{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.289347100094979,
  "global_step": 140000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "learning_rate": 5.999999999999999e-06,
      "loss": 0.9345,
      "step": 500
    },
    {
      "epoch": 0.03,
      "learning_rate": 1.1999999999999999e-05,
      "loss": 0.7514,
      "step": 1000
    },
    {
      "epoch": 0.03,
      "eval_loss": 0.9052248597145081,
      "eval_runtime": 0.5073,
      "eval_samples_per_second": 1971.284,
      "eval_steps_per_second": 31.541,
      "step": 1000
    },
    {
      "epoch": 0.05,
      "learning_rate": 1.7999999999999997e-05,
      "loss": 0.7408,
      "step": 1500
    },
    {
      "epoch": 0.06,
      "learning_rate": 2.3999999999999997e-05,
      "loss": 0.74,
      "step": 2000
    },
    {
      "epoch": 0.06,
      "eval_loss": 0.9058064818382263,
      "eval_runtime": 0.5123,
      "eval_samples_per_second": 1952.096,
      "eval_steps_per_second": 31.234,
      "step": 2000
    },
    {
      "epoch": 0.08,
      "learning_rate": 2.9999999999999997e-05,
      "loss": 0.7398,
      "step": 2500
    },
    {
      "epoch": 0.09,
      "learning_rate": 3.5999999999999994e-05,
      "loss": 0.7395,
      "step": 3000
    },
    {
      "epoch": 0.09,
      "eval_loss": 0.9068936705589294,
      "eval_runtime": 0.499,
      "eval_samples_per_second": 2003.855,
      "eval_steps_per_second": 32.062,
      "step": 3000
    },
    {
      "epoch": 0.11,
      "learning_rate": 4.2e-05,
      "loss": 0.7394,
      "step": 3500
    },
    {
      "epoch": 0.12,
      "learning_rate": 4.7999999999999994e-05,
      "loss": 0.7392,
      "step": 4000
    },
    {
      "epoch": 0.12,
      "eval_loss": 0.9032273888587952,
      "eval_runtime": 0.5086,
      "eval_samples_per_second": 1966.372,
      "eval_steps_per_second": 31.462,
      "step": 4000
    },
    {
      "epoch": 0.14,
      "learning_rate": 5.399999999999999e-05,
      "loss": 0.7389,
      "step": 4500
    },
    {
      "epoch": 0.15,
      "learning_rate": 5.9999999999999995e-05,
      "loss": 0.7386,
      "step": 5000
    },
    {
      "epoch": 0.15,
      "eval_loss": 0.9034351706504822,
      "eval_runtime": 0.5101,
      "eval_samples_per_second": 1960.24,
      "eval_steps_per_second": 31.364,
      "step": 5000
    },
    {
      "epoch": 0.17,
      "learning_rate": 6.599999999999999e-05,
      "loss": 0.7382,
      "step": 5500
    },
    {
      "epoch": 0.18,
      "learning_rate": 7.199999999999999e-05,
      "loss": 0.7377,
      "step": 6000
    },
    {
      "epoch": 0.18,
      "eval_loss": 0.8661078810691833,
      "eval_runtime": 0.5119,
      "eval_samples_per_second": 1953.649,
      "eval_steps_per_second": 31.258,
      "step": 6000
    },
    {
      "epoch": 0.2,
      "learning_rate": 7.8e-05,
      "loss": 0.7375,
      "step": 6500
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.4e-05,
      "loss": 0.7373,
      "step": 7000
    },
    {
      "epoch": 0.21,
      "eval_loss": 0.8658460974693298,
      "eval_runtime": 0.5198,
      "eval_samples_per_second": 1923.698,
      "eval_steps_per_second": 30.779,
      "step": 7000
    },
    {
      "epoch": 0.23,
      "learning_rate": 8.999999999999999e-05,
      "loss": 0.7367,
      "step": 7500
    },
    {
      "epoch": 0.25,
      "learning_rate": 9.599999999999999e-05,
      "loss": 0.7207,
      "step": 8000
    },
    {
      "epoch": 0.25,
      "eval_loss": 0.8675529956817627,
      "eval_runtime": 0.5261,
      "eval_samples_per_second": 1900.917,
      "eval_steps_per_second": 30.415,
      "step": 8000
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000102,
      "loss": 0.6905,
      "step": 8500
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00010799999999999998,
      "loss": 0.6746,
      "step": 9000
    },
    {
      "epoch": 0.28,
      "eval_loss": 0.8773286938667297,
      "eval_runtime": 0.5159,
      "eval_samples_per_second": 1938.548,
      "eval_steps_per_second": 31.017,
      "step": 9000
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00011399999999999999,
      "loss": 0.6617,
      "step": 9500
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00011999999999999999,
      "loss": 0.6406,
      "step": 10000
    },
    {
      "epoch": 0.31,
      "eval_loss": 0.8803548812866211,
      "eval_runtime": 0.5084,
      "eval_samples_per_second": 1966.828,
      "eval_steps_per_second": 31.469,
      "step": 10000
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00012599999999999997,
      "loss": 0.6275,
      "step": 10500
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00013199999999999998,
      "loss": 0.6079,
      "step": 11000
    },
    {
      "epoch": 0.34,
      "eval_loss": 0.8707549571990967,
      "eval_runtime": 0.5111,
      "eval_samples_per_second": 1956.647,
      "eval_steps_per_second": 31.306,
      "step": 11000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000138,
      "loss": 0.5977,
      "step": 11500
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00014399999999999998,
      "loss": 0.5907,
      "step": 12000
    },
    {
      "epoch": 0.37,
      "eval_loss": 0.8676372766494751,
      "eval_runtime": 0.5179,
      "eval_samples_per_second": 1930.992,
      "eval_steps_per_second": 30.896,
      "step": 12000
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00015,
      "loss": 0.5834,
      "step": 12500
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000156,
      "loss": 0.5764,
      "step": 13000
    },
    {
      "epoch": 0.4,
      "eval_loss": 0.8676320910453796,
      "eval_runtime": 0.5238,
      "eval_samples_per_second": 1909.011,
      "eval_steps_per_second": 30.544,
      "step": 13000
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000162,
      "loss": 0.5696,
      "step": 13500
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000168,
      "loss": 0.5648,
      "step": 14000
    },
    {
      "epoch": 0.43,
      "eval_loss": 0.860458254814148,
      "eval_runtime": 0.5396,
      "eval_samples_per_second": 1853.225,
      "eval_steps_per_second": 29.652,
      "step": 14000
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00017399999999999997,
      "loss": 0.5602,
      "step": 14500
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00017999999999999998,
      "loss": 0.556,
      "step": 15000
    },
    {
      "epoch": 0.46,
      "eval_loss": 0.8584501147270203,
      "eval_runtime": 0.5323,
      "eval_samples_per_second": 1878.593,
      "eval_steps_per_second": 30.057,
      "step": 15000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000186,
      "loss": 0.5519,
      "step": 15500
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00019199999999999998,
      "loss": 0.5459,
      "step": 16000
    },
    {
      "epoch": 0.49,
      "eval_loss": 0.8554501533508301,
      "eval_runtime": 0.5107,
      "eval_samples_per_second": 1957.988,
      "eval_steps_per_second": 31.328,
      "step": 16000
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000198,
      "loss": 0.5412,
      "step": 16500
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000204,
      "loss": 0.5374,
      "step": 17000
    },
    {
      "epoch": 0.52,
      "eval_loss": 0.8554509282112122,
      "eval_runtime": 0.5145,
      "eval_samples_per_second": 1943.819,
      "eval_steps_per_second": 31.101,
      "step": 17000
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00020999999999999998,
      "loss": 0.5338,
      "step": 17500
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00021599999999999996,
      "loss": 0.5305,
      "step": 18000
    },
    {
      "epoch": 0.55,
      "eval_loss": 0.8557196259498596,
      "eval_runtime": 0.5163,
      "eval_samples_per_second": 1936.818,
      "eval_steps_per_second": 30.989,
      "step": 18000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00022199999999999998,
      "loss": 0.5273,
      "step": 18500
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00022799999999999999,
      "loss": 0.5243,
      "step": 19000
    },
    {
      "epoch": 0.58,
      "eval_loss": 0.8606428503990173,
      "eval_runtime": 0.5087,
      "eval_samples_per_second": 1965.911,
      "eval_steps_per_second": 31.455,
      "step": 19000
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000234,
      "loss": 0.5215,
      "step": 19500
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00023999999999999998,
      "loss": 0.5188,
      "step": 20000
    },
    {
      "epoch": 0.61,
      "eval_loss": 0.8626545071601868,
      "eval_runtime": 0.4996,
      "eval_samples_per_second": 2001.549,
      "eval_steps_per_second": 32.025,
      "step": 20000
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00024599999999999996,
      "loss": 0.516,
      "step": 20500
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025199999999999995,
      "loss": 0.5136,
      "step": 21000
    },
    {
      "epoch": 0.64,
      "eval_loss": 0.8466567993164062,
      "eval_runtime": 0.521,
      "eval_samples_per_second": 1919.499,
      "eval_steps_per_second": 30.712,
      "step": 21000
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.000258,
      "loss": 0.5112,
      "step": 21500
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00026399999999999997,
      "loss": 0.5089,
      "step": 22000
    },
    {
      "epoch": 0.67,
      "eval_loss": 0.8465701937675476,
      "eval_runtime": 0.5015,
      "eval_samples_per_second": 1994.167,
      "eval_steps_per_second": 31.907,
      "step": 22000
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00027,
      "loss": 0.5067,
      "step": 22500
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000276,
      "loss": 0.5048,
      "step": 23000
    },
    {
      "epoch": 0.7,
      "eval_loss": 0.8367106914520264,
      "eval_runtime": 0.5157,
      "eval_samples_per_second": 1939.06,
      "eval_steps_per_second": 31.025,
      "step": 23000
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00028199999999999997,
      "loss": 0.5029,
      "step": 23500
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00028799999999999995,
      "loss": 0.5009,
      "step": 24000
    },
    {
      "epoch": 0.74,
      "eval_loss": 0.8282895088195801,
      "eval_runtime": 0.5191,
      "eval_samples_per_second": 1926.314,
      "eval_steps_per_second": 30.821,
      "step": 24000
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.000294,
      "loss": 0.4986,
      "step": 24500
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0003,
      "loss": 0.4968,
      "step": 25000
    },
    {
      "epoch": 0.77,
      "eval_loss": 0.8393005132675171,
      "eval_runtime": 0.5106,
      "eval_samples_per_second": 1958.371,
      "eval_steps_per_second": 31.334,
      "step": 25000
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00029999920715161553,
      "loss": 0.495,
      "step": 25500
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002999968286151326,
      "loss": 0.4932,
      "step": 26000
    },
    {
      "epoch": 0.8,
      "eval_loss": 0.8276752829551697,
      "eval_runtime": 0.5157,
      "eval_samples_per_second": 1939.179,
      "eval_steps_per_second": 31.027,
      "step": 26000
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002999928644165624,
      "loss": 0.4916,
      "step": 26500
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002999873145992569,
      "loss": 0.4899,
      "step": 27000
    },
    {
      "epoch": 0.83,
      "eval_loss": 0.8151644468307495,
      "eval_runtime": 0.5093,
      "eval_samples_per_second": 1963.529,
      "eval_steps_per_second": 31.416,
      "step": 27000
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.000299980179223908,
      "loss": 0.4883,
      "step": 27500
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002999714583685469,
      "loss": 0.4868,
      "step": 28000
    },
    {
      "epoch": 0.86,
      "eval_loss": 0.8103233575820923,
      "eval_runtime": 0.5126,
      "eval_samples_per_second": 1950.786,
      "eval_steps_per_second": 31.213,
      "step": 28000
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.00029996115212854366,
      "loss": 0.4855,
      "step": 28500
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.00029994926061660554,
      "loss": 0.4841,
      "step": 29000
    },
    {
      "epoch": 0.89,
      "eval_loss": 0.8084909319877625,
      "eval_runtime": 0.5298,
      "eval_samples_per_second": 1887.571,
      "eval_steps_per_second": 30.201,
      "step": 29000
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002999357839627762,
      "loss": 0.4823,
      "step": 29500
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.00029992072231443425,
      "loss": 0.4806,
      "step": 30000
    },
    {
      "epoch": 0.92,
      "eval_loss": 0.8131279945373535,
      "eval_runtime": 0.5066,
      "eval_samples_per_second": 1974.102,
      "eval_steps_per_second": 31.586,
      "step": 30000
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002999040758362914,
      "loss": 0.4791,
      "step": 30500
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.00029988584471039094,
      "loss": 0.4776,
      "step": 31000
    },
    {
      "epoch": 0.95,
      "eval_loss": 0.8034030795097351,
      "eval_runtime": 0.4996,
      "eval_samples_per_second": 2001.487,
      "eval_steps_per_second": 32.024,
      "step": 31000
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002998660291361054,
      "loss": 0.476,
      "step": 31500
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002998446293301349,
      "loss": 0.4741,
      "step": 32000
    },
    {
      "epoch": 0.98,
      "eval_loss": 0.8003847599029541,
      "eval_runtime": 0.5033,
      "eval_samples_per_second": 1986.864,
      "eval_steps_per_second": 31.79,
      "step": 32000
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002998216455265042,
      "loss": 0.4724,
      "step": 32500
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00029979707797656046,
      "loss": 0.4709,
      "step": 33000
    },
    {
      "epoch": 1.01,
      "eval_loss": 0.7997156977653503,
      "eval_runtime": 0.5174,
      "eval_samples_per_second": 1932.776,
      "eval_steps_per_second": 30.924,
      "step": 33000
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00029977092694897053,
      "loss": 0.4691,
      "step": 33500
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002997431927297178,
      "loss": 0.467,
      "step": 34000
    },
    {
      "epoch": 1.04,
      "eval_loss": 0.803820013999939,
      "eval_runtime": 0.5367,
      "eval_samples_per_second": 1863.102,
      "eval_steps_per_second": 29.81,
      "step": 34000
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00029971387562209936,
      "loss": 0.4653,
      "step": 34500
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00029968297594672226,
      "loss": 0.4638,
      "step": 35000
    },
    {
      "epoch": 1.07,
      "eval_loss": 0.8011023998260498,
      "eval_runtime": 0.5429,
      "eval_samples_per_second": 1842.037,
      "eval_steps_per_second": 29.473,
      "step": 35000
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002996504940415005,
      "loss": 0.462,
      "step": 35500
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.00029961643026165096,
      "loss": 0.4601,
      "step": 36000
    },
    {
      "epoch": 1.1,
      "eval_loss": 0.8029471635818481,
      "eval_runtime": 0.5067,
      "eval_samples_per_second": 1973.42,
      "eval_steps_per_second": 31.575,
      "step": 36000
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.00029958078497968973,
      "loss": 0.4587,
      "step": 36500
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002995435585854278,
      "loss": 0.457,
      "step": 37000
    },
    {
      "epoch": 1.13,
      "eval_loss": 0.7997317910194397,
      "eval_runtime": 0.5216,
      "eval_samples_per_second": 1917.135,
      "eval_steps_per_second": 30.674,
      "step": 37000
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002995047514859671,
      "loss": 0.4556,
      "step": 37500
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002994643641056959,
      "loss": 0.4541,
      "step": 38000
    },
    {
      "epoch": 1.16,
      "eval_loss": 0.8004159927368164,
      "eval_runtime": 0.5088,
      "eval_samples_per_second": 1965.38,
      "eval_steps_per_second": 31.446,
      "step": 38000
    },
    {
      "epoch": 1.18,
      "learning_rate": 0.000299422396886284,
      "loss": 0.4524,
      "step": 38500
    },
    {
      "epoch": 1.19,
      "learning_rate": 0.0002993788502866783,
      "loss": 0.4514,
      "step": 39000
    },
    {
      "epoch": 1.19,
      "eval_loss": 0.797309160232544,
      "eval_runtime": 0.5211,
      "eval_samples_per_second": 1918.834,
      "eval_steps_per_second": 30.701,
      "step": 39000
    },
    {
      "epoch": 1.21,
      "learning_rate": 0.00029933372478309746,
      "loss": 0.45,
      "step": 39500
    },
    {
      "epoch": 1.23,
      "learning_rate": 0.00029928702086902664,
      "loss": 0.4485,
      "step": 40000
    },
    {
      "epoch": 1.23,
      "eval_loss": 0.7876560688018799,
      "eval_runtime": 0.5381,
      "eval_samples_per_second": 1858.547,
      "eval_steps_per_second": 29.737,
      "step": 40000
    },
    {
      "epoch": 1.24,
      "learning_rate": 0.00029923873905521244,
      "loss": 0.4471,
      "step": 40500
    },
    {
      "epoch": 1.26,
      "learning_rate": 0.000299188879869657,
      "loss": 0.446,
      "step": 41000
    },
    {
      "epoch": 1.26,
      "eval_loss": 0.7949715852737427,
      "eval_runtime": 0.5157,
      "eval_samples_per_second": 1939.022,
      "eval_steps_per_second": 31.024,
      "step": 41000
    },
    {
      "epoch": 1.27,
      "learning_rate": 0.00029913744385761244,
      "loss": 0.4446,
      "step": 41500
    },
    {
      "epoch": 1.29,
      "learning_rate": 0.00029908443158157465,
      "loss": 0.4437,
      "step": 42000
    },
    {
      "epoch": 1.29,
      "eval_loss": 0.7941656112670898,
      "eval_runtime": 0.5261,
      "eval_samples_per_second": 1900.807,
      "eval_steps_per_second": 30.413,
      "step": 42000
    },
    {
      "epoch": 1.3,
      "learning_rate": 0.0002990298436212775,
      "loss": 0.4422,
      "step": 42500
    },
    {
      "epoch": 1.32,
      "learning_rate": 0.0002989736805736861,
      "loss": 0.4413,
      "step": 43000
    },
    {
      "epoch": 1.32,
      "eval_loss": 0.7861095666885376,
      "eval_runtime": 0.5221,
      "eval_samples_per_second": 1915.408,
      "eval_steps_per_second": 30.647,
      "step": 43000
    },
    {
      "epoch": 1.33,
      "learning_rate": 0.00029891594305299065,
      "loss": 0.44,
      "step": 43500
    },
    {
      "epoch": 1.35,
      "learning_rate": 0.00029885663169059926,
      "loss": 0.4391,
      "step": 44000
    },
    {
      "epoch": 1.35,
      "eval_loss": 0.7935870289802551,
      "eval_runtime": 0.5329,
      "eval_samples_per_second": 1876.586,
      "eval_steps_per_second": 30.025,
      "step": 44000
    },
    {
      "epoch": 1.36,
      "learning_rate": 0.0002987957471351316,
      "loss": 0.4378,
      "step": 44500
    },
    {
      "epoch": 1.38,
      "learning_rate": 0.00029873329005241137,
      "loss": 0.4366,
      "step": 45000
    },
    {
      "epoch": 1.38,
      "eval_loss": 0.7878534197807312,
      "eval_runtime": 0.5149,
      "eval_samples_per_second": 1941.994,
      "eval_steps_per_second": 31.072,
      "step": 45000
    },
    {
      "epoch": 1.39,
      "learning_rate": 0.00029866926112545925,
      "loss": 0.4355,
      "step": 45500
    },
    {
      "epoch": 1.41,
      "learning_rate": 0.00029860366105448534,
      "loss": 0.4342,
      "step": 46000
    },
    {
      "epoch": 1.41,
      "eval_loss": 0.7861126661300659,
      "eval_runtime": 0.5143,
      "eval_samples_per_second": 1944.264,
      "eval_steps_per_second": 31.108,
      "step": 46000
    },
    {
      "epoch": 1.42,
      "learning_rate": 0.00029853649055688143,
      "loss": 0.4333,
      "step": 46500
    },
    {
      "epoch": 1.44,
      "learning_rate": 0.00029846775036721337,
      "loss": 0.432,
      "step": 47000
    },
    {
      "epoch": 1.44,
      "eval_loss": 0.7865832448005676,
      "eval_runtime": 0.5183,
      "eval_samples_per_second": 1929.412,
      "eval_steps_per_second": 30.871,
      "step": 47000
    },
    {
      "epoch": 1.46,
      "learning_rate": 0.0002983974412372129,
      "loss": 0.4308,
      "step": 47500
    },
    {
      "epoch": 1.47,
      "learning_rate": 0.00029832556393576934,
      "loss": 0.4297,
      "step": 48000
    },
    {
      "epoch": 1.47,
      "eval_loss": 0.777747392654419,
      "eval_runtime": 0.527,
      "eval_samples_per_second": 1897.653,
      "eval_steps_per_second": 30.362,
      "step": 48000
    },
    {
      "epoch": 1.49,
      "learning_rate": 0.0002982521192489214,
      "loss": 0.4288,
      "step": 48500
    },
    {
      "epoch": 1.5,
      "learning_rate": 0.0002981771079798483,
      "loss": 0.4278,
      "step": 49000
    },
    {
      "epoch": 1.5,
      "eval_loss": 0.7763716578483582,
      "eval_runtime": 0.5211,
      "eval_samples_per_second": 1919.193,
      "eval_steps_per_second": 30.707,
      "step": 49000
    },
    {
      "epoch": 1.52,
      "learning_rate": 0.00029810053094886136,
      "loss": 0.4269,
      "step": 49500
    },
    {
      "epoch": 1.53,
      "learning_rate": 0.00029802238899339473,
      "loss": 0.4258,
      "step": 50000
    },
    {
      "epoch": 1.53,
      "eval_loss": 0.7902368307113647,
      "eval_runtime": 0.5206,
      "eval_samples_per_second": 1920.725,
      "eval_steps_per_second": 30.732,
      "step": 50000
    },
    {
      "epoch": 1.55,
      "learning_rate": 0.0002979426829679962,
      "loss": 0.4247,
      "step": 50500
    },
    {
      "epoch": 1.56,
      "learning_rate": 0.0002978614137443183,
      "loss": 0.424,
      "step": 51000
    },
    {
      "epoch": 1.56,
      "eval_loss": 0.7871042490005493,
      "eval_runtime": 0.5261,
      "eval_samples_per_second": 1900.754,
      "eval_steps_per_second": 30.412,
      "step": 51000
    },
    {
      "epoch": 1.58,
      "learning_rate": 0.000297778582211108,
      "loss": 0.4232,
      "step": 51500
    },
    {
      "epoch": 1.59,
      "learning_rate": 0.00029769418927419786,
      "loss": 0.422,
      "step": 52000
    },
    {
      "epoch": 1.59,
      "eval_loss": 0.7916954159736633,
      "eval_runtime": 0.5311,
      "eval_samples_per_second": 1882.884,
      "eval_steps_per_second": 30.126,
      "step": 52000
    },
    {
      "epoch": 1.61,
      "learning_rate": 0.0002976082358564954,
      "loss": 0.4211,
      "step": 52500
    },
    {
      "epoch": 1.62,
      "learning_rate": 0.00029752072289797353,
      "loss": 0.4202,
      "step": 53000
    },
    {
      "epoch": 1.62,
      "eval_loss": 0.7859019637107849,
      "eval_runtime": 0.5358,
      "eval_samples_per_second": 1866.426,
      "eval_steps_per_second": 29.863,
      "step": 53000
    },
    {
      "epoch": 1.64,
      "learning_rate": 0.00029743165135565986,
      "loss": 0.4194,
      "step": 53500
    },
    {
      "epoch": 1.65,
      "learning_rate": 0.00029734102220362654,
      "loss": 0.4184,
      "step": 54000
    },
    {
      "epoch": 1.65,
      "eval_loss": 0.7869405150413513,
      "eval_runtime": 0.5363,
      "eval_samples_per_second": 1864.609,
      "eval_steps_per_second": 29.834,
      "step": 54000
    },
    {
      "epoch": 1.67,
      "learning_rate": 0.00029724883643297937,
      "loss": 0.4177,
      "step": 54500
    },
    {
      "epoch": 1.69,
      "learning_rate": 0.0002971550950518473,
      "loss": 0.4165,
      "step": 55000
    },
    {
      "epoch": 1.69,
      "eval_loss": 0.7873055338859558,
      "eval_runtime": 0.5249,
      "eval_samples_per_second": 1905.112,
      "eval_steps_per_second": 30.482,
      "step": 55000
    },
    {
      "epoch": 1.7,
      "learning_rate": 0.000297059799085371,
      "loss": 0.4156,
      "step": 55500
    },
    {
      "epoch": 1.72,
      "learning_rate": 0.00029696294957569196,
      "loss": 0.4149,
      "step": 56000
    },
    {
      "epoch": 1.72,
      "eval_loss": 0.7851352095603943,
      "eval_runtime": 0.5685,
      "eval_samples_per_second": 1758.943,
      "eval_steps_per_second": 28.143,
      "step": 56000
    },
    {
      "epoch": 1.73,
      "learning_rate": 0.00029686454758194076,
      "loss": 0.4141,
      "step": 56500
    },
    {
      "epoch": 1.75,
      "learning_rate": 0.00029676459418022594,
      "loss": 0.4136,
      "step": 57000
    },
    {
      "epoch": 1.75,
      "eval_loss": 0.7847647666931152,
      "eval_runtime": 0.5348,
      "eval_samples_per_second": 1869.958,
      "eval_steps_per_second": 29.919,
      "step": 57000
    },
    {
      "epoch": 1.76,
      "learning_rate": 0.0002966630904636219,
      "loss": 0.4126,
      "step": 57500
    },
    {
      "epoch": 1.78,
      "learning_rate": 0.0002965600375421569,
      "loss": 0.4119,
      "step": 58000
    },
    {
      "epoch": 1.78,
      "eval_loss": 0.7846025228500366,
      "eval_runtime": 0.5441,
      "eval_samples_per_second": 1837.795,
      "eval_steps_per_second": 29.405,
      "step": 58000
    },
    {
      "epoch": 1.79,
      "learning_rate": 0.0002964554365428013,
      "loss": 0.4107,
      "step": 58500
    },
    {
      "epoch": 1.81,
      "learning_rate": 0.00029634928860945486,
      "loss": 0.4103,
      "step": 59000
    },
    {
      "epoch": 1.81,
      "eval_loss": 0.7957924604415894,
      "eval_runtime": 0.5171,
      "eval_samples_per_second": 1934.049,
      "eval_steps_per_second": 30.945,
      "step": 59000
    },
    {
      "epoch": 1.82,
      "learning_rate": 0.0002962415949029343,
      "loss": 0.4099,
      "step": 59500
    },
    {
      "epoch": 1.84,
      "learning_rate": 0.00029613235660096084,
      "loss": 0.409,
      "step": 60000
    },
    {
      "epoch": 1.84,
      "eval_loss": 0.7877106070518494,
      "eval_runtime": 0.5384,
      "eval_samples_per_second": 1857.26,
      "eval_steps_per_second": 29.716,
      "step": 60000
    },
    {
      "epoch": 1.85,
      "learning_rate": 0.00029602157489814693,
      "loss": 0.4084,
      "step": 60500
    },
    {
      "epoch": 1.87,
      "learning_rate": 0.00029590925100598365,
      "loss": 0.4077,
      "step": 61000
    },
    {
      "epoch": 1.87,
      "eval_loss": 0.7871591448783875,
      "eval_runtime": 0.5252,
      "eval_samples_per_second": 1904.035,
      "eval_steps_per_second": 30.465,
      "step": 61000
    },
    {
      "epoch": 1.88,
      "learning_rate": 0.000295795386152827,
      "loss": 0.4068,
      "step": 61500
    },
    {
      "epoch": 1.9,
      "learning_rate": 0.0002956799815838848,
      "loss": 0.4061,
      "step": 62000
    },
    {
      "epoch": 1.9,
      "eval_loss": 0.7920863032341003,
      "eval_runtime": 0.5271,
      "eval_samples_per_second": 1897.169,
      "eval_steps_per_second": 30.355,
      "step": 62000
    },
    {
      "epoch": 1.91,
      "learning_rate": 0.0002955630385612029,
      "loss": 0.4056,
      "step": 62500
    },
    {
      "epoch": 1.93,
      "learning_rate": 0.0002954445583636515,
      "loss": 0.405,
      "step": 63000
    },
    {
      "epoch": 1.93,
      "eval_loss": 0.7836641073226929,
      "eval_runtime": 0.5362,
      "eval_samples_per_second": 1864.939,
      "eval_steps_per_second": 29.839,
      "step": 63000
    },
    {
      "epoch": 1.95,
      "learning_rate": 0.00029532454228691103,
      "loss": 0.4043,
      "step": 63500
    },
    {
      "epoch": 1.96,
      "learning_rate": 0.0002952029916434581,
      "loss": 0.4036,
      "step": 64000
    },
    {
      "epoch": 1.96,
      "eval_loss": 0.7819440960884094,
      "eval_runtime": 0.5223,
      "eval_samples_per_second": 1914.709,
      "eval_steps_per_second": 30.635,
      "step": 64000
    },
    {
      "epoch": 1.98,
      "learning_rate": 0.00029507990776255107,
      "loss": 0.403,
      "step": 64500
    },
    {
      "epoch": 1.99,
      "learning_rate": 0.00029495529199021555,
      "loss": 0.4023,
      "step": 65000
    },
    {
      "epoch": 1.99,
      "eval_loss": 0.7944305539131165,
      "eval_runtime": 0.5274,
      "eval_samples_per_second": 1896.133,
      "eval_steps_per_second": 30.338,
      "step": 65000
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0002948291456892296,
      "loss": 0.4018,
      "step": 65500
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00029470147023910907,
      "loss": 0.4009,
      "step": 66000
    },
    {
      "epoch": 2.02,
      "eval_loss": 0.7896050214767456,
      "eval_runtime": 0.514,
      "eval_samples_per_second": 1945.63,
      "eval_steps_per_second": 31.13,
      "step": 66000
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0002945722670360921,
      "loss": 0.4006,
      "step": 66500
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0002944415374931243,
      "loss": 0.4003,
      "step": 67000
    },
    {
      "epoch": 2.05,
      "eval_loss": 0.7898439168930054,
      "eval_runtime": 0.5343,
      "eval_samples_per_second": 1871.754,
      "eval_steps_per_second": 29.948,
      "step": 67000
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00029430928303984295,
      "loss": 0.3993,
      "step": 67500
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0002941755051225616,
      "loss": 0.3988,
      "step": 68000
    },
    {
      "epoch": 2.08,
      "eval_loss": 0.7866422533988953,
      "eval_runtime": 0.5273,
      "eval_samples_per_second": 1896.599,
      "eval_steps_per_second": 30.346,
      "step": 68000
    },
    {
      "epoch": 2.1,
      "learning_rate": 0.00029404020520425417,
      "loss": 0.3984,
      "step": 68500
    },
    {
      "epoch": 2.11,
      "learning_rate": 0.0002939033847645388,
      "loss": 0.3979,
      "step": 69000
    },
    {
      "epoch": 2.11,
      "eval_loss": 0.7866395711898804,
      "eval_runtime": 0.5223,
      "eval_samples_per_second": 1914.777,
      "eval_steps_per_second": 30.636,
      "step": 69000
    },
    {
      "epoch": 2.13,
      "learning_rate": 0.00029376504529966195,
      "loss": 0.3974,
      "step": 69500
    },
    {
      "epoch": 2.14,
      "learning_rate": 0.00029362518832248184,
      "loss": 0.3968,
      "step": 70000
    },
    {
      "epoch": 2.14,
      "eval_loss": 0.7869418263435364,
      "eval_runtime": 0.52,
      "eval_samples_per_second": 1922.99,
      "eval_steps_per_second": 30.768,
      "step": 70000
    },
    {
      "epoch": 2.16,
      "learning_rate": 0.0002934838153624519,
      "loss": 0.3962,
      "step": 70500
    },
    {
      "epoch": 2.18,
      "learning_rate": 0.00029334092796560427,
      "loss": 0.3958,
      "step": 71000
    },
    {
      "epoch": 2.18,
      "eval_loss": 0.7898754477500916,
      "eval_runtime": 0.5162,
      "eval_samples_per_second": 1937.227,
      "eval_steps_per_second": 30.996,
      "step": 71000
    },
    {
      "epoch": 2.19,
      "learning_rate": 0.0002931965276945326,
      "loss": 0.3951,
      "step": 71500
    },
    {
      "epoch": 2.21,
      "learning_rate": 0.0002930506161283751,
      "loss": 0.3947,
      "step": 72000
    },
    {
      "epoch": 2.21,
      "eval_loss": 0.7828860878944397,
      "eval_runtime": 0.5502,
      "eval_samples_per_second": 1817.401,
      "eval_steps_per_second": 29.078,
      "step": 72000
    },
    {
      "epoch": 2.22,
      "learning_rate": 0.00029290319486279724,
      "loss": 0.3944,
      "step": 72500
    },
    {
      "epoch": 2.24,
      "learning_rate": 0.0002927542655099744,
      "loss": 0.3935,
      "step": 73000
    },
    {
      "epoch": 2.24,
      "eval_loss": 0.7836081385612488,
      "eval_runtime": 0.518,
      "eval_samples_per_second": 1930.433,
      "eval_steps_per_second": 30.887,
      "step": 73000
    },
    {
      "epoch": 2.25,
      "learning_rate": 0.00029260382969857417,
      "loss": 0.3936,
      "step": 73500
    },
    {
      "epoch": 2.27,
      "learning_rate": 0.00029245188907373845,
      "loss": 0.393,
      "step": 74000
    },
    {
      "epoch": 2.27,
      "eval_loss": 0.7833809852600098,
      "eval_runtime": 0.5265,
      "eval_samples_per_second": 1899.502,
      "eval_steps_per_second": 30.392,
      "step": 74000
    },
    {
      "epoch": 2.28,
      "learning_rate": 0.0002922984452970655,
      "loss": 0.3923,
      "step": 74500
    },
    {
      "epoch": 2.3,
      "learning_rate": 0.000292143500046592,
      "loss": 0.392,
      "step": 75000
    },
    {
      "epoch": 2.3,
      "eval_loss": 0.7891106009483337,
      "eval_runtime": 0.549,
      "eval_samples_per_second": 1821.565,
      "eval_steps_per_second": 29.145,
      "step": 75000
    },
    {
      "epoch": 2.31,
      "learning_rate": 0.0002919870550167743,
      "loss": 0.3917,
      "step": 75500
    },
    {
      "epoch": 2.33,
      "learning_rate": 0.0002918291119184702,
      "loss": 0.3913,
      "step": 76000
    },
    {
      "epoch": 2.33,
      "eval_loss": 0.7858054041862488,
      "eval_runtime": 0.5293,
      "eval_samples_per_second": 1889.27,
      "eval_steps_per_second": 30.228,
      "step": 76000
    },
    {
      "epoch": 2.34,
      "learning_rate": 0.0002916696724789201,
      "loss": 0.3906,
      "step": 76500
    },
    {
      "epoch": 2.36,
      "learning_rate": 0.00029150873844172823,
      "loss": 0.3904,
      "step": 77000
    },
    {
      "epoch": 2.36,
      "eval_loss": 0.7869090437889099,
      "eval_runtime": 0.5095,
      "eval_samples_per_second": 1962.553,
      "eval_steps_per_second": 31.401,
      "step": 77000
    },
    {
      "epoch": 2.37,
      "learning_rate": 0.00029134631156684334,
      "loss": 0.3898,
      "step": 77500
    },
    {
      "epoch": 2.39,
      "learning_rate": 0.0002911823936305398,
      "loss": 0.3893,
      "step": 78000
    },
    {
      "epoch": 2.39,
      "eval_loss": 0.7837140560150146,
      "eval_runtime": 0.5128,
      "eval_samples_per_second": 1950.24,
      "eval_steps_per_second": 31.204,
      "step": 78000
    },
    {
      "epoch": 2.41,
      "learning_rate": 0.0002910169864253979,
      "loss": 0.3892,
      "step": 78500
    },
    {
      "epoch": 2.42,
      "learning_rate": 0.0002908500917602842,
      "loss": 0.3886,
      "step": 79000
    },
    {
      "epoch": 2.42,
      "eval_loss": 0.787075936794281,
      "eval_runtime": 0.5289,
      "eval_samples_per_second": 1890.813,
      "eval_steps_per_second": 30.253,
      "step": 79000
    },
    {
      "epoch": 2.44,
      "learning_rate": 0.00029068171146033226,
      "loss": 0.3883,
      "step": 79500
    },
    {
      "epoch": 2.45,
      "learning_rate": 0.0002905118473669218,
      "loss": 0.3877,
      "step": 80000
    },
    {
      "epoch": 2.45,
      "eval_loss": 0.7844729423522949,
      "eval_runtime": 0.512,
      "eval_samples_per_second": 1953.312,
      "eval_steps_per_second": 31.253,
      "step": 80000
    },
    {
      "epoch": 2.47,
      "learning_rate": 0.00029034050133765947,
      "loss": 0.3873,
      "step": 80500
    },
    {
      "epoch": 2.48,
      "learning_rate": 0.00029016767524635804,
      "loss": 0.3872,
      "step": 81000
    },
    {
      "epoch": 2.48,
      "eval_loss": 0.7831702828407288,
      "eval_runtime": 0.5564,
      "eval_samples_per_second": 1797.202,
      "eval_steps_per_second": 28.755,
      "step": 81000
    },
    {
      "epoch": 2.5,
      "learning_rate": 0.00028999337098301585,
      "loss": 0.3865,
      "step": 81500
    },
    {
      "epoch": 2.51,
      "learning_rate": 0.0002898175904537964,
      "loss": 0.3858,
      "step": 82000
    },
    {
      "epoch": 2.51,
      "eval_loss": 0.7806258797645569,
      "eval_runtime": 0.5186,
      "eval_samples_per_second": 1928.174,
      "eval_steps_per_second": 30.851,
      "step": 82000
    },
    {
      "epoch": 2.53,
      "learning_rate": 0.0002896403355810075,
      "loss": 0.3859,
      "step": 82500
    },
    {
      "epoch": 2.54,
      "learning_rate": 0.00028946160830307997,
      "loss": 0.3854,
      "step": 83000
    },
    {
      "epoch": 2.54,
      "eval_loss": 0.7836428880691528,
      "eval_runtime": 0.545,
      "eval_samples_per_second": 1834.936,
      "eval_steps_per_second": 29.359,
      "step": 83000
    },
    {
      "epoch": 2.56,
      "learning_rate": 0.00028928141057454665,
      "loss": 0.3849,
      "step": 83500
    },
    {
      "epoch": 2.57,
      "learning_rate": 0.0002890997443660211,
      "loss": 0.3848,
      "step": 84000
    },
    {
      "epoch": 2.57,
      "eval_loss": 0.7852834463119507,
      "eval_runtime": 0.5158,
      "eval_samples_per_second": 1938.61,
      "eval_steps_per_second": 31.018,
      "step": 84000
    },
    {
      "epoch": 2.59,
      "learning_rate": 0.00028891661166417586,
      "loss": 0.3844,
      "step": 84500
    },
    {
      "epoch": 2.6,
      "learning_rate": 0.00028873201447172074,
      "loss": 0.3838,
      "step": 85000
    },
    {
      "epoch": 2.6,
      "eval_loss": 0.7852866053581238,
      "eval_runtime": 0.5265,
      "eval_samples_per_second": 1899.356,
      "eval_steps_per_second": 30.39,
      "step": 85000
    },
    {
      "epoch": 2.62,
      "learning_rate": 0.0002885459548073812,
      "loss": 0.3837,
      "step": 85500
    },
    {
      "epoch": 2.63,
      "learning_rate": 0.0002883584347058758,
      "loss": 0.383,
      "step": 86000
    },
    {
      "epoch": 2.63,
      "eval_loss": 0.7794146537780762,
      "eval_runtime": 0.5219,
      "eval_samples_per_second": 1916.094,
      "eval_steps_per_second": 30.657,
      "step": 86000
    },
    {
      "epoch": 2.65,
      "learning_rate": 0.00028816945621789437,
      "loss": 0.3831,
      "step": 86500
    },
    {
      "epoch": 2.67,
      "learning_rate": 0.0002879790214100753,
      "loss": 0.3823,
      "step": 87000
    },
    {
      "epoch": 2.67,
      "eval_loss": 0.7845750451087952,
      "eval_runtime": 0.5135,
      "eval_samples_per_second": 1947.28,
      "eval_steps_per_second": 31.156,
      "step": 87000
    },
    {
      "epoch": 2.68,
      "learning_rate": 0.0002877871323649833,
      "loss": 0.382,
      "step": 87500
    },
    {
      "epoch": 2.7,
      "learning_rate": 0.0002875937911810861,
      "loss": 0.3818,
      "step": 88000
    },
    {
      "epoch": 2.7,
      "eval_loss": 0.7793565988540649,
      "eval_runtime": 0.5472,
      "eval_samples_per_second": 1827.624,
      "eval_steps_per_second": 29.242,
      "step": 88000
    },
    {
      "epoch": 2.71,
      "learning_rate": 0.000287398999972732,
      "loss": 0.3817,
      "step": 88500
    },
    {
      "epoch": 2.73,
      "learning_rate": 0.00028720276087012636,
      "loss": 0.3809,
      "step": 89000
    },
    {
      "epoch": 2.73,
      "eval_loss": 0.786903440952301,
      "eval_runtime": 0.522,
      "eval_samples_per_second": 1915.886,
      "eval_steps_per_second": 30.654,
      "step": 89000
    },
    {
      "epoch": 2.74,
      "learning_rate": 0.0002870050760193086,
      "loss": 0.3806,
      "step": 89500
    },
    {
      "epoch": 2.76,
      "learning_rate": 0.00028680594758212854,
      "loss": 0.3805,
      "step": 90000
    },
    {
      "epoch": 2.76,
      "eval_loss": 0.7848221659660339,
      "eval_runtime": 0.5114,
      "eval_samples_per_second": 1955.578,
      "eval_steps_per_second": 31.289,
      "step": 90000
    },
    {
      "epoch": 2.77,
      "learning_rate": 0.00028660537773622294,
      "loss": 0.3803,
      "step": 90500
    },
    {
      "epoch": 2.79,
      "learning_rate": 0.00028640336867499143,
      "loss": 0.3796,
      "step": 91000
    },
    {
      "epoch": 2.79,
      "eval_loss": 0.7904797196388245,
      "eval_runtime": 0.5244,
      "eval_samples_per_second": 1906.796,
      "eval_steps_per_second": 30.509,
      "step": 91000
    },
    {
      "epoch": 2.8,
      "learning_rate": 0.0002861999226075728,
      "loss": 0.3798,
      "step": 91500
    },
    {
      "epoch": 2.82,
      "learning_rate": 0.0002859950417588206,
      "loss": 0.3792,
      "step": 92000
    },
    {
      "epoch": 2.82,
      "eval_loss": 0.7905736565589905,
      "eval_runtime": 0.5306,
      "eval_samples_per_second": 1884.737,
      "eval_steps_per_second": 30.156,
      "step": 92000
    },
    {
      "epoch": 2.83,
      "learning_rate": 0.00028578872836927904,
      "loss": 0.3788,
      "step": 92500
    },
    {
      "epoch": 2.85,
      "learning_rate": 0.0002855809846951582,
      "loss": 0.3789,
      "step": 93000
    },
    {
      "epoch": 2.85,
      "eval_loss": 0.7809098958969116,
      "eval_runtime": 0.5414,
      "eval_samples_per_second": 1847.16,
      "eval_steps_per_second": 29.555,
      "step": 93000
    },
    {
      "epoch": 2.86,
      "learning_rate": 0.00028537181300830963,
      "loss": 0.3782,
      "step": 93500
    },
    {
      "epoch": 2.88,
      "learning_rate": 0.0002851612155962014,
      "loss": 0.3784,
      "step": 94000
    },
    {
      "epoch": 2.88,
      "eval_loss": 0.7892218232154846,
      "eval_runtime": 0.516,
      "eval_samples_per_second": 1938.108,
      "eval_steps_per_second": 31.01,
      "step": 94000
    },
    {
      "epoch": 2.9,
      "learning_rate": 0.0002849491947618932,
      "loss": 0.3778,
      "step": 94500
    },
    {
      "epoch": 2.91,
      "learning_rate": 0.0002847357528240107,
      "loss": 0.3775,
      "step": 95000
    },
    {
      "epoch": 2.91,
      "eval_loss": 0.7847021818161011,
      "eval_runtime": 0.5181,
      "eval_samples_per_second": 1930.051,
      "eval_steps_per_second": 30.881,
      "step": 95000
    },
    {
      "epoch": 2.93,
      "learning_rate": 0.0002845208921167208,
      "loss": 0.3773,
      "step": 95500
    },
    {
      "epoch": 2.94,
      "learning_rate": 0.00028430461498970584,
      "loss": 0.3768,
      "step": 96000
    },
    {
      "epoch": 2.94,
      "eval_loss": 0.78525710105896,
      "eval_runtime": 0.5373,
      "eval_samples_per_second": 1861.031,
      "eval_steps_per_second": 29.776,
      "step": 96000
    },
    {
      "epoch": 2.96,
      "learning_rate": 0.00028408692380813775,
      "loss": 0.3768,
      "step": 96500
    },
    {
      "epoch": 2.97,
      "learning_rate": 0.00028386782095265247,
      "loss": 0.3769,
      "step": 97000
    },
    {
      "epoch": 2.97,
      "eval_loss": 0.7837897539138794,
      "eval_runtime": 0.5288,
      "eval_samples_per_second": 1891.025,
      "eval_steps_per_second": 30.256,
      "step": 97000
    },
    {
      "epoch": 2.99,
      "learning_rate": 0.0002836473088193237,
      "loss": 0.3761,
      "step": 97500
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00028342538981963677,
      "loss": 0.3797,
      "step": 98000
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.7850324511528015,
      "eval_runtime": 0.5376,
      "eval_samples_per_second": 1860.245,
      "eval_steps_per_second": 29.764,
      "step": 98000
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0002832020663804624,
      "loss": 0.3752,
      "step": 98500
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00028297734094402986,
      "loss": 0.3747,
      "step": 99000
    },
    {
      "epoch": 3.03,
      "eval_loss": 0.780020534992218,
      "eval_runtime": 0.5281,
      "eval_samples_per_second": 1893.464,
      "eval_steps_per_second": 30.295,
      "step": 99000
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0002827512159679005,
      "loss": 0.375,
      "step": 99500
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00028252369392494086,
      "loss": 0.3746,
      "step": 100000
    },
    {
      "epoch": 3.06,
      "eval_loss": 0.7855916619300842,
      "eval_runtime": 0.5271,
      "eval_samples_per_second": 1897.09,
      "eval_steps_per_second": 30.353,
      "step": 100000
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0002822947773032956,
      "loss": 0.3743,
      "step": 100500
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0002820644686063602,
      "loss": 0.3743,
      "step": 101000
    },
    {
      "epoch": 3.09,
      "eval_loss": 0.7854102253913879,
      "eval_runtime": 0.541,
      "eval_samples_per_second": 1848.349,
      "eval_steps_per_second": 29.574,
      "step": 101000
    },
    {
      "epoch": 3.11,
      "learning_rate": 0.00028183277035275363,
      "loss": 0.374,
      "step": 101500
    },
    {
      "epoch": 3.13,
      "learning_rate": 0.0002815996850762909,
      "loss": 0.374,
      "step": 102000
    },
    {
      "epoch": 3.13,
      "eval_loss": 0.7880498170852661,
      "eval_runtime": 0.5283,
      "eval_samples_per_second": 1892.948,
      "eval_steps_per_second": 30.287,
      "step": 102000
    },
    {
      "epoch": 3.14,
      "learning_rate": 0.00028136521532595515,
      "loss": 0.3737,
      "step": 102500
    },
    {
      "epoch": 3.16,
      "learning_rate": 0.00028112936366587023,
      "loss": 0.3734,
      "step": 103000
    },
    {
      "epoch": 3.16,
      "eval_loss": 0.7864383459091187,
      "eval_runtime": 0.5228,
      "eval_samples_per_second": 1912.628,
      "eval_steps_per_second": 30.602,
      "step": 103000
    },
    {
      "epoch": 3.17,
      "learning_rate": 0.00028089213267527184,
      "loss": 0.3733,
      "step": 103500
    },
    {
      "epoch": 3.19,
      "learning_rate": 0.0002806535249484803,
      "loss": 0.3731,
      "step": 104000
    },
    {
      "epoch": 3.19,
      "eval_loss": 0.783866822719574,
      "eval_runtime": 0.5321,
      "eval_samples_per_second": 1879.482,
      "eval_steps_per_second": 30.072,
      "step": 104000
    },
    {
      "epoch": 3.2,
      "learning_rate": 0.00028041354309487135,
      "loss": 0.3728,
      "step": 104500
    },
    {
      "epoch": 3.22,
      "learning_rate": 0.0002801721897388482,
      "loss": 0.3723,
      "step": 105000
    },
    {
      "epoch": 3.22,
      "eval_loss": 0.784348726272583,
      "eval_runtime": 0.5158,
      "eval_samples_per_second": 1938.703,
      "eval_steps_per_second": 31.019,
      "step": 105000
    },
    {
      "epoch": 3.23,
      "learning_rate": 0.0002799294675198124,
      "loss": 0.3723,
      "step": 105500
    },
    {
      "epoch": 3.25,
      "learning_rate": 0.00027968537909213524,
      "loss": 0.372,
      "step": 106000
    },
    {
      "epoch": 3.25,
      "eval_loss": 0.7889582514762878,
      "eval_runtime": 0.5199,
      "eval_samples_per_second": 1923.296,
      "eval_steps_per_second": 30.773,
      "step": 106000
    },
    {
      "epoch": 3.26,
      "learning_rate": 0.0002794399271251287,
      "loss": 0.3721,
      "step": 106500
    },
    {
      "epoch": 3.28,
      "learning_rate": 0.0002791931143030162,
      "loss": 0.3714,
      "step": 107000
    },
    {
      "epoch": 3.28,
      "eval_loss": 0.7859266400337219,
      "eval_runtime": 0.5189,
      "eval_samples_per_second": 1926.987,
      "eval_steps_per_second": 30.832,
      "step": 107000
    },
    {
      "epoch": 3.29,
      "learning_rate": 0.00027894494332490315,
      "loss": 0.3715,
      "step": 107500
    },
    {
      "epoch": 3.31,
      "learning_rate": 0.0002786954169047476,
      "loss": 0.3711,
      "step": 108000
    },
    {
      "epoch": 3.31,
      "eval_loss": 0.7798612117767334,
      "eval_runtime": 0.5242,
      "eval_samples_per_second": 1907.581,
      "eval_steps_per_second": 30.521,
      "step": 108000
    },
    {
      "epoch": 3.32,
      "learning_rate": 0.0002784445377713306,
      "loss": 0.3711,
      "step": 108500
    },
    {
      "epoch": 3.34,
      "learning_rate": 0.0002781923086682261,
      "loss": 0.371,
      "step": 109000
    },
    {
      "epoch": 3.34,
      "eval_loss": 0.7840728759765625,
      "eval_runtime": 0.5137,
      "eval_samples_per_second": 1946.758,
      "eval_steps_per_second": 31.148,
      "step": 109000
    },
    {
      "epoch": 3.35,
      "learning_rate": 0.0002779387323537711,
      "loss": 0.3704,
      "step": 109500
    },
    {
      "epoch": 3.37,
      "learning_rate": 0.0002776838116010356,
      "loss": 0.3705,
      "step": 110000
    },
    {
      "epoch": 3.37,
      "eval_loss": 0.7811622023582458,
      "eval_runtime": 0.5403,
      "eval_samples_per_second": 1850.736,
      "eval_steps_per_second": 29.612,
      "step": 110000
    },
    {
      "epoch": 3.39,
      "learning_rate": 0.0002774275491977922,
      "loss": 0.3703,
      "step": 110500
    },
    {
      "epoch": 3.4,
      "learning_rate": 0.0002771699479464853,
      "loss": 0.3701,
      "step": 111000
    },
    {
      "epoch": 3.4,
      "eval_loss": 0.7903389930725098,
      "eval_runtime": 0.5256,
      "eval_samples_per_second": 1902.556,
      "eval_steps_per_second": 30.441,
      "step": 111000
    },
    {
      "epoch": 3.42,
      "learning_rate": 0.00027691101066420104,
      "loss": 0.3694,
      "step": 111500
    },
    {
      "epoch": 3.43,
      "learning_rate": 0.0002766507401826361,
      "loss": 0.3692,
      "step": 112000
    },
    {
      "epoch": 3.43,
      "eval_loss": 0.7830519080162048,
      "eval_runtime": 0.5162,
      "eval_samples_per_second": 1937.313,
      "eval_steps_per_second": 30.997,
      "step": 112000
    },
    {
      "epoch": 3.45,
      "learning_rate": 0.0002763891393480666,
      "loss": 0.3694,
      "step": 112500
    },
    {
      "epoch": 3.46,
      "learning_rate": 0.0002761262110213175,
      "loss": 0.3686,
      "step": 113000
    },
    {
      "epoch": 3.46,
      "eval_loss": 0.7800700664520264,
      "eval_runtime": 0.5106,
      "eval_samples_per_second": 1958.342,
      "eval_steps_per_second": 31.333,
      "step": 113000
    },
    {
      "epoch": 3.48,
      "learning_rate": 0.00027586195807773083,
      "loss": 0.3689,
      "step": 113500
    },
    {
      "epoch": 3.49,
      "learning_rate": 0.00027559638340713435,
      "loss": 0.3685,
      "step": 114000
    },
    {
      "epoch": 3.49,
      "eval_loss": 0.7914212942123413,
      "eval_runtime": 0.5261,
      "eval_samples_per_second": 1900.85,
      "eval_steps_per_second": 30.414,
      "step": 114000
    },
    {
      "epoch": 3.51,
      "learning_rate": 0.00027532948991381025,
      "loss": 0.3686,
      "step": 114500
    },
    {
      "epoch": 3.52,
      "learning_rate": 0.00027506128051646287,
      "loss": 0.3681,
      "step": 115000
    },
    {
      "epoch": 3.52,
      "eval_loss": 0.7931650280952454,
      "eval_runtime": 0.5617,
      "eval_samples_per_second": 1780.334,
      "eval_steps_per_second": 28.485,
      "step": 115000
    },
    {
      "epoch": 3.54,
      "learning_rate": 0.00027479175814818733,
      "loss": 0.368,
      "step": 115500
    },
    {
      "epoch": 3.55,
      "learning_rate": 0.000274520925756437,
      "loss": 0.3679,
      "step": 116000
    },
    {
      "epoch": 3.55,
      "eval_loss": 0.7921015620231628,
      "eval_runtime": 0.5338,
      "eval_samples_per_second": 1873.268,
      "eval_steps_per_second": 29.972,
      "step": 116000
    },
    {
      "epoch": 3.57,
      "learning_rate": 0.00027424878630299157,
      "loss": 0.3673,
      "step": 116500
    },
    {
      "epoch": 3.58,
      "learning_rate": 0.0002739753427639244,
      "loss": 0.3673,
      "step": 117000
    },
    {
      "epoch": 3.58,
      "eval_loss": 0.7863946557044983,
      "eval_runtime": 0.528,
      "eval_samples_per_second": 1893.847,
      "eval_steps_per_second": 30.302,
      "step": 117000
    },
    {
      "epoch": 3.6,
      "learning_rate": 0.0002737005981295704,
      "loss": 0.3673,
      "step": 117500
    },
    {
      "epoch": 3.62,
      "learning_rate": 0.0002734245554044927,
      "loss": 0.3668,
      "step": 118000
    },
    {
      "epoch": 3.62,
      "eval_loss": 0.7858835458755493,
      "eval_runtime": 0.5221,
      "eval_samples_per_second": 1915.181,
      "eval_steps_per_second": 30.643,
      "step": 118000
    },
    {
      "epoch": 3.63,
      "learning_rate": 0.0002731472176074504,
      "loss": 0.3667,
      "step": 118500
    },
    {
      "epoch": 3.65,
      "learning_rate": 0.0002728685877713653,
      "loss": 0.3669,
      "step": 119000
    },
    {
      "epoch": 3.65,
      "eval_loss": 0.786138653755188,
      "eval_runtime": 0.5283,
      "eval_samples_per_second": 1893.025,
      "eval_steps_per_second": 30.288,
      "step": 119000
    },
    {
      "epoch": 3.66,
      "learning_rate": 0.0002725886689432884,
      "loss": 0.3664,
      "step": 119500
    },
    {
      "epoch": 3.68,
      "learning_rate": 0.0002723074641843674,
      "loss": 0.3664,
      "step": 120000
    },
    {
      "epoch": 3.68,
      "eval_loss": 0.7753082513809204,
      "eval_runtime": 0.539,
      "eval_samples_per_second": 1855.209,
      "eval_steps_per_second": 29.683,
      "step": 120000
    },
    {
      "epoch": 3.69,
      "learning_rate": 0.0002720249765698123,
      "loss": 0.3661,
      "step": 120500
    },
    {
      "epoch": 3.71,
      "learning_rate": 0.0002717412091888626,
      "loss": 0.3663,
      "step": 121000
    },
    {
      "epoch": 3.71,
      "eval_loss": 0.7859072089195251,
      "eval_runtime": 0.5189,
      "eval_samples_per_second": 1927.188,
      "eval_steps_per_second": 30.835,
      "step": 121000
    },
    {
      "epoch": 3.72,
      "learning_rate": 0.00027145616514475274,
      "loss": 0.3656,
      "step": 121500
    },
    {
      "epoch": 3.74,
      "learning_rate": 0.0002711698475546788,
      "loss": 0.3656,
      "step": 122000
    },
    {
      "epoch": 3.74,
      "eval_loss": 0.7800880074501038,
      "eval_runtime": 0.5426,
      "eval_samples_per_second": 1842.936,
      "eval_steps_per_second": 29.487,
      "step": 122000
    },
    {
      "epoch": 3.75,
      "learning_rate": 0.00027088225954976407,
      "loss": 0.3653,
      "step": 122500
    },
    {
      "epoch": 3.77,
      "learning_rate": 0.0002705934042750249,
      "loss": 0.3653,
      "step": 123000
    },
    {
      "epoch": 3.77,
      "eval_loss": 0.7793842554092407,
      "eval_runtime": 0.5318,
      "eval_samples_per_second": 1880.356,
      "eval_steps_per_second": 30.086,
      "step": 123000
    },
    {
      "epoch": 3.78,
      "learning_rate": 0.00027030328488933625,
      "loss": 0.3649,
      "step": 123500
    },
    {
      "epoch": 3.8,
      "learning_rate": 0.00027001190456539726,
      "loss": 0.3648,
      "step": 124000
    },
    {
      "epoch": 3.8,
      "eval_loss": 0.7881133556365967,
      "eval_runtime": 0.507,
      "eval_samples_per_second": 1972.529,
      "eval_steps_per_second": 31.56,
      "step": 124000
    },
    {
      "epoch": 3.81,
      "learning_rate": 0.0002697192664896965,
      "loss": 0.3648,
      "step": 124500
    },
    {
      "epoch": 3.83,
      "learning_rate": 0.00026942537386247706,
      "loss": 0.3644,
      "step": 125000
    },
    {
      "epoch": 3.83,
      "eval_loss": 0.7862609624862671,
      "eval_runtime": 0.5236,
      "eval_samples_per_second": 1909.846,
      "eval_steps_per_second": 30.558,
      "step": 125000
    },
    {
      "epoch": 3.85,
      "learning_rate": 0.0002691302298977016,
      "loss": 0.3643,
      "step": 125500
    },
    {
      "epoch": 3.86,
      "learning_rate": 0.0002688338378230173,
      "loss": 0.3642,
      "step": 126000
    },
    {
      "epoch": 3.86,
      "eval_loss": 0.7819104194641113,
      "eval_runtime": 0.5098,
      "eval_samples_per_second": 1961.492,
      "eval_steps_per_second": 31.384,
      "step": 126000
    },
    {
      "epoch": 3.88,
      "learning_rate": 0.00026853620087972035,
      "loss": 0.3636,
      "step": 126500
    },
    {
      "epoch": 3.89,
      "learning_rate": 0.00026823732232272065,
      "loss": 0.3638,
      "step": 127000
    },
    {
      "epoch": 3.89,
      "eval_loss": 0.7886289358139038,
      "eval_runtime": 0.5488,
      "eval_samples_per_second": 1822.211,
      "eval_steps_per_second": 29.155,
      "step": 127000
    },
    {
      "epoch": 3.91,
      "learning_rate": 0.0002679372054205063,
      "loss": 0.3637,
      "step": 127500
    },
    {
      "epoch": 3.92,
      "learning_rate": 0.0002676358534551076,
      "loss": 0.3637,
      "step": 128000
    },
    {
      "epoch": 3.92,
      "eval_loss": 0.7856259942054749,
      "eval_runtime": 0.5448,
      "eval_samples_per_second": 1835.504,
      "eval_steps_per_second": 29.368,
      "step": 128000
    },
    {
      "epoch": 3.94,
      "learning_rate": 0.00026733326972206133,
      "loss": 0.3631,
      "step": 128500
    },
    {
      "epoch": 3.95,
      "learning_rate": 0.0002670294575303748,
      "loss": 0.3631,
      "step": 129000
    },
    {
      "epoch": 3.95,
      "eval_loss": 0.7774640917778015,
      "eval_runtime": 0.5451,
      "eval_samples_per_second": 1834.446,
      "eval_steps_per_second": 29.351,
      "step": 129000
    },
    {
      "epoch": 3.97,
      "learning_rate": 0.0002667244202024894,
      "loss": 0.3629,
      "step": 129500
    },
    {
      "epoch": 3.98,
      "learning_rate": 0.00026641816107424453,
      "loss": 0.363,
      "step": 130000
    },
    {
      "epoch": 3.98,
      "eval_loss": 0.7824729084968567,
      "eval_runtime": 0.5158,
      "eval_samples_per_second": 1938.903,
      "eval_steps_per_second": 31.022,
      "step": 130000
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0002661106834948409,
      "loss": 0.3626,
      "step": 130500
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0002658019908268041,
      "loss": 0.3625,
      "step": 131000
    },
    {
      "epoch": 4.01,
      "eval_loss": 0.7918245792388916,
      "eval_runtime": 0.521,
      "eval_samples_per_second": 1919.333,
      "eval_steps_per_second": 30.709,
      "step": 131000
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00026549208644594766,
      "loss": 0.3622,
      "step": 131500
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00026518097374133627,
      "loss": 0.3622,
      "step": 132000
    },
    {
      "epoch": 4.04,
      "eval_loss": 0.7886719107627869,
      "eval_runtime": 0.5166,
      "eval_samples_per_second": 1935.893,
      "eval_steps_per_second": 30.974,
      "step": 132000
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00026486865611524853,
      "loss": 0.3622,
      "step": 132500
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00026455513698314003,
      "loss": 0.3616,
      "step": 133000
    },
    {
      "epoch": 4.07,
      "eval_loss": 0.7906731963157654,
      "eval_runtime": 0.5202,
      "eval_samples_per_second": 1922.502,
      "eval_steps_per_second": 30.76,
      "step": 133000
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002642404197736058,
      "loss": 0.3618,
      "step": 133500
    },
    {
      "epoch": 4.11,
      "learning_rate": 0.0002639245079283428,
      "loss": 0.3646,
      "step": 134000
    },
    {
      "epoch": 4.11,
      "eval_loss": 0.7901037931442261,
      "eval_runtime": 0.5304,
      "eval_samples_per_second": 1885.342,
      "eval_steps_per_second": 30.165,
      "step": 134000
    },
    {
      "epoch": 4.12,
      "learning_rate": 0.00026360740490211234,
      "loss": 0.3612,
      "step": 134500
    },
    {
      "epoch": 4.14,
      "learning_rate": 0.0002632891141627023,
      "loss": 0.3604,
      "step": 135000
    },
    {
      "epoch": 4.14,
      "eval_loss": 0.7916610836982727,
      "eval_runtime": 0.5131,
      "eval_samples_per_second": 1948.858,
      "eval_steps_per_second": 31.182,
      "step": 135000
    },
    {
      "epoch": 4.15,
      "learning_rate": 0.00026296963919088923,
      "loss": 0.3603,
      "step": 135500
    },
    {
      "epoch": 4.17,
      "learning_rate": 0.00026264898348040024,
      "loss": 0.3604,
      "step": 136000
    },
    {
      "epoch": 4.17,
      "eval_loss": 0.7712200284004211,
      "eval_runtime": 0.519,
      "eval_samples_per_second": 1926.697,
      "eval_steps_per_second": 30.827,
      "step": 136000
    },
    {
      "epoch": 4.18,
      "learning_rate": 0.0002623271505378748,
      "loss": 0.3608,
      "step": 136500
    },
    {
      "epoch": 4.2,
      "learning_rate": 0.00026200414388282637,
      "loss": 0.361,
      "step": 137000
    },
    {
      "epoch": 4.2,
      "eval_loss": 0.7850333452224731,
      "eval_runtime": 0.5415,
      "eval_samples_per_second": 1846.815,
      "eval_steps_per_second": 29.549,
      "step": 137000
    },
    {
      "epoch": 4.21,
      "learning_rate": 0.00026167996704760406,
      "loss": 0.3603,
      "step": 137500
    },
    {
      "epoch": 4.23,
      "learning_rate": 0.00026135462357735375,
      "loss": 0.3604,
      "step": 138000
    },
    {
      "epoch": 4.23,
      "eval_loss": 0.7824327349662781,
      "eval_runtime": 0.5307,
      "eval_samples_per_second": 1884.218,
      "eval_steps_per_second": 30.147,
      "step": 138000
    },
    {
      "epoch": 4.24,
      "learning_rate": 0.0002610281170299795,
      "loss": 0.3602,
      "step": 138500
    },
    {
      "epoch": 4.26,
      "learning_rate": 0.00026070045097610465,
      "loss": 0.36,
      "step": 139000
    },
    {
      "epoch": 4.26,
      "eval_loss": 0.787578821182251,
      "eval_runtime": 0.5306,
      "eval_samples_per_second": 1884.712,
      "eval_steps_per_second": 30.155,
      "step": 139000
    },
    {
      "epoch": 4.27,
      "learning_rate": 0.0002603716289990326,
      "loss": 0.3601,
      "step": 139500
    },
    {
      "epoch": 4.29,
      "learning_rate": 0.00026004165469470787,
      "loss": 0.3597,
      "step": 140000
    },
    {
      "epoch": 4.29,
      "eval_loss": 0.7803857326507568,
      "eval_runtime": 0.5242,
      "eval_samples_per_second": 1907.849,
      "eval_steps_per_second": 30.526,
      "step": 140000
    }
  ],
  "max_steps": 500000,
  "num_train_epochs": 16,
  "total_flos": 4.4727879910350966e+21,
  "trial_name": null,
  "trial_params": null
}