850000
    },
    {
      "epoch": 9.12,
      "eval_loss": 0.17425018548965454,
      "eval_runtime": 2.6779,
      "eval_samples_per_second": 857.76,
      "eval_steps_per_second": 13.443,
      "step": 850000
    },
    {
      "epoch": 9.13,
      "learning_rate": 1.838183407087156e-05,
      "loss": 0.1843,
      "step": 850500
    },
    {
      "epoch": 9.13,
      "learning_rate": 1.8326999187910095e-05,
      "loss": 0.1843,
      "step": 851000
    },
    {
      "epoch": 9.13,
      "eval_loss": 0.17367926239967346,
      "eval_runtime": 2.6854,
      "eval_samples_per_second": 855.351,
      "eval_steps_per_second": 13.406,
      "step": 851000
    },
    {
      "epoch": 9.14,
      "learning_rate": 1.8272332916577875e-05,
      "loss": 0.1846,
      "step": 851500
    },
    {
      "epoch": 9.15,
      "learning_rate": 1.8217835406330415e-05,
      "loss": 0.1844,
      "step": 852000
    },
    {
      "epoch": 9.15,
      "eval_loss": 0.17384441196918488,
      "eval_runtime": 2.6384,
      "eval_samples_per_second": 870.617,
      "eval_steps_per_second": 13.645,
      "step": 852000
    },
    {
      "epoch": 9.15,
      "learning_rate": 1.81635068061618e-05,
      "loss": 0.1844,
      "step": 852500
    },
    {
      "epoch": 9.16,
      "learning_rate": 1.810934726460436e-05,
      "loss": 0.1845,
      "step": 853000
    },
    {
      "epoch": 9.16,
      "eval_loss": 0.17389260232448578,
      "eval_runtime": 2.6711,
      "eval_samples_per_second": 859.942,
      "eval_steps_per_second": 13.478,
      "step": 853000
    },
    {
      "epoch": 9.16,
      "learning_rate": 1.80553569297282e-05,
      "loss": 0.1843,
      "step": 853500
    },
    {
      "epoch": 9.17,
      "learning_rate": 1.800153594914084e-05,
      "loss": 0.1843,
      "step": 854000
    },
    {
      "epoch": 9.17,
      "eval_loss": 0.17477978765964508,
      "eval_runtime": 2.681,
      "eval_samples_per_second": 856.769,
      "eval_steps_per_second": 13.428,
      "step": 854000
    },
    {
      "epoch": 9.17,
      "learning_rate": 1.7947884469986816e-05,
      "loss": 0.1841,
      "step": 854500
    },
    {
      "epoch": 9.18,
      "learning_rate": 1.7894402638947176e-05,
      "loss": 0.1841,
      "step": 855000
    },
    {
      "epoch": 9.18,
      "eval_loss": 0.1744370311498642,
      "eval_runtime": 2.6199,
      "eval_samples_per_second": 876.754,
      "eval_steps_per_second": 13.741,
      "step": 855000
    },
    {
      "epoch": 9.18,
      "learning_rate": 1.7841090602239237e-05,
      "loss": 0.1841,
      "step": 855500
    },
    {
      "epoch": 9.19,
      "learning_rate": 1.778794850561604e-05,
      "loss": 0.1844,
      "step": 856000
    },
    {
      "epoch": 9.19,
      "eval_loss": 0.17599613964557648,
      "eval_runtime": 2.6232,
      "eval_samples_per_second": 875.655,
      "eval_steps_per_second": 13.724,
      "step": 856000
    },
    {
      "epoch": 9.2,
      "learning_rate": 1.7734976494366073e-05,
      "loss": 0.1837,
      "step": 856500
    },
    {
      "epoch": 9.2,
      "learning_rate": 1.7682174713312805e-05,
      "loss": 0.1843,
      "step": 857000
    },
    {
      "epoch": 9.2,
      "eval_loss": 0.17385347187519073,
      "eval_runtime": 2.7089,
      "eval_samples_per_second": 847.948,
      "eval_steps_per_second": 13.29,
      "step": 857000
    },
    {
      "epoch": 9.21,
      "learning_rate": 1.7629543306814255e-05,
      "loss": 0.1838,
      "step": 857500
    },
    {
      "epoch": 9.21,
      "learning_rate": 1.75770824187627e-05,
      "loss": 0.1839,
      "step": 858000
    },
    {
      "epoch": 9.21,
      "eval_loss": 0.17458127439022064,
      "eval_runtime": 2.75,
      "eval_samples_per_second": 835.274,
      "eval_steps_per_second": 13.091,
      "step": 858000
    },
    {
      "epoch": 9.22,
      "learning_rate": 1.7524792192584186e-05,
      "loss": 0.1843,
      "step": 858500
    },
    {
      "epoch": 9.22,
      "learning_rate": 1.747267277123821e-05,
      "loss": 0.1839,
      "step": 859000
    },
    {
      "epoch": 9.22,
      "eval_loss": 0.1746589094400406,
      "eval_runtime": 2.6359,
      "eval_samples_per_second": 871.432,
      "eval_steps_per_second": 13.658,
      "step": 859000
    },
    {
      "epoch": 9.23,
      "learning_rate": 1.74207242972173e-05,
      "loss": 0.1837,
      "step": 859500
    },
    {
      "epoch": 9.23,
      "learning_rate": 1.7368946912546556e-05,
      "loss": 0.1836,
      "step": 860000
    },
    {
      "epoch": 9.23,
      "eval_loss": 0.1775263249874115,
      "eval_runtime": 2.6639,
      "eval_samples_per_second": 862.261,
      "eval_steps_per_second": 13.514,
      "step": 860000
    },
    {
      "epoch": 9.24,
      "learning_rate": 1.7317340758783407e-05,
      "loss": 0.1835,
      "step": 860500
    },
    {
      "epoch": 9.25,
      "learning_rate": 1.726590597701708e-05,
      "loss": 0.1842,
      "step": 861000
    },
    {
      "epoch": 9.25,
      "eval_loss": 0.17551767826080322,
      "eval_runtime": 2.6099,
      "eval_samples_per_second": 880.108,
      "eval_steps_per_second": 13.794,
      "step": 861000
    },
    {
      "epoch": 9.25,
      "learning_rate": 1.7214642707868325e-05,
      "loss": 0.1839,
      "step": 861500
    },
    {
      "epoch": 9.26,
      "learning_rate": 1.7163551091488952e-05,
      "loss": 0.1839,
      "step": 862000
    },
    {
      "epoch": 9.26,
      "eval_loss": 0.17372268438339233,
      "eval_runtime": 2.641,
      "eval_samples_per_second": 869.76,
      "eval_steps_per_second": 13.631,
      "step": 862000
    },
    {
      "epoch": 9.26,
      "learning_rate": 1.711263126756148e-05,
      "loss": 0.1841,
      "step": 862500
    },
    {
      "epoch": 9.27,
      "learning_rate": 1.7061883375298788e-05,
      "loss": 0.1834,
      "step": 863000
    },
    {
      "epoch": 9.27,
      "eval_loss": 0.17352163791656494,
      "eval_runtime": 2.6082,
      "eval_samples_per_second": 880.7,
      "eval_steps_per_second": 13.803,
      "step": 863000
    },
    {
      "epoch": 9.27,
      "learning_rate": 1.7011307553443647e-05,
      "loss": 0.1837,
      "step": 863500
    },
    {
      "epoch": 9.28,
      "learning_rate": 1.6960903940268456e-05,
      "loss": 0.1836,
      "step": 864000
    },
    {
      "epoch": 9.28,
      "eval_loss": 0.17270448803901672,
      "eval_runtime": 2.6998,
      "eval_samples_per_second": 850.79,
      "eval_steps_per_second": 13.334,
      "step": 864000
    },
    {
      "epoch": 9.28,
      "learning_rate": 1.6910672673574746e-05,
      "loss": 0.1838,
      "step": 864500
    },
    {
      "epoch": 9.29,
      "learning_rate": 1.6860613890692876e-05,
      "loss": 0.1832,
      "step": 865000
    },
    {
      "epoch": 9.29,
      "eval_loss": 0.1736259162425995,
      "eval_runtime": 2.674,
      "eval_samples_per_second": 859.019,
      "eval_steps_per_second": 13.463,
      "step": 865000
    },
    {
      "epoch": 9.3,
      "learning_rate": 1.6810727728481673e-05,
      "loss": 0.1833,
      "step": 865500
    },
    {
      "epoch": 9.3,
      "learning_rate": 1.6761014323327962e-05,
      "loss": 0.1835,
      "step": 866000
    },
    {
      "epoch": 9.3,
      "eval_loss": 0.17491458356380463,
      "eval_runtime": 2.6701,
      "eval_samples_per_second": 860.268,
      "eval_steps_per_second": 13.483,
      "step": 866000
    },
    {
      "epoch": 9.31,
      "learning_rate": 1.6711473811146333e-05,
      "loss": 0.1836,
      "step": 866500
    },
    {
      "epoch": 9.31,
      "learning_rate": 1.6662106327378645e-05,
      "loss": 0.1837,
      "step": 867000
    },
    {
      "epoch": 9.31,
      "eval_loss": 0.17311297357082367,
      "eval_runtime": 2.5749,
      "eval_samples_per_second": 892.083,
      "eval_steps_per_second": 13.981,
      "step": 867000
    },
    {
      "epoch": 9.32,
      "learning_rate": 1.6612912006993688e-05,
      "loss": 0.1835,
      "step": 867500
    },
    {
      "epoch": 9.32,
      "learning_rate": 1.6563890984486884e-05,
      "loss": 0.1834,
      "step": 868000
    },
    {
      "epoch": 9.32,
      "eval_loss": 0.1740087866783142,
      "eval_runtime": 2.6738,
      "eval_samples_per_second": 859.085,
      "eval_steps_per_second": 13.464,
      "step": 868000
    },
    {
      "epoch": 9.33,
      "learning_rate": 1.6515043393879825e-05,
      "loss": 0.1837,
      "step": 868500
    },
    {
      "epoch": 9.33,
      "learning_rate": 1.6466369368719955e-05,
      "loss": 0.1834,
      "step": 869000
    },
    {
      "epoch": 9.33,
      "eval_loss": 0.17570127546787262,
      "eval_runtime": 2.6358,
      "eval_samples_per_second": 871.447,
      "eval_steps_per_second": 13.658,
      "step": 869000
    },
    {
      "epoch": 9.34,
      "learning_rate": 1.641786904208022e-05,
      "loss": 0.1833,
      "step": 869500
    },
    {
      "epoch": 9.35,
      "learning_rate": 1.6369542546558626e-05,
      "loss": 0.1835,
      "step": 870000
    },
    {
      "epoch": 9.35,
      "eval_loss": 0.17456747591495514,
      "eval_runtime": 2.7306,
      "eval_samples_per_second": 841.196,
      "eval_steps_per_second": 13.184,
      "step": 870000
    },
    {
      "epoch": 9.35,
      "learning_rate": 1.6321390014277996e-05,
      "loss": 0.1831,
      "step": 870500
    },
    {
      "epoch": 9.36,
      "learning_rate": 1.6273411576885517e-05,
      "loss": 0.1836,
      "step": 871000
    },
    {
      "epoch": 9.36,
      "eval_loss": 0.17539818584918976,
      "eval_runtime": 2.6712,
      "eval_samples_per_second": 859.901,
      "eval_steps_per_second": 13.477,
      "step": 871000
    },
    {
      "epoch": 9.36,
      "learning_rate": 1.6225607365552378e-05,
      "loss": 0.1831,
      "step": 871500
    },
    {
      "epoch": 9.37,
      "learning_rate": 1.617797751097349e-05,
      "loss": 0.1832,
      "step": 872000
    },
    {
      "epoch": 9.37,
      "eval_loss": 0.1717691868543625,
      "eval_runtime": 2.6798,
      "eval_samples_per_second": 857.157,
      "eval_steps_per_second": 13.434,
      "step": 872000
    },
    {
      "epoch": 9.37,
      "learning_rate": 1.6130522143367032e-05,
      "loss": 0.1832,
      "step": 872500
    },
    {
      "epoch": 9.38,
      "learning_rate": 1.608324139247421e-05,
      "loss": 0.1835,
      "step": 873000
    },
    {
      "epoch": 9.38,
      "eval_loss": 0.1719122976064682,
      "eval_runtime": 2.6225,
      "eval_samples_per_second": 875.898,
      "eval_steps_per_second": 13.728,
      "step": 873000
    },
    {
      "epoch": 9.38,
      "learning_rate": 1.6036135387558756e-05,
      "loss": 0.1831,
      "step": 873500
    },
    {
      "epoch": 9.39,
      "learning_rate": 1.5989204257406693e-05,
      "loss": 0.1833,
      "step": 874000
    },
    {
      "epoch": 9.39,
      "eval_loss": 0.17478306591510773,
      "eval_runtime": 2.6101,
      "eval_samples_per_second": 880.046,
      "eval_steps_per_second": 13.793,
      "step": 874000
    },
    {
      "epoch": 9.4,
      "learning_rate": 1.594244813032595e-05,
      "loss": 0.1829,
      "step": 874500
    },
    {
      "epoch": 9.4,
      "learning_rate": 1.5895867134145974e-05,
      "loss": 0.1829,
      "step": 875000
    },
    {
      "epoch": 9.4,
      "eval_loss": 0.17394264042377472,
      "eval_runtime": 2.5878,
      "eval_samples_per_second": 887.623,
      "eval_steps_per_second": 13.911,
      "step": 875000
    },
    {
      "epoch": 9.41,
      "learning_rate": 1.5849461396217467e-05,
      "loss": 0.1834,
      "step": 875500
    },
    {
      "epoch": 9.41,
      "learning_rate": 1.5803231043411912e-05,
      "loss": 0.1827,
      "step": 876000
    },
    {
      "epoch": 9.41,
      "eval_loss": 0.17351944744586945,
      "eval_runtime": 2.6686,
      "eval_samples_per_second": 860.761,
      "eval_steps_per_second": 13.49,
      "step": 876000
    },
    {
      "epoch": 9.42,
      "learning_rate": 1.575717620212132e-05,
      "loss": 0.183,
      "step": 876500
    },
    {
      "epoch": 9.42,
      "learning_rate": 1.5711296998257902e-05,
      "loss": 0.1832,
      "step": 877000
    },
    {
      "epoch": 9.42,
      "eval_loss": 0.17347006499767303,
      "eval_runtime": 2.7428,
      "eval_samples_per_second": 837.47,
      "eval_steps_per_second": 13.125,
      "step": 877000
    },
    {
      "epoch": 9.43,
      "learning_rate": 1.5665593557253623e-05,
      "loss": 0.1833,
      "step": 877500
    },
    {
      "epoch": 9.44,
      "learning_rate": 1.562006600405996e-05,
      "loss": 0.1829,
      "step": 878000
    },
    {
      "epoch": 9.44,
      "eval_loss": 0.1734461635351181,
      "eval_runtime": 2.6113,
      "eval_samples_per_second": 879.646,
      "eval_steps_per_second": 13.786,
      "step": 878000
    },
    {
      "epoch": 9.44,
      "learning_rate": 1.5574714463147512e-05,
      "loss": 0.1831,
      "step": 878500
    },
    {
      "epoch": 9.45,
      "learning_rate": 1.5529539058505624e-05,
      "loss": 0.183,
      "step": 879000
    },
    {
      "epoch": 9.45,
      "eval_loss": 0.17375677824020386,
      "eval_runtime": 2.5315,
      "eval_samples_per_second": 907.374,
      "eval_steps_per_second": 14.221,
      "step": 879000
    },
    {
      "epoch": 9.45,
      "learning_rate": 1.5484539913642175e-05,
      "loss": 0.1826,
      "step": 879500
    },
    {
      "epoch": 9.46,
      "learning_rate": 1.543971715158307e-05,
      "loss": 0.1828,
      "step": 880000
    },
    {
      "epoch": 9.46,
      "eval_loss": 0.17431409657001495,
      "eval_runtime": 2.6398,
      "eval_samples_per_second": 870.144,
      "eval_steps_per_second": 13.637,
      "step": 880000
    },
    {
      "epoch": 9.46,
      "learning_rate": 1.539507089487205e-05,
      "loss": 0.183,
      "step": 880500
    },
    {
      "epoch": 9.47,
      "learning_rate": 1.535060126557028e-05,
      "loss": 0.1829,
      "step": 881000
    },
    {
      "epoch": 9.47,
      "eval_loss": 0.17408204078674316,
      "eval_runtime": 2.6439,
      "eval_samples_per_second": 868.809,
      "eval_steps_per_second": 13.617,
      "step": 881000
    },
    {
      "epoch": 9.47,
      "learning_rate": 1.5306308385255997e-05,
      "loss": 0.1827,
      "step": 881500
    },
    {
      "epoch": 9.48,
      "learning_rate": 1.5262192375024284e-05,
      "loss": 0.1827,
      "step": 882000
    },
    {
      "epoch": 9.48,
      "eval_loss": 0.17428572475910187,
      "eval_runtime": 2.6251,
      "eval_samples_per_second": 875.0,
      "eval_steps_per_second": 13.714,
      "step": 882000
    },
    {
      "epoch": 9.49,
      "learning_rate": 1.521825335548661e-05,
      "loss": 0.1832,
      "step": 882500
    },
    {
      "epoch": 9.49,
      "learning_rate": 1.5174491446770566e-05,
      "loss": 0.1827,
      "step": 883000
    },
    {
      "epoch": 9.49,
      "eval_loss": 0.17153075337409973,
      "eval_runtime": 2.6515,
      "eval_samples_per_second": 866.317,
      "eval_steps_per_second": 13.577,
      "step": 883000
    },
    {
      "epoch": 9.5,
      "learning_rate": 1.5130906768519563e-05,
      "loss": 0.1827,
      "step": 883500
    },
    {
      "epoch": 9.5,
      "learning_rate": 1.508749943989242e-05,
      "loss": 0.183,
      "step": 884000
    },
    {
      "epoch": 9.5,
      "eval_loss": 0.17301537096500397,
      "eval_runtime": 2.656,
      "eval_samples_per_second": 864.819,
      "eval_steps_per_second": 13.554,
      "step": 884000
    },
    {
      "epoch": 9.51,
      "learning_rate": 1.5044269579563144e-05,
      "loss": 0.1825,
      "step": 884500
    },
    {
      "epoch": 9.51,
      "learning_rate": 1.500121730572051e-05,
      "loss": 0.183,
      "step": 885000
    },
    {
      "epoch": 9.51,
      "eval_loss": 0.17374014854431152,
      "eval_runtime": 2.719,
      "eval_samples_per_second": 844.787,
      "eval_steps_per_second": 13.24,
      "step": 885000
    },
    {
      "epoch": 9.52,
      "learning_rate": 1.4958342736067783e-05,
      "loss": 0.1829,
      "step": 885500
    },
    {
      "epoch": 9.52,
      "learning_rate": 1.4915645987822406e-05,
      "loss": 0.1829,
      "step": 886000
    },
    {
      "epoch": 9.52,
      "eval_loss": 0.17604438960552216,
      "eval_runtime": 2.7026,
      "eval_samples_per_second": 849.921,
      "eval_steps_per_second": 13.32,
      "step": 886000
    },
    {
      "epoch": 9.53,
      "learning_rate": 1.4873127177715653e-05,
      "loss": 0.1827,
      "step": 886500
    },
    {
      "epoch": 9.54,
      "learning_rate": 1.4830786421992347e-05,
      "loss": 0.1829,
      "step": 887000
    },
    {
      "epoch": 9.54,
      "eval_loss": 0.17339639365673065,
      "eval_runtime": 2.7392,
      "eval_samples_per_second": 838.573,
      "eval_steps_per_second": 13.143,
      "step": 887000
    },
    {
      "epoch": 9.54,
      "learning_rate": 1.4788623836410479e-05,
      "loss": 0.1823,
      "step": 887500
    },
    {
      "epoch": 9.55,
      "learning_rate": 1.4746639536240942e-05,
      "loss": 0.1824,
      "step": 888000
    },
    {
      "epoch": 9.55,
      "eval_loss": 0.17382191121578217,
      "eval_runtime": 2.7435,
      "eval_samples_per_second": 837.254,
      "eval_steps_per_second": 13.122,
      "step": 888000
    },
    {
      "epoch": 9.55,
      "learning_rate": 1.4704833636267232e-05,
      "loss": 0.1825,
      "step": 888500
    },
    {
      "epoch": 9.56,
      "learning_rate": 1.4663206250785055e-05,
      "loss": 0.1824,
      "step": 889000
    },
    {
      "epoch": 9.56,
      "eval_loss": 0.17390523850917816,
      "eval_runtime": 2.7145,
      "eval_samples_per_second": 846.211,
      "eval_steps_per_second": 13.262,
      "step": 889000
    },
    {
      "epoch": 9.56,
      "learning_rate": 1.4621757493602125e-05,
      "loss": 0.1826,
      "step": 889500
    },
    {
      "epoch": 9.57,
      "learning_rate": 1.4580487478037748e-05,
      "loss": 0.1826,
      "step": 890000
    },
    {
      "epoch": 9.57,
      "eval_loss": 0.17268939316272736,
      "eval_runtime": 2.6865,
      "eval_samples_per_second": 855.002,
      "eval_steps_per_second": 13.4,
      "step": 890000
    },
    {
      "epoch": 9.57,
      "learning_rate": 1.4539396316922552e-05,
      "loss": 0.1822,
      "step": 890500
    },
    {
      "epoch": 9.58,
      "learning_rate": 1.4498484122598232e-05,
      "loss": 0.1824,
      "step": 891000
    },
    {
      "epoch": 9.58,
      "eval_loss": 0.17072877287864685,
      "eval_runtime": 2.7539,
      "eval_samples_per_second": 834.091,
      "eval_steps_per_second": 13.072,
      "step": 891000
    },
    {
      "epoch": 9.59,
      "learning_rate": 1.4457751006917137e-05,
      "loss": 0.1822,
      "step": 891500
    },
    {
      "epoch": 9.59,
      "learning_rate": 1.4417197081242083e-05,
      "loss": 0.1829,
      "step": 892000
    },
    {
      "epoch": 9.59,
      "eval_loss": 0.17135965824127197,
      "eval_runtime": 2.7156,
      "eval_samples_per_second": 845.864,
      "eval_steps_per_second": 13.257,
      "step": 892000
    },
    {
      "epoch": 9.6,
      "learning_rate": 1.4376822456445926e-05,
      "loss": 0.1825,
      "step": 892500
    },
    {
      "epoch": 9.6,
      "learning_rate": 1.433662724291136e-05,
      "loss": 0.1828,
      "step": 893000
    },
    {
      "epoch": 9.6,
      "eval_loss": 0.17201776802539825,
      "eval_runtime": 2.6531,
      "eval_samples_per_second": 865.779,
      "eval_steps_per_second": 13.569,
      "step": 893000
    },
    {
      "epoch": 9.61,
      "learning_rate": 1.4296611550530563e-05,
      "loss": 0.1823,
      "step": 893500
    },
    {
      "epoch": 9.61,
      "learning_rate": 1.4256775488704904e-05,
      "loss": 0.1825,
      "step": 894000
    },
    {
      "epoch": 9.61,
      "eval_loss": 0.17262525856494904,
      "eval_runtime": 2.6435,
      "eval_samples_per_second": 868.939,
      "eval_steps_per_second": 13.619,
      "step": 894000
    },
    {
      "epoch": 9.62,
      "learning_rate": 1.4217119166344665e-05,
      "loss": 0.1822,
      "step": 894500
    },
    {
      "epoch": 9.62,
      "learning_rate": 1.4177642691868717e-05,
      "loss": 0.1819,
      "step": 895000
    },
    {
      "epoch": 9.62,
      "eval_loss": 0.1716921329498291,
      "eval_runtime": 2.6607,
      "eval_samples_per_second": 863.304,
      "eval_steps_per_second": 13.53,
      "step": 895000
    },
    {
      "epoch": 9.63,
      "learning_rate": 1.4138346173204218e-05,
      "loss": 0.1823,
      "step": 895500
    },
    {
      "epoch": 9.64,
      "learning_rate": 1.4099229717786368e-05,
      "loss": 0.1825,
      "step": 896000
    },
    {
      "epoch": 9.64,
      "eval_loss": 0.1738402545452118,
      "eval_runtime": 2.7328,
      "eval_samples_per_second": 840.538,
      "eval_steps_per_second": 13.173,
      "step": 896000
    },
    {
      "epoch": 9.64,
      "learning_rate": 1.406029343255806e-05,
      "loss": 0.1823,
      "step": 896500
    },
    {
      "epoch": 9.65,
      "learning_rate": 1.4021537423969588e-05,
      "loss": 0.1823,
      "step": 897000
    },
    {
      "epoch": 9.65,
      "eval_loss": 0.1751101166009903,
      "eval_runtime": 2.7169,
      "eval_samples_per_second": 845.459,
      "eval_steps_per_second": 13.251,
      "step": 897000
    },
    {
      "epoch": 9.65,
      "learning_rate": 1.3982961797978431e-05,
      "loss": 0.1819,
      "step": 897500
    },
    {
      "epoch": 9.66,
      "learning_rate": 1.3944566660048863e-05,
      "loss": 0.1823,
      "step": 898000
    },
    {
      "epoch": 9.66,
      "eval_loss": 0.1745595484972,
      "eval_runtime": 2.78,
      "eval_samples_per_second": 826.25,
      "eval_steps_per_second": 12.95,
      "step": 898000
    },
    {
      "epoch": 9.66,
      "learning_rate": 1.3906352115151725e-05,
      "loss": 0.1821,
      "step": 898500
    },
    {
      "epoch": 9.67,
      "learning_rate": 1.3868318267764128e-05,
      "loss": 0.1821,
      "step": 899000
    },
    {
      "epoch": 9.67,
      "eval_loss": 0.17514048516750336,
      "eval_runtime": 2.6917,
      "eval_samples_per_second": 853.365,
      "eval_steps_per_second": 13.374,
      "step": 899000
    },
    {
      "epoch": 9.67,
      "learning_rate": 1.3830465221869146e-05,
      "loss": 0.1821,
      "step": 899500
    },
    {
      "epoch": 9.68,
      "learning_rate": 1.3792793080955574e-05,
      "loss": 0.1819,
      "step": 900000
    },
    {
      "epoch": 9.68,
      "eval_loss": 0.17348013818264008,
      "eval_runtime": 2.6872,
      "eval_samples_per_second": 854.803,
      "eval_steps_per_second": 13.397,
      "step": 900000
    },
    {
      "epoch": 9.69,
      "learning_rate": 1.3755301948017599e-05,
      "loss": 0.1822,
      "step": 900500
    },
    {
      "epoch": 9.69,
      "learning_rate": 1.3717991925554562e-05,
      "loss": 0.1821,
      "step": 901000
    },
    {
      "epoch": 9.69,
      "eval_loss": 0.17332369089126587,
      "eval_runtime": 2.5587,
      "eval_samples_per_second": 897.731,
      "eval_steps_per_second": 14.07,
      "step": 901000
    },
    {
      "epoch": 9.7,
      "learning_rate": 1.368086311557062e-05,
      "loss": 0.1821,
      "step": 901500
    },
    {
      "epoch": 9.7,
      "learning_rate": 1.3643915619574529e-05,
      "loss": 0.1821,
      "step": 902000
    },
    {
      "epoch": 9.7,
      "eval_loss": 0.1713598072528839,
      "eval_runtime": 2.7303,
      "eval_samples_per_second": 841.295,
      "eval_steps_per_second": 13.185,
      "step": 902000
    },
    {
      "epoch": 9.71,
      "learning_rate": 1.3607149538579341e-05,
      "loss": 0.182,
      "step": 902500
    },
    {
      "epoch": 9.71,
      "learning_rate": 1.35705649731021e-05,
      "loss": 0.1816,
      "step": 903000
    },
    {
      "epoch": 9.71,
      "eval_loss": 0.17178992927074432,
      "eval_runtime": 2.6316,
      "eval_samples_per_second": 872.853,
      "eval_steps_per_second": 13.68,
      "step": 903000
    },
    {
      "epoch": 9.72,
      "learning_rate": 1.3534162023163642e-05,
      "loss": 0.1821,
      "step": 903500
    },
    {
      "epoch": 9.73,
      "learning_rate": 1.3497940788288195e-05,
      "loss": 0.182,
      "step": 904000
    },
    {
      "epoch": 9.73,
      "eval_loss": 0.17401227355003357,
      "eval_runtime": 2.6109,
      "eval_samples_per_second": 879.785,
      "eval_steps_per_second": 13.789,
      "step": 904000
    },
    {
      "epoch": 9.73,
      "learning_rate": 1.3461901367503262e-05,
      "loss": 0.1816,
      "step": 904500
    },
    {
      "epoch": 9.74,
      "learning_rate": 1.3426043859339253e-05,
      "loss": 0.1822,
      "step": 905000
    },
    {
      "epoch": 9.74,
      "eval_loss": 0.16998076438903809,
      "eval_runtime": 2.6306,
      "eval_samples_per_second": 873.188,
      "eval_steps_per_second": 13.685,
      "step": 905000
    },
    {
      "epoch": 9.74,
      "learning_rate": 1.3390368361829197e-05,
      "loss": 0.1818,
      "step": 905500
    },
    {
      "epoch": 9.75,
      "learning_rate": 1.3354874972508582e-05,
      "loss": 0.1815,
      "step": 906000
    },
    {
      "epoch": 9.75,
      "eval_loss": 0.17435437440872192,
      "eval_runtime": 2.7558,
      "eval_samples_per_second": 833.52,
      "eval_steps_per_second": 13.063,
      "step": 906000
    },
    {
      "epoch": 9.75,
      "learning_rate": 1.3319563788414934e-05,
      "loss": 0.182,
      "step": 906500
    },
    {
      "epoch": 9.76,
      "learning_rate": 1.3284434906087695e-05,
      "loss": 0.1823,
      "step": 907000
    },
    {
      "epoch": 9.76,
      "eval_loss": 0.17397646605968475,
      "eval_runtime": 2.6707,
      "eval_samples_per_second": 860.067,
      "eval_steps_per_second": 13.479,
      "step": 907000
    },
    {
      "epoch": 9.76,
      "learning_rate": 1.3249488421567911e-05,
      "loss": 0.1818,
      "step": 907500
    },
    {
      "epoch": 9.77,
      "learning_rate": 1.3214724430397915e-05,
      "loss": 0.1817,
      "step": 908000
    },
    {
      "epoch": 9.77,
      "eval_loss": 0.17298473417758942,
      "eval_runtime": 2.659,
      "eval_samples_per_second": 863.865,
      "eval_steps_per_second": 13.539,
      "step": 908000
    },
    {
      "epoch": 9.78,
      "learning_rate": 1.3180143027621145e-05,
      "loss": 0.1819,
      "step": 908500
    },
    {
      "epoch": 9.78,
      "learning_rate": 1.314574430778182e-05,
      "loss": 0.1817,
      "step": 909000
    },
    {
      "epoch": 9.78,
      "eval_loss": 0.17148981988430023,
      "eval_runtime": 2.7078,
      "eval_samples_per_second": 848.287,
      "eval_steps_per_second": 13.295,
      "step": 909000
    },
    {
      "epoch": 9.79,
      "learning_rate": 1.311152836492473e-05,
      "loss": 0.1817,
      "step": 909500
    },
    {
      "epoch": 9.79,
      "learning_rate": 1.3077495292594966e-05,
      "loss": 0.1817,
      "step": 910000
    },
    {
      "epoch": 9.79,
      "eval_loss": 0.17322474718093872,
      "eval_runtime": 2.6678,
      "eval_samples_per_second": 861.024,
      "eval_steps_per_second": 13.495,
      "step": 910000
    },
    {
      "epoch": 9.8,
      "learning_rate": 1.3043645183837645e-05,
      "loss": 0.1817,
      "step": 910500
    },
    {
      "epoch": 9.8,
      "learning_rate": 1.3009978131197669e-05,
      "loss": 0.1821,
      "step": 911000
    },
    {
      "epoch": 9.8,
      "eval_loss": 0.17097479104995728,
      "eval_runtime": 2.7046,
      "eval_samples_per_second": 849.303,
      "eval_steps_per_second": 13.311,
      "step": 911000
    },
    {
      "epoch": 9.81,
      "learning_rate": 1.297649422671947e-05,
      "loss": 0.1815,
      "step": 911500
    },
    {
      "epoch": 9.81,
      "learning_rate": 1.2943193561946762e-05,
      "loss": 0.1814,
      "step": 912000
    },
    {
      "epoch": 9.81,
      "eval_loss": 0.1718176305294037,
      "eval_runtime": 2.6645,
      "eval_samples_per_second": 862.075,
      "eval_steps_per_second": 13.511,
      "step": 912000
    },
    {
      "epoch": 9.82,
      "learning_rate": 1.291007622792231e-05,
      "loss": 0.1816,
      "step": 912500
    },
    {
      "epoch": 9.83,
      "learning_rate": 1.2877142315187628e-05,
      "loss": 0.1818,
      "step": 913000
    },
    {
      "epoch": 9.83,
      "eval_loss": 0.1710846871137619,
      "eval_runtime": 2.7026,
      "eval_samples_per_second": 849.924,
      "eval_steps_per_second": 13.321,
      "step": 913000
    },
    {
      "epoch": 9.83,
      "learning_rate": 1.2844391913782773e-05,
      "loss": 0.182,
      "step": 913500
    },
    {
      "epoch": 9.84,
      "learning_rate": 1.28118251132461e-05,
      "loss": 0.1815,
      "step": 914000
    },
    {
      "epoch": 9.84,
      "eval_loss": 0.17333008348941803,
      "eval_runtime": 2.5502,
      "eval_samples_per_second": 900.72,
      "eval_steps_per_second": 14.117,
      "step": 914000
    },
    {
      "epoch": 9.84,
      "learning_rate": 1.2779442002613984e-05,
      "loss": 0.1814,
      "step": 914500
    },
    {
      "epoch": 9.85,
      "learning_rate": 1.274724267042063e-05,
      "loss": 0.1814,
      "step": 915000
    },
    {
      "epoch": 9.85,
      "eval_loss": 0.1717572808265686,
      "eval_runtime": 2.636,
      "eval_samples_per_second": 871.395,
      "eval_steps_per_second": 13.657,
      "step": 915000
    },
    {
      "epoch": 9.85,
      "learning_rate": 1.2715227204697775e-05,
      "loss": 0.1814,
      "step": 915500
    },
    {
      "epoch": 9.86,
      "learning_rate": 1.2683395692974472e-05,
      "loss": 0.1819,
      "step": 916000
    },
    {
      "epoch": 9.86,
      "eval_loss": 0.17225094139575958,
      "eval_runtime": 2.6373,
      "eval_samples_per_second": 870.95,
      "eval_steps_per_second": 13.65,
      "step": 916000
    },
    {
      "epoch": 9.86,
      "learning_rate": 1.2651748222276879e-05,
      "loss": 0.1815,
      "step": 916500
    },
    {
      "epoch": 9.87,
      "learning_rate": 1.2620284879127947e-05,
      "loss": 0.1816,
      "step": 917000
    },
    {
      "epoch": 9.87,
      "eval_loss": 0.17146611213684082,
      "eval_runtime": 2.7493,
      "eval_samples_per_second": 835.499,
      "eval_steps_per_second": 13.094,
      "step": 917000
    },
    {
      "epoch": 9.88,
      "learning_rate": 1.2589005749547281e-05,
      "loss": 0.1815,
      "step": 917500
    },
    {
      "epoch": 9.88,
      "learning_rate": 1.2557910919050803e-05,
      "loss": 0.1813,
      "step": 918000
    },
    {
      "epoch": 9.88,
      "eval_loss": 0.16947948932647705,
      "eval_runtime": 2.7503,
      "eval_samples_per_second": 835.189,
      "eval_steps_per_second": 13.09,
      "step": 918000
    },
    {
      "epoch": 9.89,
      "learning_rate": 1.2527000472650597e-05,
      "loss": 0.1815,
      "step": 918500
    },
    {
      "epoch": 9.89,
      "learning_rate": 1.2496274494854666e-05,
      "loss": 0.1812,
      "step": 919000
    },
    {
      "epoch": 9.89,
      "eval_loss": 0.17127934098243713,
      "eval_runtime": 2.6734,
      "eval_samples_per_second": 859.205,
      "eval_steps_per_second": 13.466,
      "step": 919000
    },
    {
      "epoch": 9.9,
      "learning_rate": 1.2465733069666629e-05,
      "loss": 0.1813,
      "step": 919500
    },
    {
      "epoch": 9.9,
      "learning_rate": 1.24353762805856e-05,
      "loss": 0.1814,
      "step": 920000
    },
    {
      "epoch": 9.9,
      "eval_loss": 0.17229017615318298,
      "eval_runtime": 2.6708,
      "eval_samples_per_second": 860.053,
      "eval_steps_per_second": 13.479,
      "step": 920000
    },
    {
      "epoch": 9.91,
      "learning_rate": 1.240520421060586e-05,
      "loss": 0.1818,
      "step": 920500
    },
    {
      "epoch": 9.91,
      "learning_rate": 1.2375216942216713e-05,
      "loss": 0.1817,
      "step": 921000
    },
    {
      "epoch": 9.91,
      "eval_loss": 0.17294897139072418,
      "eval_runtime": 2.7216,
      "eval_samples_per_second": 843.974,
      "eval_steps_per_second": 13.227,
      "step": 921000
    },
    {
      "epoch": 9.92,
      "learning_rate": 1.2345414557402198e-05,
      "loss": 0.1811,
      "step": 921500
    },
    {
      "epoch": 9.93,
      "learning_rate": 1.2315797137640906e-05,
      "loss": 0.1814,
      "step": 922000
    },
    {
      "epoch": 9.93,
      "eval_loss": 0.1728929877281189,
      "eval_runtime": 2.6867,
      "eval_samples_per_second": 854.941,
      "eval_steps_per_second": 13.399,
      "step": 922000
    },
    {
      "epoch": 9.93,
      "learning_rate": 1.2286364763905723e-05,
      "loss": 0.1813,
      "step": 922500
    },
    {
      "epoch": 9.94,
      "learning_rate": 1.225711751666363e-05,
      "loss": 0.1815,
      "step": 923000
    },
    {
      "epoch": 9.94,
      "eval_loss": 0.17273372411727905,
      "eval_runtime": 2.6851,
      "eval_samples_per_second": 855.459,
      "eval_steps_per_second": 13.407,
      "step": 923000
    },
    {
      "epoch": 9.94,
      "learning_rate": 1.2228055475875488e-05,
      "loss": 0.1811,
      "step": 923500
    },
    {
      "epoch": 9.95,
      "learning_rate": 1.2199178720995825e-05,
      "loss": 0.1814,
      "step": 924000
    },
    {
      "epoch": 9.95,
      "eval_loss": 0.17107011377811432,
      "eval_runtime": 2.7626,
      "eval_samples_per_second": 831.466,
      "eval_steps_per_second": 13.031,
      "step": 924000
    },
    {
      "epoch": 9.95,
      "learning_rate": 1.217048733097256e-05,
      "loss": 0.181,
      "step": 924500
    },
    {
      "epoch": 9.96,
      "learning_rate": 1.2141981384246874e-05,
      "loss": 0.1808,
      "step": 925000
    },
    {
      "epoch": 9.96,
      "eval_loss": 0.17153695225715637,
      "eval_runtime": 2.6072,
      "eval_samples_per_second": 881.021,
      "eval_steps_per_second": 13.808,
      "step": 925000
    },
    {
      "epoch": 9.96,
      "learning_rate": 1.211366095875293e-05,
      "loss": 0.1811,
      "step": 925500
    },
    {
      "epoch": 9.97,
      "learning_rate": 1.2085526131917685e-05,
      "loss": 0.181,
      "step": 926000
    },
    {
      "epoch": 9.97,
      "eval_loss": 0.1717982143163681,
      "eval_runtime": 2.6346,
      "eval_samples_per_second": 871.846,
      "eval_steps_per_second": 13.664,
      "step": 926000
    },
    {
      "epoch": 9.98,
      "learning_rate": 1.2057576980660691e-05,
      "loss": 0.181,
      "step": 926500
    },
    {
      "epoch": 9.98,
      "learning_rate": 1.2029813581393866e-05,
      "loss": 0.1813,
      "step": 927000
    },
    {
      "epoch": 9.98,
      "eval_loss": 0.17178404331207275,
      "eval_runtime": 2.6835,
      "eval_samples_per_second": 855.987,
      "eval_steps_per_second": 13.416,
      "step": 927000
    },
    {
      "epoch": 9.99,
      "learning_rate": 1.2002236010021269e-05,
      "loss": 0.1811,
      "step": 927500
    },
    {
      "epoch": 9.99,
      "learning_rate": 1.197484434193893e-05,
      "loss": 0.1811,
      "step": 928000
    },
    {
      "epoch": 9.99,
      "eval_loss": 0.17153653502464294,
      "eval_runtime": 2.6987,
      "eval_samples_per_second": 851.135,
      "eval_steps_per_second": 13.34,
      "step": 928000
    },
    {
      "epoch": 10.0,
      "learning_rate": 1.1947638652034617e-05,
      "loss": 0.1809,
      "step": 928500
    },
    {
      "epoch": 10.0,
      "learning_rate": 1.192061901468768e-05,
      "loss": 0.1811,
      "step": 929000
    },
    {
      "epoch": 10.0,
      "eval_loss": 0.17235822975635529,
      "eval_runtime": 2.6438,
      "eval_samples_per_second": 868.809,
      "eval_steps_per_second": 13.617,
      "step": 929000
    },
    {
      "epoch": 10.01,
      "learning_rate": 1.1893785503768736e-05,
      "loss": 0.1812,
      "step": 929500
    },
    {
      "epoch": 10.02,
      "learning_rate": 1.1867138192639601e-05,
      "loss": 0.1809,
      "step": 930000
    },
    {
      "epoch": 10.02,
      "eval_loss": 0.1709609031677246,
      "eval_runtime": 2.6178,
      "eval_samples_per_second": 877.459,
      "eval_steps_per_second": 13.752,
      "step": 930000
    },
    {
      "epoch": 10.02,
      "learning_rate": 1.1840677154152987e-05,
      "loss": 0.181,
      "step": 930500
    },
    {
      "epoch": 10.03,
      "learning_rate": 1.1814402460652382e-05,
      "loss": 0.1811,
      "step": 931000
    },
    {
      "epoch": 10.03,
      "eval_loss": 0.17144934833049774,
      "eval_runtime": 2.7383,
      "eval_samples_per_second": 838.849,
      "eval_steps_per_second": 13.147,
      "step": 931000
    },
    {
      "epoch": 10.03,
      "learning_rate": 1.178831418397181e-05,
      "loss": 0.181,
      "step": 931500
    },
    {
      "epoch": 10.04,
      "learning_rate": 1.176241239543558e-05,
      "loss": 0.181,
      "step": 932000
    },
    {
      "epoch": 10.04,
      "eval_loss": 0.17210912704467773,
      "eval_runtime": 2.6312,
      "eval_samples_per_second": 872.975,
      "eval_steps_per_second": 13.682,
      "step": 932000
    },
    {
      "epoch": 10.04,
      "learning_rate": 1.173669716585822e-05,
      "loss": 0.1809,
      "step": 932500
    },
    {
      "epoch": 10.05,
      "learning_rate": 1.171116856554418e-05,
      "loss": 0.1809,
      "step": 933000
    },
    {
      "epoch": 10.05,
      "eval_loss": 0.17279262840747833,
      "eval_runtime": 2.687,
      "eval_samples_per_second": 854.858,
      "eval_steps_per_second": 13.398,
      "step": 933000
    },
    {
      "epoch": 10.05,
      "learning_rate": 1.168582666428768e-05,
      "loss": 0.1809,
      "step": 933500
    },
    {
      "epoch": 10.06,
      "learning_rate": 1.1660671531372517e-05,
      "loss": 0.1807,
      "step": 934000
    },
    {
      "epoch": 10.06,
      "eval_loss": 0.17214839160442352,
      "eval_runtime": 2.6862,
      "eval_samples_per_second": 855.103,
      "eval_steps_per_second": 13.402,
      "step": 934000
    },
    {
      "epoch": 10.07,
      "learning_rate": 1.1635703235571846e-05,
      "loss": 0.181,
      "step": 934500
    },
    {
      "epoch": 10.07,
      "learning_rate": 1.1610921845148052e-05,
      "loss": 0.1805,
      "step": 935000
    },
    {
      "epoch": 10.07,
      "eval_loss": 0.17261220514774323,
      "eval_runtime": 2.7622,
      "eval_samples_per_second": 831.593,
      "eval_steps_per_second": 13.033,
      "step": 935000
    },
    {
      "epoch": 10.08,
      "learning_rate": 1.1586327427852503e-05,
      "loss": 0.1805,
      "step": 935500
    },
    {
      "epoch": 10.08,
      "learning_rate": 1.156192005092539e-05,
      "loss": 0.1807,
      "step": 936000
    },
    {
      "epoch": 10.08,
      "eval_loss": 0.17041905224323273,
      "eval_runtime": 2.643,
      "eval_samples_per_second": 869.101,
      "eval_steps_per_second": 13.621,
      "step": 936000
    },
    {
      "epoch": 10.09,
      "learning_rate": 1.153769978109557e-05,
      "loss": 0.1806,
      "step": 936500
    },
    {
      "epoch": 10.09,
      "learning_rate": 1.1513666684580308e-05,
      "loss": 0.1809,
      "step": 937000
    },
    {
      "epoch": 10.09,
      "eval_loss": 0.1718713790178299,
      "eval_runtime": 2.6411,
      "eval_samples_per_second": 869.707,
      "eval_steps_per_second": 13.631,
      "step": 937000
    },
    {
      "epoch": 10.1,
      "learning_rate": 1.1489820827085185e-05,
      "loss": 0.1808,
      "step": 937500
    },
    {
      "epoch": 10.1,
      "learning_rate": 1.1466162273803876e-05,
      "loss": 0.1809,
      "step": 938000
    },
    {
      "epoch": 10.1,
      "eval_loss": 0.17236891388893127,
      "eval_runtime": 2.4881,
      "eval_samples_per_second": 923.197,
      "eval_steps_per_second": 14.469,
      "step": 938000
    },
    {
      "epoch": 10.11,
      "learning_rate": 1.144269108941795e-05,
      "loss": 0.1808,
      "step": 938500
    },
    {
      "epoch": 10.12,
      "learning_rate": 1.1419407338096732e-05,
      "loss": 0.1807,
      "step": 939000
    },
    {
      "epoch": 10.12,
      "eval_loss": 0.17213864624500275,
      "eval_runtime": 2.697,
      "eval_samples_per_second": 851.681,
      "eval_steps_per_second": 13.348,
      "step": 939000
    },
    {
      "epoch": 10.12,
      "learning_rate": 1.1396311083497103e-05,
      "loss": 0.1808,
      "step": 939500
    },
    {
      "epoch": 10.13,
      "learning_rate": 1.1373402388763346e-05,
      "loss": 0.1806,
      "step": 940000
    },
    {
      "epoch": 10.13,
      "eval_loss": 0.17225030064582825,
      "eval_runtime": 2.5852,
      "eval_samples_per_second": 888.512,
      "eval_steps_per_second": 13.925,
      "step": 940000
    },
    {
      "epoch": 10.13,
      "learning_rate": 1.1350681316526965e-05,
      "loss": 0.1805,
      "step": 940500
    },
    {
      "epoch": 10.14,
      "learning_rate": 1.1328147928906494e-05,
      "loss": 0.1809,
      "step": 941000
    },
    {
      "epoch": 10.14,
      "eval_loss": 0.1728110909461975,
      "eval_runtime": 2.6803,
      "eval_samples_per_second": 857.002,
      "eval_steps_per_second": 13.431,
      "step": 941000
    },
    {
      "epoch": 10.14,
      "learning_rate": 1.1305802287507358e-05,
      "loss": 0.1806,
      "step": 941500
    },
    {
      "epoch": 10.15,
      "learning_rate": 1.1283644453421678e-05,
      "loss": 0.1806,
      "step": 942000
    },
    {
      "epoch": 10.15,
      "eval_loss": 0.17073097825050354,
      "eval_runtime": 2.5804,
      "eval_samples_per_second": 890.183,
      "eval_steps_per_second": 13.951,
      "step": 942000
    },
    {
      "epoch": 10.15,
      "learning_rate": 1.1261674487228149e-05,
      "loss": 0.1805,
      "step": 942500
    },
    {
      "epoch": 10.16,
      "learning_rate": 1.1239892448991798e-05,
      "loss": 0.1806,
      "step": 943000
    },
    {
      "epoch": 10.16,
      "eval_loss": 0.17256046831607819,
      "eval_runtime": 2.657,
      "eval_samples_per_second": 864.515,
      "eval_steps_per_second": 13.549,
      "step": 943000
    },
    {
      "epoch": 10.17,
      "learning_rate": 1.1218298398263894e-05,
      "loss": 0.1808,
      "step": 943500
    },
    {
      "epoch": 10.17,
      "learning_rate": 1.1196892394081743e-05,
      "loss": 0.1803,
      "step": 944000
    },
    {
      "epoch": 10.17,
      "eval_loss": 0.1697072833776474,
      "eval_runtime": 2.5585,
      "eval_samples_per_second": 897.801,
      "eval_steps_per_second": 14.071,
      "step": 944000
    },
    {
      "epoch": 10.18,
      "learning_rate": 1.1175674494968552e-05,
      "loss": 0.1803,
      "step": 944500
    },
    {
      "epoch": 10.18,
      "learning_rate": 1.1154644758933235e-05,
      "loss": 0.1807,
      "step": 945000
    },
    {
      "epoch": 10.18,
      "eval_loss": 0.17261387407779694,
      "eval_runtime": 2.6868,
      "eval_samples_per_second": 854.925,
      "eval_steps_per_second": 13.399,
      "step": 945000
    },
    {
      "epoch": 10.19,
      "learning_rate": 1.11338032434703e-05,
      "loss": 0.1804,
      "step": 945500
    },
    {
      "epoch": 10.19,
      "learning_rate": 1.1113150005559644e-05,
      "loss": 0.1808,
      "step": 946000
    },
    {
      "epoch": 10.19,
      "eval_loss": 0.17092828452587128,
      "eval_runtime": 2.6611,
      "eval_samples_per_second": 863.19,
      "eval_steps_per_second": 13.528,
      "step": 946000
    },
    {
      "epoch": 10.2,
      "learning_rate": 1.1092685101666438e-05,
      "loss": 0.1806,
      "step": 946500
    },
    {
      "epoch": 10.2,
      "learning_rate": 1.1072408587740942e-05,
      "loss": 0.1804,
      "step": 947000
    },
    {
      "epoch": 10.2,
      "eval_loss": 0.17135068774223328,
      "eval_runtime": 2.7301,
      "eval_samples_per_second": 841.37,
      "eval_steps_per_second": 13.186,
      "step": 947000
    },
    {
      "epoch": 10.21,
      "learning_rate": 1.1052320519218383e-05,
      "loss": 0.1804,
      "step": 947500
    },
    {
      "epoch": 10.22,
      "learning_rate": 1.1032420951018755e-05,
      "loss": 0.1806,
      "step": 948000
    },
    {
      "epoch": 10.22,
      "eval_loss": 0.16970402002334595,
      "eval_runtime": 2.5839,
      "eval_samples_per_second": 888.964,
      "eval_steps_per_second": 13.932,
      "step": 948000
    },
    {
      "epoch": 10.22,
      "learning_rate": 1.1012709937546722e-05,
      "loss": 0.1805,
      "step": 948500
    },
    {
      "epoch": 10.23,
      "learning_rate": 1.0993187532691458e-05,
      "loss": 0.1804,
      "step": 949000
    },
    {
      "epoch": 10.23,
      "eval_loss": 0.17099051177501678,
      "eval_runtime": 2.7073,
      "eval_samples_per_second": 848.444,
      "eval_steps_per_second": 13.297,
      "step": 949000
    },
    {
      "epoch": 10.23,
      "learning_rate": 1.0973853789826454e-05,
      "loss": 0.1804,
      "step": 949500
    },
    {
      "epoch": 10.24,
      "learning_rate": 1.0954708761809438e-05,
      "loss": 0.1806,
      "step": 950000
    },
    {
      "epoch": 10.24,
      "eval_loss": 0.1725110560655594,
      "eval_runtime": 2.6133,
      "eval_samples_per_second": 878.965,
      "eval_steps_per_second": 13.776,
      "step": 950000
    },
    {
      "epoch": 10.24,
      "learning_rate": 1.0935752500982175e-05,
      "loss": 0.1805,
      "step": 950500
    },
    {
      "epoch": 10.25,
      "learning_rate": 1.091698505917036e-05,
      "loss": 0.1804,
      "step": 951000
    },
    {
      "epoch": 10.25,
      "eval_loss": 0.1698637306690216,
      "eval_runtime": 2.5965,
      "eval_samples_per_second": 884.656,
      "eval_steps_per_second": 13.865,
      "step": 951000
    },
    {
      "epoch": 10.25,
      "learning_rate": 1.0898406487683472e-05,
      "loss": 0.1805,
      "step": 951500
    },
    {
      "epoch": 10.26,
      "learning_rate": 1.0880016837314599e-05,
      "loss": 0.1803,
      "step": 952000
    },
    {
      "epoch": 10.26,
      "eval_loss": 0.17085076868534088,
      "eval_runtime": 2.596,
      "eval_samples_per_second": 884.806,
      "eval_steps_per_second": 13.867,
      "step": 952000
    },
    {
      "epoch": 10.27,
      "learning_rate": 1.0861816158340365e-05,
      "loss": 0.1807,
      "step": 952500
    },
    {
      "epoch": 10.27,
      "learning_rate": 1.084380450052071e-05,
      "loss": 0.1803,
      "step": 953000
    },
    {
      "epoch": 10.27,
      "eval_loss": 0.17190536856651306,
      "eval_runtime": 2.595,
      "eval_samples_per_second": 885.153,
      "eval_steps_per_second": 13.873,
      "step": 953000
    },
    {
      "epoch": 10.28,
      "learning_rate": 1.0825981913098828e-05,
      "loss": 0.1799,
      "step": 953500
    },
    {
      "epoch": 10.28,
      "learning_rate": 1.0808348444801e-05,
      "loss": 0.1802,
      "step": 954000
    },
    {
      "epoch": 10.28,
      "eval_loss": 0.16949187219142914,
      "eval_runtime": 2.6166,
      "eval_samples_per_second": 877.869,
      "eval_steps_per_second": 13.758,
      "step": 954000
    },
    {
      "epoch": 10.29,
      "learning_rate": 1.0790904143836438e-05,
      "loss": 0.1804,
      "step": 954500
    },
    {
      "epoch": 10.29,
      "learning_rate": 1.0773649057897206e-05,
      "loss": 0.1802,
      "step": 955000
    },
    {
      "epoch": 10.29,
      "eval_loss": 0.16995471715927124,
      "eval_runtime": 2.6165,
      "eval_samples_per_second": 877.886,
      "eval_steps_per_second": 13.759,
      "step": 955000
    },
    {
      "epoch": 10.3,
      "learning_rate": 1.0756583234158057e-05,
      "loss": 0.1799,
      "step": 955500
    },
    {
      "epoch": 10.31,
      "learning_rate": 1.073970671927628e-05,
      "loss": 0.1802,
      "step": 956000
    },
    {
      "epoch": 10.31,
      "eval_loss": 0.17191793024539948,
      "eval_runtime": 2.6164,
      "eval_samples_per_second": 877.931,
      "eval_steps_per_second": 13.759,
      "step": 956000
    },
    {
      "epoch": 10.31,
      "learning_rate": 1.0723019559391643e-05,
      "loss": 0.1804,
      "step": 956500
    },
    {
      "epoch": 10.32,
      "learning_rate": 1.0706521800126198e-05,
      "loss": 0.18,
      "step": 957000
    },
    {
      "epoch": 10.32,
      "eval_loss": 0.17065568268299103,
      "eval_runtime": 2.6326,
      "eval_samples_per_second": 872.521,
      "eval_steps_per_second": 13.675,
      "step": 957000
    },
    {
      "epoch": 10.32,
      "learning_rate": 1.0690213486584175e-05,
      "loss": 0.18,
      "step": 957500
    },
    {
      "epoch": 10.33,
      "learning_rate": 1.0674094663351906e-05,
      "loss": 0.18,
      "step": 958000
    },
    {
      "epoch": 10.33,
      "eval_loss": 0.1698225736618042,
      "eval_runtime": 2.6744,
      "eval_samples_per_second": 858.898,
      "eval_steps_per_second": 13.461,
      "step": 958000
    },
    {
      "epoch": 10.33,
      "learning_rate": 1.0658165374497611e-05,
      "loss": 0.1804,
      "step": 958500
    },
    {
      "epoch": 10.34,
      "learning_rate": 1.0642425663571383e-05,
      "loss": 0.1802,
      "step": 959000
    },
    {
      "epoch": 10.34,
      "eval_loss": 0.17188780009746552,
      "eval_runtime": 2.6352,
      "eval_samples_per_second": 871.663,
      "eval_steps_per_second": 13.661,
      "step": 959000
    },
    {
      "epoch": 10.34,
      "learning_rate": 1.062687557360497e-05,
      "loss": 0.1802,
      "step": 959500
    },
    {
      "epoch": 10.35,
      "learning_rate": 1.0611515147111736e-05,
      "loss": 0.1802,
      "step": 960000
    },
    {
      "epoch": 10.35,
      "eval_loss": 0.16846837103366852,
      "eval_runtime": 2.7425,
      "eval_samples_per_second": 837.549,
      "eval_steps_per_second": 13.127,
      "step": 960000
    },
    {
      "epoch": 10.36,
      "learning_rate": 1.0596344426086501e-05,
      "loss": 0.1798,
      "step": 960500
    },
    {
      "epoch": 10.36,
      "learning_rate": 1.0581363452005424e-05,
      "loss": 0.1805,
      "step": 961000
    },
    {
      "epoch": 10.36,
      "eval_loss": 0.17073865234851837,
      "eval_runtime": 2.5938,
      "eval_samples_per_second": 885.585,
      "eval_steps_per_second": 13.879,
      "step": 961000
    },
    {
      "epoch": 10.37,
      "learning_rate": 1.0566572265825932e-05,
      "loss": 0.18,
      "step": 961500
    },
    {
      "epoch": 10.37,
      "learning_rate": 1.0551970907986557e-05,
      "loss": 0.1801,
      "step": 962000
    },
    {
      "epoch": 10.37,
      "eval_loss": 0.17134888470172882,
      "eval_runtime": 2.5813,
      "eval_samples_per_second": 889.852,
      "eval_steps_per_second": 13.946,
      "step": 962000
    },
    {
      "epoch": 10.38,
      "learning_rate": 1.0537559418406849e-05,
      "loss": 0.18,
      "step": 962500
    },
    {
      "epoch": 10.38,
      "learning_rate": 1.0523337836487271e-05,
      "loss": 0.1799,
      "step": 963000
    },
    {
      "epoch": 10.38,
      "eval_loss": 0.17050015926361084,
      "eval_runtime": 2.6391,
      "eval_samples_per_second": 870.378,
      "eval_steps_per_second": 13.641,
      "step": 963000
    },
    {
      "epoch": 10.39,
      "learning_rate": 1.0509306201109092e-05,
      "loss": 0.1801,
      "step": 963500
    },
    {
      "epoch": 10.39,
      "learning_rate": 1.0495464550634267e-05,
      "loss": 0.18,
      "step": 964000
    },
    {
      "epoch": 10.39,
      "eval_loss": 0.17047521471977234,
      "eval_runtime": 2.6548,
      "eval_samples_per_second": 865.215,
      "eval_steps_per_second": 13.56,
      "step": 964000
    },
    {
      "epoch": 10.4,
      "learning_rate": 1.0481812922905339e-05,
      "loss": 0.1805,
      "step": 964500
    },
    {
      "epoch": 10.41,
      "learning_rate": 1.046835135524533e-05,
      "loss": 0.1798,
      "step": 965000
    },
    {
      "epoch": 10.41,
      "eval_loss": 0.17172271013259888,
      "eval_runtime": 2.5812,
      "eval_samples_per_second": 889.895,
      "eval_steps_per_second": 13.947,
      "step": 965000
    },
    {
      "epoch": 10.41,
      "learning_rate": 1.0455079884457653e-05,
      "loss": 0.1801,
      "step": 965500
    },
    {
      "epoch": 10.42,
      "learning_rate": 1.044199854682601e-05,
      "loss": 0.1797,
      "step": 966000
    },
    {
      "epoch": 10.42,
      "eval_loss": 0.16956347227096558,
      "eval_runtime": 2.6699,
      "eval_samples_per_second": 860.34,
      "eval_steps_per_second": 13.484,
      "step": 966000
    },
    {
      "epoch": 10.42,
      "learning_rate": 1.0429107378114277e-05,
      "loss": 0.1802,
      "step": 966500
    },
    {
      "epoch": 10.43,
      "learning_rate": 1.0416406413566414e-05,
      "loss": 0.1802,
      "step": 967000
    },
    {
      "epoch": 10.43,
      "eval_loss": 0.17151953279972076,
      "eval_runtime": 2.5629,
      "eval_samples_per_second": 896.252,
      "eval_steps_per_second": 14.047,
      "step": 967000
    },
    {
      "epoch": 10.43,
      "learning_rate": 1.0403895687906366e-05,
      "loss": 0.1803,
      "step": 967500
    },
    {
      "epoch": 10.44,
      "learning_rate": 1.0391575235337991e-05,
      "loss": 0.1798,
      "step": 968000
    },
    {
      "epoch": 10.44,
      "eval_loss": 0.17273712158203125,
      "eval_runtime": 2.6967,
      "eval_samples_per_second": 851.792,
      "eval_steps_per_second": 13.35,
      "step": 968000
    },
    {
      "epoch": 10.44,
      "learning_rate": 1.0379445089544929e-05,
      "loss": 0.1799,
      "step": 968500
    },
    {
      "epoch": 10.45,
      "learning_rate": 1.0367505283690547e-05,
      "loss": 0.1797,
      "step": 969000
    },
    {
      "epoch": 10.45,
      "eval_loss": 0.17085492610931396,
      "eval_runtime": 2.6519,
      "eval_samples_per_second": 866.18,
      "eval_steps_per_second": 13.575,
      "step": 969000
    },
    {
      "epoch": 10.46,
      "learning_rate": 1.0355755850417803e-05,
      "loss": 0.1797,
      "step": 969500
    },
    {
      "epoch": 10.46,
      "learning_rate": 1.0344196821849202e-05,
      "loss": 0.1799,
      "step": 970000
    },
    {
      "epoch": 10.46,
      "eval_loss": 0.1711302548646927,
      "eval_runtime": 2.5979,
      "eval_samples_per_second": 884.178,
      "eval_steps_per_second": 13.857,
      "step": 970000
    },
    {
      "epoch": 10.82,
      "learning_rate": 1.0332828229586692e-05,
      "loss": 0.1799,
      "step": 970500
    },
    {
      "epoch": 10.83,
      "learning_rate": 1.032165010471157e-05,
      "loss": 0.1796,
      "step": 971000
    },
    {
      "epoch": 10.83,
      "eval_loss": 0.17119638621807098,
      "eval_runtime": 2.5911,
      "eval_samples_per_second": 886.512,
      "eval_steps_per_second": 13.894,
      "step": 971000
    },
    {
      "epoch": 10.84,
      "learning_rate": 1.0310662477784401e-05,
      "loss": 0.1804,
      "step": 971500
    },
    {
      "epoch": 10.84,
      "learning_rate": 1.0299865378844936e-05,
      "loss": 0.1798,
      "step": 972000
    },
    {
      "epoch": 10.84,
      "eval_loss": 0.1710081547498703,
      "eval_runtime": 2.5437,
      "eval_samples_per_second": 903.014,
      "eval_steps_per_second": 14.153,
      "step": 972000
    },
    {
      "epoch": 10.85,
      "learning_rate": 1.028925883741203e-05,
      "loss": 0.18,
      "step": 972500
    },
    {
      "epoch": 10.85,
      "learning_rate": 1.0278842882483569e-05,
      "loss": 0.1797,
      "step": 973000
    },
    {
      "epoch": 10.85,
      "eval_loss": 0.17146818339824677,
      "eval_runtime": 2.5692,
      "eval_samples_per_second": 894.045,
      "eval_steps_per_second": 14.012,
      "step": 973000
    },
    {
      "epoch": 10.86,
      "learning_rate": 1.026861754253637e-05,
      "loss": 0.1796,
      "step": 973500
    },
    {
      "epoch": 10.86,
      "learning_rate": 1.025858284552612e-05,
      "loss": 0.1797,
      "step": 974000
    },
    {
      "epoch": 10.86,
      "eval_loss": 0.1706797480583191,
      "eval_runtime": 2.6865,
      "eval_samples_per_second": 855.008,
      "eval_steps_per_second": 13.4,
      "step": 974000
    },
    {
      "epoch": 10.87,
      "learning_rate": 1.0248738818887307e-05,
      "loss": 0.1799,
      "step": 974500
    },
    {
      "epoch": 10.87,
      "learning_rate": 1.023908548953311e-05,
      "loss": 0.1799,
      "step": 975000
    },
    {
      "epoch": 10.87,
      "eval_loss": 0.1708817481994629,
      "eval_runtime": 2.5759,
      "eval_samples_per_second": 891.738,
      "eval_steps_per_second": 13.976,
      "step": 975000
    },
    {
      "epoch": 10.88,
      "learning_rate": 1.0229622883855378e-05,
      "loss": 0.1798,
      "step": 975500
    },
    {
      "epoch": 10.89,
      "learning_rate": 1.02203510277245e-05,
      "loss": 0.1796,
      "step": 976000
    },
    {
      "epoch": 10.89,
      "eval_loss": 0.1709393560886383,
      "eval_runtime": 2.6094,
      "eval_samples_per_second": 880.296,
      "eval_steps_per_second": 13.797,
      "step": 976000
    },
    {
      "epoch": 10.89,
      "learning_rate": 1.021126994648939e-05,
      "loss": 0.1801,
      "step": 976500
    },
    {
      "epoch": 10.9,
      "learning_rate": 1.0202379664977364e-05,
      "loss": 0.1799,
      "step": 977000
    },
    {
      "epoch": 10.9,
      "eval_loss": 0.17174768447875977,
      "eval_runtime": 2.6289,
      "eval_samples_per_second": 873.739,
      "eval_steps_per_second": 13.694,
      "step": 977000
    },
    {
      "epoch": 10.9,
      "learning_rate": 1.019368020749412e-05,
      "loss": 0.1797,
      "step": 977500
    },
    {
      "epoch": 10.91,
      "learning_rate": 1.018517159782365e-05,
      "loss": 0.1797,
      "step": 978000
    },
    {
      "epoch": 10.91,
      "eval_loss": 0.16800174117088318,
      "eval_runtime": 2.57,
      "eval_samples_per_second": 893.767,
      "eval_steps_per_second": 14.008,
      "step": 978000
    },
    {
      "epoch": 10.91,
      "learning_rate": 1.0176853859228149e-05,
      "loss": 0.1794,
      "step": 978500
    },
    {
      "epoch": 10.92,
      "learning_rate": 1.0168727014448004e-05,
      "loss": 0.1794,
      "step": 979000
    },
    {
      "epoch": 10.92,
      "eval_loss": 0.16953879594802856,
      "eval_runtime": 2.6173,
      "eval_samples_per_second": 877.629,
      "eval_steps_per_second": 13.755,
      "step": 979000
    },
    {
      "epoch": 10.92,
      "learning_rate": 1.0160791085701714e-05,
      "loss": 0.1798,
      "step": 979500
    },
    {
      "epoch": 10.93,
      "learning_rate": 1.0153046094685783e-05,
      "loss": 0.1794,
      "step": 980000
    },
    {
      "epoch": 10.93,
      "eval_loss": 0.1709355264902115,
      "eval_runtime": 2.587,
      "eval_samples_per_second": 887.915,
      "eval_steps_per_second": 13.916,
      "step": 980000
    }
  ],
  "max_steps": 1000000,
  "num_train_epochs": 12,
  "total_flos": 6.869770816498864e+22,
  "trial_name": null,
  "trial_params": null
}