{
  "best_metric": 2.2634286880493164,
  "best_model_checkpoint": "./model_tweets_2020_Q4_25/checkpoint-2016000",
  "epoch": 6.736955569778018,
  "eval_steps": 8000,
  "global_step": 2400000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "eval_loss": 2.592593193054199,
      "eval_runtime": 340.1804,
      "eval_samples_per_second": 881.867,
      "eval_steps_per_second": 55.118,
      "step": 8000
    },
    {
      "epoch": 0.04,
      "learning_rate": 4.0726666666666665e-07,
      "loss": 2.7864,
      "step": 16000
    },
    {
      "epoch": 0.04,
      "eval_loss": 2.5070722103118896,
      "eval_runtime": 334.504,
      "eval_samples_per_second": 896.832,
      "eval_steps_per_second": 56.053,
      "step": 16000
    },
    {
      "epoch": 0.07,
      "eval_loss": 2.468975782394409,
      "eval_runtime": 337.9946,
      "eval_samples_per_second": 887.57,
      "eval_steps_per_second": 55.474,
      "step": 24000
    },
    {
      "epoch": 0.09,
      "learning_rate": 4.0453333333333336e-07,
      "loss": 2.5937,
      "step": 32000
    },
    {
      "epoch": 0.09,
      "eval_loss": 2.435482978820801,
      "eval_runtime": 338.3798,
      "eval_samples_per_second": 886.56,
      "eval_steps_per_second": 55.411,
      "step": 32000
    },
    {
      "epoch": 0.11,
      "eval_loss": 2.4124932289123535,
      "eval_runtime": 330.9228,
      "eval_samples_per_second": 906.538,
      "eval_steps_per_second": 56.66,
      "step": 40000
    },
    {
      "epoch": 0.13,
      "learning_rate": 4.018e-07,
      "loss": 2.55,
      "step": 48000
    },
    {
      "epoch": 0.13,
      "eval_loss": 2.400876760482788,
      "eval_runtime": 334.2849,
      "eval_samples_per_second": 897.42,
      "eval_steps_per_second": 56.09,
      "step": 48000
    },
    {
      "epoch": 0.16,
      "eval_loss": 2.389232873916626,
      "eval_runtime": 333.216,
      "eval_samples_per_second": 900.299,
      "eval_steps_per_second": 56.27,
      "step": 56000
    },
    {
      "epoch": 0.18,
      "learning_rate": 3.9906666666666667e-07,
      "loss": 2.5159,
      "step": 64000
    },
    {
      "epoch": 0.18,
      "eval_loss": 2.3736019134521484,
      "eval_runtime": 333.0778,
      "eval_samples_per_second": 900.672,
      "eval_steps_per_second": 56.293,
      "step": 64000
    },
    {
      "epoch": 0.2,
      "eval_loss": 2.3712990283966064,
      "eval_runtime": 332.3155,
      "eval_samples_per_second": 902.738,
      "eval_steps_per_second": 56.422,
      "step": 72000
    },
    {
      "epoch": 0.22,
      "learning_rate": 3.963333333333333e-07,
      "loss": 2.495,
      "step": 80000
    },
    {
      "epoch": 0.22,
      "eval_loss": 2.364145278930664,
      "eval_runtime": 329.5546,
      "eval_samples_per_second": 910.301,
      "eval_steps_per_second": 56.895,
      "step": 80000
    },
    {
      "epoch": 0.25,
      "eval_loss": 2.357400417327881,
      "eval_runtime": 330.227,
      "eval_samples_per_second": 908.448,
      "eval_steps_per_second": 56.779,
      "step": 88000
    },
    {
      "epoch": 0.27,
      "learning_rate": 3.936e-07,
      "loss": 2.4845,
      "step": 96000
    },
    {
      "epoch": 0.27,
      "eval_loss": 2.349066734313965,
      "eval_runtime": 332.2917,
      "eval_samples_per_second": 902.803,
      "eval_steps_per_second": 56.426,
      "step": 96000
    },
    {
      "epoch": 0.29,
      "eval_loss": 2.3484702110290527,
      "eval_runtime": 330.9315,
      "eval_samples_per_second": 906.514,
      "eval_steps_per_second": 56.658,
      "step": 104000
    },
    {
      "epoch": 0.31,
      "learning_rate": 3.908666666666667e-07,
      "loss": 2.4765,
      "step": 112000
    },
    {
      "epoch": 0.31,
      "eval_loss": 2.3432772159576416,
      "eval_runtime": 331.9574,
      "eval_samples_per_second": 903.712,
      "eval_steps_per_second": 56.483,
      "step": 112000
    },
    {
      "epoch": 0.34,
      "eval_loss": 2.337597131729126,
      "eval_runtime": 328.3652,
      "eval_samples_per_second": 913.599,
      "eval_steps_per_second": 57.101,
      "step": 120000
    },
    {
      "epoch": 0.36,
      "learning_rate": 3.8813333333333334e-07,
      "loss": 2.472,
      "step": 128000
    },
    {
      "epoch": 0.36,
      "eval_loss": 2.339606285095215,
      "eval_runtime": 331.483,
      "eval_samples_per_second": 905.006,
      "eval_steps_per_second": 56.564,
      "step": 128000
    },
    {
      "epoch": 0.38,
      "eval_loss": 2.3326175212860107,
      "eval_runtime": 331.9588,
      "eval_samples_per_second": 903.709,
      "eval_steps_per_second": 56.483,
      "step": 136000
    },
    {
      "epoch": 0.4,
      "learning_rate": 3.854e-07,
      "loss": 2.467,
      "step": 144000
    },
    {
      "epoch": 0.4,
      "eval_loss": 2.3383891582489014,
      "eval_runtime": 330.5258,
      "eval_samples_per_second": 907.626,
      "eval_steps_per_second": 56.728,
      "step": 144000
    },
    {
      "epoch": 0.43,
      "eval_loss": 2.3349926471710205,
      "eval_runtime": 331.6599,
      "eval_samples_per_second": 904.523,
      "eval_steps_per_second": 56.534,
      "step": 152000
    },
    {
      "epoch": 0.45,
      "learning_rate": 3.8266666666666665e-07,
      "loss": 2.46,
      "step": 160000
    },
    {
      "epoch": 0.45,
      "eval_loss": 2.3262743949890137,
      "eval_runtime": 332.3863,
      "eval_samples_per_second": 902.546,
      "eval_steps_per_second": 56.41,
      "step": 160000
    },
    {
      "epoch": 0.47,
      "eval_loss": 2.323078155517578,
      "eval_runtime": 328.7829,
      "eval_samples_per_second": 912.438,
      "eval_steps_per_second": 57.029,
      "step": 168000
    },
    {
      "epoch": 0.49,
      "learning_rate": 3.799333333333333e-07,
      "loss": 2.4593,
      "step": 176000
    },
    {
      "epoch": 0.49,
      "eval_loss": 2.32228946685791,
      "eval_runtime": 329.1085,
      "eval_samples_per_second": 911.535,
      "eval_steps_per_second": 56.972,
      "step": 176000
    },
    {
      "epoch": 0.52,
      "eval_loss": 2.32503604888916,
      "eval_runtime": 329.9498,
      "eval_samples_per_second": 909.211,
      "eval_steps_per_second": 56.827,
      "step": 184000
    },
    {
      "epoch": 0.54,
      "learning_rate": 3.772e-07,
      "loss": 2.4552,
      "step": 192000
    },
    {
      "epoch": 0.54,
      "eval_loss": 2.319518566131592,
      "eval_runtime": 328.4451,
      "eval_samples_per_second": 913.376,
      "eval_steps_per_second": 57.087,
      "step": 192000
    },
    {
      "epoch": 0.56,
      "eval_loss": 2.3236451148986816,
      "eval_runtime": 329.6628,
      "eval_samples_per_second": 910.003,
      "eval_steps_per_second": 56.876,
      "step": 200000
    },
    {
      "epoch": 0.58,
      "learning_rate": 3.7446666666666667e-07,
      "loss": 2.4558,
      "step": 208000
    },
    {
      "epoch": 0.58,
      "eval_loss": 2.322141170501709,
      "eval_runtime": 328.8408,
      "eval_samples_per_second": 912.277,
      "eval_steps_per_second": 57.018,
      "step": 208000
    },
    {
      "epoch": 0.61,
      "eval_loss": 2.3194377422332764,
      "eval_runtime": 329.1871,
      "eval_samples_per_second": 911.318,
      "eval_steps_per_second": 56.958,
      "step": 216000
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7173333333333333e-07,
      "loss": 2.4487,
      "step": 224000
    },
    {
      "epoch": 0.63,
      "eval_loss": 2.3224573135375977,
      "eval_runtime": 328.2655,
      "eval_samples_per_second": 913.876,
      "eval_steps_per_second": 57.118,
      "step": 224000
    },
    {
      "epoch": 0.65,
      "eval_loss": 2.3220534324645996,
      "eval_runtime": 329.5297,
      "eval_samples_per_second": 910.37,
      "eval_steps_per_second": 56.899,
      "step": 232000
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.69e-07,
      "loss": 2.4485,
      "step": 240000
    },
    {
      "epoch": 0.67,
      "eval_loss": 2.3135385513305664,
      "eval_runtime": 328.6734,
      "eval_samples_per_second": 912.742,
      "eval_steps_per_second": 57.048,
      "step": 240000
    },
    {
      "epoch": 0.7,
      "eval_loss": 2.3109307289123535,
      "eval_runtime": 330.0109,
      "eval_samples_per_second": 909.043,
      "eval_steps_per_second": 56.816,
      "step": 248000
    },
    {
      "epoch": 0.72,
      "learning_rate": 3.6626666666666664e-07,
      "loss": 2.4461,
      "step": 256000
    },
    {
      "epoch": 0.72,
      "eval_loss": 2.3133792877197266,
      "eval_runtime": 330.0512,
      "eval_samples_per_second": 908.932,
      "eval_steps_per_second": 56.809,
      "step": 256000
    },
    {
      "epoch": 0.74,
      "eval_loss": 2.317667007446289,
      "eval_runtime": 328.6469,
      "eval_samples_per_second": 912.816,
      "eval_steps_per_second": 57.052,
      "step": 264000
    },
    {
      "epoch": 0.76,
      "learning_rate": 3.6353333333333335e-07,
      "loss": 2.4513,
      "step": 272000
    },
    {
      "epoch": 0.76,
      "eval_loss": 2.310190200805664,
      "eval_runtime": 329.3646,
      "eval_samples_per_second": 910.826,
      "eval_steps_per_second": 56.928,
      "step": 272000
    },
    {
      "epoch": 0.79,
      "eval_loss": 2.3051719665527344,
      "eval_runtime": 329.4854,
      "eval_samples_per_second": 910.492,
      "eval_steps_per_second": 56.907,
      "step": 280000
    },
    {
      "epoch": 0.81,
      "learning_rate": 3.608e-07,
      "loss": 2.4488,
      "step": 288000
    },
    {
      "epoch": 0.81,
      "eval_loss": 2.3044047355651855,
      "eval_runtime": 329.2769,
      "eval_samples_per_second": 911.069,
      "eval_steps_per_second": 56.943,
      "step": 288000
    },
    {
      "epoch": 0.83,
      "eval_loss": 2.3117146492004395,
      "eval_runtime": 328.6955,
      "eval_samples_per_second": 912.681,
      "eval_steps_per_second": 57.044,
      "step": 296000
    },
    {
      "epoch": 0.85,
      "learning_rate": 3.5806666666666666e-07,
      "loss": 2.4447,
      "step": 304000
    },
    {
      "epoch": 0.85,
      "eval_loss": 2.3050835132598877,
      "eval_runtime": 329.2136,
      "eval_samples_per_second": 911.244,
      "eval_steps_per_second": 56.954,
      "step": 304000
    },
    {
      "epoch": 0.88,
      "eval_loss": 2.3112313747406006,
      "eval_runtime": 329.9672,
      "eval_samples_per_second": 909.163,
      "eval_steps_per_second": 56.824,
      "step": 312000
    },
    {
      "epoch": 0.9,
      "learning_rate": 3.553333333333333e-07,
      "loss": 2.4485,
      "step": 320000
    },
    {
      "epoch": 0.9,
      "eval_loss": 2.3064165115356445,
      "eval_runtime": 328.707,
      "eval_samples_per_second": 912.649,
      "eval_steps_per_second": 57.042,
      "step": 320000
    },
    {
      "epoch": 0.92,
      "eval_loss": 2.3099164962768555,
      "eval_runtime": 329.8653,
      "eval_samples_per_second": 909.444,
      "eval_steps_per_second": 56.841,
      "step": 328000
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.5259999999999997e-07,
      "loss": 2.4475,
      "step": 336000
    },
    {
      "epoch": 0.94,
      "eval_loss": 2.310988664627075,
      "eval_runtime": 329.0328,
      "eval_samples_per_second": 911.745,
      "eval_steps_per_second": 56.985,
      "step": 336000
    },
    {
      "epoch": 0.97,
      "eval_loss": 2.3013858795166016,
      "eval_runtime": 329.2936,
      "eval_samples_per_second": 911.023,
      "eval_steps_per_second": 56.94,
      "step": 344000
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.498666666666667e-07,
      "loss": 2.4464,
      "step": 352000
    },
    {
      "epoch": 0.99,
      "eval_loss": 2.3032288551330566,
      "eval_runtime": 329.4779,
      "eval_samples_per_second": 910.513,
      "eval_steps_per_second": 56.908,
      "step": 352000
    },
    {
      "epoch": 1.01,
      "eval_loss": 2.3035871982574463,
      "eval_runtime": 329.3314,
      "eval_samples_per_second": 910.918,
      "eval_steps_per_second": 56.934,
      "step": 360000
    },
    {
      "epoch": 1.03,
      "learning_rate": 3.4713333333333333e-07,
      "loss": 2.4478,
      "step": 368000
    },
    {
      "epoch": 1.03,
      "eval_loss": 2.305039882659912,
      "eval_runtime": 330.1998,
      "eval_samples_per_second": 908.523,
      "eval_steps_per_second": 56.784,
      "step": 368000
    },
    {
      "epoch": 1.06,
      "eval_loss": 2.3078055381774902,
      "eval_runtime": 329.8754,
      "eval_samples_per_second": 909.416,
      "eval_steps_per_second": 56.84,
      "step": 376000
    },
    {
      "epoch": 1.08,
      "learning_rate": 3.444e-07,
      "loss": 2.4416,
      "step": 384000
    },
    {
      "epoch": 1.08,
      "eval_loss": 2.302764892578125,
      "eval_runtime": 328.8958,
      "eval_samples_per_second": 912.125,
      "eval_steps_per_second": 57.009,
      "step": 384000
    },
    {
      "epoch": 1.1,
      "eval_loss": 2.3016645908355713,
      "eval_runtime": 329.422,
      "eval_samples_per_second": 910.668,
      "eval_steps_per_second": 56.918,
      "step": 392000
    },
    {
      "epoch": 1.12,
      "learning_rate": 3.416666666666667e-07,
      "loss": 2.4374,
      "step": 400000
    },
    {
      "epoch": 1.12,
      "eval_loss": 2.3012008666992188,
      "eval_runtime": 329.1475,
      "eval_samples_per_second": 911.427,
      "eval_steps_per_second": 56.965,
      "step": 400000
    },
    {
      "epoch": 1.15,
      "eval_loss": 2.3017375469207764,
      "eval_runtime": 329.6739,
      "eval_samples_per_second": 909.972,
      "eval_steps_per_second": 56.874,
      "step": 408000
    },
    {
      "epoch": 1.17,
      "learning_rate": 3.3893333333333335e-07,
      "loss": 2.4406,
      "step": 416000
    },
    {
      "epoch": 1.17,
      "eval_loss": 2.3042545318603516,
      "eval_runtime": 329.4772,
      "eval_samples_per_second": 910.515,
      "eval_steps_per_second": 56.908,
      "step": 416000
    },
    {
      "epoch": 1.19,
      "eval_loss": 2.3058371543884277,
      "eval_runtime": 329.3774,
      "eval_samples_per_second": 910.791,
      "eval_steps_per_second": 56.926,
      "step": 424000
    },
    {
      "epoch": 1.21,
      "learning_rate": 3.3619999999999995e-07,
      "loss": 2.4434,
      "step": 432000
    },
    {
      "epoch": 1.21,
      "eval_loss": 2.2937967777252197,
      "eval_runtime": 328.8166,
      "eval_samples_per_second": 912.345,
      "eval_steps_per_second": 57.023,
      "step": 432000
    },
    {
      "epoch": 1.24,
      "eval_loss": 2.2971138954162598,
      "eval_runtime": 330.0079,
      "eval_samples_per_second": 909.051,
      "eval_steps_per_second": 56.817,
      "step": 440000
    },
    {
      "epoch": 1.26,
      "learning_rate": 3.3346666666666666e-07,
      "loss": 2.4421,
      "step": 448000
    },
    {
      "epoch": 1.26,
      "eval_loss": 2.3025450706481934,
      "eval_runtime": 329.6817,
      "eval_samples_per_second": 909.95,
      "eval_steps_per_second": 56.873,
      "step": 448000
    },
    {
      "epoch": 1.28,
      "eval_loss": 2.2950313091278076,
      "eval_runtime": 329.2997,
      "eval_samples_per_second": 911.006,
      "eval_steps_per_second": 56.939,
      "step": 456000
    },
    {
      "epoch": 1.3,
      "learning_rate": 3.307333333333333e-07,
      "loss": 2.443,
      "step": 464000
    },
    {
      "epoch": 1.3,
      "eval_loss": 2.2986950874328613,
      "eval_runtime": 329.2432,
      "eval_samples_per_second": 911.162,
      "eval_steps_per_second": 56.949,
      "step": 464000
    },
    {
      "epoch": 1.32,
      "eval_loss": 2.294912099838257,
      "eval_runtime": 329.1309,
      "eval_samples_per_second": 911.473,
      "eval_steps_per_second": 56.968,
      "step": 472000
    },
    {
      "epoch": 1.35,
      "learning_rate": 3.28e-07,
      "loss": 2.4357,
      "step": 480000
    },
    {
      "epoch": 1.35,
      "eval_loss": 2.3026058673858643,
      "eval_runtime": 329.2084,
      "eval_samples_per_second": 911.259,
      "eval_steps_per_second": 56.955,
      "step": 480000
    },
    {
      "epoch": 1.37,
      "eval_loss": 2.2961277961730957,
      "eval_runtime": 329.3643,
      "eval_samples_per_second": 910.827,
      "eval_steps_per_second": 56.928,
      "step": 488000
    },
    {
      "epoch": 1.39,
      "learning_rate": 3.252666666666667e-07,
      "loss": 2.4366,
      "step": 496000
    },
    {
      "epoch": 1.39,
      "eval_loss": 2.3002982139587402,
      "eval_runtime": 331.8417,
      "eval_samples_per_second": 904.027,
      "eval_steps_per_second": 56.503,
      "step": 496000
    },
    {
      "epoch": 1.41,
      "eval_loss": 2.2953805923461914,
      "eval_runtime": 330.1726,
      "eval_samples_per_second": 908.598,
      "eval_steps_per_second": 56.788,
      "step": 504000
    },
    {
      "epoch": 1.44,
      "learning_rate": 3.2253333333333334e-07,
      "loss": 2.4528,
      "step": 512000
    },
    {
      "epoch": 1.44,
      "eval_loss": 2.2882533073425293,
      "eval_runtime": 332.2968,
      "eval_samples_per_second": 902.789,
      "eval_steps_per_second": 56.425,
      "step": 512000
    },
    {
      "epoch": 1.46,
      "eval_loss": 2.2999649047851562,
      "eval_runtime": 329.9715,
      "eval_samples_per_second": 909.151,
      "eval_steps_per_second": 56.823,
      "step": 520000
    },
    {
      "epoch": 1.48,
      "learning_rate": 3.198e-07,
      "loss": 2.4389,
      "step": 528000
    },
    {
      "epoch": 1.48,
      "eval_loss": 2.2938716411590576,
      "eval_runtime": 330.7123,
      "eval_samples_per_second": 907.115,
      "eval_steps_per_second": 56.696,
      "step": 528000
    },
    {
      "epoch": 1.5,
      "eval_loss": 2.29899525642395,
      "eval_runtime": 329.622,
      "eval_samples_per_second": 910.115,
      "eval_steps_per_second": 56.883,
      "step": 536000
    },
    {
      "epoch": 1.53,
      "learning_rate": 3.1706666666666665e-07,
      "loss": 2.441,
      "step": 544000
    },
    {
      "epoch": 1.53,
      "eval_loss": 2.2915515899658203,
      "eval_runtime": 331.8875,
      "eval_samples_per_second": 903.903,
      "eval_steps_per_second": 56.495,
      "step": 544000
    },
    {
      "epoch": 1.55,
      "eval_loss": 2.2906155586242676,
      "eval_runtime": 331.8429,
      "eval_samples_per_second": 904.024,
      "eval_steps_per_second": 56.503,
      "step": 552000
    },
    {
      "epoch": 1.57,
      "learning_rate": 3.1433333333333336e-07,
      "loss": 2.4372,
      "step": 560000
    },
    {
      "epoch": 1.57,
      "eval_loss": 2.2884891033172607,
      "eval_runtime": 329.9724,
      "eval_samples_per_second": 909.149,
      "eval_steps_per_second": 56.823,
      "step": 560000
    },
    {
      "epoch": 1.59,
      "eval_loss": 2.300299882888794,
      "eval_runtime": 329.9035,
      "eval_samples_per_second": 909.339,
      "eval_steps_per_second": 56.835,
      "step": 568000
    },
    {
      "epoch": 1.62,
      "learning_rate": 3.116e-07,
      "loss": 2.4379,
      "step": 576000
    },
    {
      "epoch": 1.62,
      "eval_loss": 2.298779249191284,
      "eval_runtime": 329.3281,
      "eval_samples_per_second": 910.927,
      "eval_steps_per_second": 56.934,
      "step": 576000
    },
    {
      "epoch": 1.64,
      "eval_loss": 2.2923216819763184,
      "eval_runtime": 330.0728,
      "eval_samples_per_second": 908.872,
      "eval_steps_per_second": 56.806,
      "step": 584000
    },
    {
      "epoch": 1.66,
      "learning_rate": 3.0886666666666667e-07,
      "loss": 2.4347,
      "step": 592000
    },
    {
      "epoch": 1.66,
      "eval_loss": 2.2936880588531494,
      "eval_runtime": 331.2101,
      "eval_samples_per_second": 905.751,
      "eval_steps_per_second": 56.611,
      "step": 592000
    },
    {
      "epoch": 1.68,
      "eval_loss": 2.2957868576049805,
      "eval_runtime": 330.8976,
      "eval_samples_per_second": 906.607,
      "eval_steps_per_second": 56.664,
      "step": 600000
    },
    {
      "epoch": 1.71,
      "learning_rate": 3.061333333333333e-07,
      "loss": 2.4311,
      "step": 608000
    },
    {
      "epoch": 1.71,
      "eval_loss": 2.299522638320923,
      "eval_runtime": 330.1467,
      "eval_samples_per_second": 908.669,
      "eval_steps_per_second": 56.793,
      "step": 608000
    },
    {
      "epoch": 1.73,
      "eval_loss": 2.2941486835479736,
      "eval_runtime": 329.8116,
      "eval_samples_per_second": 909.592,
      "eval_steps_per_second": 56.851,
      "step": 616000
    },
    {
      "epoch": 1.75,
      "learning_rate": 3.034e-07,
      "loss": 2.4437,
      "step": 624000
    },
    {
      "epoch": 1.75,
      "eval_loss": 2.2949397563934326,
      "eval_runtime": 330.2069,
      "eval_samples_per_second": 908.503,
      "eval_steps_per_second": 56.783,
      "step": 624000
    },
    {
      "epoch": 1.77,
      "eval_loss": 2.2877914905548096,
      "eval_runtime": 331.6123,
      "eval_samples_per_second": 904.653,
      "eval_steps_per_second": 56.542,
      "step": 632000
    },
    {
      "epoch": 1.8,
      "learning_rate": 3.0066666666666663e-07,
      "loss": 2.4306,
      "step": 640000
    },
    {
      "epoch": 1.8,
      "eval_loss": 2.28951096534729,
      "eval_runtime": 331.7636,
      "eval_samples_per_second": 904.24,
      "eval_steps_per_second": 56.516,
      "step": 640000
    },
    {
      "epoch": 1.82,
      "eval_loss": 2.2930004596710205,
      "eval_runtime": 332.1218,
      "eval_samples_per_second": 903.265,
      "eval_steps_per_second": 56.455,
      "step": 648000
    },
    {
      "epoch": 1.84,
      "learning_rate": 2.9793333333333334e-07,
      "loss": 2.4341,
      "step": 656000
    },
    {
      "epoch": 1.84,
      "eval_loss": 2.2894575595855713,
      "eval_runtime": 330.5104,
      "eval_samples_per_second": 907.669,
      "eval_steps_per_second": 56.73,
      "step": 656000
    },
    {
      "epoch": 1.86,
      "eval_loss": 2.290764808654785,
      "eval_runtime": 333.5634,
      "eval_samples_per_second": 899.361,
      "eval_steps_per_second": 56.211,
      "step": 664000
    },
    {
      "epoch": 1.89,
      "learning_rate": 2.952e-07,
      "loss": 2.4333,
      "step": 672000
    },
    {
      "epoch": 1.89,
      "eval_loss": 2.284210681915283,
      "eval_runtime": 331.1722,
      "eval_samples_per_second": 905.855,
      "eval_steps_per_second": 56.617,
      "step": 672000
    },
    {
      "epoch": 1.91,
      "eval_loss": 2.29123592376709,
      "eval_runtime": 331.1751,
      "eval_samples_per_second": 905.847,
      "eval_steps_per_second": 56.617,
      "step": 680000
    },
    {
      "epoch": 1.93,
      "learning_rate": 2.9246666666666665e-07,
      "loss": 2.4403,
      "step": 688000
    },
    {
      "epoch": 1.93,
      "eval_loss": 2.290036916732788,
      "eval_runtime": 331.0957,
      "eval_samples_per_second": 906.064,
      "eval_steps_per_second": 56.63,
      "step": 688000
    },
    {
      "epoch": 1.95,
      "eval_loss": 2.2862110137939453,
      "eval_runtime": 332.3754,
      "eval_samples_per_second": 902.576,
      "eval_steps_per_second": 56.412,
      "step": 696000
    },
    {
      "epoch": 1.98,
      "learning_rate": 2.897333333333333e-07,
      "loss": 2.4396,
      "step": 704000
    },
    {
      "epoch": 1.98,
      "eval_loss": 2.287149667739868,
      "eval_runtime": 332.2121,
      "eval_samples_per_second": 903.019,
      "eval_steps_per_second": 56.44,
      "step": 704000
    },
    {
      "epoch": 2.0,
      "eval_loss": 2.294781446456909,
      "eval_runtime": 331.5538,
      "eval_samples_per_second": 904.813,
      "eval_steps_per_second": 56.552,
      "step": 712000
    },
    {
      "epoch": 2.02,
      "learning_rate": 2.8699999999999996e-07,
      "loss": 2.441,
      "step": 720000
    },
    {
      "epoch": 2.02,
      "eval_loss": 2.294177293777466,
      "eval_runtime": 332.0516,
      "eval_samples_per_second": 903.456,
      "eval_steps_per_second": 56.467,
      "step": 720000
    },
    {
      "epoch": 2.04,
      "eval_loss": 2.2828338146209717,
      "eval_runtime": 331.6742,
      "eval_samples_per_second": 904.484,
      "eval_steps_per_second": 56.531,
      "step": 728000
    },
    {
      "epoch": 2.07,
      "learning_rate": 2.8426666666666667e-07,
      "loss": 2.434,
      "step": 736000
    },
    {
      "epoch": 2.07,
      "eval_loss": 2.2808279991149902,
      "eval_runtime": 332.944,
      "eval_samples_per_second": 901.034,
      "eval_steps_per_second": 56.316,
      "step": 736000
    },
    {
      "epoch": 2.09,
      "eval_loss": 2.2883412837982178,
      "eval_runtime": 334.8461,
      "eval_samples_per_second": 895.916,
      "eval_steps_per_second": 55.996,
      "step": 744000
    },
    {
      "epoch": 2.11,
      "learning_rate": 2.815333333333333e-07,
      "loss": 2.4387,
      "step": 752000
    },
    {
      "epoch": 2.11,
      "eval_loss": 2.292271375656128,
      "eval_runtime": 332.2552,
      "eval_samples_per_second": 902.902,
      "eval_steps_per_second": 56.433,
      "step": 752000
    },
    {
      "epoch": 2.13,
      "eval_loss": 2.284794569015503,
      "eval_runtime": 335.1059,
      "eval_samples_per_second": 895.222,
      "eval_steps_per_second": 55.952,
      "step": 760000
    },
    {
      "epoch": 2.16,
      "learning_rate": 2.7880000000000003e-07,
      "loss": 2.4342,
      "step": 768000
    },
    {
      "epoch": 2.16,
      "eval_loss": 2.2847986221313477,
      "eval_runtime": 331.6425,
      "eval_samples_per_second": 904.57,
      "eval_steps_per_second": 56.537,
      "step": 768000
    },
    {
      "epoch": 2.18,
      "eval_loss": 2.2864720821380615,
      "eval_runtime": 332.0862,
      "eval_samples_per_second": 903.362,
      "eval_steps_per_second": 56.461,
      "step": 776000
    },
    {
      "epoch": 2.2,
      "learning_rate": 2.7606666666666664e-07,
      "loss": 2.4389,
      "step": 784000
    },
    {
      "epoch": 2.2,
      "eval_loss": 2.288520574569702,
      "eval_runtime": 332.0566,
      "eval_samples_per_second": 903.442,
      "eval_steps_per_second": 56.466,
      "step": 784000
    },
    {
      "epoch": 2.22,
      "eval_loss": 2.2794013023376465,
      "eval_runtime": 331.7528,
      "eval_samples_per_second": 904.27,
      "eval_steps_per_second": 56.518,
      "step": 792000
    },
    {
      "epoch": 2.25,
      "learning_rate": 2.733333333333333e-07,
      "loss": 2.4318,
      "step": 800000
    },
    {
      "epoch": 2.25,
      "eval_loss": 2.2861220836639404,
      "eval_runtime": 332.6258,
      "eval_samples_per_second": 901.896,
      "eval_steps_per_second": 56.37,
      "step": 800000
    },
    {
      "epoch": 2.27,
      "eval_loss": 2.2875726222991943,
      "eval_runtime": 332.0265,
      "eval_samples_per_second": 903.524,
      "eval_steps_per_second": 56.471,
      "step": 808000
    },
    {
      "epoch": 2.29,
      "learning_rate": 2.706e-07,
      "loss": 2.4343,
      "step": 816000
    },
    {
      "epoch": 2.29,
      "eval_loss": 2.2820258140563965,
      "eval_runtime": 332.3446,
      "eval_samples_per_second": 902.659,
      "eval_steps_per_second": 56.417,
      "step": 816000
    },
    {
      "epoch": 2.31,
      "eval_loss": 2.283487319946289,
      "eval_runtime": 332.0052,
      "eval_samples_per_second": 903.582,
      "eval_steps_per_second": 56.475,
      "step": 824000
    },
    {
      "epoch": 2.34,
      "learning_rate": 2.6786666666666666e-07,
      "loss": 2.4335,
      "step": 832000
    },
    {
      "epoch": 2.34,
      "eval_loss": 2.278824806213379,
      "eval_runtime": 334.3682,
      "eval_samples_per_second": 897.196,
      "eval_steps_per_second": 56.076,
      "step": 832000
    },
    {
      "epoch": 2.36,
      "eval_loss": 2.281332015991211,
      "eval_runtime": 334.2594,
      "eval_samples_per_second": 897.489,
      "eval_steps_per_second": 56.094,
      "step": 840000
    },
    {
      "epoch": 2.38,
      "learning_rate": 2.651333333333333e-07,
      "loss": 2.4428,
      "step": 848000
    },
    {
      "epoch": 2.38,
      "eval_loss": 2.2788710594177246,
      "eval_runtime": 335.0984,
      "eval_samples_per_second": 895.242,
      "eval_steps_per_second": 55.954,
      "step": 848000
    },
    {
      "epoch": 2.4,
      "eval_loss": 2.285792589187622,
      "eval_runtime": 333.8076,
      "eval_samples_per_second": 898.703,
      "eval_steps_per_second": 56.17,
      "step": 856000
    },
    {
      "epoch": 2.43,
      "learning_rate": 2.624e-07,
      "loss": 2.4272,
      "step": 864000
    },
    {
      "epoch": 2.43,
      "eval_loss": 2.288302183151245,
      "eval_runtime": 334.3033,
      "eval_samples_per_second": 897.371,
      "eval_steps_per_second": 56.087,
      "step": 864000
    },
    {
      "epoch": 2.45,
      "eval_loss": 2.280890941619873,
      "eval_runtime": 335.6751,
      "eval_samples_per_second": 893.704,
      "eval_steps_per_second": 55.858,
      "step": 872000
    },
    {
      "epoch": 2.47,
      "learning_rate": 2.596666666666667e-07,
      "loss": 2.4331,
      "step": 880000
    },
    {
      "epoch": 2.47,
      "eval_loss": 2.288017988204956,
      "eval_runtime": 340.6851,
      "eval_samples_per_second": 880.561,
      "eval_steps_per_second": 55.036,
      "step": 880000
    },
    {
      "epoch": 2.49,
      "eval_loss": 2.283820867538452,
      "eval_runtime": 339.1018,
      "eval_samples_per_second": 884.672,
      "eval_steps_per_second": 55.293,
      "step": 888000
    },
    {
      "epoch": 2.52,
      "learning_rate": 2.5693333333333333e-07,
      "loss": 2.4326,
      "step": 896000
    },
    {
      "epoch": 2.52,
      "eval_loss": 2.2804486751556396,
      "eval_runtime": 335.9214,
      "eval_samples_per_second": 893.048,
      "eval_steps_per_second": 55.817,
      "step": 896000
    },
    {
      "epoch": 2.54,
      "eval_loss": 2.2831339836120605,
      "eval_runtime": 337.559,
      "eval_samples_per_second": 888.716,
      "eval_steps_per_second": 55.546,
      "step": 904000
    },
    {
      "epoch": 2.56,
      "learning_rate": 2.542e-07,
      "loss": 2.436,
      "step": 912000
    },
    {
      "epoch": 2.56,
      "eval_loss": 2.286670446395874,
      "eval_runtime": 334.8087,
      "eval_samples_per_second": 896.016,
      "eval_steps_per_second": 56.002,
      "step": 912000
    },
    {
      "epoch": 2.58,
      "eval_loss": 2.28481125831604,
      "eval_runtime": 336.2362,
      "eval_samples_per_second": 892.212,
      "eval_steps_per_second": 55.764,
      "step": 920000
    },
    {
      "epoch": 2.6,
      "learning_rate": 2.5146666666666664e-07,
      "loss": 2.435,
      "step": 928000
    },
    {
      "epoch": 2.6,
      "eval_loss": 2.287050485610962,
      "eval_runtime": 335.1109,
      "eval_samples_per_second": 895.208,
      "eval_steps_per_second": 55.952,
      "step": 928000
    },
    {
      "epoch": 2.63,
      "eval_loss": 2.2828164100646973,
      "eval_runtime": 333.979,
      "eval_samples_per_second": 898.242,
      "eval_steps_per_second": 56.141,
      "step": 936000
    },
    {
      "epoch": 2.65,
      "learning_rate": 2.4873333333333335e-07,
      "loss": 2.44,
      "step": 944000
    },
    {
      "epoch": 2.65,
      "eval_loss": 2.2807881832122803,
      "eval_runtime": 334.0843,
      "eval_samples_per_second": 897.959,
      "eval_steps_per_second": 56.124,
      "step": 944000
    },
    {
      "epoch": 2.67,
      "eval_loss": 2.2852513790130615,
      "eval_runtime": 334.6701,
      "eval_samples_per_second": 896.387,
      "eval_steps_per_second": 56.025,
      "step": 952000
    },
    {
      "epoch": 2.69,
      "learning_rate": 2.46e-07,
      "loss": 2.4285,
      "step": 960000
    },
    {
      "epoch": 2.69,
      "eval_loss": 2.279860258102417,
      "eval_runtime": 338.1471,
      "eval_samples_per_second": 887.17,
      "eval_steps_per_second": 55.449,
      "step": 960000
    },
    {
      "epoch": 2.72,
      "eval_loss": 2.282912015914917,
      "eval_runtime": 335.4921,
      "eval_samples_per_second": 894.191,
      "eval_steps_per_second": 55.888,
      "step": 968000
    },
    {
      "epoch": 2.74,
      "learning_rate": 2.4326666666666666e-07,
      "loss": 2.423,
      "step": 976000
    },
    {
      "epoch": 2.74,
      "eval_loss": 2.2761270999908447,
      "eval_runtime": 335.1528,
      "eval_samples_per_second": 895.096,
      "eval_steps_per_second": 55.945,
      "step": 976000
    },
    {
      "epoch": 2.76,
      "eval_loss": 2.276808738708496,
      "eval_runtime": 336.6204,
      "eval_samples_per_second": 891.194,
      "eval_steps_per_second": 55.701,
      "step": 984000
    },
    {
      "epoch": 2.78,
      "learning_rate": 2.405333333333333e-07,
      "loss": 2.4353,
      "step": 992000
    },
    {
      "epoch": 2.78,
      "eval_loss": 2.2844138145446777,
      "eval_runtime": 335.677,
      "eval_samples_per_second": 893.698,
      "eval_steps_per_second": 55.857,
      "step": 992000
    },
    {
      "epoch": 2.81,
      "eval_loss": 2.2828099727630615,
      "eval_runtime": 335.9499,
      "eval_samples_per_second": 892.972,
      "eval_steps_per_second": 55.812,
      "step": 1000000
    },
    {
      "epoch": 2.83,
      "learning_rate": 2.3779999999999997e-07,
      "loss": 2.4301,
      "step": 1008000
    },
    {
      "epoch": 2.83,
      "eval_loss": 2.2806167602539062,
      "eval_runtime": 334.273,
      "eval_samples_per_second": 897.452,
      "eval_steps_per_second": 56.092,
      "step": 1008000
    },
    {
      "epoch": 2.85,
      "eval_loss": 2.281301975250244,
      "eval_runtime": 335.4442,
      "eval_samples_per_second": 894.319,
      "eval_steps_per_second": 55.896,
      "step": 1016000
    },
    {
      "epoch": 2.87,
      "learning_rate": 2.3506666666666668e-07,
      "loss": 2.4284,
      "step": 1024000
    },
    {
      "epoch": 2.87,
      "eval_loss": 2.2789454460144043,
      "eval_runtime": 334.3,
      "eval_samples_per_second": 897.38,
      "eval_steps_per_second": 56.087,
      "step": 1024000
    },
    {
      "epoch": 2.9,
      "eval_loss": 2.2769646644592285,
      "eval_runtime": 333.9542,
      "eval_samples_per_second": 898.309,
      "eval_steps_per_second": 56.145,
      "step": 1032000
    },
    {
      "epoch": 2.92,
      "learning_rate": 2.3233333333333334e-07,
      "loss": 2.4252,
      "step": 1040000
    },
    {
      "epoch": 2.92,
      "eval_loss": 2.2762739658355713,
      "eval_runtime": 334.5705,
      "eval_samples_per_second": 896.654,
      "eval_steps_per_second": 56.042,
      "step": 1040000
    },
    {
      "epoch": 2.94,
      "eval_loss": 2.276264190673828,
      "eval_runtime": 334.7693,
      "eval_samples_per_second": 896.122,
      "eval_steps_per_second": 56.009,
      "step": 1048000
    },
    {
      "epoch": 2.96,
      "learning_rate": 2.2960000000000002e-07,
      "loss": 2.4289,
      "step": 1056000
    },
    {
      "epoch": 2.96,
      "eval_loss": 2.277852773666382,
      "eval_runtime": 334.9735,
      "eval_samples_per_second": 895.575,
      "eval_steps_per_second": 55.975,
      "step": 1056000
    },
    {
      "epoch": 2.99,
      "eval_loss": 2.2811880111694336,
      "eval_runtime": 334.7215,
      "eval_samples_per_second": 896.25,
      "eval_steps_per_second": 56.017,
      "step": 1064000
    },
    {
      "epoch": 3.01,
      "learning_rate": 2.2686666666666667e-07,
      "loss": 2.4349,
      "step": 1072000
    },
    {
      "epoch": 3.01,
      "eval_loss": 2.2881336212158203,
      "eval_runtime": 335.0398,
      "eval_samples_per_second": 895.398,
      "eval_steps_per_second": 55.963,
      "step": 1072000
    },
    {
      "epoch": 3.03,
      "eval_loss": 2.2804529666900635,
      "eval_runtime": 336.2168,
      "eval_samples_per_second": 892.264,
      "eval_steps_per_second": 55.768,
      "step": 1080000
    },
    {
      "epoch": 3.05,
      "learning_rate": 2.2413333333333333e-07,
      "loss": 2.4365,
      "step": 1088000
    },
    {
      "epoch": 3.05,
      "eval_loss": 2.2758219242095947,
      "eval_runtime": 335.853,
      "eval_samples_per_second": 893.23,
      "eval_steps_per_second": 55.828,
      "step": 1088000
    },
    {
      "epoch": 3.08,
      "eval_loss": 2.2732620239257812,
      "eval_runtime": 334.8017,
      "eval_samples_per_second": 896.035,
      "eval_steps_per_second": 56.003,
      "step": 1096000
    },
    {
      "epoch": 3.1,
      "learning_rate": 2.214e-07,
      "loss": 2.4274,
      "step": 1104000
    },
    {
      "epoch": 3.1,
      "eval_loss": 2.2842442989349365,
      "eval_runtime": 337.9471,
      "eval_samples_per_second": 887.695,
      "eval_steps_per_second": 55.482,
      "step": 1104000
    },
    {
      "epoch": 3.12,
      "eval_loss": 2.280796766281128,
      "eval_runtime": 336.7832,
      "eval_samples_per_second": 890.763,
      "eval_steps_per_second": 55.674,
      "step": 1112000
    },
    {
      "epoch": 3.14,
      "learning_rate": 2.1866666666666667e-07,
      "loss": 2.4326,
      "step": 1120000
    },
    {
      "epoch": 3.14,
      "eval_loss": 2.2752950191497803,
      "eval_runtime": 335.1287,
      "eval_samples_per_second": 895.161,
      "eval_steps_per_second": 55.949,
      "step": 1120000
    },
    {
      "epoch": 3.17,
      "eval_loss": 2.2791523933410645,
      "eval_runtime": 335.4799,
      "eval_samples_per_second": 894.224,
      "eval_steps_per_second": 55.89,
      "step": 1128000
    },
    {
      "epoch": 3.19,
      "learning_rate": 2.1593333333333332e-07,
      "loss": 2.4244,
      "step": 1136000
    },
    {
      "epoch": 3.19,
      "eval_loss": 2.27884578704834,
      "eval_runtime": 335.4217,
      "eval_samples_per_second": 894.379,
      "eval_steps_per_second": 55.9,
      "step": 1136000
    },
    {
      "epoch": 3.21,
      "eval_loss": 2.282371997833252,
      "eval_runtime": 336.2876,
      "eval_samples_per_second": 892.076,
      "eval_steps_per_second": 55.756,
      "step": 1144000
    },
    {
      "epoch": 3.23,
      "learning_rate": 2.132e-07,
      "loss": 2.4285,
      "step": 1152000
    },
    {
      "epoch": 3.23,
      "eval_loss": 2.2799980640411377,
      "eval_runtime": 335.867,
      "eval_samples_per_second": 893.193,
      "eval_steps_per_second": 55.826,
      "step": 1152000
    },
    {
      "epoch": 3.26,
      "eval_loss": 2.2783775329589844,
      "eval_runtime": 335.6089,
      "eval_samples_per_second": 893.88,
      "eval_steps_per_second": 55.869,
      "step": 1160000
    },
    {
      "epoch": 3.28,
      "learning_rate": 2.1046666666666666e-07,
      "loss": 2.4371,
      "step": 1168000
    },
    {
      "epoch": 3.28,
      "eval_loss": 2.2675371170043945,
      "eval_runtime": 335.7808,
      "eval_samples_per_second": 893.422,
      "eval_steps_per_second": 55.84,
      "step": 1168000
    },
    {
      "epoch": 3.3,
      "eval_loss": 2.2739932537078857,
      "eval_runtime": 336.4769,
      "eval_samples_per_second": 891.574,
      "eval_steps_per_second": 55.724,
      "step": 1176000
    },
    {
      "epoch": 3.32,
      "learning_rate": 2.0773333333333334e-07,
      "loss": 2.4273,
      "step": 1184000
    },
    {
      "epoch": 3.32,
      "eval_loss": 2.2804715633392334,
      "eval_runtime": 339.219,
      "eval_samples_per_second": 884.367,
      "eval_steps_per_second": 55.274,
      "step": 1184000
    },
    {
      "epoch": 3.35,
      "eval_loss": 2.2848641872406006,
      "eval_runtime": 336.5161,
      "eval_samples_per_second": 891.47,
      "eval_steps_per_second": 55.718,
      "step": 1192000
    },
    {
      "epoch": 3.37,
      "learning_rate": 2.05e-07,
      "loss": 2.4359,
      "step": 1200000
    },
    {
      "epoch": 3.37,
      "eval_loss": 2.2807633876800537,
      "eval_runtime": 336.9191,
      "eval_samples_per_second": 890.404,
      "eval_steps_per_second": 55.651,
      "step": 1200000
    },
    {
      "epoch": 3.39,
      "eval_loss": 2.2790510654449463,
      "eval_runtime": 336.1666,
      "eval_samples_per_second": 892.397,
      "eval_steps_per_second": 55.776,
      "step": 1208000
    },
    {
      "epoch": 3.41,
      "learning_rate": 2.0226666666666668e-07,
      "loss": 2.4303,
      "step": 1216000
    },
    {
      "epoch": 3.41,
      "eval_loss": 2.2729713916778564,
      "eval_runtime": 337.949,
      "eval_samples_per_second": 887.69,
      "eval_steps_per_second": 55.482,
      "step": 1216000
    },
    {
      "epoch": 3.44,
      "eval_loss": 2.273223638534546,
      "eval_runtime": 338.6051,
      "eval_samples_per_second": 885.97,
      "eval_steps_per_second": 55.374,
      "step": 1224000
    },
    {
      "epoch": 3.46,
      "learning_rate": 1.9953333333333333e-07,
      "loss": 2.4306,
      "step": 1232000
    },
    {
      "epoch": 3.46,
      "eval_loss": 2.2784602642059326,
      "eval_runtime": 338.5063,
      "eval_samples_per_second": 886.229,
      "eval_steps_per_second": 55.39,
      "step": 1232000
    },
    {
      "epoch": 3.48,
      "eval_loss": 2.2763924598693848,
      "eval_runtime": 336.2284,
      "eval_samples_per_second": 892.233,
      "eval_steps_per_second": 55.766,
      "step": 1240000
    },
    {
      "epoch": 3.5,
      "learning_rate": 1.968e-07,
      "loss": 2.4267,
      "step": 1248000
    },
    {
      "epoch": 3.5,
      "eval_loss": 2.2739803791046143,
      "eval_runtime": 337.8728,
      "eval_samples_per_second": 887.89,
      "eval_steps_per_second": 55.494,
      "step": 1248000
    },
    {
      "epoch": 3.53,
      "eval_loss": 2.2789418697357178,
      "eval_runtime": 337.6123,
      "eval_samples_per_second": 888.576,
      "eval_steps_per_second": 55.537,
      "step": 1256000
    },
    {
      "epoch": 3.55,
      "learning_rate": 1.9406666666666667e-07,
      "loss": 2.4271,
      "step": 1264000
    },
    {
      "epoch": 3.55,
      "eval_loss": 2.277411937713623,
      "eval_runtime": 337.0558,
      "eval_samples_per_second": 890.043,
      "eval_steps_per_second": 55.629,
      "step": 1264000
    },
    {
      "epoch": 3.57,
      "eval_loss": 2.276827335357666,
      "eval_runtime": 337.1407,
      "eval_samples_per_second": 889.818,
      "eval_steps_per_second": 55.615,
      "step": 1272000
    },
    {
      "epoch": 3.59,
      "learning_rate": 1.9133333333333333e-07,
      "loss": 2.4263,
      "step": 1280000
    },
    {
      "epoch": 3.59,
      "eval_loss": 2.279576539993286,
      "eval_runtime": 338.6094,
      "eval_samples_per_second": 885.959,
      "eval_steps_per_second": 55.374,
      "step": 1280000
    },
    {
      "epoch": 3.62,
      "eval_loss": 2.2759058475494385,
      "eval_runtime": 337.2844,
      "eval_samples_per_second": 889.439,
      "eval_steps_per_second": 55.591,
      "step": 1288000
    },
    {
      "epoch": 3.64,
      "learning_rate": 1.886e-07,
      "loss": 2.431,
      "step": 1296000
    },
    {
      "epoch": 3.64,
      "eval_loss": 2.274071216583252,
      "eval_runtime": 337.5903,
      "eval_samples_per_second": 888.633,
      "eval_steps_per_second": 55.541,
      "step": 1296000
    },
    {
      "epoch": 3.66,
      "eval_loss": 2.282141923904419,
      "eval_runtime": 337.5885,
      "eval_samples_per_second": 888.638,
      "eval_steps_per_second": 55.541,
      "step": 1304000
    },
    {
      "epoch": 3.68,
      "learning_rate": 1.8586666666666666e-07,
      "loss": 2.4273,
      "step": 1312000
    },
    {
      "epoch": 3.68,
      "eval_loss": 2.2739663124084473,
      "eval_runtime": 337.7029,
      "eval_samples_per_second": 888.337,
      "eval_steps_per_second": 55.522,
      "step": 1312000
    },
    {
      "epoch": 3.71,
      "eval_loss": 2.2712557315826416,
      "eval_runtime": 337.356,
      "eval_samples_per_second": 889.251,
      "eval_steps_per_second": 55.579,
      "step": 1320000
    },
    {
      "epoch": 3.73,
      "learning_rate": 1.8313333333333332e-07,
      "loss": 2.4371,
      "step": 1328000
    },
    {
      "epoch": 3.73,
      "eval_loss": 2.2704272270202637,
      "eval_runtime": 339.5591,
      "eval_samples_per_second": 883.481,
      "eval_steps_per_second": 55.219,
      "step": 1328000
    },
    {
      "epoch": 3.75,
      "eval_loss": 2.273430109024048,
      "eval_runtime": 339.8184,
      "eval_samples_per_second": 882.807,
      "eval_steps_per_second": 55.177,
      "step": 1336000
    },
    {
      "epoch": 3.77,
      "learning_rate": 1.804e-07,
      "loss": 2.4273,
      "step": 1344000
    },
    {
      "epoch": 3.77,
      "eval_loss": 2.2745580673217773,
      "eval_runtime": 338.4911,
      "eval_samples_per_second": 886.268,
      "eval_steps_per_second": 55.393,
      "step": 1344000
    },
    {
      "epoch": 3.8,
      "eval_loss": 2.284024953842163,
      "eval_runtime": 337.8579,
      "eval_samples_per_second": 887.929,
      "eval_steps_per_second": 55.497,
      "step": 1352000
    },
    {
      "epoch": 3.82,
      "learning_rate": 1.7766666666666666e-07,
      "loss": 2.4246,
      "step": 1360000
    },
    {
      "epoch": 3.82,
      "eval_loss": 2.2764360904693604,
      "eval_runtime": 337.6899,
      "eval_samples_per_second": 888.371,
      "eval_steps_per_second": 55.524,
      "step": 1360000
    },
    {
      "epoch": 3.84,
      "eval_loss": 2.274030923843384,
      "eval_runtime": 338.4156,
      "eval_samples_per_second": 886.466,
      "eval_steps_per_second": 55.405,
      "step": 1368000
    },
    {
      "epoch": 3.86,
      "learning_rate": 1.7493333333333334e-07,
      "loss": 2.4308,
      "step": 1376000
    },
    {
      "epoch": 3.86,
      "eval_loss": 2.2730188369750977,
      "eval_runtime": 338.8191,
      "eval_samples_per_second": 885.411,
      "eval_steps_per_second": 55.339,
      "step": 1376000
    },
    {
      "epoch": 3.88,
      "eval_loss": 2.2750706672668457,
      "eval_runtime": 338.5243,
      "eval_samples_per_second": 886.182,
      "eval_steps_per_second": 55.387,
      "step": 1384000
    },
    {
      "epoch": 3.91,
      "learning_rate": 1.722e-07,
      "loss": 2.4341,
      "step": 1392000
    },
    {
      "epoch": 3.91,
      "eval_loss": 2.277709484100342,
      "eval_runtime": 338.1761,
      "eval_samples_per_second": 887.094,
      "eval_steps_per_second": 55.444,
      "step": 1392000
    },
    {
      "epoch": 3.93,
      "eval_loss": 2.267932653427124,
      "eval_runtime": 340.4821,
      "eval_samples_per_second": 881.086,
      "eval_steps_per_second": 55.069,
      "step": 1400000
    },
    {
      "epoch": 3.95,
      "learning_rate": 1.6946666666666668e-07,
      "loss": 2.4266,
      "step": 1408000
    },
    {
      "epoch": 3.95,
      "eval_loss": 2.277672052383423,
      "eval_runtime": 338.3287,
      "eval_samples_per_second": 886.694,
      "eval_steps_per_second": 55.419,
      "step": 1408000
    },
    {
      "epoch": 3.97,
      "eval_loss": 2.2783212661743164,
      "eval_runtime": 338.5149,
      "eval_samples_per_second": 886.206,
      "eval_steps_per_second": 55.389,
      "step": 1416000
    },
    {
      "epoch": 4.0,
      "learning_rate": 1.6673333333333333e-07,
      "loss": 2.4344,
      "step": 1424000
    },
    {
      "epoch": 4.0,
      "eval_loss": 2.2742812633514404,
      "eval_runtime": 339.2997,
      "eval_samples_per_second": 884.156,
      "eval_steps_per_second": 55.261,
      "step": 1424000
    },
    {
      "epoch": 4.02,
      "eval_loss": 2.2690906524658203,
      "eval_runtime": 338.4672,
      "eval_samples_per_second": 886.331,
      "eval_steps_per_second": 55.397,
      "step": 1432000
    },
    {
      "epoch": 4.04,
      "learning_rate": 1.64e-07,
      "loss": 2.431,
      "step": 1440000
    },
    {
      "epoch": 4.04,
      "eval_loss": 2.2713911533355713,
      "eval_runtime": 339.3847,
      "eval_samples_per_second": 883.935,
      "eval_steps_per_second": 55.247,
      "step": 1440000
    },
    {
      "epoch": 4.06,
      "eval_loss": 2.2693912982940674,
      "eval_runtime": 339.8425,
      "eval_samples_per_second": 882.744,
      "eval_steps_per_second": 55.173,
      "step": 1448000
    },
    {
      "epoch": 4.09,
      "learning_rate": 1.6126666666666667e-07,
      "loss": 2.4296,
      "step": 1456000
    },
    {
      "epoch": 4.09,
      "eval_loss": 2.274890422821045,
      "eval_runtime": 341.4559,
      "eval_samples_per_second": 878.573,
      "eval_steps_per_second": 54.912,
      "step": 1456000
    },
    {
      "epoch": 4.11,
      "eval_loss": 2.280956268310547,
      "eval_runtime": 340.0142,
      "eval_samples_per_second": 882.298,
      "eval_steps_per_second": 55.145,
      "step": 1464000
    },
    {
      "epoch": 4.13,
      "learning_rate": 1.5853333333333332e-07,
      "loss": 2.4265,
      "step": 1472000
    },
    {
      "epoch": 4.13,
      "eval_loss": 2.2743868827819824,
      "eval_runtime": 341.6626,
      "eval_samples_per_second": 878.042,
      "eval_steps_per_second": 54.879,
      "step": 1472000
    },
    {
      "epoch": 4.15,
      "eval_loss": 2.271418571472168,
      "eval_runtime": 339.1644,
      "eval_samples_per_second": 884.509,
      "eval_steps_per_second": 55.283,
      "step": 1480000
    },
    {
      "epoch": 4.18,
      "learning_rate": 1.558e-07,
      "loss": 2.4266,
      "step": 1488000
    },
    {
      "epoch": 4.18,
      "eval_loss": 2.273254871368408,
      "eval_runtime": 339.6628,
      "eval_samples_per_second": 883.211,
      "eval_steps_per_second": 55.202,
      "step": 1488000
    },
    {
      "epoch": 4.2,
      "eval_loss": 2.278977632522583,
      "eval_runtime": 339.3751,
      "eval_samples_per_second": 883.96,
      "eval_steps_per_second": 55.249,
      "step": 1496000
    },
    {
      "epoch": 4.22,
      "learning_rate": 1.5306666666666666e-07,
      "loss": 2.4253,
      "step": 1504000
    },
    {
      "epoch": 4.22,
      "eval_loss": 2.27659273147583,
      "eval_runtime": 339.267,
      "eval_samples_per_second": 884.242,
      "eval_steps_per_second": 55.266,
      "step": 1504000
    },
    {
      "epoch": 4.24,
      "eval_loss": 2.276425361633301,
      "eval_runtime": 339.521,
      "eval_samples_per_second": 883.58,
      "eval_steps_per_second": 55.225,
      "step": 1512000
    },
    {
      "epoch": 4.27,
      "learning_rate": 1.5033333333333332e-07,
      "loss": 2.4303,
      "step": 1520000
    },
    {
      "epoch": 4.27,
      "eval_loss": 2.269226312637329,
      "eval_runtime": 340.2122,
      "eval_samples_per_second": 881.785,
      "eval_steps_per_second": 55.113,
      "step": 1520000
    },
    {
      "epoch": 4.29,
      "eval_loss": 2.268404006958008,
      "eval_runtime": 340.6392,
      "eval_samples_per_second": 880.68,
      "eval_steps_per_second": 55.044,
      "step": 1528000
    },
    {
      "epoch": 4.31,
      "learning_rate": 1.476e-07,
      "loss": 2.4373,
      "step": 1536000
    },
    {
      "epoch": 4.31,
      "eval_loss": 2.275193929672241,
      "eval_runtime": 341.5541,
      "eval_samples_per_second": 878.321,
      "eval_steps_per_second": 54.896,
      "step": 1536000
    },
    {
      "epoch": 4.33,
      "eval_loss": 2.270094633102417,
      "eval_runtime": 341.2941,
      "eval_samples_per_second": 878.99,
      "eval_steps_per_second": 54.938,
      "step": 1544000
    },
    {
      "epoch": 4.36,
      "learning_rate": 1.4486666666666665e-07,
      "loss": 2.4346,
      "step": 1552000
    },
    {
      "epoch": 4.36,
      "eval_loss": 2.2757863998413086,
      "eval_runtime": 340.0188,
      "eval_samples_per_second": 882.287,
      "eval_steps_per_second": 55.144,
      "step": 1552000
    },
    {
      "epoch": 4.38,
      "eval_loss": 2.2727184295654297,
      "eval_runtime": 340.3133,
      "eval_samples_per_second": 881.523,
      "eval_steps_per_second": 55.096,
      "step": 1560000
    },
    {
      "epoch": 4.4,
      "learning_rate": 1.4213333333333334e-07,
      "loss": 2.4294,
      "step": 1568000
    },
    {
      "epoch": 4.4,
      "eval_loss": 2.2752585411071777,
      "eval_runtime": 340.0045,
      "eval_samples_per_second": 882.324,
      "eval_steps_per_second": 55.146,
      "step": 1568000
    },
    {
      "epoch": 4.42,
      "eval_loss": 2.2686994075775146,
      "eval_runtime": 340.2528,
      "eval_samples_per_second": 881.68,
      "eval_steps_per_second": 55.106,
      "step": 1576000
    },
    {
      "epoch": 4.45,
      "learning_rate": 1.3940000000000002e-07,
      "loss": 2.439,
      "step": 1584000
    },
    {
      "epoch": 4.45,
      "eval_loss": 2.2775542736053467,
      "eval_runtime": 342.6447,
      "eval_samples_per_second": 875.525,
      "eval_steps_per_second": 54.721,
      "step": 1584000
    },
    {
      "epoch": 4.47,
      "eval_loss": 2.274559497833252,
      "eval_runtime": 342.3787,
      "eval_samples_per_second": 876.205,
      "eval_steps_per_second": 54.764,
      "step": 1592000
    },
    {
      "epoch": 4.49,
      "learning_rate": 1.3666666666666665e-07,
      "loss": 2.4337,
      "step": 1600000
    },
    {
      "epoch": 4.49,
      "eval_loss": 2.2730839252471924,
      "eval_runtime": 340.9451,
      "eval_samples_per_second": 879.889,
      "eval_steps_per_second": 54.994,
      "step": 1600000
    },
    {
      "epoch": 4.51,
      "eval_loss": 2.2721805572509766,
      "eval_runtime": 342.0012,
      "eval_samples_per_second": 877.172,
      "eval_steps_per_second": 54.824,
      "step": 1608000
    },
    {
      "epoch": 4.54,
      "learning_rate": 1.3393333333333333e-07,
      "loss": 2.4273,
      "step": 1616000
    },
    {
      "epoch": 4.54,
      "eval_loss": 2.270340919494629,
      "eval_runtime": 341.2582,
      "eval_samples_per_second": 879.082,
      "eval_steps_per_second": 54.944,
      "step": 1616000
    },
    {
      "epoch": 4.56,
      "eval_loss": 2.2802205085754395,
      "eval_runtime": 341.9475,
      "eval_samples_per_second": 877.31,
      "eval_steps_per_second": 54.833,
      "step": 1624000
    },
    {
      "epoch": 4.58,
      "learning_rate": 1.312e-07,
      "loss": 2.4275,
      "step": 1632000
    },
    {
      "epoch": 4.58,
      "eval_loss": 2.2707149982452393,
      "eval_runtime": 341.3433,
      "eval_samples_per_second": 878.863,
      "eval_steps_per_second": 54.93,
      "step": 1632000
    },
    {
      "epoch": 4.6,
      "eval_loss": 2.270657777786255,
      "eval_runtime": 342.3544,
      "eval_samples_per_second": 876.267,
      "eval_steps_per_second": 54.768,
      "step": 1640000
    },
    {
      "epoch": 4.63,
      "learning_rate": 1.2846666666666667e-07,
      "loss": 2.4201,
      "step": 1648000
    },
    {
      "epoch": 4.63,
      "eval_loss": 2.268555164337158,
      "eval_runtime": 343.0641,
      "eval_samples_per_second": 874.455,
      "eval_steps_per_second": 54.655,
      "step": 1648000
    },
    {
      "epoch": 4.65,
      "eval_loss": 2.2706844806671143,
      "eval_runtime": 343.4762,
      "eval_samples_per_second": 873.405,
      "eval_steps_per_second": 54.589,
      "step": 1656000
    },
    {
      "epoch": 4.67,
      "learning_rate": 1.2573333333333332e-07,
      "loss": 2.4319,
      "step": 1664000
    },
    {
      "epoch": 4.67,
      "eval_loss": 2.2739858627319336,
      "eval_runtime": 346.4999,
      "eval_samples_per_second": 865.784,
      "eval_steps_per_second": 54.113,
      "step": 1664000
    },
    {
      "epoch": 4.69,
      "eval_loss": 2.2696831226348877,
      "eval_runtime": 344.1357,
      "eval_samples_per_second": 871.732,
      "eval_steps_per_second": 54.484,
      "step": 1672000
    },
    {
      "epoch": 4.72,
      "learning_rate": 1.23e-07,
      "loss": 2.4314,
      "step": 1680000
    },
    {
      "epoch": 4.72,
      "eval_loss": 2.2746658325195312,
      "eval_runtime": 342.9467,
      "eval_samples_per_second": 874.754,
      "eval_steps_per_second": 54.673,
      "step": 1680000
    },
    {
      "epoch": 4.74,
      "eval_loss": 2.2693660259246826,
      "eval_runtime": 345.5224,
      "eval_samples_per_second": 868.233,
      "eval_steps_per_second": 54.266,
      "step": 1688000
    },
    {
      "epoch": 4.76,
      "learning_rate": 1.2026666666666666e-07,
      "loss": 2.4242,
      "step": 1696000
    },
    {
      "epoch": 4.76,
      "eval_loss": 2.2731966972351074,
      "eval_runtime": 346.3015,
      "eval_samples_per_second": 866.28,
      "eval_steps_per_second": 54.144,
      "step": 1696000
    },
    {
      "epoch": 4.78,
      "eval_loss": 2.272595167160034,
      "eval_runtime": 348.6322,
      "eval_samples_per_second": 860.488,
      "eval_steps_per_second": 53.782,
      "step": 1704000
    },
    {
      "epoch": 4.81,
      "learning_rate": 1.1753333333333334e-07,
      "loss": 2.4302,
      "step": 1712000
    },
    {
      "epoch": 4.81,
      "eval_loss": 2.2703990936279297,
      "eval_runtime": 344.7163,
      "eval_samples_per_second": 870.263,
      "eval_steps_per_second": 54.393,
      "step": 1712000
    },
    {
      "epoch": 4.83,
      "eval_loss": 2.2755091190338135,
      "eval_runtime": 342.7534,
      "eval_samples_per_second": 875.247,
      "eval_steps_per_second": 54.704,
      "step": 1720000
    },
    {
      "epoch": 4.85,
      "learning_rate": 1.1480000000000001e-07,
      "loss": 2.4375,
      "step": 1728000
    },
    {
      "epoch": 4.85,
      "eval_loss": 2.270075559616089,
      "eval_runtime": 342.9559,
      "eval_samples_per_second": 874.73,
      "eval_steps_per_second": 54.672,
      "step": 1728000
    },
    {
      "epoch": 4.87,
      "eval_loss": 2.2719573974609375,
      "eval_runtime": 342.2711,
      "eval_samples_per_second": 876.481,
      "eval_steps_per_second": 54.781,
      "step": 1736000
    },
    {
      "epoch": 4.9,
      "learning_rate": 1.1206666666666666e-07,
      "loss": 2.4305,
      "step": 1744000
    },
    {
      "epoch": 4.9,
      "eval_loss": 2.2697696685791016,
      "eval_runtime": 342.9739,
      "eval_samples_per_second": 874.685,
      "eval_steps_per_second": 54.669,
      "step": 1744000
    },
    {
      "epoch": 4.92,
      "eval_loss": 2.272111415863037,
      "eval_runtime": 344.9176,
      "eval_samples_per_second": 869.755,
      "eval_steps_per_second": 54.361,
      "step": 1752000
    },
    {
      "epoch": 4.94,
      "learning_rate": 1.0933333333333333e-07,
      "loss": 2.4353,
      "step": 1760000
    },
    {
      "epoch": 4.94,
      "eval_loss": 2.2751970291137695,
      "eval_runtime": 344.6056,
      "eval_samples_per_second": 870.543,
      "eval_steps_per_second": 54.41,
      "step": 1760000
    },
    {
      "epoch": 4.96,
      "eval_loss": 2.2763051986694336,
      "eval_runtime": 344.4917,
      "eval_samples_per_second": 870.831,
      "eval_steps_per_second": 54.428,
      "step": 1768000
    },
    {
      "epoch": 4.99,
      "learning_rate": 1.066e-07,
      "loss": 2.4274,
      "step": 1776000
    },
    {
      "epoch": 4.99,
      "eval_loss": 2.2746589183807373,
      "eval_runtime": 345.8714,
      "eval_samples_per_second": 867.357,
      "eval_steps_per_second": 54.211,
      "step": 1776000
    },
    {
      "epoch": 5.01,
      "eval_loss": 2.277564764022827,
      "eval_runtime": 344.7831,
      "eval_samples_per_second": 870.095,
      "eval_steps_per_second": 54.382,
      "step": 1784000
    },
    {
      "epoch": 5.03,
      "learning_rate": 1.0386666666666667e-07,
      "loss": 2.4234,
      "step": 1792000
    },
    {
      "epoch": 5.03,
      "eval_loss": 2.2705652713775635,
      "eval_runtime": 345.9447,
      "eval_samples_per_second": 867.173,
      "eval_steps_per_second": 54.199,
      "step": 1792000
    },
    {
      "epoch": 5.05,
      "eval_loss": 2.2719192504882812,
      "eval_runtime": 345.7061,
      "eval_samples_per_second": 867.772,
      "eval_steps_per_second": 54.237,
      "step": 1800000
    },
    {
      "epoch": 5.08,
      "learning_rate": 1.0113333333333334e-07,
      "loss": 2.4304,
      "step": 1808000
    },
    {
      "epoch": 5.08,
      "eval_loss": 2.2667484283447266,
      "eval_runtime": 344.5475,
      "eval_samples_per_second": 870.69,
      "eval_steps_per_second": 54.419,
      "step": 1808000
    },
    {
      "epoch": 5.1,
      "eval_loss": 2.276196241378784,
      "eval_runtime": 342.6831,
      "eval_samples_per_second": 875.427,
      "eval_steps_per_second": 54.715,
      "step": 1816000
    },
    {
      "epoch": 5.12,
      "learning_rate": 9.84e-08,
      "loss": 2.4308,
      "step": 1824000
    },
    {
      "epoch": 5.12,
      "eval_loss": 2.27565860748291,
      "eval_runtime": 344.3771,
      "eval_samples_per_second": 871.121,
      "eval_steps_per_second": 54.446,
      "step": 1824000
    },
    {
      "epoch": 5.14,
      "eval_loss": 2.27123761177063,
      "eval_runtime": 343.6662,
      "eval_samples_per_second": 872.923,
      "eval_steps_per_second": 54.559,
      "step": 1832000
    },
    {
      "epoch": 5.16,
      "learning_rate": 9.566666666666666e-08,
      "loss": 2.4342,
      "step": 1840000
    },
    {
      "epoch": 5.16,
      "eval_loss": 2.267634868621826,
      "eval_runtime": 343.6527,
      "eval_samples_per_second": 872.957,
      "eval_steps_per_second": 54.561,
      "step": 1840000
    },
    {
      "epoch": 5.19,
      "eval_loss": 2.273836851119995,
      "eval_runtime": 346.2597,
      "eval_samples_per_second": 866.384,
      "eval_steps_per_second": 54.15,
      "step": 1848000
    },
    {
      "epoch": 5.21,
      "learning_rate": 9.293333333333333e-08,
      "loss": 2.4342,
      "step": 1856000
    },
    {
      "epoch": 5.21,
      "eval_loss": 2.2754852771759033,
      "eval_runtime": 343.1416,
      "eval_samples_per_second": 874.257,
      "eval_steps_per_second": 54.642,
      "step": 1856000
    },
    {
      "epoch": 5.23,
      "eval_loss": 2.274082899093628,
      "eval_runtime": 343.2282,
      "eval_samples_per_second": 874.037,
      "eval_steps_per_second": 54.628,
      "step": 1864000
    },
    {
      "epoch": 5.25,
      "learning_rate": 9.02e-08,
      "loss": 2.4329,
      "step": 1872000
    },
    {
      "epoch": 5.25,
      "eval_loss": 2.2734124660491943,
      "eval_runtime": 346.8274,
      "eval_samples_per_second": 864.966,
      "eval_steps_per_second": 54.061,
      "step": 1872000
    },
    {
      "epoch": 5.28,
      "eval_loss": 2.27142596244812,
      "eval_runtime": 344.6072,
      "eval_samples_per_second": 870.539,
      "eval_steps_per_second": 54.41,
      "step": 1880000
    },
    {
      "epoch": 5.3,
      "learning_rate": 8.746666666666667e-08,
      "loss": 2.4306,
      "step": 1888000
    },
    {
      "epoch": 5.3,
      "eval_loss": 2.272188663482666,
      "eval_runtime": 345.5379,
      "eval_samples_per_second": 868.194,
      "eval_steps_per_second": 54.263,
      "step": 1888000
    },
    {
      "epoch": 5.32,
      "eval_loss": 2.270195484161377,
      "eval_runtime": 344.8128,
      "eval_samples_per_second": 870.02,
      "eval_steps_per_second": 54.377,
      "step": 1896000
    },
    {
      "epoch": 5.34,
      "learning_rate": 8.473333333333334e-08,
      "loss": 2.4302,
      "step": 1904000
    },
    {
      "epoch": 5.34,
      "eval_loss": 2.276052713394165,
      "eval_runtime": 343.8201,
      "eval_samples_per_second": 872.532,
      "eval_steps_per_second": 54.534,
      "step": 1904000
    },
    {
      "epoch": 5.37,
      "eval_loss": 2.2747642993927,
      "eval_runtime": 343.5822,
      "eval_samples_per_second": 873.136,
      "eval_steps_per_second": 54.572,
      "step": 1912000
    },
    {
      "epoch": 5.39,
      "learning_rate": 8.2e-08,
      "loss": 2.4303,
      "step": 1920000
    },
    {
      "epoch": 5.39,
      "eval_loss": 2.2763144969940186,
      "eval_runtime": 343.8699,
      "eval_samples_per_second": 872.406,
      "eval_steps_per_second": 54.526,
      "step": 1920000
    },
    {
      "epoch": 5.41,
      "eval_loss": 2.2730941772460938,
      "eval_runtime": 343.83,
      "eval_samples_per_second": 872.507,
      "eval_steps_per_second": 54.533,
      "step": 1928000
    },
    {
      "epoch": 5.43,
      "learning_rate": 7.926666666666666e-08,
      "loss": 2.4234,
      "step": 1936000
    },
    {
      "epoch": 5.43,
      "eval_loss": 2.2676327228546143,
      "eval_runtime": 346.6045,
      "eval_samples_per_second": 865.523,
      "eval_steps_per_second": 54.096,
      "step": 1936000
    },
    {
      "epoch": 5.46,
      "eval_loss": 2.275022268295288,
      "eval_runtime": 343.8317,
      "eval_samples_per_second": 872.502,
      "eval_steps_per_second": 54.532,
      "step": 1944000
    },
    {
      "epoch": 5.48,
      "learning_rate": 7.653333333333333e-08,
      "loss": 2.4349,
      "step": 1952000
    },
    {
      "epoch": 5.48,
      "eval_loss": 2.276860475540161,
      "eval_runtime": 344.8812,
      "eval_samples_per_second": 869.847,
      "eval_steps_per_second": 54.367,
      "step": 1952000
    },
    {
      "epoch": 5.5,
      "eval_loss": 2.2728497982025146,
      "eval_runtime": 345.0042,
      "eval_samples_per_second": 869.537,
      "eval_steps_per_second": 54.347,
      "step": 1960000
    },
    {
      "epoch": 5.52,
      "learning_rate": 7.38e-08,
      "loss": 2.4295,
      "step": 1968000
    },
    {
      "epoch": 5.52,
      "eval_loss": 2.275022506713867,
      "eval_runtime": 344.3706,
      "eval_samples_per_second": 871.137,
      "eval_steps_per_second": 54.447,
      "step": 1968000
    },
    {
      "epoch": 5.55,
      "eval_loss": 2.270230531692505,
      "eval_runtime": 344.3401,
      "eval_samples_per_second": 871.214,
      "eval_steps_per_second": 54.452,
      "step": 1976000
    },
    {
      "epoch": 5.57,
      "learning_rate": 7.106666666666667e-08,
      "loss": 2.428,
      "step": 1984000
    },
    {
      "epoch": 5.57,
      "eval_loss": 2.2729129791259766,
      "eval_runtime": 346.4482,
      "eval_samples_per_second": 865.913,
      "eval_steps_per_second": 54.121,
      "step": 1984000
    },
    {
      "epoch": 5.59,
      "eval_loss": 2.2706665992736816,
      "eval_runtime": 344.2826,
      "eval_samples_per_second": 871.36,
      "eval_steps_per_second": 54.461,
      "step": 1992000
    },
    {
      "epoch": 5.61,
      "learning_rate": 6.833333333333332e-08,
      "loss": 2.4336,
      "step": 2000000
    },
    {
      "epoch": 5.61,
      "eval_loss": 2.277449607849121,
      "eval_runtime": 346.3633,
      "eval_samples_per_second": 866.125,
      "eval_steps_per_second": 54.134,
      "step": 2000000
    },
    {
      "epoch": 5.64,
      "eval_loss": 2.273486375808716,
      "eval_runtime": 345.4177,
      "eval_samples_per_second": 868.496,
      "eval_steps_per_second": 54.282,
      "step": 2008000
    },
    {
      "epoch": 5.66,
      "learning_rate": 6.56e-08,
      "loss": 2.4332,
      "step": 2016000
    },
    {
      "epoch": 5.66,
      "eval_loss": 2.2634286880493164,
      "eval_runtime": 344.8659,
      "eval_samples_per_second": 869.886,
      "eval_steps_per_second": 54.369,
      "step": 2016000
    },
    {
      "epoch": 5.68,
      "eval_loss": 2.2678945064544678,
      "eval_runtime": 345.3228,
      "eval_samples_per_second": 868.735,
      "eval_steps_per_second": 54.297,
      "step": 2024000
    },
    {
      "epoch": 5.7,
      "learning_rate": 6.286666666666666e-08,
      "loss": 2.4342,
      "step": 2032000
    },
    {
      "epoch": 5.7,
      "eval_loss": 2.2753427028656006,
      "eval_runtime": 345.298,
      "eval_samples_per_second": 868.797,
      "eval_steps_per_second": 54.301,
      "step": 2032000
    },
    {
      "epoch": 5.73,
      "eval_loss": 2.271911382675171,
      "eval_runtime": 346.7418,
      "eval_samples_per_second": 865.18,
      "eval_steps_per_second": 54.075,
      "step": 2040000
    },
    {
      "epoch": 5.75,
      "learning_rate": 6.013333333333333e-08,
      "loss": 2.4279,
      "step": 2048000
    },
    {
      "epoch": 5.75,
      "eval_loss": 2.271139621734619,
      "eval_runtime": 345.3244,
      "eval_samples_per_second": 868.731,
      "eval_steps_per_second": 54.297,
      "step": 2048000
    },
    {
      "epoch": 5.77,
      "eval_loss": 2.277822256088257,
      "eval_runtime": 346.0842,
      "eval_samples_per_second": 866.824,
      "eval_steps_per_second": 54.178,
      "step": 2056000
    },
    {
      "epoch": 5.79,
      "learning_rate": 5.7400000000000004e-08,
      "loss": 2.4281,
      "step": 2064000
    },
    {
      "epoch": 5.79,
      "eval_loss": 2.2693228721618652,
      "eval_runtime": 345.3065,
      "eval_samples_per_second": 868.776,
      "eval_steps_per_second": 54.3,
      "step": 2064000
    },
    {
      "epoch": 5.82,
      "eval_loss": 2.271515369415283,
      "eval_runtime": 346.2536,
      "eval_samples_per_second": 866.4,
      "eval_steps_per_second": 54.151,
      "step": 2072000
    },
    {
      "epoch": 5.84,
      "learning_rate": 5.4666666666666666e-08,
      "loss": 2.4246,
      "step": 2080000
    },
    {
      "epoch": 5.84,
      "eval_loss": 2.2674171924591064,
      "eval_runtime": 345.4527,
      "eval_samples_per_second": 868.408,
      "eval_steps_per_second": 54.277,
      "step": 2080000
    },
    {
      "epoch": 5.86,
      "eval_loss": 2.2699599266052246,
      "eval_runtime": 345.7915,
      "eval_samples_per_second": 867.557,
      "eval_steps_per_second": 54.223,
      "step": 2088000
    },
    {
      "epoch": 5.88,
      "learning_rate": 5.1933333333333335e-08,
      "loss": 2.4235,
      "step": 2096000
    },
    {
      "epoch": 5.88,
      "eval_loss": 2.270324230194092,
      "eval_runtime": 347.3105,
      "eval_samples_per_second": 863.763,
      "eval_steps_per_second": 53.986,
      "step": 2096000
    },
    {
      "epoch": 5.91,
      "eval_loss": 2.272321939468384,
      "eval_runtime": 347.0148,
      "eval_samples_per_second": 864.499,
      "eval_steps_per_second": 54.032,
      "step": 2104000
    },
    {
      "epoch": 5.93,
      "learning_rate": 4.92e-08,
      "loss": 2.4388,
      "step": 2112000
    },
    {
      "epoch": 5.93,
      "eval_loss": 2.268273115158081,
      "eval_runtime": 346.0854,
      "eval_samples_per_second": 866.821,
      "eval_steps_per_second": 54.177,
      "step": 2112000
    },
    {
      "epoch": 5.95,
      "eval_loss": 2.2712411880493164,
      "eval_runtime": 346.9622,
      "eval_samples_per_second": 864.63,
      "eval_steps_per_second": 54.04,
      "step": 2120000
    },
    {
      "epoch": 5.97,
      "learning_rate": 4.6466666666666666e-08,
      "loss": 2.431,
      "step": 2128000
    },
    {
      "epoch": 5.97,
      "eval_loss": 2.27392578125,
      "eval_runtime": 346.2653,
      "eval_samples_per_second": 866.37,
      "eval_steps_per_second": 54.149,
      "step": 2128000
    },
    {
      "epoch": 6.0,
      "eval_loss": 2.2757456302642822,
      "eval_runtime": 346.0244,
      "eval_samples_per_second": 866.973,
      "eval_steps_per_second": 54.187,
      "step": 2136000
    },
    {
      "epoch": 6.02,
      "learning_rate": 4.3733333333333335e-08,
      "loss": 2.4329,
      "step": 2144000
    },
    {
      "epoch": 6.02,
      "eval_loss": 2.2785327434539795,
      "eval_runtime": 346.851,
      "eval_samples_per_second": 864.907,
      "eval_steps_per_second": 54.058,
      "step": 2144000
    },
    {
      "epoch": 6.04,
      "eval_loss": 2.2720842361450195,
      "eval_runtime": 346.7603,
      "eval_samples_per_second": 865.134,
      "eval_steps_per_second": 54.072,
      "step": 2152000
    },
    {
      "epoch": 6.06,
      "learning_rate": 4.1e-08,
      "loss": 2.4266,
      "step": 2160000
    },
    {
      "epoch": 6.06,
      "eval_loss": 2.274451971054077,
      "eval_runtime": 346.0729,
      "eval_samples_per_second": 866.852,
      "eval_steps_per_second": 54.179,
      "step": 2160000
    },
    {
      "epoch": 6.09,
      "eval_loss": 2.2738053798675537,
      "eval_runtime": 346.4358,
      "eval_samples_per_second": 865.944,
      "eval_steps_per_second": 54.123,
      "step": 2168000
    },
    {
      "epoch": 6.11,
      "learning_rate": 3.8266666666666665e-08,
      "loss": 2.4255,
      "step": 2176000
    },
    {
      "epoch": 6.11,
      "eval_loss": 2.273524284362793,
      "eval_runtime": 346.5246,
      "eval_samples_per_second": 865.722,
      "eval_steps_per_second": 54.109,
      "step": 2176000
    },
    {
      "epoch": 6.13,
      "eval_loss": 2.2667336463928223,
      "eval_runtime": 347.6074,
      "eval_samples_per_second": 863.025,
      "eval_steps_per_second": 53.94,
      "step": 2184000
    },
    {
      "epoch": 6.15,
      "learning_rate": 3.5533333333333334e-08,
      "loss": 2.4263,
      "step": 2192000
    },
    {
      "epoch": 6.15,
      "eval_loss": 2.2765865325927734,
      "eval_runtime": 347.5449,
      "eval_samples_per_second": 863.181,
      "eval_steps_per_second": 53.95,
      "step": 2192000
    },
    {
      "epoch": 6.18,
      "eval_loss": 2.2754104137420654,
      "eval_runtime": 349.1889,
      "eval_samples_per_second": 859.117,
      "eval_steps_per_second": 53.696,
      "step": 2200000
    },
    {
      "epoch": 6.2,
      "learning_rate": 3.28e-08,
      "loss": 2.4388,
      "step": 2208000
    },
    {
      "epoch": 6.2,
      "eval_loss": 2.269387722015381,
      "eval_runtime": 347.4241,
      "eval_samples_per_second": 863.481,
      "eval_steps_per_second": 53.969,
      "step": 2208000
    },
    {
      "epoch": 6.22,
      "eval_loss": 2.267467737197876,
      "eval_runtime": 348.3026,
      "eval_samples_per_second": 861.303,
      "eval_steps_per_second": 53.832,
      "step": 2216000
    },
    {
      "epoch": 6.24,
      "learning_rate": 3.0066666666666665e-08,
      "loss": 2.4293,
      "step": 2224000
    },
    {
      "epoch": 6.24,
      "eval_loss": 2.26993465423584,
      "eval_runtime": 347.7333,
      "eval_samples_per_second": 862.713,
      "eval_steps_per_second": 53.921,
      "step": 2224000
    },
    {
      "epoch": 6.27,
      "eval_loss": 2.271167039871216,
      "eval_runtime": 348.7054,
      "eval_samples_per_second": 860.308,
      "eval_steps_per_second": 53.77,
      "step": 2232000
    },
    {
      "epoch": 6.29,
      "learning_rate": 2.7333333333333333e-08,
      "loss": 2.428,
      "step": 2240000
    },
    {
      "epoch": 6.29,
      "eval_loss": 2.270735740661621,
      "eval_runtime": 348.9551,
      "eval_samples_per_second": 859.692,
      "eval_steps_per_second": 53.732,
      "step": 2240000
    },
    {
      "epoch": 6.31,
      "eval_loss": 2.273216962814331,
      "eval_runtime": 349.8006,
      "eval_samples_per_second": 857.614,
      "eval_steps_per_second": 53.602,
      "step": 2248000
    },
    {
      "epoch": 6.33,
      "learning_rate": 2.46e-08,
      "loss": 2.4247,
      "step": 2256000
    },
    {
      "epoch": 6.33,
      "eval_loss": 2.275233745574951,
      "eval_runtime": 347.1846,
      "eval_samples_per_second": 864.076,
      "eval_steps_per_second": 54.006,
      "step": 2256000
    },
    {
      "epoch": 6.36,
      "eval_loss": 2.2703025341033936,
      "eval_runtime": 347.2363,
      "eval_samples_per_second": 863.948,
      "eval_steps_per_second": 53.998,
      "step": 2264000
    },
    {
      "epoch": 6.38,
      "learning_rate": 2.1866666666666667e-08,
      "loss": 2.4272,
      "step": 2272000
    },
    {
      "epoch": 6.38,
      "eval_loss": 2.268977403640747,
      "eval_runtime": 347.4129,
      "eval_samples_per_second": 863.509,
      "eval_steps_per_second": 53.97,
      "step": 2272000
    },
    {
      "epoch": 6.4,
      "eval_loss": 2.2775228023529053,
      "eval_runtime": 348.4338,
      "eval_samples_per_second": 860.979,
      "eval_steps_per_second": 53.812,
      "step": 2280000
    },
    {
      "epoch": 6.42,
      "learning_rate": 1.9133333333333333e-08,
      "loss": 2.4297,
      "step": 2288000
    },
    {
      "epoch": 6.42,
      "eval_loss": 2.2680258750915527,
      "eval_runtime": 349.0209,
      "eval_samples_per_second": 859.53,
      "eval_steps_per_second": 53.722,
      "step": 2288000
    },
    {
      "epoch": 6.45,
      "eval_loss": 2.271179676055908,
      "eval_runtime": 349.3722,
      "eval_samples_per_second": 858.666,
      "eval_steps_per_second": 53.668,
      "step": 2296000
    },
    {
      "epoch": 6.47,
      "learning_rate": 1.64e-08,
      "loss": 2.4268,
      "step": 2304000
    },
    {
      "epoch": 6.47,
      "eval_loss": 2.2815394401550293,
      "eval_runtime": 347.9244,
      "eval_samples_per_second": 862.239,
      "eval_steps_per_second": 53.891,
      "step": 2304000
    },
    {
      "epoch": 6.49,
      "eval_loss": 2.269704818725586,
      "eval_runtime": 348.7638,
      "eval_samples_per_second": 860.164,
      "eval_steps_per_second": 53.761,
      "step": 2312000
    },
    {
      "epoch": 6.51,
      "learning_rate": 1.3666666666666667e-08,
      "loss": 2.4248,
      "step": 2320000
    },
    {
      "epoch": 6.51,
      "eval_loss": 2.2793667316436768,
      "eval_runtime": 349.717,
      "eval_samples_per_second": 857.819,
      "eval_steps_per_second": 53.615,
      "step": 2320000
    },
    {
      "epoch": 6.53,
      "eval_loss": 2.2721614837646484,
      "eval_runtime": 349.5657,
      "eval_samples_per_second": 858.191,
      "eval_steps_per_second": 53.638,
      "step": 2328000
    },
    {
      "epoch": 6.56,
      "learning_rate": 1.0933333333333334e-08,
      "loss": 2.4285,
      "step": 2336000
    },
    {
      "epoch": 6.56,
      "eval_loss": 2.2685911655426025,
      "eval_runtime": 348.4666,
      "eval_samples_per_second": 860.897,
      "eval_steps_per_second": 53.807,
      "step": 2336000
    },
    {
      "epoch": 6.58,
      "eval_loss": 2.274138927459717,
      "eval_runtime": 348.1529,
      "eval_samples_per_second": 861.673,
      "eval_steps_per_second": 53.856,
      "step": 2344000
    },
    {
      "epoch": 6.6,
      "learning_rate": 8.2e-09,
      "loss": 2.4318,
      "step": 2352000
    },
    {
      "epoch": 6.6,
      "eval_loss": 2.267868995666504,
      "eval_runtime": 349.9266,
      "eval_samples_per_second": 857.305,
      "eval_steps_per_second": 53.583,
      "step": 2352000
    },
    {
      "epoch": 6.62,
      "eval_loss": 2.272257089614868,
      "eval_runtime": 352.1942,
      "eval_samples_per_second": 851.786,
      "eval_steps_per_second": 53.238,
      "step": 2360000
    },
    {
      "epoch": 6.65,
      "learning_rate": 5.466666666666667e-09,
      "loss": 2.4269,
      "step": 2368000
    },
    {
      "epoch": 6.65,
      "eval_loss": 2.274069309234619,
      "eval_runtime": 348.3964,
      "eval_samples_per_second": 861.071,
      "eval_steps_per_second": 53.818,
      "step": 2368000
    },
    {
      "epoch": 6.67,
      "eval_loss": 2.2739241123199463,
      "eval_runtime": 348.8166,
      "eval_samples_per_second": 860.034,
      "eval_steps_per_second": 53.753,
      "step": 2376000
    },
    {
      "epoch": 6.69,
      "learning_rate": 2.7333333333333334e-09,
      "loss": 2.4275,
      "step": 2384000
    },
    {
      "epoch": 6.69,
      "eval_loss": 2.27441143989563,
      "eval_runtime": 348.6995,
      "eval_samples_per_second": 860.322,
      "eval_steps_per_second": 53.771,
      "step": 2384000
    },
    {
      "epoch": 6.71,
      "eval_loss": 2.2764933109283447,
      "eval_runtime": 350.1503,
      "eval_samples_per_second": 856.758,
      "eval_steps_per_second": 53.548,
      "step": 2392000
    },
    {
      "epoch": 6.74,
      "learning_rate": 0.0,
      "loss": 2.4259,
      "step": 2400000
    },
    {
      "epoch": 6.74,
      "eval_loss": 2.278808116912842,
      "eval_runtime": 349.1911,
      "eval_samples_per_second": 859.111,
      "eval_steps_per_second": 53.696,
      "step": 2400000
    },
    {
      "epoch": 6.74,
      "step": 2400000,
      "total_flos": 7.565202754943813e+17,
      "train_loss": 2.4400800537109375,
      "train_runtime": 255136.6894,
      "train_samples_per_second": 150.508,
      "train_steps_per_second": 9.407
    }
  ],
  "logging_steps": 16000,
  "max_steps": 2400000,
  "num_train_epochs": 7,
  "save_steps": 32000,
  "total_flos": 7.565202754943813e+17,
  "trial_name": null,
  "trial_params": null
}