{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 153.84615384615384,
  "eval_steps": 100,
  "global_step": 2000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0,
      "num_input_tokens_seen": 0,
      "step": 0,
      "train_accuracy": 0.001953125
    },
    {
      "epoch": 0.07692307692307693,
      "grad_norm": 113.0272216796875,
      "learning_rate": 0.0,
      "loss": 6.546019554138184,
      "num_input_tokens_seen": 4096,
      "step": 1,
      "train_runtime": 3.5976,
      "train_tokens_per_second": 1138.522
    },
    {
      "epoch": 0.07692307692307693,
      "num_input_tokens_seen": 4096,
      "step": 1,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 0.15384615384615385,
      "grad_norm": 113.77076721191406,
      "learning_rate": 0.001,
      "loss": 6.613107204437256,
      "num_input_tokens_seen": 8192,
      "step": 2,
      "train_runtime": 4.3771,
      "train_tokens_per_second": 1871.568
    },
    {
      "epoch": 0.15384615384615385,
      "num_input_tokens_seen": 8192,
      "step": 2,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 0.23076923076923078,
      "grad_norm": 116.1823959350586,
      "learning_rate": 0.002,
      "loss": 6.340221405029297,
      "num_input_tokens_seen": 12288,
      "step": 3,
      "train_runtime": 5.168,
      "train_tokens_per_second": 2377.707
    },
    {
      "epoch": 0.23076923076923078,
      "num_input_tokens_seen": 12288,
      "step": 3,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 0.3076923076923077,
      "grad_norm": 27.13140106201172,
      "learning_rate": 0.003,
      "loss": 8.609485626220703,
      "num_input_tokens_seen": 16384,
      "step": 4,
      "train_runtime": 5.9553,
      "train_tokens_per_second": 2751.168
    },
    {
      "epoch": 0.3076923076923077,
      "num_input_tokens_seen": 16384,
      "step": 4,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 0.38461538461538464,
      "grad_norm": 24.710792541503906,
      "learning_rate": 0.004,
      "loss": 9.139598846435547,
      "num_input_tokens_seen": 20480,
      "step": 5,
      "train_runtime": 6.7415,
      "train_tokens_per_second": 3037.921
    },
    {
      "epoch": 0.38461538461538464,
      "num_input_tokens_seen": 20480,
      "step": 5,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 0.46153846153846156,
      "grad_norm": 35.96022033691406,
      "learning_rate": 0.005,
      "loss": 8.047222137451172,
      "num_input_tokens_seen": 24576,
      "step": 6,
      "train_runtime": 7.5328,
      "train_tokens_per_second": 3262.536
    },
    {
      "epoch": 0.46153846153846156,
      "num_input_tokens_seen": 24576,
      "step": 6,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 0.5384615384615384,
      "grad_norm": 35.30043029785156,
      "learning_rate": 0.006,
      "loss": 7.312543869018555,
      "num_input_tokens_seen": 28672,
      "step": 7,
      "train_runtime": 8.3242,
      "train_tokens_per_second": 3444.404
    },
    {
      "epoch": 0.5384615384615384,
      "num_input_tokens_seen": 28672,
      "step": 7,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 0.6153846153846154,
      "grad_norm": 59.60608673095703,
      "learning_rate": 0.006999999999999999,
      "loss": 8.326001167297363,
      "num_input_tokens_seen": 32768,
      "step": 8,
      "train_runtime": 9.1124,
      "train_tokens_per_second": 3595.981
    },
    {
      "epoch": 0.6153846153846154,
      "num_input_tokens_seen": 32768,
      "step": 8,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 0.6923076923076923,
      "grad_norm": 50.62687683105469,
      "learning_rate": 0.008,
      "loss": 11.003422737121582,
      "num_input_tokens_seen": 36864,
      "step": 9,
      "train_runtime": 9.9012,
      "train_tokens_per_second": 3723.169
    },
    {
      "epoch": 0.6923076923076923,
      "num_input_tokens_seen": 36864,
      "step": 9,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 0.7692307692307693,
      "grad_norm": 89.37097930908203,
      "learning_rate": 0.009000000000000001,
      "loss": 13.825056076049805,
      "num_input_tokens_seen": 40960,
      "step": 10,
      "train_runtime": 10.6904,
      "train_tokens_per_second": 3831.459
    },
    {
      "epoch": 0.7692307692307693,
      "num_input_tokens_seen": 40960,
      "step": 10,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 0.8461538461538461,
      "grad_norm": 90.11058807373047,
      "learning_rate": 0.01,
      "loss": 20.330724716186523,
      "num_input_tokens_seen": 45056,
      "step": 11,
      "train_runtime": 11.4806,
      "train_tokens_per_second": 3924.525
    },
    {
      "epoch": 0.8461538461538461,
      "num_input_tokens_seen": 45056,
      "step": 11,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 0.9230769230769231,
      "grad_norm": 98.63053894042969,
      "learning_rate": 0.01,
      "loss": 32.64373016357422,
      "num_input_tokens_seen": 49152,
      "step": 12,
      "train_runtime": 12.2648,
      "train_tokens_per_second": 4007.559
    },
    {
      "epoch": 0.9230769230769231,
      "num_input_tokens_seen": 49152,
      "step": 12,
      "train_accuracy": 0.019543973729014397
    },
    {
      "epoch": 1.0,
      "grad_norm": 59.70159149169922,
      "learning_rate": 0.01,
      "loss": 27.997303009033203,
      "num_input_tokens_seen": 51608,
      "step": 13,
      "train_runtime": 12.7652,
      "train_tokens_per_second": 4042.877
    },
    {
      "epoch": 1.0,
      "num_input_tokens_seen": 51608,
      "step": 13,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 1.0769230769230769,
      "grad_norm": 59.93867111206055,
      "learning_rate": 0.01,
      "loss": 26.742982864379883,
      "num_input_tokens_seen": 55704,
      "step": 14,
      "train_runtime": 13.5676,
      "train_tokens_per_second": 4105.655
    },
    {
      "epoch": 1.0769230769230769,
      "num_input_tokens_seen": 55704,
      "step": 14,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 1.1538461538461537,
      "grad_norm": 101.93472290039062,
      "learning_rate": 0.01,
      "loss": 30.33885955810547,
      "num_input_tokens_seen": 59800,
      "step": 15,
      "train_runtime": 14.3572,
      "train_tokens_per_second": 4165.166
    },
    {
      "epoch": 1.1538461538461537,
      "num_input_tokens_seen": 59800,
      "step": 15,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 1.2307692307692308,
      "grad_norm": 69.6849594116211,
      "learning_rate": 0.01,
      "loss": 33.29315948486328,
      "num_input_tokens_seen": 63896,
      "step": 16,
      "train_runtime": 15.153,
      "train_tokens_per_second": 4216.713
    },
    {
      "epoch": 1.2307692307692308,
      "num_input_tokens_seen": 63896,
      "step": 16,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 1.3076923076923077,
      "grad_norm": 104.34686279296875,
      "learning_rate": 0.01,
      "loss": 43.05381774902344,
      "num_input_tokens_seen": 67992,
      "step": 17,
      "train_runtime": 15.9433,
      "train_tokens_per_second": 4264.624
    },
    {
      "epoch": 1.3076923076923077,
      "num_input_tokens_seen": 67992,
      "step": 17,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 1.3846153846153846,
      "grad_norm": 102.8227767944336,
      "learning_rate": 0.01,
      "loss": 44.411582946777344,
      "num_input_tokens_seen": 72088,
      "step": 18,
      "train_runtime": 16.738,
      "train_tokens_per_second": 4306.855
    },
    {
      "epoch": 1.3846153846153846,
      "num_input_tokens_seen": 72088,
      "step": 18,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 1.4615384615384617,
      "grad_norm": 71.9914779663086,
      "learning_rate": 0.01,
      "loss": 41.01606750488281,
      "num_input_tokens_seen": 76184,
      "step": 19,
      "train_runtime": 17.5345,
      "train_tokens_per_second": 4344.796
    },
    {
      "epoch": 1.4615384615384617,
      "num_input_tokens_seen": 76184,
      "step": 19,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 1.5384615384615383,
      "grad_norm": 64.82003784179688,
      "learning_rate": 0.01,
      "loss": 39.845672607421875,
      "num_input_tokens_seen": 80280,
      "step": 20,
      "train_runtime": 18.3368,
      "train_tokens_per_second": 4378.077
    },
    {
      "epoch": 1.5384615384615383,
      "num_input_tokens_seen": 80280,
      "step": 20,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 1.6153846153846154,
      "grad_norm": 89.51396179199219,
      "learning_rate": 0.01,
      "loss": 43.63105010986328,
      "num_input_tokens_seen": 84376,
      "step": 21,
      "train_runtime": 19.132,
      "train_tokens_per_second": 4410.199
    },
    {
      "epoch": 1.6153846153846154,
      "num_input_tokens_seen": 84376,
      "step": 21,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 1.6923076923076923,
      "grad_norm": 84.98103332519531,
      "learning_rate": 0.01,
      "loss": 49.0236701965332,
      "num_input_tokens_seen": 88472,
      "step": 22,
      "train_runtime": 19.9279,
      "train_tokens_per_second": 4439.601
    },
    {
      "epoch": 1.6923076923076923,
      "num_input_tokens_seen": 88472,
      "step": 22,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 1.7692307692307692,
      "grad_norm": 55.64690017700195,
      "learning_rate": 0.01,
      "loss": 44.23743438720703,
      "num_input_tokens_seen": 92568,
      "step": 23,
      "train_runtime": 20.7215,
      "train_tokens_per_second": 4467.243
    },
    {
      "epoch": 1.7692307692307692,
      "num_input_tokens_seen": 92568,
      "step": 23,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 1.8461538461538463,
      "grad_norm": 41.61865234375,
      "learning_rate": 0.01,
      "loss": 41.33677673339844,
      "num_input_tokens_seen": 96664,
      "step": 24,
      "train_runtime": 21.5143,
      "train_tokens_per_second": 4493.016
    },
    {
      "epoch": 1.8461538461538463,
      "num_input_tokens_seen": 96664,
      "step": 24,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 1.9230769230769231,
      "grad_norm": 83.00218200683594,
      "learning_rate": 0.01,
      "loss": 35.780399322509766,
      "num_input_tokens_seen": 100760,
      "step": 25,
      "train_runtime": 22.3026,
      "train_tokens_per_second": 4517.851
    },
    {
      "epoch": 1.9230769230769231,
      "num_input_tokens_seen": 100760,
      "step": 25,
      "train_accuracy": 0.016286645084619522
    },
    {
      "epoch": 2.0,
      "grad_norm": 103.5835952758789,
      "learning_rate": 0.01,
      "loss": 60.61474609375,
      "num_input_tokens_seen": 103216,
      "step": 26,
      "train_runtime": 22.8053,
      "train_tokens_per_second": 4525.968
    },
    {
      "epoch": 2.0,
      "num_input_tokens_seen": 103216,
      "step": 26,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 2.076923076923077,
      "grad_norm": 89.98833465576172,
      "learning_rate": 0.01,
      "loss": 53.00630187988281,
      "num_input_tokens_seen": 107312,
      "step": 27,
      "train_runtime": 23.6097,
      "train_tokens_per_second": 4545.247
    },
    {
      "epoch": 2.076923076923077,
      "num_input_tokens_seen": 107312,
      "step": 27,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 2.1538461538461537,
      "grad_norm": 65.72506713867188,
      "learning_rate": 0.01,
      "loss": 58.71630859375,
      "num_input_tokens_seen": 111408,
      "step": 28,
      "train_runtime": 24.402,
      "train_tokens_per_second": 4565.527
    },
    {
      "epoch": 2.1538461538461537,
      "num_input_tokens_seen": 111408,
      "step": 28,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 2.230769230769231,
      "grad_norm": 100.10186004638672,
      "learning_rate": 0.01,
      "loss": 62.033809661865234,
      "num_input_tokens_seen": 115504,
      "step": 29,
      "train_runtime": 25.194,
      "train_tokens_per_second": 4584.588
    },
    {
      "epoch": 2.230769230769231,
      "num_input_tokens_seen": 115504,
      "step": 29,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 2.3076923076923075,
      "grad_norm": 72.38996124267578,
      "learning_rate": 0.01,
      "loss": 65.41529846191406,
      "num_input_tokens_seen": 119600,
      "step": 30,
      "train_runtime": 25.9882,
      "train_tokens_per_second": 4602.084
    },
    {
      "epoch": 2.3076923076923075,
      "num_input_tokens_seen": 119600,
      "step": 30,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 2.3846153846153846,
      "grad_norm": 94.2196273803711,
      "learning_rate": 0.01,
      "loss": 68.32504272460938,
      "num_input_tokens_seen": 123696,
      "step": 31,
      "train_runtime": 26.7814,
      "train_tokens_per_second": 4618.721
    },
    {
      "epoch": 2.3846153846153846,
      "num_input_tokens_seen": 123696,
      "step": 31,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 2.4615384615384617,
      "grad_norm": 96.7938232421875,
      "learning_rate": 0.01,
      "loss": 66.63648223876953,
      "num_input_tokens_seen": 127792,
      "step": 32,
      "train_runtime": 27.575,
      "train_tokens_per_second": 4634.345
    },
    {
      "epoch": 2.4615384615384617,
      "num_input_tokens_seen": 127792,
      "step": 32,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 2.5384615384615383,
      "grad_norm": 72.12036895751953,
      "learning_rate": 0.01,
      "loss": 66.71937561035156,
      "num_input_tokens_seen": 131888,
      "step": 33,
      "train_runtime": 28.3696,
      "train_tokens_per_second": 4648.921
    },
    {
      "epoch": 2.5384615384615383,
      "num_input_tokens_seen": 131888,
      "step": 33,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 2.6153846153846154,
      "grad_norm": 75.74029541015625,
      "learning_rate": 0.01,
      "loss": 67.44764709472656,
      "num_input_tokens_seen": 135984,
      "step": 34,
      "train_runtime": 29.1643,
      "train_tokens_per_second": 4662.687
    },
    {
      "epoch": 2.6153846153846154,
      "num_input_tokens_seen": 135984,
      "step": 34,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 2.6923076923076925,
      "grad_norm": 66.4405517578125,
      "learning_rate": 0.01,
      "loss": 59.08979034423828,
      "num_input_tokens_seen": 140080,
      "step": 35,
      "train_runtime": 29.958,
      "train_tokens_per_second": 4675.875
    },
    {
      "epoch": 2.6923076923076925,
      "num_input_tokens_seen": 140080,
      "step": 35,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 2.769230769230769,
      "grad_norm": 66.96763610839844,
      "learning_rate": 0.01,
      "loss": 47.858978271484375,
      "num_input_tokens_seen": 144176,
      "step": 36,
      "train_runtime": 30.7536,
      "train_tokens_per_second": 4688.099
    },
    {
      "epoch": 2.769230769230769,
      "num_input_tokens_seen": 144176,
      "step": 36,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 2.8461538461538463,
      "grad_norm": 74.87189483642578,
      "learning_rate": 0.01,
      "loss": 49.067626953125,
      "num_input_tokens_seen": 148272,
      "step": 37,
      "train_runtime": 31.5492,
      "train_tokens_per_second": 4699.708
    },
    {
      "epoch": 2.8461538461538463,
      "num_input_tokens_seen": 148272,
      "step": 37,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 2.9230769230769234,
      "grad_norm": 96.19770050048828,
      "learning_rate": 0.01,
      "loss": 50.87727355957031,
      "num_input_tokens_seen": 152368,
      "step": 38,
      "train_runtime": 32.3421,
      "train_tokens_per_second": 4711.129
    },
    {
      "epoch": 2.9230769230769234,
      "num_input_tokens_seen": 152368,
      "step": 38,
      "train_accuracy": 0.0032573288772255182
    },
    {
      "epoch": 3.0,
      "grad_norm": 99.37137603759766,
      "learning_rate": 0.01,
      "loss": 60.30193328857422,
      "num_input_tokens_seen": 154824,
      "step": 39,
      "train_runtime": 32.8477,
      "train_tokens_per_second": 4713.394
    },
    {
      "epoch": 3.0,
      "num_input_tokens_seen": 154824,
      "step": 39,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 3.076923076923077,
      "grad_norm": 62.05854415893555,
      "learning_rate": 0.01,
      "loss": 67.73375701904297,
      "num_input_tokens_seen": 158920,
      "step": 40,
      "train_runtime": 33.6537,
      "train_tokens_per_second": 4722.218
    },
    {
      "epoch": 3.076923076923077,
      "num_input_tokens_seen": 158920,
      "step": 40,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 3.1538461538461537,
      "grad_norm": 89.07279205322266,
      "learning_rate": 0.01,
      "loss": 68.0074462890625,
      "num_input_tokens_seen": 163016,
      "step": 41,
      "train_runtime": 34.4465,
      "train_tokens_per_second": 4732.441
    },
    {
      "epoch": 3.1538461538461537,
      "num_input_tokens_seen": 163016,
      "step": 41,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 3.230769230769231,
      "grad_norm": 90.51819610595703,
      "learning_rate": 0.01,
      "loss": 66.1943588256836,
      "num_input_tokens_seen": 167112,
      "step": 42,
      "train_runtime": 35.2426,
      "train_tokens_per_second": 4741.766
    },
    {
      "epoch": 3.230769230769231,
      "num_input_tokens_seen": 167112,
      "step": 42,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 3.3076923076923075,
      "grad_norm": 83.96027374267578,
      "learning_rate": 0.01,
      "loss": 60.99195098876953,
      "num_input_tokens_seen": 171208,
      "step": 43,
      "train_runtime": 36.0415,
      "train_tokens_per_second": 4750.296
    },
    {
      "epoch": 3.3076923076923075,
      "num_input_tokens_seen": 171208,
      "step": 43,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 3.3846153846153846,
      "grad_norm": 78.46168518066406,
      "learning_rate": 0.01,
      "loss": 59.39600372314453,
      "num_input_tokens_seen": 175304,
      "step": 44,
      "train_runtime": 36.8398,
      "train_tokens_per_second": 4758.543
    },
    {
      "epoch": 3.3846153846153846,
      "num_input_tokens_seen": 175304,
      "step": 44,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 3.4615384615384617,
      "grad_norm": 85.69463348388672,
      "learning_rate": 0.01,
      "loss": 70.74111938476562,
      "num_input_tokens_seen": 179400,
      "step": 45,
      "train_runtime": 37.6417,
      "train_tokens_per_second": 4765.991
    },
    {
      "epoch": 3.4615384615384617,
      "num_input_tokens_seen": 179400,
      "step": 45,
      "train_accuracy": 0.001953125
    },
    {
      "epoch": 3.5384615384615383,
      "grad_norm": 96.39139556884766,
      "learning_rate": 0.01,
      "loss": 88.665771484375,
      "num_input_tokens_seen": 183496,
      "step": 46,
      "train_runtime": 38.4405,
      "train_tokens_per_second": 4773.507
    },
    {
      "epoch": 3.5384615384615383,
      "num_input_tokens_seen": 183496,
      "step": 46,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 3.6153846153846154,
      "grad_norm": 57.92229080200195,
      "learning_rate": 0.01,
      "loss": 84.55277252197266,
      "num_input_tokens_seen": 187592,
      "step": 47,
      "train_runtime": 39.2363,
      "train_tokens_per_second": 4781.083
    },
    {
      "epoch": 3.6153846153846154,
      "num_input_tokens_seen": 187592,
      "step": 47,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 3.6923076923076925,
      "grad_norm": 77.67723846435547,
      "learning_rate": 0.01,
      "loss": 76.61539459228516,
      "num_input_tokens_seen": 191688,
      "step": 48,
      "train_runtime": 40.0342,
      "train_tokens_per_second": 4788.109
    },
    {
      "epoch": 3.6923076923076925,
      "num_input_tokens_seen": 191688,
      "step": 48,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 3.769230769230769,
      "grad_norm": 94.53797149658203,
      "learning_rate": 0.01,
      "loss": 90.10026550292969,
      "num_input_tokens_seen": 195784,
      "step": 49,
      "train_runtime": 40.8358,
      "train_tokens_per_second": 4794.421
    },
    {
      "epoch": 3.769230769230769,
      "num_input_tokens_seen": 195784,
      "step": 49,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 3.8461538461538463,
      "grad_norm": 88.79261779785156,
      "learning_rate": 0.01,
      "loss": 93.755126953125,
      "num_input_tokens_seen": 199880,
      "step": 50,
      "train_runtime": 41.6387,
      "train_tokens_per_second": 4800.338
    },
    {
      "epoch": 3.8461538461538463,
      "num_input_tokens_seen": 199880,
      "step": 50,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 3.9230769230769234,
      "grad_norm": 63.9875602722168,
      "learning_rate": 0.01,
      "loss": 87.74848175048828,
      "num_input_tokens_seen": 203976,
      "step": 51,
      "train_runtime": 42.4304,
      "train_tokens_per_second": 4807.307
    },
    {
      "epoch": 3.9230769230769234,
      "num_input_tokens_seen": 203976,
      "step": 51,
      "train_accuracy": 0.013029315508902073
    },
    {
      "epoch": 4.0,
      "grad_norm": 55.67214584350586,
      "learning_rate": 0.01,
      "loss": 92.73106384277344,
      "num_input_tokens_seen": 206432,
      "step": 52,
      "train_runtime": 42.9329,
      "train_tokens_per_second": 4808.249
    },
    {
      "epoch": 4.0,
      "num_input_tokens_seen": 206432,
      "step": 52,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 4.076923076923077,
      "grad_norm": 91.23143768310547,
      "learning_rate": 0.01,
      "loss": 100.60653686523438,
      "num_input_tokens_seen": 210528,
      "step": 53,
      "train_runtime": 43.7405,
      "train_tokens_per_second": 4813.117
    },
    {
      "epoch": 4.076923076923077,
      "num_input_tokens_seen": 210528,
      "step": 53,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 4.153846153846154,
      "grad_norm": 60.825382232666016,
      "learning_rate": 0.01,
      "loss": 92.57173156738281,
      "num_input_tokens_seen": 214624,
      "step": 54,
      "train_runtime": 44.5422,
      "train_tokens_per_second": 4818.44
    },
    {
      "epoch": 4.153846153846154,
      "num_input_tokens_seen": 214624,
      "step": 54,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 4.230769230769231,
      "grad_norm": 91.31449127197266,
      "learning_rate": 0.01,
      "loss": 99.74507904052734,
      "num_input_tokens_seen": 218720,
      "step": 55,
      "train_runtime": 45.3388,
      "train_tokens_per_second": 4824.119
    },
    {
      "epoch": 4.230769230769231,
      "num_input_tokens_seen": 218720,
      "step": 55,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 4.3076923076923075,
      "grad_norm": 65.0865707397461,
      "learning_rate": 0.01,
      "loss": 97.59388732910156,
      "num_input_tokens_seen": 222816,
      "step": 56,
      "train_runtime": 46.1387,
      "train_tokens_per_second": 4829.261
    },
    {
      "epoch": 4.3076923076923075,
      "num_input_tokens_seen": 222816,
      "step": 56,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 4.384615384615385,
      "grad_norm": 56.028629302978516,
      "learning_rate": 0.01,
      "loss": 94.15910339355469,
      "num_input_tokens_seen": 226912,
      "step": 57,
      "train_runtime": 46.9361,
      "train_tokens_per_second": 4834.485
    },
    {
      "epoch": 4.384615384615385,
      "num_input_tokens_seen": 226912,
      "step": 57,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 4.461538461538462,
      "grad_norm": 90.6352767944336,
      "learning_rate": 0.01,
      "loss": 79.20365142822266,
      "num_input_tokens_seen": 231008,
      "step": 58,
      "train_runtime": 47.7351,
      "train_tokens_per_second": 4839.373
    },
    {
      "epoch": 4.461538461538462,
      "num_input_tokens_seen": 231008,
      "step": 58,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 4.538461538461538,
      "grad_norm": 73.205810546875,
      "learning_rate": 0.01,
      "loss": 82.34382629394531,
      "num_input_tokens_seen": 235104,
      "step": 59,
      "train_runtime": 48.5416,
      "train_tokens_per_second": 4843.35
    },
    {
      "epoch": 4.538461538461538,
      "num_input_tokens_seen": 235104,
      "step": 59,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 4.615384615384615,
      "grad_norm": 89.18135070800781,
      "learning_rate": 0.01,
      "loss": 81.22946166992188,
      "num_input_tokens_seen": 239200,
      "step": 60,
      "train_runtime": 49.34,
      "train_tokens_per_second": 4847.996
    },
    {
      "epoch": 4.615384615384615,
      "num_input_tokens_seen": 239200,
      "step": 60,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 4.6923076923076925,
      "grad_norm": 82.80770111083984,
      "learning_rate": 0.01,
      "loss": 89.39828491210938,
      "num_input_tokens_seen": 243296,
      "step": 61,
      "train_runtime": 50.1404,
      "train_tokens_per_second": 4852.295
    },
    {
      "epoch": 4.6923076923076925,
      "num_input_tokens_seen": 243296,
      "step": 61,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 4.769230769230769,
      "grad_norm": 80.53831481933594,
      "learning_rate": 0.01,
      "loss": 93.91964721679688,
      "num_input_tokens_seen": 247392,
      "step": 62,
      "train_runtime": 50.941,
      "train_tokens_per_second": 4856.442
    },
    {
      "epoch": 4.769230769230769,
      "num_input_tokens_seen": 247392,
      "step": 62,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 4.846153846153846,
      "grad_norm": 87.39923858642578,
      "learning_rate": 0.01,
      "loss": 95.6671371459961,
      "num_input_tokens_seen": 251488,
      "step": 63,
      "train_runtime": 51.7384,
      "train_tokens_per_second": 4860.759
    },
    {
      "epoch": 4.846153846153846,
      "num_input_tokens_seen": 251488,
      "step": 63,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 4.923076923076923,
      "grad_norm": 80.83548736572266,
      "learning_rate": 0.01,
      "loss": 70.75057983398438,
      "num_input_tokens_seen": 255584,
      "step": 64,
      "train_runtime": 52.5359,
      "train_tokens_per_second": 4864.94
    },
    {
      "epoch": 4.923076923076923,
      "num_input_tokens_seen": 255584,
      "step": 64,
      "train_accuracy": 0.016286645084619522
    },
    {
      "epoch": 5.0,
      "grad_norm": 73.11419677734375,
      "learning_rate": 0.01,
      "loss": 68.34355926513672,
      "num_input_tokens_seen": 258040,
      "step": 65,
      "train_runtime": 53.0416,
      "train_tokens_per_second": 4864.857
    },
    {
      "epoch": 5.0,
      "num_input_tokens_seen": 258040,
      "step": 65,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 5.076923076923077,
      "grad_norm": 87.57298278808594,
      "learning_rate": 0.01,
      "loss": 68.34878540039062,
      "num_input_tokens_seen": 262136,
      "step": 66,
      "train_runtime": 53.8523,
      "train_tokens_per_second": 4867.689
    },
    {
      "epoch": 5.076923076923077,
      "num_input_tokens_seen": 262136,
      "step": 66,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 5.153846153846154,
      "grad_norm": 87.19989013671875,
      "learning_rate": 0.01,
      "loss": 71.25694274902344,
      "num_input_tokens_seen": 266232,
      "step": 67,
      "train_runtime": 54.6483,
      "train_tokens_per_second": 4871.735
    },
    {
      "epoch": 5.153846153846154,
      "num_input_tokens_seen": 266232,
      "step": 67,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 5.230769230769231,
      "grad_norm": 88.10555267333984,
      "learning_rate": 0.01,
      "loss": 75.13947296142578,
      "num_input_tokens_seen": 270328,
      "step": 68,
      "train_runtime": 55.4461,
      "train_tokens_per_second": 4875.512
    },
    {
      "epoch": 5.230769230769231,
      "num_input_tokens_seen": 270328,
      "step": 68,
      "train_accuracy": 0.0
    },
    {
      "epoch": 5.3076923076923075,
      "grad_norm": 86.66246795654297,
      "learning_rate": 0.01,
      "loss": 73.39552307128906,
      "num_input_tokens_seen": 274424,
      "step": 69,
      "train_runtime": 56.2446,
      "train_tokens_per_second": 4879.117
    },
    {
      "epoch": 5.3076923076923075,
      "num_input_tokens_seen": 274424,
      "step": 69,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 5.384615384615385,
      "grad_norm": 61.33372116088867,
      "learning_rate": 0.01,
      "loss": 74.055419921875,
      "num_input_tokens_seen": 278520,
      "step": 70,
      "train_runtime": 57.0431,
      "train_tokens_per_second": 4882.626
    },
    {
      "epoch": 5.384615384615385,
      "num_input_tokens_seen": 278520,
      "step": 70,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 5.461538461538462,
      "grad_norm": 75.86312866210938,
      "learning_rate": 0.01,
      "loss": 74.29896545410156,
      "num_input_tokens_seen": 282616,
      "step": 71,
      "train_runtime": 57.8433,
      "train_tokens_per_second": 4885.888
    },
    {
      "epoch": 5.461538461538462,
      "num_input_tokens_seen": 282616,
      "step": 71,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 5.538461538461538,
      "grad_norm": 53.104637145996094,
      "learning_rate": 0.01,
      "loss": 61.737548828125,
      "num_input_tokens_seen": 286712,
      "step": 72,
      "train_runtime": 58.6464,
      "train_tokens_per_second": 4888.829
    },
    {
      "epoch": 5.538461538461538,
      "num_input_tokens_seen": 286712,
      "step": 72,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 5.615384615384615,
      "grad_norm": 81.82807922363281,
      "learning_rate": 0.01,
      "loss": 61.38819885253906,
      "num_input_tokens_seen": 290808,
      "step": 73,
      "train_runtime": 59.4482,
      "train_tokens_per_second": 4891.788
    },
    {
      "epoch": 5.615384615384615,
      "num_input_tokens_seen": 290808,
      "step": 73,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 5.6923076923076925,
      "grad_norm": 84.85137176513672,
      "learning_rate": 0.01,
      "loss": 70.53759002685547,
      "num_input_tokens_seen": 294904,
      "step": 74,
      "train_runtime": 60.2482,
      "train_tokens_per_second": 4894.816
    },
    {
      "epoch": 5.6923076923076925,
      "num_input_tokens_seen": 294904,
      "step": 74,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 5.769230769230769,
      "grad_norm": 68.4859390258789,
      "learning_rate": 0.01,
      "loss": 67.85568237304688,
      "num_input_tokens_seen": 299000,
      "step": 75,
      "train_runtime": 61.0482,
      "train_tokens_per_second": 4897.773
    },
    {
      "epoch": 5.769230769230769,
      "num_input_tokens_seen": 299000,
      "step": 75,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 5.846153846153846,
      "grad_norm": 83.89737701416016,
      "learning_rate": 0.01,
      "loss": 63.241065979003906,
      "num_input_tokens_seen": 303096,
      "step": 76,
      "train_runtime": 61.8504,
      "train_tokens_per_second": 4900.466
    },
    {
      "epoch": 5.846153846153846,
      "num_input_tokens_seen": 303096,
      "step": 76,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 5.923076923076923,
      "grad_norm": 84.33633422851562,
      "learning_rate": 0.01,
      "loss": 65.82176208496094,
      "num_input_tokens_seen": 307192,
      "step": 77,
      "train_runtime": 62.6481,
      "train_tokens_per_second": 4903.451
    },
    {
      "epoch": 5.923076923076923,
      "num_input_tokens_seen": 307192,
      "step": 77,
      "train_accuracy": 0.013029315508902073
    },
    {
      "epoch": 6.0,
      "grad_norm": 58.69049072265625,
      "learning_rate": 0.01,
      "loss": 58.87933349609375,
      "num_input_tokens_seen": 309648,
      "step": 78,
      "train_runtime": 63.1559,
      "train_tokens_per_second": 4902.914
    },
    {
      "epoch": 6.0,
      "num_input_tokens_seen": 309648,
      "step": 78,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 6.076923076923077,
      "grad_norm": 58.454830169677734,
      "learning_rate": 0.01,
      "loss": 60.6256217956543,
      "num_input_tokens_seen": 313744,
      "step": 79,
      "train_runtime": 63.9682,
      "train_tokens_per_second": 4904.684
    },
    {
      "epoch": 6.076923076923077,
      "num_input_tokens_seen": 313744,
      "step": 79,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 6.153846153846154,
      "grad_norm": 57.2540168762207,
      "learning_rate": 0.01,
      "loss": 58.05278396606445,
      "num_input_tokens_seen": 317840,
      "step": 80,
      "train_runtime": 64.7688,
      "train_tokens_per_second": 4907.303
    },
    {
      "epoch": 6.153846153846154,
      "num_input_tokens_seen": 317840,
      "step": 80,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 6.230769230769231,
      "grad_norm": 50.83211898803711,
      "learning_rate": 0.01,
      "loss": 51.120361328125,
      "num_input_tokens_seen": 321936,
      "step": 81,
      "train_runtime": 65.5698,
      "train_tokens_per_second": 4909.818
    },
    {
      "epoch": 6.230769230769231,
      "num_input_tokens_seen": 321936,
      "step": 81,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 6.3076923076923075,
      "grad_norm": 60.46956253051758,
      "learning_rate": 0.01,
      "loss": 52.14616394042969,
      "num_input_tokens_seen": 326032,
      "step": 82,
      "train_runtime": 66.371,
      "train_tokens_per_second": 4912.266
    },
    {
      "epoch": 6.3076923076923075,
      "num_input_tokens_seen": 326032,
      "step": 82,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 6.384615384615385,
      "grad_norm": 76.81056213378906,
      "learning_rate": 0.01,
      "loss": 54.78953170776367,
      "num_input_tokens_seen": 330128,
      "step": 83,
      "train_runtime": 67.1735,
      "train_tokens_per_second": 4914.559
    },
    {
      "epoch": 6.384615384615385,
      "num_input_tokens_seen": 330128,
      "step": 83,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 6.461538461538462,
      "grad_norm": 82.9794692993164,
      "learning_rate": 0.01,
      "loss": 65.67823028564453,
      "num_input_tokens_seen": 334224,
      "step": 84,
      "train_runtime": 67.9749,
      "train_tokens_per_second": 4916.872
    },
    {
      "epoch": 6.461538461538462,
      "num_input_tokens_seen": 334224,
      "step": 84,
      "train_accuracy": 0.0
    },
    {
      "epoch": 6.538461538461538,
      "grad_norm": 83.5008773803711,
      "learning_rate": 0.01,
      "loss": 68.018798828125,
      "num_input_tokens_seen": 338320,
      "step": 85,
      "train_runtime": 68.7761,
      "train_tokens_per_second": 4919.147
    },
    {
      "epoch": 6.538461538461538,
      "num_input_tokens_seen": 338320,
      "step": 85,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 6.615384615384615,
      "grad_norm": 69.92463684082031,
      "learning_rate": 0.01,
      "loss": 55.379913330078125,
      "num_input_tokens_seen": 342416,
      "step": 86,
      "train_runtime": 69.5783,
      "train_tokens_per_second": 4921.307
    },
    {
      "epoch": 6.615384615384615,
      "num_input_tokens_seen": 342416,
      "step": 86,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 6.6923076923076925,
      "grad_norm": 73.4237060546875,
      "learning_rate": 0.01,
      "loss": 67.15409851074219,
      "num_input_tokens_seen": 346512,
      "step": 87,
      "train_runtime": 70.3801,
      "train_tokens_per_second": 4923.435
    },
    {
      "epoch": 6.6923076923076925,
      "num_input_tokens_seen": 346512,
      "step": 87,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 6.769230769230769,
      "grad_norm": 60.570884704589844,
      "learning_rate": 0.01,
      "loss": 66.96223449707031,
      "num_input_tokens_seen": 350608,
      "step": 88,
      "train_runtime": 71.182,
      "train_tokens_per_second": 4925.515
    },
    {
      "epoch": 6.769230769230769,
      "num_input_tokens_seen": 350608,
      "step": 88,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 6.846153846153846,
      "grad_norm": 37.7768669128418,
      "learning_rate": 0.01,
      "loss": 66.39012145996094,
      "num_input_tokens_seen": 354704,
      "step": 89,
      "train_runtime": 71.9835,
      "train_tokens_per_second": 4927.574
    },
    {
      "epoch": 6.846153846153846,
      "num_input_tokens_seen": 354704,
      "step": 89,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 6.923076923076923,
      "grad_norm": 58.63819885253906,
      "learning_rate": 0.01,
      "loss": 64.41407775878906,
      "num_input_tokens_seen": 358800,
      "step": 90,
      "train_runtime": 72.7812,
      "train_tokens_per_second": 4929.841
    },
    {
      "epoch": 6.923076923076923,
      "num_input_tokens_seen": 358800,
      "step": 90,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 7.0,
      "grad_norm": 54.507652282714844,
      "learning_rate": 0.01,
      "loss": 68.82718658447266,
      "num_input_tokens_seen": 361256,
      "step": 91,
      "train_runtime": 73.2899,
      "train_tokens_per_second": 4929.136
    },
    {
      "epoch": 7.0,
      "num_input_tokens_seen": 361256,
      "step": 91,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 7.076923076923077,
      "grad_norm": 68.39746856689453,
      "learning_rate": 0.01,
      "loss": 76.84484100341797,
      "num_input_tokens_seen": 365352,
      "step": 92,
      "train_runtime": 74.1037,
      "train_tokens_per_second": 4930.283
    },
    {
      "epoch": 7.076923076923077,
      "num_input_tokens_seen": 365352,
      "step": 92,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 7.153846153846154,
      "grad_norm": 68.6640853881836,
      "learning_rate": 0.01,
      "loss": 64.662353515625,
      "num_input_tokens_seen": 369448,
      "step": 93,
      "train_runtime": 74.9052,
      "train_tokens_per_second": 4932.207
    },
    {
      "epoch": 7.153846153846154,
      "num_input_tokens_seen": 369448,
      "step": 93,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 7.230769230769231,
      "grad_norm": 77.38482666015625,
      "learning_rate": 0.01,
      "loss": 70.11073303222656,
      "num_input_tokens_seen": 373544,
      "step": 94,
      "train_runtime": 75.7079,
      "train_tokens_per_second": 4934.018
    },
    {
      "epoch": 7.230769230769231,
      "num_input_tokens_seen": 373544,
      "step": 94,
      "train_accuracy": 0.0
    },
    {
      "epoch": 7.3076923076923075,
      "grad_norm": 76.43072509765625,
      "learning_rate": 0.01,
      "loss": 68.50315856933594,
      "num_input_tokens_seen": 377640,
      "step": 95,
      "train_runtime": 76.5101,
      "train_tokens_per_second": 4935.819
    },
    {
      "epoch": 7.3076923076923075,
      "num_input_tokens_seen": 377640,
      "step": 95,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 7.384615384615385,
      "grad_norm": 36.90632247924805,
      "learning_rate": 0.01,
      "loss": 58.06911849975586,
      "num_input_tokens_seen": 381736,
      "step": 96,
      "train_runtime": 77.3124,
      "train_tokens_per_second": 4937.578
    },
    {
      "epoch": 7.384615384615385,
      "num_input_tokens_seen": 381736,
      "step": 96,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 7.461538461538462,
      "grad_norm": 78.21217346191406,
      "learning_rate": 0.01,
      "loss": 75.7255630493164,
      "num_input_tokens_seen": 385832,
      "step": 97,
      "train_runtime": 78.115,
      "train_tokens_per_second": 4939.282
    },
    {
      "epoch": 7.461538461538462,
      "num_input_tokens_seen": 385832,
      "step": 97,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 7.538461538461538,
      "grad_norm": 76.51519012451172,
      "learning_rate": 0.01,
      "loss": 70.77012634277344,
      "num_input_tokens_seen": 389928,
      "step": 98,
      "train_runtime": 78.9183,
      "train_tokens_per_second": 4940.909
    },
    {
      "epoch": 7.538461538461538,
      "num_input_tokens_seen": 389928,
      "step": 98,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 7.615384615384615,
      "grad_norm": 67.02277374267578,
      "learning_rate": 0.01,
      "loss": 57.284114837646484,
      "num_input_tokens_seen": 394024,
      "step": 99,
      "train_runtime": 79.7215,
      "train_tokens_per_second": 4942.508
    },
    {
      "epoch": 7.615384615384615,
      "num_input_tokens_seen": 394024,
      "step": 99,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 7.6923076923076925,
      "grad_norm": 69.74555206298828,
      "learning_rate": 0.01,
      "loss": 59.708221435546875,
      "num_input_tokens_seen": 398120,
      "step": 100,
      "train_runtime": 80.5242,
      "train_tokens_per_second": 4944.101
    },
    {
      "epoch": 7.6923076923076925,
      "eval_CMD_3_branch_eval_accuracy": 0.011211573236889693,
      "eval_CMD_3_branch_eval_loss": 63.77542495727539,
      "eval_CMD_3_branch_eval_runtime": 1.1333,
      "eval_CMD_3_branch_eval_samples_per_second": 2439.689,
      "eval_CMD_3_branch_eval_steps_per_second": 5.294,
      "num_input_tokens_seen": 398120,
      "step": 100
    },
    {
      "epoch": 7.6923076923076925,
      "num_input_tokens_seen": 398120,
      "step": 100,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 7.769230769230769,
      "grad_norm": 66.39517211914062,
      "learning_rate": 0.01,
      "loss": 63.51918029785156,
      "num_input_tokens_seen": 402216,
      "step": 101,
      "train_runtime": 82.4657,
      "train_tokens_per_second": 4877.374
    },
    {
      "epoch": 7.769230769230769,
      "num_input_tokens_seen": 402216,
      "step": 101,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 7.846153846153846,
      "grad_norm": 71.41845703125,
      "learning_rate": 0.01,
      "loss": 58.58770751953125,
      "num_input_tokens_seen": 406312,
      "step": 102,
      "train_runtime": 83.2683,
      "train_tokens_per_second": 4879.551
    },
    {
      "epoch": 7.846153846153846,
      "num_input_tokens_seen": 406312,
      "step": 102,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 7.923076923076923,
      "grad_norm": 63.093082427978516,
      "learning_rate": 0.01,
      "loss": 49.53801727294922,
      "num_input_tokens_seen": 410408,
      "step": 103,
      "train_runtime": 84.0684,
      "train_tokens_per_second": 4881.833
    },
    {
      "epoch": 7.923076923076923,
      "num_input_tokens_seen": 410408,
      "step": 103,
      "train_accuracy": 0.026058631017804146
    },
    {
      "epoch": 8.0,
      "grad_norm": 63.079898834228516,
      "learning_rate": 0.01,
      "loss": 62.99091339111328,
      "num_input_tokens_seen": 412864,
      "step": 104,
      "train_runtime": 84.5763,
      "train_tokens_per_second": 4881.559
    },
    {
      "epoch": 8.0,
      "num_input_tokens_seen": 412864,
      "step": 104,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 8.076923076923077,
      "grad_norm": 50.123836517333984,
      "learning_rate": 0.01,
      "loss": 59.246253967285156,
      "num_input_tokens_seen": 416960,
      "step": 105,
      "train_runtime": 85.3899,
      "train_tokens_per_second": 4883.013
    },
    {
      "epoch": 8.076923076923077,
      "num_input_tokens_seen": 416960,
      "step": 105,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 8.153846153846153,
      "grad_norm": 72.22982025146484,
      "learning_rate": 0.01,
      "loss": 57.421173095703125,
      "num_input_tokens_seen": 421056,
      "step": 106,
      "train_runtime": 86.1927,
      "train_tokens_per_second": 4885.055
    },
    {
      "epoch": 8.153846153846153,
      "num_input_tokens_seen": 421056,
      "step": 106,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 8.23076923076923,
      "grad_norm": 72.39375305175781,
      "learning_rate": 0.01,
      "loss": 55.41351318359375,
      "num_input_tokens_seen": 425152,
      "step": 107,
      "train_runtime": 86.9963,
      "train_tokens_per_second": 4887.013
    },
    {
      "epoch": 8.23076923076923,
      "num_input_tokens_seen": 425152,
      "step": 107,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 8.307692307692308,
      "grad_norm": 50.16789627075195,
      "learning_rate": 0.01,
      "loss": 59.79741668701172,
      "num_input_tokens_seen": 429248,
      "step": 108,
      "train_runtime": 87.8002,
      "train_tokens_per_second": 4888.92
    },
    {
      "epoch": 8.307692307692308,
      "num_input_tokens_seen": 429248,
      "step": 108,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 8.384615384615385,
      "grad_norm": 72.01069641113281,
      "learning_rate": 0.01,
      "loss": 60.06975173950195,
      "num_input_tokens_seen": 433344,
      "step": 109,
      "train_runtime": 88.6049,
      "train_tokens_per_second": 4890.745
    },
    {
      "epoch": 8.384615384615385,
      "num_input_tokens_seen": 433344,
      "step": 109,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 8.461538461538462,
      "grad_norm": 55.495506286621094,
      "learning_rate": 0.01,
      "loss": 57.364906311035156,
      "num_input_tokens_seen": 437440,
      "step": 110,
      "train_runtime": 89.4096,
      "train_tokens_per_second": 4892.539
    },
    {
      "epoch": 8.461538461538462,
      "num_input_tokens_seen": 437440,
      "step": 110,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 8.538461538461538,
      "grad_norm": 67.98399353027344,
      "learning_rate": 0.01,
      "loss": 66.41151428222656,
      "num_input_tokens_seen": 441536,
      "step": 111,
      "train_runtime": 90.2139,
      "train_tokens_per_second": 4894.325
    },
    {
      "epoch": 8.538461538461538,
      "num_input_tokens_seen": 441536,
      "step": 111,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 8.615384615384615,
      "grad_norm": 68.21734619140625,
      "learning_rate": 0.01,
      "loss": 65.20812225341797,
      "num_input_tokens_seen": 445632,
      "step": 112,
      "train_runtime": 91.0187,
      "train_tokens_per_second": 4896.051
    },
    {
      "epoch": 8.615384615384615,
      "num_input_tokens_seen": 445632,
      "step": 112,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 8.692307692307692,
      "grad_norm": 53.03721237182617,
      "learning_rate": 0.01,
      "loss": 56.20685577392578,
      "num_input_tokens_seen": 449728,
      "step": 113,
      "train_runtime": 91.8234,
      "train_tokens_per_second": 4897.748
    },
    {
      "epoch": 8.692307692307692,
      "num_input_tokens_seen": 449728,
      "step": 113,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 8.76923076923077,
      "grad_norm": 50.524742126464844,
      "learning_rate": 0.01,
      "loss": 61.24467849731445,
      "num_input_tokens_seen": 453824,
      "step": 114,
      "train_runtime": 92.628,
      "train_tokens_per_second": 4899.427
    },
    {
      "epoch": 8.76923076923077,
      "num_input_tokens_seen": 453824,
      "step": 114,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 8.846153846153847,
      "grad_norm": 56.96110916137695,
      "learning_rate": 0.01,
      "loss": 57.87140655517578,
      "num_input_tokens_seen": 457920,
      "step": 115,
      "train_runtime": 93.4345,
      "train_tokens_per_second": 4900.976
    },
    {
      "epoch": 8.846153846153847,
      "num_input_tokens_seen": 457920,
      "step": 115,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 8.923076923076923,
      "grad_norm": 40.74517822265625,
      "learning_rate": 0.01,
      "loss": 46.19894027709961,
      "num_input_tokens_seen": 462016,
      "step": 116,
      "train_runtime": 94.2377,
      "train_tokens_per_second": 4902.669
    },
    {
      "epoch": 8.923076923076923,
      "num_input_tokens_seen": 462016,
      "step": 116,
      "train_accuracy": 0.013029315508902073
    },
    {
      "epoch": 9.0,
      "grad_norm": 44.081607818603516,
      "learning_rate": 0.01,
      "loss": 46.22126388549805,
      "num_input_tokens_seen": 464472,
      "step": 117,
      "train_runtime": 94.7473,
      "train_tokens_per_second": 4902.221
    },
    {
      "epoch": 9.0,
      "num_input_tokens_seen": 464472,
      "step": 117,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 9.076923076923077,
      "grad_norm": 65.67768859863281,
      "learning_rate": 0.01,
      "loss": 51.726722717285156,
      "num_input_tokens_seen": 468568,
      "step": 118,
      "train_runtime": 95.5645,
      "train_tokens_per_second": 4903.162
    },
    {
      "epoch": 9.076923076923077,
      "num_input_tokens_seen": 468568,
      "step": 118,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 9.153846153846153,
      "grad_norm": 67.71527099609375,
      "learning_rate": 0.01,
      "loss": 52.51469421386719,
      "num_input_tokens_seen": 472664,
      "step": 119,
      "train_runtime": 96.3683,
      "train_tokens_per_second": 4904.764
    },
    {
      "epoch": 9.153846153846153,
      "num_input_tokens_seen": 472664,
      "step": 119,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 9.23076923076923,
      "grad_norm": 35.15571212768555,
      "learning_rate": 0.01,
      "loss": 48.70310592651367,
      "num_input_tokens_seen": 476760,
      "step": 120,
      "train_runtime": 97.1726,
      "train_tokens_per_second": 4906.322
    },
    {
      "epoch": 9.23076923076923,
      "num_input_tokens_seen": 476760,
      "step": 120,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 9.307692307692308,
      "grad_norm": 64.75183868408203,
      "learning_rate": 0.01,
      "loss": 52.49262237548828,
      "num_input_tokens_seen": 480856,
      "step": 121,
      "train_runtime": 97.9774,
      "train_tokens_per_second": 4907.825
    },
    {
      "epoch": 9.307692307692308,
      "num_input_tokens_seen": 480856,
      "step": 121,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 9.384615384615385,
      "grad_norm": 58.96760940551758,
      "learning_rate": 0.01,
      "loss": 48.93315124511719,
      "num_input_tokens_seen": 484952,
      "step": 122,
      "train_runtime": 98.7806,
      "train_tokens_per_second": 4909.384
    },
    {
      "epoch": 9.384615384615385,
      "num_input_tokens_seen": 484952,
      "step": 122,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 9.461538461538462,
      "grad_norm": 63.780548095703125,
      "learning_rate": 0.01,
      "loss": 44.274940490722656,
      "num_input_tokens_seen": 489048,
      "step": 123,
      "train_runtime": 99.5838,
      "train_tokens_per_second": 4910.922
    },
    {
      "epoch": 9.461538461538462,
      "num_input_tokens_seen": 489048,
      "step": 123,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 9.538461538461538,
      "grad_norm": 61.71628952026367,
      "learning_rate": 0.01,
      "loss": 49.18743133544922,
      "num_input_tokens_seen": 493144,
      "step": 124,
      "train_runtime": 100.3871,
      "train_tokens_per_second": 4912.423
    },
    {
      "epoch": 9.538461538461538,
      "num_input_tokens_seen": 493144,
      "step": 124,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 9.615384615384615,
      "grad_norm": 62.674835205078125,
      "learning_rate": 0.01,
      "loss": 49.97962951660156,
      "num_input_tokens_seen": 497240,
      "step": 125,
      "train_runtime": 101.1903,
      "train_tokens_per_second": 4913.908
    },
    {
      "epoch": 9.615384615384615,
      "num_input_tokens_seen": 497240,
      "step": 125,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 9.692307692307692,
      "grad_norm": 62.50501251220703,
      "learning_rate": 0.01,
      "loss": 50.329891204833984,
      "num_input_tokens_seen": 501336,
      "step": 126,
      "train_runtime": 101.9935,
      "train_tokens_per_second": 4915.373
    },
    {
      "epoch": 9.692307692307692,
      "num_input_tokens_seen": 501336,
      "step": 126,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 9.76923076923077,
      "grad_norm": 63.707183837890625,
      "learning_rate": 0.01,
      "loss": 52.13093566894531,
      "num_input_tokens_seen": 505432,
      "step": 127,
      "train_runtime": 102.7966,
      "train_tokens_per_second": 4916.816
    },
    {
      "epoch": 9.76923076923077,
      "num_input_tokens_seen": 505432,
      "step": 127,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 9.846153846153847,
      "grad_norm": 62.29219436645508,
      "learning_rate": 0.01,
      "loss": 55.159767150878906,
      "num_input_tokens_seen": 509528,
      "step": 128,
      "train_runtime": 103.6009,
      "train_tokens_per_second": 4918.18
    },
    {
      "epoch": 9.846153846153847,
      "num_input_tokens_seen": 509528,
      "step": 128,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 9.923076923076923,
      "grad_norm": 63.7969856262207,
      "learning_rate": 0.01,
      "loss": 59.77757263183594,
      "num_input_tokens_seen": 513624,
      "step": 129,
      "train_runtime": 104.4001,
      "train_tokens_per_second": 4919.765
    },
    {
      "epoch": 9.923076923076923,
      "num_input_tokens_seen": 513624,
      "step": 129,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 10.0,
      "grad_norm": 61.75998306274414,
      "learning_rate": 0.01,
      "loss": 56.43214416503906,
      "num_input_tokens_seen": 516080,
      "step": 130,
      "train_runtime": 104.9091,
      "train_tokens_per_second": 4919.304
    },
    {
      "epoch": 10.0,
      "num_input_tokens_seen": 516080,
      "step": 130,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 10.076923076923077,
      "grad_norm": 45.13194274902344,
      "learning_rate": 0.01,
      "loss": 48.59112548828125,
      "num_input_tokens_seen": 520176,
      "step": 131,
      "train_runtime": 105.7249,
      "train_tokens_per_second": 4920.09
    },
    {
      "epoch": 10.076923076923077,
      "num_input_tokens_seen": 520176,
      "step": 131,
      "train_accuracy": 0.001953125
    },
    {
      "epoch": 10.153846153846153,
      "grad_norm": 43.682220458984375,
      "learning_rate": 0.01,
      "loss": 45.91192626953125,
      "num_input_tokens_seen": 524272,
      "step": 132,
      "train_runtime": 106.5299,
      "train_tokens_per_second": 4921.358
    },
    {
      "epoch": 10.153846153846153,
      "num_input_tokens_seen": 524272,
      "step": 132,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 10.23076923076923,
      "grad_norm": 62.21281814575195,
      "learning_rate": 0.01,
      "loss": 51.57979202270508,
      "num_input_tokens_seen": 528368,
      "step": 133,
      "train_runtime": 107.3341,
      "train_tokens_per_second": 4922.648
    },
    {
      "epoch": 10.23076923076923,
      "num_input_tokens_seen": 528368,
      "step": 133,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 10.307692307692308,
      "grad_norm": 51.02768325805664,
      "learning_rate": 0.01,
      "loss": 43.974334716796875,
      "num_input_tokens_seen": 532464,
      "step": 134,
      "train_runtime": 108.1456,
      "train_tokens_per_second": 4923.585
    },
    {
      "epoch": 10.307692307692308,
      "num_input_tokens_seen": 532464,
      "step": 134,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 10.384615384615385,
      "grad_norm": 54.074371337890625,
      "learning_rate": 0.01,
      "loss": 42.32614517211914,
      "num_input_tokens_seen": 536560,
      "step": 135,
      "train_runtime": 108.9535,
      "train_tokens_per_second": 4924.669
    },
    {
      "epoch": 10.384615384615385,
      "num_input_tokens_seen": 536560,
      "step": 135,
      "train_accuracy": 0.001953125
    },
    {
      "epoch": 10.461538461538462,
      "grad_norm": 44.661739349365234,
      "learning_rate": 0.01,
      "loss": 50.82025146484375,
      "num_input_tokens_seen": 540656,
      "step": 136,
      "train_runtime": 109.7591,
      "train_tokens_per_second": 4925.844
    },
    {
      "epoch": 10.461538461538462,
      "num_input_tokens_seen": 540656,
      "step": 136,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 10.538461538461538,
      "grad_norm": 53.57250213623047,
      "learning_rate": 0.01,
      "loss": 42.54605484008789,
      "num_input_tokens_seen": 544752,
      "step": 137,
      "train_runtime": 110.5639,
      "train_tokens_per_second": 4927.035
    },
    {
      "epoch": 10.538461538461538,
      "num_input_tokens_seen": 544752,
      "step": 137,
      "train_accuracy": 0.001953125
    },
    {
      "epoch": 10.615384615384615,
      "grad_norm": 52.8401985168457,
      "learning_rate": 0.01,
      "loss": 39.991355895996094,
      "num_input_tokens_seen": 548848,
      "step": 138,
      "train_runtime": 111.3689,
      "train_tokens_per_second": 4928.198
    },
    {
      "epoch": 10.615384615384615,
      "num_input_tokens_seen": 548848,
      "step": 138,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 10.692307692307692,
      "grad_norm": 37.74988555908203,
      "learning_rate": 0.01,
      "loss": 40.31428909301758,
      "num_input_tokens_seen": 552944,
      "step": 139,
      "train_runtime": 112.1732,
      "train_tokens_per_second": 4929.376
    },
    {
      "epoch": 10.692307692307692,
      "num_input_tokens_seen": 552944,
      "step": 139,
      "train_accuracy": 0.001953125
    },
    {
      "epoch": 10.76923076923077,
      "grad_norm": 33.514705657958984,
      "learning_rate": 0.01,
      "loss": 42.767417907714844,
      "num_input_tokens_seen": 557040,
      "step": 140,
      "train_runtime": 112.978,
      "train_tokens_per_second": 4930.515
    },
    {
      "epoch": 10.76923076923077,
      "num_input_tokens_seen": 557040,
      "step": 140,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 10.846153846153847,
      "grad_norm": 43.23638916015625,
      "learning_rate": 0.01,
      "loss": 34.514678955078125,
      "num_input_tokens_seen": 561136,
      "step": 141,
      "train_runtime": 113.782,
      "train_tokens_per_second": 4931.678
    },
    {
      "epoch": 10.846153846153847,
      "num_input_tokens_seen": 561136,
      "step": 141,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 10.923076923076923,
      "grad_norm": 23.037885665893555,
      "learning_rate": 0.01,
      "loss": 30.216079711914062,
      "num_input_tokens_seen": 565232,
      "step": 142,
      "train_runtime": 114.5826,
      "train_tokens_per_second": 4932.967
    },
    {
      "epoch": 10.923076923076923,
      "num_input_tokens_seen": 565232,
      "step": 142,
      "train_accuracy": 0.026058631017804146
    },
    {
      "epoch": 11.0,
      "grad_norm": 53.58565139770508,
      "learning_rate": 0.01,
      "loss": 36.06959533691406,
      "num_input_tokens_seen": 567688,
      "step": 143,
      "train_runtime": 115.0933,
      "train_tokens_per_second": 4932.417
    },
    {
      "epoch": 11.0,
      "num_input_tokens_seen": 567688,
      "step": 143,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 11.076923076923077,
      "grad_norm": 53.94941711425781,
      "learning_rate": 0.01,
      "loss": 35.18943786621094,
      "num_input_tokens_seen": 571784,
      "step": 144,
      "train_runtime": 115.9083,
      "train_tokens_per_second": 4933.071
    },
    {
      "epoch": 11.076923076923077,
      "num_input_tokens_seen": 571784,
      "step": 144,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 11.153846153846153,
      "grad_norm": 42.16376495361328,
      "learning_rate": 0.01,
      "loss": 40.448516845703125,
      "num_input_tokens_seen": 575880,
      "step": 145,
      "train_runtime": 116.7126,
      "train_tokens_per_second": 4934.173
    },
    {
      "epoch": 11.153846153846153,
      "num_input_tokens_seen": 575880,
      "step": 145,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 11.23076923076923,
      "grad_norm": 50.17839813232422,
      "learning_rate": 0.01,
      "loss": 36.233680725097656,
      "num_input_tokens_seen": 579976,
      "step": 146,
      "train_runtime": 117.5175,
      "train_tokens_per_second": 4935.23
    },
    {
      "epoch": 11.23076923076923,
      "num_input_tokens_seen": 579976,
      "step": 146,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 11.307692307692308,
      "grad_norm": 26.932775497436523,
      "learning_rate": 0.01,
      "loss": 40.6314697265625,
      "num_input_tokens_seen": 584072,
      "step": 147,
      "train_runtime": 118.324,
      "train_tokens_per_second": 4936.21
    },
    {
      "epoch": 11.307692307692308,
      "num_input_tokens_seen": 584072,
      "step": 147,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 11.384615384615385,
      "grad_norm": 40.751590728759766,
      "learning_rate": 0.01,
      "loss": 38.37248992919922,
      "num_input_tokens_seen": 588168,
      "step": 148,
      "train_runtime": 119.1292,
      "train_tokens_per_second": 4937.226
    },
    {
      "epoch": 11.384615384615385,
      "num_input_tokens_seen": 588168,
      "step": 148,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 11.461538461538462,
      "grad_norm": 43.54330062866211,
      "learning_rate": 0.01,
      "loss": 36.76280975341797,
      "num_input_tokens_seen": 592264,
      "step": 149,
      "train_runtime": 119.9338,
      "train_tokens_per_second": 4938.259
    },
    {
      "epoch": 11.461538461538462,
      "num_input_tokens_seen": 592264,
      "step": 149,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 11.538461538461538,
      "grad_norm": 27.936599731445312,
      "learning_rate": 0.01,
      "loss": 33.58572769165039,
      "num_input_tokens_seen": 596360,
      "step": 150,
      "train_runtime": 120.7438,
      "train_tokens_per_second": 4939.053
    },
    {
      "epoch": 11.538461538461538,
      "num_input_tokens_seen": 596360,
      "step": 150,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 11.615384615384615,
      "grad_norm": 31.514619827270508,
      "learning_rate": 0.01,
      "loss": 33.77937316894531,
      "num_input_tokens_seen": 600456,
      "step": 151,
      "train_runtime": 121.551,
      "train_tokens_per_second": 4939.95
    },
    {
      "epoch": 11.615384615384615,
      "num_input_tokens_seen": 600456,
      "step": 151,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 11.692307692307692,
      "grad_norm": 28.991971969604492,
      "learning_rate": 0.01,
      "loss": 30.876453399658203,
      "num_input_tokens_seen": 604552,
      "step": 152,
      "train_runtime": 122.3575,
      "train_tokens_per_second": 4940.865
    },
    {
      "epoch": 11.692307692307692,
      "num_input_tokens_seen": 604552,
      "step": 152,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 11.76923076923077,
      "grad_norm": 17.685880661010742,
      "learning_rate": 0.01,
      "loss": 27.308061599731445,
      "num_input_tokens_seen": 608648,
      "step": 153,
      "train_runtime": 123.163,
      "train_tokens_per_second": 4941.807
    },
    {
      "epoch": 11.76923076923077,
      "num_input_tokens_seen": 608648,
      "step": 153,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 11.846153846153847,
      "grad_norm": 45.63172149658203,
      "learning_rate": 0.01,
      "loss": 32.146240234375,
      "num_input_tokens_seen": 612744,
      "step": 154,
      "train_runtime": 123.9664,
      "train_tokens_per_second": 4942.825
    },
    {
      "epoch": 11.846153846153847,
      "num_input_tokens_seen": 612744,
      "step": 154,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 11.923076923076923,
      "grad_norm": 18.520980834960938,
      "learning_rate": 0.01,
      "loss": 28.883460998535156,
      "num_input_tokens_seen": 616840,
      "step": 155,
      "train_runtime": 124.767,
      "train_tokens_per_second": 4943.935
    },
    {
      "epoch": 11.923076923076923,
      "num_input_tokens_seen": 616840,
      "step": 155,
      "train_accuracy": 0.016286645084619522
    },
    {
      "epoch": 12.0,
      "grad_norm": 25.246355056762695,
      "learning_rate": 0.01,
      "loss": 27.419111251831055,
      "num_input_tokens_seen": 619296,
      "step": 156,
      "train_runtime": 125.2769,
      "train_tokens_per_second": 4943.418
    },
    {
      "epoch": 12.0,
      "num_input_tokens_seen": 619296,
      "step": 156,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 12.076923076923077,
      "grad_norm": 31.113773345947266,
      "learning_rate": 0.01,
      "loss": 26.06835174560547,
      "num_input_tokens_seen": 623392,
      "step": 157,
      "train_runtime": 126.0934,
      "train_tokens_per_second": 4943.89
    },
    {
      "epoch": 12.076923076923077,
      "num_input_tokens_seen": 623392,
      "step": 157,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 12.153846153846153,
      "grad_norm": 41.25708770751953,
      "learning_rate": 0.01,
      "loss": 28.284313201904297,
      "num_input_tokens_seen": 627488,
      "step": 158,
      "train_runtime": 126.8996,
      "train_tokens_per_second": 4944.759
    },
    {
      "epoch": 12.153846153846153,
      "num_input_tokens_seen": 627488,
      "step": 158,
      "train_accuracy": 0.001953125
    },
    {
      "epoch": 12.23076923076923,
      "grad_norm": 23.370925903320312,
      "learning_rate": 0.01,
      "loss": 28.755294799804688,
      "num_input_tokens_seen": 631584,
      "step": 159,
      "train_runtime": 127.7051,
      "train_tokens_per_second": 4945.644
    },
    {
      "epoch": 12.23076923076923,
      "num_input_tokens_seen": 631584,
      "step": 159,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 12.307692307692308,
      "grad_norm": 21.386932373046875,
      "learning_rate": 0.01,
      "loss": 27.405820846557617,
      "num_input_tokens_seen": 635680,
      "step": 160,
      "train_runtime": 128.5108,
      "train_tokens_per_second": 4946.509
    },
    {
      "epoch": 12.307692307692308,
      "num_input_tokens_seen": 635680,
      "step": 160,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 12.384615384615385,
      "grad_norm": 23.117815017700195,
      "learning_rate": 0.01,
      "loss": 26.803756713867188,
      "num_input_tokens_seen": 639776,
      "step": 161,
      "train_runtime": 129.3174,
      "train_tokens_per_second": 4947.332
    },
    {
      "epoch": 12.384615384615385,
      "num_input_tokens_seen": 639776,
      "step": 161,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 12.461538461538462,
      "grad_norm": 17.937519073486328,
      "learning_rate": 0.01,
      "loss": 21.780914306640625,
      "num_input_tokens_seen": 643872,
      "step": 162,
      "train_runtime": 130.1236,
      "train_tokens_per_second": 4948.155
    },
    {
      "epoch": 12.461538461538462,
      "num_input_tokens_seen": 643872,
      "step": 162,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 12.538461538461538,
      "grad_norm": 19.52775764465332,
      "learning_rate": 0.01,
      "loss": 20.647687911987305,
      "num_input_tokens_seen": 647968,
      "step": 163,
      "train_runtime": 130.9334,
      "train_tokens_per_second": 4948.837
    },
    {
      "epoch": 12.538461538461538,
      "num_input_tokens_seen": 647968,
      "step": 163,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 12.615384615384615,
      "grad_norm": 30.54848861694336,
      "learning_rate": 0.01,
      "loss": 21.44734001159668,
      "num_input_tokens_seen": 652064,
      "step": 164,
      "train_runtime": 131.746,
      "train_tokens_per_second": 4949.403
    },
    {
      "epoch": 12.615384615384615,
      "num_input_tokens_seen": 652064,
      "step": 164,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 12.692307692307692,
      "grad_norm": 28.124404907226562,
      "learning_rate": 0.01,
      "loss": 20.266860961914062,
      "num_input_tokens_seen": 656160,
      "step": 165,
      "train_runtime": 132.5566,
      "train_tokens_per_second": 4950.038
    },
    {
      "epoch": 12.692307692307692,
      "num_input_tokens_seen": 656160,
      "step": 165,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 12.76923076923077,
      "grad_norm": 17.651201248168945,
      "learning_rate": 0.01,
      "loss": 18.141666412353516,
      "num_input_tokens_seen": 660256,
      "step": 166,
      "train_runtime": 133.3617,
      "train_tokens_per_second": 4950.866
    },
    {
      "epoch": 12.76923076923077,
      "num_input_tokens_seen": 660256,
      "step": 166,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 12.846153846153847,
      "grad_norm": 26.800312042236328,
      "learning_rate": 0.01,
      "loss": 19.31981658935547,
      "num_input_tokens_seen": 664352,
      "step": 167,
      "train_runtime": 134.1667,
      "train_tokens_per_second": 4951.691
    },
    {
      "epoch": 12.846153846153847,
      "num_input_tokens_seen": 664352,
      "step": 167,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 12.923076923076923,
      "grad_norm": 14.499151229858398,
      "learning_rate": 0.01,
      "loss": 18.218852996826172,
      "num_input_tokens_seen": 668448,
      "step": 168,
      "train_runtime": 134.968,
      "train_tokens_per_second": 4952.641
    },
    {
      "epoch": 12.923076923076923,
      "num_input_tokens_seen": 668448,
      "step": 168,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 13.0,
      "grad_norm": 12.949419975280762,
      "learning_rate": 0.01,
      "loss": 17.60304832458496,
      "num_input_tokens_seen": 670904,
      "step": 169,
      "train_runtime": 135.4781,
      "train_tokens_per_second": 4952.12
    },
    {
      "epoch": 13.0,
      "num_input_tokens_seen": 670904,
      "step": 169,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 13.076923076923077,
      "grad_norm": 12.970791816711426,
      "learning_rate": 0.01,
      "loss": 16.74079704284668,
      "num_input_tokens_seen": 675000,
      "step": 170,
      "train_runtime": 136.2963,
      "train_tokens_per_second": 4952.447
    },
    {
      "epoch": 13.076923076923077,
      "num_input_tokens_seen": 675000,
      "step": 170,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 13.153846153846153,
      "grad_norm": 12.812407493591309,
      "learning_rate": 0.01,
      "loss": 15.077863693237305,
      "num_input_tokens_seen": 679096,
      "step": 171,
      "train_runtime": 137.1012,
      "train_tokens_per_second": 4953.248
    },
    {
      "epoch": 13.153846153846153,
      "num_input_tokens_seen": 679096,
      "step": 171,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 13.23076923076923,
      "grad_norm": 13.73751163482666,
      "learning_rate": 0.01,
      "loss": 14.415550231933594,
      "num_input_tokens_seen": 683192,
      "step": 172,
      "train_runtime": 137.9067,
      "train_tokens_per_second": 4954.018
    },
    {
      "epoch": 13.23076923076923,
      "num_input_tokens_seen": 683192,
      "step": 172,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 13.307692307692308,
      "grad_norm": 9.886204719543457,
      "learning_rate": 0.01,
      "loss": 13.83185863494873,
      "num_input_tokens_seen": 687288,
      "step": 173,
      "train_runtime": 138.712,
      "train_tokens_per_second": 4954.784
    },
    {
      "epoch": 13.307692307692308,
      "num_input_tokens_seen": 687288,
      "step": 173,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 13.384615384615385,
      "grad_norm": 11.833088874816895,
      "learning_rate": 0.01,
      "loss": 13.087343215942383,
      "num_input_tokens_seen": 691384,
      "step": 174,
      "train_runtime": 139.519,
      "train_tokens_per_second": 4955.481
    },
    {
      "epoch": 13.384615384615385,
      "num_input_tokens_seen": 691384,
      "step": 174,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 13.461538461538462,
      "grad_norm": 13.609389305114746,
      "learning_rate": 0.01,
      "loss": 12.435286521911621,
      "num_input_tokens_seen": 695480,
      "step": 175,
      "train_runtime": 140.3246,
      "train_tokens_per_second": 4956.222
    },
    {
      "epoch": 13.461538461538462,
      "num_input_tokens_seen": 695480,
      "step": 175,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 13.538461538461538,
      "grad_norm": 8.842402458190918,
      "learning_rate": 0.01,
      "loss": 11.198854446411133,
      "num_input_tokens_seen": 699576,
      "step": 176,
      "train_runtime": 141.1304,
      "train_tokens_per_second": 4956.947
    },
    {
      "epoch": 13.538461538461538,
      "num_input_tokens_seen": 699576,
      "step": 176,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 13.615384615384615,
      "grad_norm": 9.526718139648438,
      "learning_rate": 0.01,
      "loss": 10.711845397949219,
      "num_input_tokens_seen": 703672,
      "step": 177,
      "train_runtime": 141.9408,
      "train_tokens_per_second": 4957.502
    },
    {
      "epoch": 13.615384615384615,
      "num_input_tokens_seen": 703672,
      "step": 177,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 13.692307692307692,
      "grad_norm": 9.6870756149292,
      "learning_rate": 0.01,
      "loss": 11.304187774658203,
      "num_input_tokens_seen": 707768,
      "step": 178,
      "train_runtime": 142.7496,
      "train_tokens_per_second": 4958.109
    },
    {
      "epoch": 13.692307692307692,
      "num_input_tokens_seen": 707768,
      "step": 178,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 13.76923076923077,
      "grad_norm": 8.219425201416016,
      "learning_rate": 0.01,
      "loss": 9.206737518310547,
      "num_input_tokens_seen": 711864,
      "step": 179,
      "train_runtime": 143.5558,
      "train_tokens_per_second": 4958.798
    },
    {
      "epoch": 13.76923076923077,
      "num_input_tokens_seen": 711864,
      "step": 179,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 13.846153846153847,
      "grad_norm": 6.607143402099609,
      "learning_rate": 0.01,
      "loss": 9.438591003417969,
      "num_input_tokens_seen": 715960,
      "step": 180,
      "train_runtime": 144.3617,
      "train_tokens_per_second": 4959.487
    },
    {
      "epoch": 13.846153846153847,
      "num_input_tokens_seen": 715960,
      "step": 180,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 13.923076923076923,
      "grad_norm": 7.296689033508301,
      "learning_rate": 0.01,
      "loss": 8.733935356140137,
      "num_input_tokens_seen": 720056,
      "step": 181,
      "train_runtime": 145.1628,
      "train_tokens_per_second": 4960.335
    },
    {
      "epoch": 13.923076923076923,
      "num_input_tokens_seen": 720056,
      "step": 181,
      "train_accuracy": 0.0065146577544510365
    },
    {
      "epoch": 14.0,
      "grad_norm": 8.992117881774902,
      "learning_rate": 0.01,
      "loss": 8.867663383483887,
      "num_input_tokens_seen": 722512,
      "step": 182,
      "train_runtime": 145.6733,
      "train_tokens_per_second": 4959.809
    },
    {
      "epoch": 14.0,
      "num_input_tokens_seen": 722512,
      "step": 182,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 14.076923076923077,
      "grad_norm": 6.970577239990234,
      "learning_rate": 0.01,
      "loss": 8.428256034851074,
      "num_input_tokens_seen": 726608,
      "step": 183,
      "train_runtime": 146.4902,
      "train_tokens_per_second": 4960.112
    },
    {
      "epoch": 14.076923076923077,
      "num_input_tokens_seen": 726608,
      "step": 183,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 14.153846153846153,
      "grad_norm": 6.69838809967041,
      "learning_rate": 0.01,
      "loss": 7.663095474243164,
      "num_input_tokens_seen": 730704,
      "step": 184,
      "train_runtime": 147.2963,
      "train_tokens_per_second": 4960.778
    },
    {
      "epoch": 14.153846153846153,
      "num_input_tokens_seen": 730704,
      "step": 184,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 14.23076923076923,
      "grad_norm": 7.49269962310791,
      "learning_rate": 0.01,
      "loss": 7.656740188598633,
      "num_input_tokens_seen": 734800,
      "step": 185,
      "train_runtime": 148.1016,
      "train_tokens_per_second": 4961.46
    },
    {
      "epoch": 14.23076923076923,
      "num_input_tokens_seen": 734800,
      "step": 185,
      "train_accuracy": 0.001953125
    },
    {
      "epoch": 14.307692307692308,
      "grad_norm": 4.829015731811523,
      "learning_rate": 0.01,
      "loss": 7.406508445739746,
      "num_input_tokens_seen": 738896,
      "step": 186,
      "train_runtime": 148.9071,
      "train_tokens_per_second": 4962.126
    },
    {
      "epoch": 14.307692307692308,
      "num_input_tokens_seen": 738896,
      "step": 186,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 14.384615384615385,
      "grad_norm": 5.506415367126465,
      "learning_rate": 0.01,
      "loss": 7.3244428634643555,
      "num_input_tokens_seen": 742992,
      "step": 187,
      "train_runtime": 149.7122,
      "train_tokens_per_second": 4962.802
    },
    {
      "epoch": 14.384615384615385,
      "num_input_tokens_seen": 742992,
      "step": 187,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 14.461538461538462,
      "grad_norm": 5.795366287231445,
      "learning_rate": 0.01,
      "loss": 7.266047954559326,
      "num_input_tokens_seen": 747088,
      "step": 188,
      "train_runtime": 150.5184,
      "train_tokens_per_second": 4963.432
    },
    {
      "epoch": 14.461538461538462,
      "num_input_tokens_seen": 747088,
      "step": 188,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 14.538461538461538,
      "grad_norm": 5.064494609832764,
      "learning_rate": 0.01,
      "loss": 6.987112998962402,
      "num_input_tokens_seen": 751184,
      "step": 189,
      "train_runtime": 151.3245,
      "train_tokens_per_second": 4964.062
    },
    {
      "epoch": 14.538461538461538,
      "num_input_tokens_seen": 751184,
      "step": 189,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 14.615384615384615,
      "grad_norm": 4.764325141906738,
      "learning_rate": 0.01,
      "loss": 7.1143083572387695,
      "num_input_tokens_seen": 755280,
      "step": 190,
      "train_runtime": 152.1302,
      "train_tokens_per_second": 4964.696
    },
    {
      "epoch": 14.615384615384615,
      "num_input_tokens_seen": 755280,
      "step": 190,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 14.692307692307692,
      "grad_norm": 5.380343437194824,
      "learning_rate": 0.01,
      "loss": 6.598371982574463,
      "num_input_tokens_seen": 759376,
      "step": 191,
      "train_runtime": 152.9372,
      "train_tokens_per_second": 4965.278
    },
    {
      "epoch": 14.692307692307692,
      "num_input_tokens_seen": 759376,
      "step": 191,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 14.76923076923077,
      "grad_norm": 4.394591808319092,
      "learning_rate": 0.01,
      "loss": 6.156771659851074,
      "num_input_tokens_seen": 763472,
      "step": 192,
      "train_runtime": 153.7488,
      "train_tokens_per_second": 4965.711
    },
    {
      "epoch": 14.76923076923077,
      "num_input_tokens_seen": 763472,
      "step": 192,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 14.846153846153847,
      "grad_norm": 4.73301887512207,
      "learning_rate": 0.01,
      "loss": 6.109883785247803,
      "num_input_tokens_seen": 767568,
      "step": 193,
      "train_runtime": 154.5554,
      "train_tokens_per_second": 4966.296
    },
    {
      "epoch": 14.846153846153847,
      "num_input_tokens_seen": 767568,
      "step": 193,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 14.923076923076923,
      "grad_norm": 5.311499118804932,
      "learning_rate": 0.01,
      "loss": 5.823588848114014,
      "num_input_tokens_seen": 771664,
      "step": 194,
      "train_runtime": 155.3568,
      "train_tokens_per_second": 4967.045
    },
    {
      "epoch": 14.923076923076923,
      "num_input_tokens_seen": 771664,
      "step": 194,
      "train_accuracy": 0.013029315508902073
    },
    {
      "epoch": 15.0,
      "grad_norm": 5.872067451477051,
      "learning_rate": 0.01,
      "loss": 6.032412052154541,
      "num_input_tokens_seen": 774120,
      "step": 195,
      "train_runtime": 155.8668,
      "train_tokens_per_second": 4966.549
    },
    {
      "epoch": 15.0,
      "num_input_tokens_seen": 774120,
      "step": 195,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 15.076923076923077,
      "grad_norm": 4.518616676330566,
      "learning_rate": 0.01,
      "loss": 5.693182468414307,
      "num_input_tokens_seen": 778216,
      "step": 196,
      "train_runtime": 156.6827,
      "train_tokens_per_second": 4966.828
    },
    {
      "epoch": 15.076923076923077,
      "num_input_tokens_seen": 778216,
      "step": 196,
      "train_accuracy": 0.001953125
    },
    {
      "epoch": 15.153846153846153,
      "grad_norm": 3.8842692375183105,
      "learning_rate": 0.01,
      "loss": 5.344891548156738,
      "num_input_tokens_seen": 782312,
      "step": 197,
      "train_runtime": 157.4886,
      "train_tokens_per_second": 4967.419
    },
    {
      "epoch": 15.153846153846153,
      "num_input_tokens_seen": 782312,
      "step": 197,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 15.23076923076923,
      "grad_norm": 3.7938222885131836,
      "learning_rate": 0.01,
      "loss": 5.419033527374268,
      "num_input_tokens_seen": 786408,
      "step": 198,
      "train_runtime": 158.2942,
      "train_tokens_per_second": 4968.014
    },
    {
      "epoch": 15.23076923076923,
      "num_input_tokens_seen": 786408,
      "step": 198,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 15.307692307692308,
      "grad_norm": 3.7228262424468994,
      "learning_rate": 0.01,
      "loss": 5.356511116027832,
      "num_input_tokens_seen": 790504,
      "step": 199,
      "train_runtime": 159.1005,
      "train_tokens_per_second": 4968.581
    },
    {
      "epoch": 15.307692307692308,
      "num_input_tokens_seen": 790504,
      "step": 199,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 15.384615384615385,
      "grad_norm": 3.548959970474243,
      "learning_rate": 0.01,
      "loss": 5.381818771362305,
      "num_input_tokens_seen": 794600,
      "step": 200,
      "train_runtime": 159.907,
      "train_tokens_per_second": 4969.139
    },
    {
      "epoch": 15.384615384615385,
      "eval_CMD_3_branch_eval_accuracy": 0.007594936708860759,
      "eval_CMD_3_branch_eval_loss": 5.368804931640625,
      "eval_CMD_3_branch_eval_runtime": 1.1297,
      "eval_CMD_3_branch_eval_samples_per_second": 2447.572,
      "eval_CMD_3_branch_eval_steps_per_second": 5.311,
      "num_input_tokens_seen": 794600,
      "step": 200
    },
    {
      "epoch": 15.384615384615385,
      "num_input_tokens_seen": 794600,
      "step": 200,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 15.461538461538462,
      "grad_norm": 3.4090161323547363,
      "learning_rate": 0.01,
      "loss": 5.308737754821777,
      "num_input_tokens_seen": 798696,
      "step": 201,
      "train_runtime": 161.8478,
      "train_tokens_per_second": 4934.86
    },
    {
      "epoch": 15.461538461538462,
      "num_input_tokens_seen": 798696,
      "step": 201,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 15.538461538461538,
      "grad_norm": 3.5405497550964355,
      "learning_rate": 0.01,
      "loss": 5.098702430725098,
      "num_input_tokens_seen": 802792,
      "step": 202,
      "train_runtime": 162.655,
      "train_tokens_per_second": 4935.55
    },
    {
      "epoch": 15.538461538461538,
      "num_input_tokens_seen": 802792,
      "step": 202,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 15.615384615384615,
      "grad_norm": 3.1300690174102783,
      "learning_rate": 0.01,
      "loss": 5.000946521759033,
      "num_input_tokens_seen": 806888,
      "step": 203,
      "train_runtime": 163.4623,
      "train_tokens_per_second": 4936.232
    },
    {
      "epoch": 15.615384615384615,
      "num_input_tokens_seen": 806888,
      "step": 203,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 15.692307692307692,
      "grad_norm": 3.1456973552703857,
      "learning_rate": 0.01,
      "loss": 5.021718502044678,
      "num_input_tokens_seen": 810984,
      "step": 204,
      "train_runtime": 164.2691,
      "train_tokens_per_second": 4936.922
    },
    {
      "epoch": 15.692307692307692,
      "num_input_tokens_seen": 810984,
      "step": 204,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 15.76923076923077,
      "grad_norm": 2.667361259460449,
      "learning_rate": 0.01,
      "loss": 4.919438362121582,
      "num_input_tokens_seen": 815080,
      "step": 205,
      "train_runtime": 165.077,
      "train_tokens_per_second": 4937.574
    },
    {
      "epoch": 15.76923076923077,
      "num_input_tokens_seen": 815080,
      "step": 205,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 15.846153846153847,
      "grad_norm": 2.7647666931152344,
      "learning_rate": 0.01,
      "loss": 4.9262847900390625,
      "num_input_tokens_seen": 819176,
      "step": 206,
      "train_runtime": 165.8829,
      "train_tokens_per_second": 4938.279
    },
    {
      "epoch": 15.846153846153847,
      "num_input_tokens_seen": 819176,
      "step": 206,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 15.923076923076923,
      "grad_norm": 3.502366065979004,
      "learning_rate": 0.01,
      "loss": 4.926443099975586,
      "num_input_tokens_seen": 823272,
      "step": 207,
      "train_runtime": 166.685,
      "train_tokens_per_second": 4939.09
    },
    {
      "epoch": 15.923076923076923,
      "num_input_tokens_seen": 823272,
      "step": 207,
      "train_accuracy": 0.0065146577544510365
    },
    {
      "epoch": 16.0,
      "grad_norm": 3.5238306522369385,
      "learning_rate": 0.01,
      "loss": 4.891132354736328,
      "num_input_tokens_seen": 825728,
      "step": 208,
      "train_runtime": 167.1958,
      "train_tokens_per_second": 4938.689
    },
    {
      "epoch": 16.0,
      "num_input_tokens_seen": 825728,
      "step": 208,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 16.076923076923077,
      "grad_norm": 3.013728141784668,
      "learning_rate": 0.01,
      "loss": 4.880594730377197,
      "num_input_tokens_seen": 829824,
      "step": 209,
      "train_runtime": 168.0142,
      "train_tokens_per_second": 4939.01
    },
    {
      "epoch": 16.076923076923077,
      "num_input_tokens_seen": 829824,
      "step": 209,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 16.153846153846153,
      "grad_norm": 2.7540860176086426,
      "learning_rate": 0.01,
      "loss": 4.831616401672363,
      "num_input_tokens_seen": 833920,
      "step": 210,
      "train_runtime": 168.82,
      "train_tokens_per_second": 4939.699
    },
    {
      "epoch": 16.153846153846153,
      "num_input_tokens_seen": 833920,
      "step": 210,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 16.23076923076923,
      "grad_norm": 2.9010918140411377,
      "learning_rate": 0.01,
      "loss": 4.833306312561035,
      "num_input_tokens_seen": 838016,
      "step": 211,
      "train_runtime": 169.6263,
      "train_tokens_per_second": 4940.366
    },
    {
      "epoch": 16.23076923076923,
      "num_input_tokens_seen": 838016,
      "step": 211,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 16.307692307692307,
      "grad_norm": 2.8716068267822266,
      "learning_rate": 0.01,
      "loss": 4.825284481048584,
      "num_input_tokens_seen": 842112,
      "step": 212,
      "train_runtime": 170.4342,
      "train_tokens_per_second": 4940.979
    },
    {
      "epoch": 16.307692307692307,
      "num_input_tokens_seen": 842112,
      "step": 212,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 16.384615384615383,
      "grad_norm": 2.4198055267333984,
      "learning_rate": 0.01,
      "loss": 4.814431667327881,
      "num_input_tokens_seen": 846208,
      "step": 213,
      "train_runtime": 171.2416,
      "train_tokens_per_second": 4941.603
    },
    {
      "epoch": 16.384615384615383,
      "num_input_tokens_seen": 846208,
      "step": 213,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 16.46153846153846,
      "grad_norm": 2.166982412338257,
      "learning_rate": 0.01,
      "loss": 4.759323596954346,
      "num_input_tokens_seen": 850304,
      "step": 214,
      "train_runtime": 172.0481,
      "train_tokens_per_second": 4942.245
    },
    {
      "epoch": 16.46153846153846,
      "num_input_tokens_seen": 850304,
      "step": 214,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 16.53846153846154,
      "grad_norm": 2.081040620803833,
      "learning_rate": 0.01,
      "loss": 4.729094505310059,
      "num_input_tokens_seen": 854400,
      "step": 215,
      "train_runtime": 172.8552,
      "train_tokens_per_second": 4942.866
    },
    {
      "epoch": 16.53846153846154,
      "num_input_tokens_seen": 854400,
      "step": 215,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 16.615384615384617,
      "grad_norm": 2.177140951156616,
      "learning_rate": 0.01,
      "loss": 4.726467132568359,
      "num_input_tokens_seen": 858496,
      "step": 216,
      "train_runtime": 173.6631,
      "train_tokens_per_second": 4943.457
    },
    {
      "epoch": 16.615384615384617,
      "num_input_tokens_seen": 858496,
      "step": 216,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 16.692307692307693,
      "grad_norm": 2.091027021408081,
      "learning_rate": 0.01,
      "loss": 4.732223987579346,
      "num_input_tokens_seen": 862592,
      "step": 217,
      "train_runtime": 174.4692,
      "train_tokens_per_second": 4944.094
    },
    {
      "epoch": 16.692307692307693,
      "num_input_tokens_seen": 862592,
      "step": 217,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 16.76923076923077,
      "grad_norm": 1.9717937707901,
      "learning_rate": 0.01,
      "loss": 4.70272159576416,
      "num_input_tokens_seen": 866688,
      "step": 218,
      "train_runtime": 175.2753,
      "train_tokens_per_second": 4944.723
    },
    {
      "epoch": 16.76923076923077,
      "num_input_tokens_seen": 866688,
      "step": 218,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 16.846153846153847,
      "grad_norm": 2.0200533866882324,
      "learning_rate": 0.01,
      "loss": 4.724699974060059,
      "num_input_tokens_seen": 870784,
      "step": 219,
      "train_runtime": 176.0818,
      "train_tokens_per_second": 4945.339
    },
    {
      "epoch": 16.846153846153847,
      "num_input_tokens_seen": 870784,
      "step": 219,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 16.923076923076923,
      "grad_norm": 2.4475533962249756,
      "learning_rate": 0.01,
      "loss": 4.796332359313965,
      "num_input_tokens_seen": 874880,
      "step": 220,
      "train_runtime": 176.8838,
      "train_tokens_per_second": 4946.071
    },
    {
      "epoch": 16.923076923076923,
      "num_input_tokens_seen": 874880,
      "step": 220,
      "train_accuracy": 0.0065146577544510365
    },
    {
      "epoch": 17.0,
      "grad_norm": 2.6400070190429688,
      "learning_rate": 0.01,
      "loss": 4.780521869659424,
      "num_input_tokens_seen": 877336,
      "step": 221,
      "train_runtime": 177.3948,
      "train_tokens_per_second": 4945.668
    },
    {
      "epoch": 17.0,
      "num_input_tokens_seen": 877336,
      "step": 221,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 17.076923076923077,
      "grad_norm": 2.774951457977295,
      "learning_rate": 0.01,
      "loss": 4.7688775062561035,
      "num_input_tokens_seen": 881432,
      "step": 222,
      "train_runtime": 178.2127,
      "train_tokens_per_second": 4945.956
    },
    {
      "epoch": 17.076923076923077,
      "num_input_tokens_seen": 881432,
      "step": 222,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 17.153846153846153,
      "grad_norm": 2.80867338180542,
      "learning_rate": 0.01,
      "loss": 4.773642539978027,
      "num_input_tokens_seen": 885528,
      "step": 223,
      "train_runtime": 179.0195,
      "train_tokens_per_second": 4946.546
    },
    {
      "epoch": 17.153846153846153,
      "num_input_tokens_seen": 885528,
      "step": 223,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 17.23076923076923,
      "grad_norm": 1.7940993309020996,
      "learning_rate": 0.01,
      "loss": 4.654668807983398,
      "num_input_tokens_seen": 889624,
      "step": 224,
      "train_runtime": 179.8267,
      "train_tokens_per_second": 4947.119
    },
    {
      "epoch": 17.23076923076923,
      "num_input_tokens_seen": 889624,
      "step": 224,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 17.307692307692307,
      "grad_norm": 2.628485679626465,
      "learning_rate": 0.01,
      "loss": 4.76577091217041,
      "num_input_tokens_seen": 893720,
      "step": 225,
      "train_runtime": 180.6336,
      "train_tokens_per_second": 4947.696
    },
    {
      "epoch": 17.307692307692307,
      "num_input_tokens_seen": 893720,
      "step": 225,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 17.384615384615383,
      "grad_norm": 1.9633597135543823,
      "learning_rate": 0.01,
      "loss": 4.685235977172852,
      "num_input_tokens_seen": 897816,
      "step": 226,
      "train_runtime": 181.441,
      "train_tokens_per_second": 4948.252
    },
    {
      "epoch": 17.384615384615383,
      "num_input_tokens_seen": 897816,
      "step": 226,
      "train_accuracy": 0.001953125
    },
    {
      "epoch": 17.46153846153846,
      "grad_norm": 3.5501697063446045,
      "learning_rate": 0.01,
      "loss": 4.778290271759033,
      "num_input_tokens_seen": 901912,
      "step": 227,
      "train_runtime": 182.2475,
      "train_tokens_per_second": 4948.831
    },
    {
      "epoch": 17.46153846153846,
      "num_input_tokens_seen": 901912,
      "step": 227,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 17.53846153846154,
      "grad_norm": 1.8580361604690552,
      "learning_rate": 0.01,
      "loss": 4.6676788330078125,
      "num_input_tokens_seen": 906008,
      "step": 228,
      "train_runtime": 183.0547,
      "train_tokens_per_second": 4949.383
    },
    {
      "epoch": 17.53846153846154,
      "num_input_tokens_seen": 906008,
      "step": 228,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 17.615384615384617,
      "grad_norm": 2.9398105144500732,
      "learning_rate": 0.01,
      "loss": 4.755415916442871,
      "num_input_tokens_seen": 910104,
      "step": 229,
      "train_runtime": 183.8622,
      "train_tokens_per_second": 4949.924
    },
    {
      "epoch": 17.615384615384617,
      "num_input_tokens_seen": 910104,
      "step": 229,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 17.692307692307693,
      "grad_norm": 2.014890193939209,
      "learning_rate": 0.01,
      "loss": 4.652470111846924,
      "num_input_tokens_seen": 914200,
      "step": 230,
      "train_runtime": 184.6689,
      "train_tokens_per_second": 4950.482
    },
    {
      "epoch": 17.692307692307693,
      "num_input_tokens_seen": 914200,
      "step": 230,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 17.76923076923077,
      "grad_norm": 3.346104621887207,
      "learning_rate": 0.01,
      "loss": 4.73101282119751,
      "num_input_tokens_seen": 918296,
      "step": 231,
      "train_runtime": 185.4748,
      "train_tokens_per_second": 4951.055
    },
    {
      "epoch": 17.76923076923077,
      "num_input_tokens_seen": 918296,
      "step": 231,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 17.846153846153847,
      "grad_norm": 2.626401901245117,
      "learning_rate": 0.01,
      "loss": 4.7478346824646,
      "num_input_tokens_seen": 922392,
      "step": 232,
      "train_runtime": 186.2809,
      "train_tokens_per_second": 4951.618
    },
    {
      "epoch": 17.846153846153847,
      "num_input_tokens_seen": 922392,
      "step": 232,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 17.923076923076923,
      "grad_norm": 2.8272528648376465,
      "learning_rate": 0.01,
      "loss": 4.785407066345215,
      "num_input_tokens_seen": 926488,
      "step": 233,
      "train_runtime": 187.0832,
      "train_tokens_per_second": 4952.278
    },
    {
      "epoch": 17.923076923076923,
      "num_input_tokens_seen": 926488,
      "step": 233,
      "train_accuracy": 0.013029315508902073
    },
    {
      "epoch": 18.0,
      "grad_norm": 2.6225979328155518,
      "learning_rate": 0.01,
      "loss": 4.660401821136475,
      "num_input_tokens_seen": 928944,
      "step": 234,
      "train_runtime": 187.5948,
      "train_tokens_per_second": 4951.865
    },
    {
      "epoch": 18.0,
      "num_input_tokens_seen": 928944,
      "step": 234,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 18.076923076923077,
      "grad_norm": 4.465658187866211,
      "learning_rate": 0.01,
      "loss": 4.832589626312256,
      "num_input_tokens_seen": 933040,
      "step": 235,
      "train_runtime": 188.4125,
      "train_tokens_per_second": 4952.114
    },
    {
      "epoch": 18.076923076923077,
      "num_input_tokens_seen": 933040,
      "step": 235,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 18.153846153846153,
      "grad_norm": 2.2223477363586426,
      "learning_rate": 0.01,
      "loss": 4.696511268615723,
      "num_input_tokens_seen": 937136,
      "step": 236,
      "train_runtime": 189.2191,
      "train_tokens_per_second": 4952.651
    },
    {
      "epoch": 18.153846153846153,
      "num_input_tokens_seen": 937136,
      "step": 236,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 18.23076923076923,
      "grad_norm": 3.487868070602417,
      "learning_rate": 0.01,
      "loss": 4.8751912117004395,
      "num_input_tokens_seen": 941232,
      "step": 237,
      "train_runtime": 190.0255,
      "train_tokens_per_second": 4953.188
    },
    {
      "epoch": 18.23076923076923,
      "num_input_tokens_seen": 941232,
      "step": 237,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 18.307692307692307,
      "grad_norm": 2.3653151988983154,
      "learning_rate": 0.01,
      "loss": 4.719097137451172,
      "num_input_tokens_seen": 945328,
      "step": 238,
      "train_runtime": 190.833,
      "train_tokens_per_second": 4953.693
    },
    {
      "epoch": 18.307692307692307,
      "num_input_tokens_seen": 945328,
      "step": 238,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 18.384615384615383,
      "grad_norm": 2.1690564155578613,
      "learning_rate": 0.01,
      "loss": 4.701546669006348,
      "num_input_tokens_seen": 949424,
      "step": 239,
      "train_runtime": 191.6405,
      "train_tokens_per_second": 4954.192
    },
    {
      "epoch": 18.384615384615383,
      "num_input_tokens_seen": 949424,
      "step": 239,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 18.46153846153846,
      "grad_norm": 2.517916679382324,
      "learning_rate": 0.01,
      "loss": 4.731206893920898,
      "num_input_tokens_seen": 953520,
      "step": 240,
      "train_runtime": 192.447,
      "train_tokens_per_second": 4954.714
    },
    {
      "epoch": 18.46153846153846,
      "num_input_tokens_seen": 953520,
      "step": 240,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 18.53846153846154,
      "grad_norm": 2.1356940269470215,
      "learning_rate": 0.01,
      "loss": 4.752324104309082,
      "num_input_tokens_seen": 957616,
      "step": 241,
      "train_runtime": 193.2532,
      "train_tokens_per_second": 4955.241
    },
    {
      "epoch": 18.53846153846154,
      "num_input_tokens_seen": 957616,
      "step": 241,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 18.615384615384617,
      "grad_norm": 2.4131526947021484,
      "learning_rate": 0.01,
      "loss": 4.7234649658203125,
      "num_input_tokens_seen": 961712,
      "step": 242,
      "train_runtime": 194.0606,
      "train_tokens_per_second": 4955.729
    },
    {
      "epoch": 18.615384615384617,
      "num_input_tokens_seen": 961712,
      "step": 242,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 18.692307692307693,
      "grad_norm": 1.9735586643218994,
      "learning_rate": 0.01,
      "loss": 4.670131683349609,
      "num_input_tokens_seen": 965808,
      "step": 243,
      "train_runtime": 194.867,
      "train_tokens_per_second": 4956.243
    },
    {
      "epoch": 18.692307692307693,
      "num_input_tokens_seen": 965808,
      "step": 243,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 18.76923076923077,
      "grad_norm": 2.581371545791626,
      "learning_rate": 0.01,
      "loss": 4.725827217102051,
      "num_input_tokens_seen": 969904,
      "step": 244,
      "train_runtime": 195.6727,
      "train_tokens_per_second": 4956.766
    },
    {
      "epoch": 18.76923076923077,
      "num_input_tokens_seen": 969904,
      "step": 244,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 18.846153846153847,
      "grad_norm": 1.9495716094970703,
      "learning_rate": 0.01,
      "loss": 4.678308963775635,
      "num_input_tokens_seen": 974000,
      "step": 245,
      "train_runtime": 196.4791,
      "train_tokens_per_second": 4957.271
    },
    {
      "epoch": 18.846153846153847,
      "num_input_tokens_seen": 974000,
      "step": 245,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 18.923076923076923,
      "grad_norm": 2.1714582443237305,
      "learning_rate": 0.01,
      "loss": 4.677157402038574,
      "num_input_tokens_seen": 978096,
      "step": 246,
      "train_runtime": 197.2814,
      "train_tokens_per_second": 4957.874
    },
    {
      "epoch": 18.923076923076923,
      "num_input_tokens_seen": 978096,
      "step": 246,
      "train_accuracy": 0.016286645084619522
    },
    {
      "epoch": 19.0,
      "grad_norm": 2.327296495437622,
      "learning_rate": 0.01,
      "loss": 4.7444257736206055,
      "num_input_tokens_seen": 980552,
      "step": 247,
      "train_runtime": 197.7923,
      "train_tokens_per_second": 4957.483
    },
    {
      "epoch": 19.0,
      "num_input_tokens_seen": 980552,
      "step": 247,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 19.076923076923077,
      "grad_norm": 2.7381880283355713,
      "learning_rate": 0.01,
      "loss": 4.704586982727051,
      "num_input_tokens_seen": 984648,
      "step": 248,
      "train_runtime": 198.6105,
      "train_tokens_per_second": 4957.683
    },
    {
      "epoch": 19.076923076923077,
      "num_input_tokens_seen": 984648,
      "step": 248,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 19.153846153846153,
      "grad_norm": 1.770967721939087,
      "learning_rate": 0.01,
      "loss": 4.686239242553711,
      "num_input_tokens_seen": 988744,
      "step": 249,
      "train_runtime": 199.4176,
      "train_tokens_per_second": 4958.159
    },
    {
      "epoch": 19.153846153846153,
      "num_input_tokens_seen": 988744,
      "step": 249,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 19.23076923076923,
      "grad_norm": 2.8549108505249023,
      "learning_rate": 0.01,
      "loss": 4.693802356719971,
      "num_input_tokens_seen": 992840,
      "step": 250,
      "train_runtime": 200.2235,
      "train_tokens_per_second": 4958.658
    },
    {
      "epoch": 19.23076923076923,
      "num_input_tokens_seen": 992840,
      "step": 250,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 19.307692307692307,
      "grad_norm": 1.802127718925476,
      "learning_rate": 0.01,
      "loss": 4.626243591308594,
      "num_input_tokens_seen": 996936,
      "step": 251,
      "train_runtime": 201.0295,
      "train_tokens_per_second": 4959.152
    },
    {
      "epoch": 19.307692307692307,
      "num_input_tokens_seen": 996936,
      "step": 251,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 19.384615384615383,
      "grad_norm": 2.309999465942383,
      "learning_rate": 0.01,
      "loss": 4.71285343170166,
      "num_input_tokens_seen": 1001032,
      "step": 252,
      "train_runtime": 201.8378,
      "train_tokens_per_second": 4959.586
    },
    {
      "epoch": 19.384615384615383,
      "num_input_tokens_seen": 1001032,
      "step": 252,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 19.46153846153846,
      "grad_norm": 2.1743581295013428,
      "learning_rate": 0.01,
      "loss": 4.666388034820557,
      "num_input_tokens_seen": 1005128,
      "step": 253,
      "train_runtime": 202.6461,
      "train_tokens_per_second": 4960.018
    },
    {
      "epoch": 19.46153846153846,
      "num_input_tokens_seen": 1005128,
      "step": 253,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 19.53846153846154,
      "grad_norm": 2.185983657836914,
      "learning_rate": 0.01,
      "loss": 4.637975692749023,
      "num_input_tokens_seen": 1009224,
      "step": 254,
      "train_runtime": 203.4532,
      "train_tokens_per_second": 4960.472
    },
    {
      "epoch": 19.53846153846154,
      "num_input_tokens_seen": 1009224,
      "step": 254,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 19.615384615384617,
      "grad_norm": 2.5004985332489014,
      "learning_rate": 0.01,
      "loss": 4.726057529449463,
      "num_input_tokens_seen": 1013320,
      "step": 255,
      "train_runtime": 204.2609,
      "train_tokens_per_second": 4960.91
    },
    {
      "epoch": 19.615384615384617,
      "num_input_tokens_seen": 1013320,
      "step": 255,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 19.692307692307693,
      "grad_norm": 1.7645424604415894,
      "learning_rate": 0.01,
      "loss": 4.638632774353027,
      "num_input_tokens_seen": 1017416,
      "step": 256,
      "train_runtime": 205.0687,
      "train_tokens_per_second": 4961.342
    },
    {
      "epoch": 19.692307692307693,
      "num_input_tokens_seen": 1017416,
      "step": 256,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 19.76923076923077,
      "grad_norm": 2.305911064147949,
      "learning_rate": 0.01,
      "loss": 4.699167251586914,
      "num_input_tokens_seen": 1021512,
      "step": 257,
      "train_runtime": 205.8757,
      "train_tokens_per_second": 4961.79
    },
    {
      "epoch": 19.76923076923077,
      "num_input_tokens_seen": 1021512,
      "step": 257,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 19.846153846153847,
      "grad_norm": 2.590080499649048,
      "learning_rate": 0.01,
      "loss": 4.735995292663574,
      "num_input_tokens_seen": 1025608,
      "step": 258,
      "train_runtime": 206.6829,
      "train_tokens_per_second": 4962.231
    },
    {
      "epoch": 19.846153846153847,
      "num_input_tokens_seen": 1025608,
      "step": 258,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 19.923076923076923,
      "grad_norm": 2.1151185035705566,
      "learning_rate": 0.01,
      "loss": 4.6779093742370605,
      "num_input_tokens_seen": 1029704,
      "step": 259,
      "train_runtime": 207.4854,
      "train_tokens_per_second": 4962.778
    },
    {
      "epoch": 19.923076923076923,
      "num_input_tokens_seen": 1029704,
      "step": 259,
      "train_accuracy": 0.0065146577544510365
    },
    {
      "epoch": 20.0,
      "grad_norm": 2.6231706142425537,
      "learning_rate": 0.01,
      "loss": 4.725683689117432,
      "num_input_tokens_seen": 1032160,
      "step": 260,
      "train_runtime": 207.9971,
      "train_tokens_per_second": 4962.377
    },
    {
      "epoch": 20.0,
      "num_input_tokens_seen": 1032160,
      "step": 260,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 20.076923076923077,
      "grad_norm": 1.8787047863006592,
      "learning_rate": 0.01,
      "loss": 4.62359619140625,
      "num_input_tokens_seen": 1036256,
      "step": 261,
      "train_runtime": 208.8157,
      "train_tokens_per_second": 4962.538
    },
    {
      "epoch": 20.076923076923077,
      "num_input_tokens_seen": 1036256,
      "step": 261,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 20.153846153846153,
      "grad_norm": 2.500467538833618,
      "learning_rate": 0.01,
      "loss": 4.736563682556152,
      "num_input_tokens_seen": 1040352,
      "step": 262,
      "train_runtime": 209.6247,
      "train_tokens_per_second": 4962.926
    },
    {
      "epoch": 20.153846153846153,
      "num_input_tokens_seen": 1040352,
      "step": 262,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 20.23076923076923,
      "grad_norm": 1.9345626831054688,
      "learning_rate": 0.01,
      "loss": 4.717474460601807,
      "num_input_tokens_seen": 1044448,
      "step": 263,
      "train_runtime": 210.4328,
      "train_tokens_per_second": 4963.332
    },
    {
      "epoch": 20.23076923076923,
      "num_input_tokens_seen": 1044448,
      "step": 263,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 20.307692307692307,
      "grad_norm": 2.108369827270508,
      "learning_rate": 0.01,
      "loss": 4.710242748260498,
      "num_input_tokens_seen": 1048544,
      "step": 264,
      "train_runtime": 211.2443,
      "train_tokens_per_second": 4963.655
    },
    {
      "epoch": 20.307692307692307,
      "num_input_tokens_seen": 1048544,
      "step": 264,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 20.384615384615383,
      "grad_norm": 1.7605409622192383,
      "learning_rate": 0.01,
      "loss": 4.638559341430664,
      "num_input_tokens_seen": 1052640,
      "step": 265,
      "train_runtime": 212.0506,
      "train_tokens_per_second": 4964.099
    },
    {
      "epoch": 20.384615384615383,
      "num_input_tokens_seen": 1052640,
      "step": 265,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 20.46153846153846,
      "grad_norm": 2.7836520671844482,
      "learning_rate": 0.01,
      "loss": 4.789233207702637,
      "num_input_tokens_seen": 1056736,
      "step": 266,
      "train_runtime": 212.8608,
      "train_tokens_per_second": 4964.446
    },
    {
      "epoch": 20.46153846153846,
      "num_input_tokens_seen": 1056736,
      "step": 266,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 20.53846153846154,
      "grad_norm": 1.9727662801742554,
      "learning_rate": 0.01,
      "loss": 4.658707618713379,
      "num_input_tokens_seen": 1060832,
      "step": 267,
      "train_runtime": 213.6681,
      "train_tokens_per_second": 4964.859
    },
    {
      "epoch": 20.53846153846154,
      "num_input_tokens_seen": 1060832,
      "step": 267,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 20.615384615384617,
      "grad_norm": 2.9147422313690186,
      "learning_rate": 0.01,
      "loss": 4.785709381103516,
      "num_input_tokens_seen": 1064928,
      "step": 268,
      "train_runtime": 214.4745,
      "train_tokens_per_second": 4965.289
    },
    {
      "epoch": 20.615384615384617,
      "num_input_tokens_seen": 1064928,
      "step": 268,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 20.692307692307693,
      "grad_norm": 2.6022112369537354,
      "learning_rate": 0.01,
      "loss": 4.660740852355957,
      "num_input_tokens_seen": 1069024,
      "step": 269,
      "train_runtime": 215.2813,
      "train_tokens_per_second": 4965.707
    },
    {
      "epoch": 20.692307692307693,
      "num_input_tokens_seen": 1069024,
      "step": 269,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 20.76923076923077,
      "grad_norm": 2.548490524291992,
      "learning_rate": 0.01,
      "loss": 4.7107157707214355,
      "num_input_tokens_seen": 1073120,
      "step": 270,
      "train_runtime": 216.0872,
      "train_tokens_per_second": 4966.143
    },
    {
      "epoch": 20.76923076923077,
      "num_input_tokens_seen": 1073120,
      "step": 270,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 20.846153846153847,
      "grad_norm": 2.851895809173584,
      "learning_rate": 0.01,
      "loss": 4.749560356140137,
      "num_input_tokens_seen": 1077216,
      "step": 271,
      "train_runtime": 216.8951,
      "train_tokens_per_second": 4966.529
    },
    {
      "epoch": 20.846153846153847,
      "num_input_tokens_seen": 1077216,
      "step": 271,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 20.923076923076923,
      "grad_norm": 2.1356582641601562,
      "learning_rate": 0.01,
      "loss": 4.705930709838867,
      "num_input_tokens_seen": 1081312,
      "step": 272,
      "train_runtime": 217.6981,
      "train_tokens_per_second": 4967.024
    },
    {
      "epoch": 20.923076923076923,
      "num_input_tokens_seen": 1081312,
      "step": 272,
      "train_accuracy": 0.0065146577544510365
    },
    {
      "epoch": 21.0,
      "grad_norm": 3.311933994293213,
      "learning_rate": 0.01,
      "loss": 4.859829902648926,
      "num_input_tokens_seen": 1083768,
      "step": 273,
      "train_runtime": 218.2094,
      "train_tokens_per_second": 4966.642
    },
    {
      "epoch": 21.0,
      "num_input_tokens_seen": 1083768,
      "step": 273,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 21.076923076923077,
      "grad_norm": 2.1029155254364014,
      "learning_rate": 0.01,
      "loss": 4.697745323181152,
      "num_input_tokens_seen": 1087864,
      "step": 274,
      "train_runtime": 219.0285,
      "train_tokens_per_second": 4966.77
    },
    {
      "epoch": 21.076923076923077,
      "num_input_tokens_seen": 1087864,
      "step": 274,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 21.153846153846153,
      "grad_norm": 2.1796224117279053,
      "learning_rate": 0.01,
      "loss": 4.717341899871826,
      "num_input_tokens_seen": 1091960,
      "step": 275,
      "train_runtime": 219.8385,
      "train_tokens_per_second": 4967.101
    },
    {
      "epoch": 21.153846153846153,
      "num_input_tokens_seen": 1091960,
      "step": 275,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 21.23076923076923,
      "grad_norm": 2.699026346206665,
      "learning_rate": 0.01,
      "loss": 4.845812797546387,
      "num_input_tokens_seen": 1096056,
      "step": 276,
      "train_runtime": 220.6475,
      "train_tokens_per_second": 4967.452
    },
    {
      "epoch": 21.23076923076923,
      "num_input_tokens_seen": 1096056,
      "step": 276,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 21.307692307692307,
      "grad_norm": 2.3528244495391846,
      "learning_rate": 0.01,
      "loss": 4.74075984954834,
      "num_input_tokens_seen": 1100152,
      "step": 277,
      "train_runtime": 221.4545,
      "train_tokens_per_second": 4967.847
    },
    {
      "epoch": 21.307692307692307,
      "num_input_tokens_seen": 1100152,
      "step": 277,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 21.384615384615383,
      "grad_norm": 2.3818016052246094,
      "learning_rate": 0.01,
      "loss": 4.771634101867676,
      "num_input_tokens_seen": 1104248,
      "step": 278,
      "train_runtime": 222.2667,
      "train_tokens_per_second": 4968.121
    },
    {
      "epoch": 21.384615384615383,
      "num_input_tokens_seen": 1104248,
      "step": 278,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 21.46153846153846,
      "grad_norm": 2.1720387935638428,
      "learning_rate": 0.01,
      "loss": 4.709393501281738,
      "num_input_tokens_seen": 1108344,
      "step": 279,
      "train_runtime": 223.0734,
      "train_tokens_per_second": 4968.517
    },
    {
      "epoch": 21.46153846153846,
      "num_input_tokens_seen": 1108344,
      "step": 279,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 21.53846153846154,
      "grad_norm": 1.855604648590088,
      "learning_rate": 0.01,
      "loss": 4.684809684753418,
      "num_input_tokens_seen": 1112440,
      "step": 280,
      "train_runtime": 223.88,
      "train_tokens_per_second": 4968.913
    },
    {
      "epoch": 21.53846153846154,
      "num_input_tokens_seen": 1112440,
      "step": 280,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 21.615384615384617,
      "grad_norm": 2.3128485679626465,
      "learning_rate": 0.01,
      "loss": 4.782155990600586,
      "num_input_tokens_seen": 1116536,
      "step": 281,
      "train_runtime": 224.6864,
      "train_tokens_per_second": 4969.307
    },
    {
      "epoch": 21.615384615384617,
      "num_input_tokens_seen": 1116536,
      "step": 281,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 21.692307692307693,
      "grad_norm": 2.5007569789886475,
      "learning_rate": 0.01,
      "loss": 4.736386775970459,
      "num_input_tokens_seen": 1120632,
      "step": 282,
      "train_runtime": 225.4937,
      "train_tokens_per_second": 4969.682
    },
    {
      "epoch": 21.692307692307693,
      "num_input_tokens_seen": 1120632,
      "step": 282,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 21.76923076923077,
      "grad_norm": 1.760748267173767,
      "learning_rate": 0.01,
      "loss": 4.67552375793457,
      "num_input_tokens_seen": 1124728,
      "step": 283,
      "train_runtime": 226.3016,
      "train_tokens_per_second": 4970.039
    },
    {
      "epoch": 21.76923076923077,
      "num_input_tokens_seen": 1124728,
      "step": 283,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 21.846153846153847,
      "grad_norm": 2.350501775741577,
      "learning_rate": 0.01,
      "loss": 4.71092414855957,
      "num_input_tokens_seen": 1128824,
      "step": 284,
      "train_runtime": 227.1093,
      "train_tokens_per_second": 4970.4
    },
    {
      "epoch": 21.846153846153847,
      "num_input_tokens_seen": 1128824,
      "step": 284,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 21.923076923076923,
      "grad_norm": 2.3892035484313965,
      "learning_rate": 0.01,
      "loss": 4.7376708984375,
      "num_input_tokens_seen": 1132920,
      "step": 285,
      "train_runtime": 227.9131,
      "train_tokens_per_second": 4970.843
    },
    {
      "epoch": 21.923076923076923,
      "num_input_tokens_seen": 1132920,
      "step": 285,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 22.0,
      "grad_norm": 3.1911215782165527,
      "learning_rate": 0.01,
      "loss": 4.834023952484131,
      "num_input_tokens_seen": 1135376,
      "step": 286,
      "train_runtime": 228.4242,
      "train_tokens_per_second": 4970.473
    },
    {
      "epoch": 22.0,
      "num_input_tokens_seen": 1135376,
      "step": 286,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 22.076923076923077,
      "grad_norm": 2.4082560539245605,
      "learning_rate": 0.01,
      "loss": 4.777586936950684,
      "num_input_tokens_seen": 1139472,
      "step": 287,
      "train_runtime": 229.2428,
      "train_tokens_per_second": 4970.59
    },
    {
      "epoch": 22.076923076923077,
      "num_input_tokens_seen": 1139472,
      "step": 287,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 22.153846153846153,
      "grad_norm": 2.413447618484497,
      "learning_rate": 0.01,
      "loss": 4.754624843597412,
      "num_input_tokens_seen": 1143568,
      "step": 288,
      "train_runtime": 230.0496,
      "train_tokens_per_second": 4970.962
    },
    {
      "epoch": 22.153846153846153,
      "num_input_tokens_seen": 1143568,
      "step": 288,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 22.23076923076923,
      "grad_norm": 2.6883504390716553,
      "learning_rate": 0.01,
      "loss": 4.817875385284424,
      "num_input_tokens_seen": 1147664,
      "step": 289,
      "train_runtime": 230.858,
      "train_tokens_per_second": 4971.299
    },
    {
      "epoch": 22.23076923076923,
      "num_input_tokens_seen": 1147664,
      "step": 289,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 22.307692307692307,
      "grad_norm": 3.0178632736206055,
      "learning_rate": 0.01,
      "loss": 4.860883712768555,
      "num_input_tokens_seen": 1151760,
      "step": 290,
      "train_runtime": 231.6672,
      "train_tokens_per_second": 4971.614
    },
    {
      "epoch": 22.307692307692307,
      "num_input_tokens_seen": 1151760,
      "step": 290,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 22.384615384615383,
      "grad_norm": 4.1211090087890625,
      "learning_rate": 0.01,
      "loss": 4.9048357009887695,
      "num_input_tokens_seen": 1155856,
      "step": 291,
      "train_runtime": 232.4733,
      "train_tokens_per_second": 4971.994
    },
    {
      "epoch": 22.384615384615383,
      "num_input_tokens_seen": 1155856,
      "step": 291,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 22.46153846153846,
      "grad_norm": 2.780254602432251,
      "learning_rate": 0.01,
      "loss": 4.758027076721191,
      "num_input_tokens_seen": 1159952,
      "step": 292,
      "train_runtime": 233.6182,
      "train_tokens_per_second": 4965.161
    },
    {
      "epoch": 22.46153846153846,
      "num_input_tokens_seen": 1159952,
      "step": 292,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 22.53846153846154,
      "grad_norm": 2.5430874824523926,
      "learning_rate": 0.01,
      "loss": 4.874720096588135,
      "num_input_tokens_seen": 1164048,
      "step": 293,
      "train_runtime": 234.4245,
      "train_tokens_per_second": 4965.556
    },
    {
      "epoch": 22.53846153846154,
      "num_input_tokens_seen": 1164048,
      "step": 293,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 22.615384615384617,
      "grad_norm": 2.681941032409668,
      "learning_rate": 0.01,
      "loss": 4.79051399230957,
      "num_input_tokens_seen": 1168144,
      "step": 294,
      "train_runtime": 235.232,
      "train_tokens_per_second": 4965.922
    },
    {
      "epoch": 22.615384615384617,
      "num_input_tokens_seen": 1168144,
      "step": 294,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 22.692307692307693,
      "grad_norm": 2.2946882247924805,
      "learning_rate": 0.01,
      "loss": 4.760282516479492,
      "num_input_tokens_seen": 1172240,
      "step": 295,
      "train_runtime": 236.0396,
      "train_tokens_per_second": 4966.286
    },
    {
      "epoch": 22.692307692307693,
      "num_input_tokens_seen": 1172240,
      "step": 295,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 22.76923076923077,
      "grad_norm": 2.588116407394409,
      "learning_rate": 0.01,
      "loss": 4.80474853515625,
      "num_input_tokens_seen": 1176336,
      "step": 296,
      "train_runtime": 236.8462,
      "train_tokens_per_second": 4966.666
    },
    {
      "epoch": 22.76923076923077,
      "num_input_tokens_seen": 1176336,
      "step": 296,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 22.846153846153847,
      "grad_norm": 2.5479612350463867,
      "learning_rate": 0.01,
      "loss": 4.77901029586792,
      "num_input_tokens_seen": 1180432,
      "step": 297,
      "train_runtime": 237.6532,
      "train_tokens_per_second": 4967.036
    },
    {
      "epoch": 22.846153846153847,
      "num_input_tokens_seen": 1180432,
      "step": 297,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 22.923076923076923,
      "grad_norm": 2.49416446685791,
      "learning_rate": 0.01,
      "loss": 4.781351089477539,
      "num_input_tokens_seen": 1184528,
      "step": 298,
      "train_runtime": 238.4562,
      "train_tokens_per_second": 4967.487
    },
    {
      "epoch": 22.923076923076923,
      "num_input_tokens_seen": 1184528,
      "step": 298,
      "train_accuracy": 0.0032573288772255182
    },
    {
      "epoch": 23.0,
      "grad_norm": 2.447633743286133,
      "learning_rate": 0.01,
      "loss": 4.748744487762451,
      "num_input_tokens_seen": 1186984,
      "step": 299,
      "train_runtime": 238.9676,
      "train_tokens_per_second": 4967.134
    },
    {
      "epoch": 23.0,
      "num_input_tokens_seen": 1186984,
      "step": 299,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 23.076923076923077,
      "grad_norm": 3.9213333129882812,
      "learning_rate": 0.01,
      "loss": 4.831058502197266,
      "num_input_tokens_seen": 1191080,
      "step": 300,
      "train_runtime": 239.7857,
      "train_tokens_per_second": 4967.269
    },
    {
      "epoch": 23.076923076923077,
      "eval_CMD_3_branch_eval_accuracy": 0.015189873417721518,
      "eval_CMD_3_branch_eval_loss": 4.984875679016113,
      "eval_CMD_3_branch_eval_runtime": 1.1324,
      "eval_CMD_3_branch_eval_samples_per_second": 2441.65,
      "eval_CMD_3_branch_eval_steps_per_second": 5.298,
      "num_input_tokens_seen": 1191080,
      "step": 300
    },
    {
      "epoch": 23.076923076923077,
      "num_input_tokens_seen": 1191080,
      "step": 300,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 23.153846153846153,
      "grad_norm": 3.7104413509368896,
      "learning_rate": 0.01,
      "loss": 4.880445957183838,
      "num_input_tokens_seen": 1195176,
      "step": 301,
      "train_runtime": 241.7296,
      "train_tokens_per_second": 4944.269
    },
    {
      "epoch": 23.153846153846153,
      "num_input_tokens_seen": 1195176,
      "step": 301,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 23.23076923076923,
      "grad_norm": 2.4190526008605957,
      "learning_rate": 0.01,
      "loss": 4.79976224899292,
      "num_input_tokens_seen": 1199272,
      "step": 302,
      "train_runtime": 242.5356,
      "train_tokens_per_second": 4944.726
    },
    {
      "epoch": 23.23076923076923,
      "num_input_tokens_seen": 1199272,
      "step": 302,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 23.307692307692307,
      "grad_norm": 2.964792013168335,
      "learning_rate": 0.01,
      "loss": 4.900379657745361,
      "num_input_tokens_seen": 1203368,
      "step": 303,
      "train_runtime": 243.3473,
      "train_tokens_per_second": 4945.065
    },
    {
      "epoch": 23.307692307692307,
      "num_input_tokens_seen": 1203368,
      "step": 303,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 23.384615384615383,
      "grad_norm": 2.890063762664795,
      "learning_rate": 0.01,
      "loss": 4.858273029327393,
      "num_input_tokens_seen": 1207464,
      "step": 304,
      "train_runtime": 244.1551,
      "train_tokens_per_second": 4945.48
    },
    {
      "epoch": 23.384615384615383,
      "num_input_tokens_seen": 1207464,
      "step": 304,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 23.46153846153846,
      "grad_norm": 2.7856032848358154,
      "learning_rate": 0.01,
      "loss": 4.86794376373291,
      "num_input_tokens_seen": 1211560,
      "step": 305,
      "train_runtime": 244.9621,
      "train_tokens_per_second": 4945.908
    },
    {
      "epoch": 23.46153846153846,
      "num_input_tokens_seen": 1211560,
      "step": 305,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 23.53846153846154,
      "grad_norm": 3.326423406600952,
      "learning_rate": 0.01,
      "loss": 4.906798362731934,
      "num_input_tokens_seen": 1215656,
      "step": 306,
      "train_runtime": 245.7683,
      "train_tokens_per_second": 4946.35
    },
    {
      "epoch": 23.53846153846154,
      "num_input_tokens_seen": 1215656,
      "step": 306,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 23.615384615384617,
      "grad_norm": 2.812473773956299,
      "learning_rate": 0.01,
      "loss": 4.856661796569824,
      "num_input_tokens_seen": 1219752,
      "step": 307,
      "train_runtime": 246.5753,
      "train_tokens_per_second": 4946.772
    },
    {
      "epoch": 23.615384615384617,
      "num_input_tokens_seen": 1219752,
      "step": 307,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 23.692307692307693,
      "grad_norm": 3.1735622882843018,
      "learning_rate": 0.01,
      "loss": 4.842247009277344,
      "num_input_tokens_seen": 1223848,
      "step": 308,
      "train_runtime": 247.3828,
      "train_tokens_per_second": 4947.183
    },
    {
      "epoch": 23.692307692307693,
      "num_input_tokens_seen": 1223848,
      "step": 308,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 23.76923076923077,
      "grad_norm": 1.9832615852355957,
      "learning_rate": 0.01,
      "loss": 4.670631408691406,
      "num_input_tokens_seen": 1227944,
      "step": 309,
      "train_runtime": 248.1893,
      "train_tokens_per_second": 4947.61
    },
    {
      "epoch": 23.76923076923077,
      "num_input_tokens_seen": 1227944,
      "step": 309,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 23.846153846153847,
      "grad_norm": 3.3848817348480225,
      "learning_rate": 0.01,
      "loss": 4.795812606811523,
      "num_input_tokens_seen": 1232040,
      "step": 310,
      "train_runtime": 248.9957,
      "train_tokens_per_second": 4948.038
    },
    {
      "epoch": 23.846153846153847,
      "num_input_tokens_seen": 1232040,
      "step": 310,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 23.923076923076923,
      "grad_norm": 2.17008638381958,
      "learning_rate": 0.01,
      "loss": 4.756107330322266,
      "num_input_tokens_seen": 1236136,
      "step": 311,
      "train_runtime": 249.7978,
      "train_tokens_per_second": 4948.547
    },
    {
      "epoch": 23.923076923076923,
      "num_input_tokens_seen": 1236136,
      "step": 311,
      "train_accuracy": 0.019543973729014397
    },
    {
      "epoch": 24.0,
      "grad_norm": 2.9384548664093018,
      "learning_rate": 0.01,
      "loss": 4.840476989746094,
      "num_input_tokens_seen": 1238592,
      "step": 312,
      "train_runtime": 250.3091,
      "train_tokens_per_second": 4948.249
    },
    {
      "epoch": 24.0,
      "num_input_tokens_seen": 1238592,
      "step": 312,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 24.076923076923077,
      "grad_norm": 2.095926523208618,
      "learning_rate": 0.01,
      "loss": 4.775547027587891,
      "num_input_tokens_seen": 1242688,
      "step": 313,
      "train_runtime": 251.1277,
      "train_tokens_per_second": 4948.431
    },
    {
      "epoch": 24.076923076923077,
      "num_input_tokens_seen": 1242688,
      "step": 313,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 24.153846153846153,
      "grad_norm": 2.1713414192199707,
      "learning_rate": 0.01,
      "loss": 4.785286903381348,
      "num_input_tokens_seen": 1246784,
      "step": 314,
      "train_runtime": 251.9347,
      "train_tokens_per_second": 4948.838
    },
    {
      "epoch": 24.153846153846153,
      "num_input_tokens_seen": 1246784,
      "step": 314,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 24.23076923076923,
      "grad_norm": 2.2279763221740723,
      "learning_rate": 0.01,
      "loss": 4.751096248626709,
      "num_input_tokens_seen": 1250880,
      "step": 315,
      "train_runtime": 252.7423,
      "train_tokens_per_second": 4949.23
    },
    {
      "epoch": 24.23076923076923,
      "num_input_tokens_seen": 1250880,
      "step": 315,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 24.307692307692307,
      "grad_norm": 2.8324530124664307,
      "learning_rate": 0.01,
      "loss": 4.821085453033447,
      "num_input_tokens_seen": 1254976,
      "step": 316,
      "train_runtime": 253.55,
      "train_tokens_per_second": 4949.62
    },
    {
      "epoch": 24.307692307692307,
      "num_input_tokens_seen": 1254976,
      "step": 316,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 24.384615384615383,
      "grad_norm": 2.314676523208618,
      "learning_rate": 0.01,
      "loss": 4.786411285400391,
      "num_input_tokens_seen": 1259072,
      "step": 317,
      "train_runtime": 254.3573,
      "train_tokens_per_second": 4950.014
    },
    {
      "epoch": 24.384615384615383,
      "num_input_tokens_seen": 1259072,
      "step": 317,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 24.46153846153846,
      "grad_norm": 2.2938666343688965,
      "learning_rate": 0.01,
      "loss": 4.797077178955078,
      "num_input_tokens_seen": 1263168,
      "step": 318,
      "train_runtime": 255.1637,
      "train_tokens_per_second": 4950.422
    },
    {
      "epoch": 24.46153846153846,
      "num_input_tokens_seen": 1263168,
      "step": 318,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 24.53846153846154,
      "grad_norm": 2.1615288257598877,
      "learning_rate": 0.01,
      "loss": 4.717720985412598,
      "num_input_tokens_seen": 1267264,
      "step": 319,
      "train_runtime": 255.9706,
      "train_tokens_per_second": 4950.819
    },
    {
      "epoch": 24.53846153846154,
      "num_input_tokens_seen": 1267264,
      "step": 319,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 24.615384615384617,
      "grad_norm": 2.3392107486724854,
      "learning_rate": 0.01,
      "loss": 4.733541488647461,
      "num_input_tokens_seen": 1271360,
      "step": 320,
      "train_runtime": 256.7778,
      "train_tokens_per_second": 4951.207
    },
    {
      "epoch": 24.615384615384617,
      "num_input_tokens_seen": 1271360,
      "step": 320,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 24.692307692307693,
      "grad_norm": 2.0798051357269287,
      "learning_rate": 0.01,
      "loss": 4.784616470336914,
      "num_input_tokens_seen": 1275456,
      "step": 321,
      "train_runtime": 257.5861,
      "train_tokens_per_second": 4951.571
    },
    {
      "epoch": 24.692307692307693,
      "num_input_tokens_seen": 1275456,
      "step": 321,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 24.76923076923077,
      "grad_norm": 1.7728190422058105,
      "learning_rate": 0.01,
      "loss": 4.739435195922852,
      "num_input_tokens_seen": 1279552,
      "step": 322,
      "train_runtime": 258.3936,
      "train_tokens_per_second": 4951.949
    },
    {
      "epoch": 24.76923076923077,
      "num_input_tokens_seen": 1279552,
      "step": 322,
      "train_accuracy": 0.0
    },
    {
      "epoch": 24.846153846153847,
      "grad_norm": 2.3809309005737305,
      "learning_rate": 0.01,
      "loss": 4.806252479553223,
      "num_input_tokens_seen": 1283648,
      "step": 323,
      "train_runtime": 259.201,
      "train_tokens_per_second": 4952.326
    },
    {
      "epoch": 24.846153846153847,
      "num_input_tokens_seen": 1283648,
      "step": 323,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 24.923076923076923,
      "grad_norm": 2.374556064605713,
      "learning_rate": 0.01,
      "loss": 4.791336536407471,
      "num_input_tokens_seen": 1287744,
      "step": 324,
      "train_runtime": 260.005,
      "train_tokens_per_second": 4952.766
    },
    {
      "epoch": 24.923076923076923,
      "num_input_tokens_seen": 1287744,
      "step": 324,
      "train_accuracy": 0.0065146577544510365
    },
    {
      "epoch": 25.0,
      "grad_norm": 2.228682518005371,
      "learning_rate": 0.01,
      "loss": 4.765847206115723,
      "num_input_tokens_seen": 1290200,
      "step": 325,
      "train_runtime": 260.5169,
      "train_tokens_per_second": 4952.463
    },
    {
      "epoch": 25.0,
      "num_input_tokens_seen": 1290200,
      "step": 325,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 25.076923076923077,
      "grad_norm": 2.6056597232818604,
      "learning_rate": 0.01,
      "loss": 4.897510528564453,
      "num_input_tokens_seen": 1294296,
      "step": 326,
      "train_runtime": 261.336,
      "train_tokens_per_second": 4952.612
    },
    {
      "epoch": 25.076923076923077,
      "num_input_tokens_seen": 1294296,
      "step": 326,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 25.153846153846153,
      "grad_norm": 2.3389954566955566,
      "learning_rate": 0.01,
      "loss": 4.76637077331543,
      "num_input_tokens_seen": 1298392,
      "step": 327,
      "train_runtime": 262.1438,
      "train_tokens_per_second": 4952.977
    },
    {
      "epoch": 25.153846153846153,
      "num_input_tokens_seen": 1298392,
      "step": 327,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 25.23076923076923,
      "grad_norm": 2.6925578117370605,
      "learning_rate": 0.01,
      "loss": 4.770495891571045,
      "num_input_tokens_seen": 1302488,
      "step": 328,
      "train_runtime": 262.9515,
      "train_tokens_per_second": 4953.34
    },
    {
      "epoch": 25.23076923076923,
      "num_input_tokens_seen": 1302488,
      "step": 328,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 25.307692307692307,
      "grad_norm": 2.749713659286499,
      "learning_rate": 0.01,
      "loss": 4.881984710693359,
      "num_input_tokens_seen": 1306584,
      "step": 329,
      "train_runtime": 263.7588,
      "train_tokens_per_second": 4953.708
    },
    {
      "epoch": 25.307692307692307,
      "num_input_tokens_seen": 1306584,
      "step": 329,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 25.384615384615383,
      "grad_norm": 2.8878915309906006,
      "learning_rate": 0.01,
      "loss": 4.886727333068848,
      "num_input_tokens_seen": 1310680,
      "step": 330,
      "train_runtime": 264.5658,
      "train_tokens_per_second": 4954.08
    },
    {
      "epoch": 25.384615384615383,
      "num_input_tokens_seen": 1310680,
      "step": 330,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 25.46153846153846,
      "grad_norm": 3.6340856552124023,
      "learning_rate": 0.01,
      "loss": 4.835007190704346,
      "num_input_tokens_seen": 1314776,
      "step": 331,
      "train_runtime": 265.3725,
      "train_tokens_per_second": 4954.456
    },
    {
      "epoch": 25.46153846153846,
      "num_input_tokens_seen": 1314776,
      "step": 331,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 25.53846153846154,
      "grad_norm": 3.042142391204834,
      "learning_rate": 0.01,
      "loss": 4.860703468322754,
      "num_input_tokens_seen": 1318872,
      "step": 332,
      "train_runtime": 266.1802,
      "train_tokens_per_second": 4954.81
    },
    {
      "epoch": 25.53846153846154,
      "num_input_tokens_seen": 1318872,
      "step": 332,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 25.615384615384617,
      "grad_norm": 2.693301200866699,
      "learning_rate": 0.01,
      "loss": 4.840606212615967,
      "num_input_tokens_seen": 1322968,
      "step": 333,
      "train_runtime": 266.9876,
      "train_tokens_per_second": 4955.167
    },
    {
      "epoch": 25.615384615384617,
      "num_input_tokens_seen": 1322968,
      "step": 333,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 25.692307692307693,
      "grad_norm": 2.5403964519500732,
      "learning_rate": 0.01,
      "loss": 4.798892498016357,
      "num_input_tokens_seen": 1327064,
      "step": 334,
      "train_runtime": 267.7941,
      "train_tokens_per_second": 4955.538
    },
    {
      "epoch": 25.692307692307693,
      "num_input_tokens_seen": 1327064,
      "step": 334,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 25.76923076923077,
      "grad_norm": 2.2730178833007812,
      "learning_rate": 0.01,
      "loss": 4.782463073730469,
      "num_input_tokens_seen": 1331160,
      "step": 335,
      "train_runtime": 268.6006,
      "train_tokens_per_second": 4955.908
    },
    {
      "epoch": 25.76923076923077,
      "num_input_tokens_seen": 1331160,
      "step": 335,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 25.846153846153847,
      "grad_norm": 2.9360063076019287,
      "learning_rate": 0.01,
      "loss": 4.909140110015869,
      "num_input_tokens_seen": 1335256,
      "step": 336,
      "train_runtime": 269.4078,
      "train_tokens_per_second": 4956.263
    },
    {
      "epoch": 25.846153846153847,
      "num_input_tokens_seen": 1335256,
      "step": 336,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 25.923076923076923,
      "grad_norm": 2.116119861602783,
      "learning_rate": 0.01,
      "loss": 4.812901496887207,
      "num_input_tokens_seen": 1339352,
      "step": 337,
      "train_runtime": 270.2115,
      "train_tokens_per_second": 4956.681
    },
    {
      "epoch": 25.923076923076923,
      "num_input_tokens_seen": 1339352,
      "step": 337,
      "train_accuracy": 0.02280130237340927
    },
    {
      "epoch": 26.0,
      "grad_norm": 3.5643439292907715,
      "learning_rate": 0.01,
      "loss": 4.895590305328369,
      "num_input_tokens_seen": 1341808,
      "step": 338,
      "train_runtime": 270.7232,
      "train_tokens_per_second": 4956.384
    },
    {
      "epoch": 26.0,
      "num_input_tokens_seen": 1341808,
      "step": 338,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 26.076923076923077,
      "grad_norm": 4.132648468017578,
      "learning_rate": 0.01,
      "loss": 5.072445869445801,
      "num_input_tokens_seen": 1345904,
      "step": 339,
      "train_runtime": 271.5418,
      "train_tokens_per_second": 4956.526
    },
    {
      "epoch": 26.076923076923077,
      "num_input_tokens_seen": 1345904,
      "step": 339,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 26.153846153846153,
      "grad_norm": 1.9905977249145508,
      "learning_rate": 0.01,
      "loss": 4.787455081939697,
      "num_input_tokens_seen": 1350000,
      "step": 340,
      "train_runtime": 272.3495,
      "train_tokens_per_second": 4956.867
    },
    {
      "epoch": 26.153846153846153,
      "num_input_tokens_seen": 1350000,
      "step": 340,
      "train_accuracy": 0.001953125
    },
    {
      "epoch": 26.23076923076923,
      "grad_norm": 2.674907684326172,
      "learning_rate": 0.01,
      "loss": 4.856403350830078,
      "num_input_tokens_seen": 1354096,
      "step": 341,
      "train_runtime": 273.1569,
      "train_tokens_per_second": 4957.21
    },
    {
      "epoch": 26.23076923076923,
      "num_input_tokens_seen": 1354096,
      "step": 341,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 26.307692307692307,
      "grad_norm": 1.9412745237350464,
      "learning_rate": 0.01,
      "loss": 4.798261642456055,
      "num_input_tokens_seen": 1358192,
      "step": 342,
      "train_runtime": 273.9636,
      "train_tokens_per_second": 4957.563
    },
    {
      "epoch": 26.307692307692307,
      "num_input_tokens_seen": 1358192,
      "step": 342,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 26.384615384615383,
      "grad_norm": 2.8431615829467773,
      "learning_rate": 0.01,
      "loss": 4.952361583709717,
      "num_input_tokens_seen": 1362288,
      "step": 343,
      "train_runtime": 274.7702,
      "train_tokens_per_second": 4957.917
    },
    {
      "epoch": 26.384615384615383,
      "num_input_tokens_seen": 1362288,
      "step": 343,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 26.46153846153846,
      "grad_norm": 2.662569761276245,
      "learning_rate": 0.01,
      "loss": 4.8351850509643555,
      "num_input_tokens_seen": 1366384,
      "step": 344,
      "train_runtime": 275.5778,
      "train_tokens_per_second": 4958.251
    },
    {
      "epoch": 26.46153846153846,
      "num_input_tokens_seen": 1366384,
      "step": 344,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 26.53846153846154,
      "grad_norm": 2.3961122035980225,
      "learning_rate": 0.01,
      "loss": 4.846569061279297,
      "num_input_tokens_seen": 1370480,
      "step": 345,
      "train_runtime": 276.3857,
      "train_tokens_per_second": 4958.578
    },
    {
      "epoch": 26.53846153846154,
      "num_input_tokens_seen": 1370480,
      "step": 345,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 26.615384615384617,
      "grad_norm": 2.540269136428833,
      "learning_rate": 0.01,
      "loss": 4.879981994628906,
      "num_input_tokens_seen": 1374576,
      "step": 346,
      "train_runtime": 277.1925,
      "train_tokens_per_second": 4958.922
    },
    {
      "epoch": 26.615384615384617,
      "num_input_tokens_seen": 1374576,
      "step": 346,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 26.692307692307693,
      "grad_norm": 2.8331077098846436,
      "learning_rate": 0.01,
      "loss": 4.893893718719482,
      "num_input_tokens_seen": 1378672,
      "step": 347,
      "train_runtime": 277.9992,
      "train_tokens_per_second": 4959.266
    },
    {
      "epoch": 26.692307692307693,
      "num_input_tokens_seen": 1378672,
      "step": 347,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 26.76923076923077,
      "grad_norm": 3.754427194595337,
      "learning_rate": 0.01,
      "loss": 5.022141933441162,
      "num_input_tokens_seen": 1382768,
      "step": 348,
      "train_runtime": 278.8062,
      "train_tokens_per_second": 4959.603
    },
    {
      "epoch": 26.76923076923077,
      "num_input_tokens_seen": 1382768,
      "step": 348,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 26.846153846153847,
      "grad_norm": 3.9744033813476562,
      "learning_rate": 0.01,
      "loss": 5.177316665649414,
      "num_input_tokens_seen": 1386864,
      "step": 349,
      "train_runtime": 279.6137,
      "train_tokens_per_second": 4959.928
    },
    {
      "epoch": 26.846153846153847,
      "num_input_tokens_seen": 1386864,
      "step": 349,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 26.923076923076923,
      "grad_norm": 3.872500419616699,
      "learning_rate": 0.01,
      "loss": 5.195061206817627,
      "num_input_tokens_seen": 1390960,
      "step": 350,
      "train_runtime": 280.4168,
      "train_tokens_per_second": 4960.331
    },
    {
      "epoch": 26.923076923076923,
      "num_input_tokens_seen": 1390960,
      "step": 350,
      "train_accuracy": 0.0032573288772255182
    },
    {
      "epoch": 27.0,
      "grad_norm": 3.184382200241089,
      "learning_rate": 0.01,
      "loss": 5.033001899719238,
      "num_input_tokens_seen": 1393416,
      "step": 351,
      "train_runtime": 280.9292,
      "train_tokens_per_second": 4960.025
    },
    {
      "epoch": 27.0,
      "num_input_tokens_seen": 1393416,
      "step": 351,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 27.076923076923077,
      "grad_norm": 5.33905553817749,
      "learning_rate": 0.01,
      "loss": 5.47437858581543,
      "num_input_tokens_seen": 1397512,
      "step": 352,
      "train_runtime": 281.7478,
      "train_tokens_per_second": 4960.153
    },
    {
      "epoch": 27.076923076923077,
      "num_input_tokens_seen": 1397512,
      "step": 352,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 27.153846153846153,
      "grad_norm": 5.295461654663086,
      "learning_rate": 0.01,
      "loss": 5.488711357116699,
      "num_input_tokens_seen": 1401608,
      "step": 353,
      "train_runtime": 282.5549,
      "train_tokens_per_second": 4960.481
    },
    {
      "epoch": 27.153846153846153,
      "num_input_tokens_seen": 1401608,
      "step": 353,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 27.23076923076923,
      "grad_norm": 5.008157730102539,
      "learning_rate": 0.01,
      "loss": 5.523514270782471,
      "num_input_tokens_seen": 1405704,
      "step": 354,
      "train_runtime": 283.3614,
      "train_tokens_per_second": 4960.817
    },
    {
      "epoch": 27.23076923076923,
      "num_input_tokens_seen": 1405704,
      "step": 354,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 27.307692307692307,
      "grad_norm": 2.9176597595214844,
      "learning_rate": 0.01,
      "loss": 5.203047275543213,
      "num_input_tokens_seen": 1409800,
      "step": 355,
      "train_runtime": 284.1688,
      "train_tokens_per_second": 4961.136
    },
    {
      "epoch": 27.307692307692307,
      "num_input_tokens_seen": 1409800,
      "step": 355,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 27.384615384615383,
      "grad_norm": 3.083993673324585,
      "learning_rate": 0.01,
      "loss": 5.156909942626953,
      "num_input_tokens_seen": 1413896,
      "step": 356,
      "train_runtime": 284.978,
      "train_tokens_per_second": 4961.422
    },
    {
      "epoch": 27.384615384615383,
      "num_input_tokens_seen": 1413896,
      "step": 356,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 27.46153846153846,
      "grad_norm": 3.0954196453094482,
      "learning_rate": 0.01,
      "loss": 5.153910160064697,
      "num_input_tokens_seen": 1417992,
      "step": 357,
      "train_runtime": 285.7855,
      "train_tokens_per_second": 4961.736
    },
    {
      "epoch": 27.46153846153846,
      "num_input_tokens_seen": 1417992,
      "step": 357,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 27.53846153846154,
      "grad_norm": 3.142883777618408,
      "learning_rate": 0.01,
      "loss": 5.057559967041016,
      "num_input_tokens_seen": 1422088,
      "step": 358,
      "train_runtime": 286.5968,
      "train_tokens_per_second": 4961.981
    },
    {
      "epoch": 27.53846153846154,
      "num_input_tokens_seen": 1422088,
      "step": 358,
      "train_accuracy": 0.001953125
    },
    {
      "epoch": 27.615384615384617,
      "grad_norm": 3.245189905166626,
      "learning_rate": 0.01,
      "loss": 5.013225555419922,
      "num_input_tokens_seen": 1426184,
      "step": 359,
      "train_runtime": 287.4051,
      "train_tokens_per_second": 4962.278
    },
    {
      "epoch": 27.615384615384617,
      "num_input_tokens_seen": 1426184,
      "step": 359,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 27.692307692307693,
      "grad_norm": 2.5022335052490234,
      "learning_rate": 0.01,
      "loss": 4.944221496582031,
      "num_input_tokens_seen": 1430280,
      "step": 360,
      "train_runtime": 288.2127,
      "train_tokens_per_second": 4962.585
    },
    {
      "epoch": 27.692307692307693,
      "num_input_tokens_seen": 1430280,
      "step": 360,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 27.76923076923077,
      "grad_norm": 2.206346273422241,
      "learning_rate": 0.01,
      "loss": 4.890231132507324,
      "num_input_tokens_seen": 1434376,
      "step": 361,
      "train_runtime": 289.0203,
      "train_tokens_per_second": 4962.891
    },
    {
      "epoch": 27.76923076923077,
      "num_input_tokens_seen": 1434376,
      "step": 361,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 27.846153846153847,
      "grad_norm": 4.545768737792969,
      "learning_rate": 0.01,
      "loss": 5.098722457885742,
      "num_input_tokens_seen": 1438472,
      "step": 362,
      "train_runtime": 289.8279,
      "train_tokens_per_second": 4963.194
    },
    {
      "epoch": 27.846153846153847,
      "num_input_tokens_seen": 1438472,
      "step": 362,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 27.923076923076923,
      "grad_norm": 5.130817413330078,
      "learning_rate": 0.01,
      "loss": 5.554351806640625,
      "num_input_tokens_seen": 1442568,
      "step": 363,
      "train_runtime": 290.6321,
      "train_tokens_per_second": 4963.554
    },
    {
      "epoch": 27.923076923076923,
      "num_input_tokens_seen": 1442568,
      "step": 363,
      "train_accuracy": 0.0032573288772255182
    },
    {
      "epoch": 28.0,
      "grad_norm": 6.362403392791748,
      "learning_rate": 0.01,
      "loss": 5.701359748840332,
      "num_input_tokens_seen": 1445024,
      "step": 364,
      "train_runtime": 291.1439,
      "train_tokens_per_second": 4963.264
    },
    {
      "epoch": 28.0,
      "num_input_tokens_seen": 1445024,
      "step": 364,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 28.076923076923077,
      "grad_norm": 2.791771650314331,
      "learning_rate": 0.01,
      "loss": 5.103519439697266,
      "num_input_tokens_seen": 1449120,
      "step": 365,
      "train_runtime": 291.9634,
      "train_tokens_per_second": 4963.362
    },
    {
      "epoch": 28.076923076923077,
      "num_input_tokens_seen": 1449120,
      "step": 365,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 28.153846153846153,
      "grad_norm": 3.010110378265381,
      "learning_rate": 0.01,
      "loss": 5.303025245666504,
      "num_input_tokens_seen": 1453216,
      "step": 366,
      "train_runtime": 292.7707,
      "train_tokens_per_second": 4963.666
    },
    {
      "epoch": 28.153846153846153,
      "num_input_tokens_seen": 1453216,
      "step": 366,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 28.23076923076923,
      "grad_norm": 2.968033790588379,
      "learning_rate": 0.01,
      "loss": 5.217446804046631,
      "num_input_tokens_seen": 1457312,
      "step": 367,
      "train_runtime": 293.5782,
      "train_tokens_per_second": 4963.966
    },
    {
      "epoch": 28.23076923076923,
      "num_input_tokens_seen": 1457312,
      "step": 367,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 28.307692307692307,
      "grad_norm": 3.1182174682617188,
      "learning_rate": 0.01,
      "loss": 5.131748199462891,
      "num_input_tokens_seen": 1461408,
      "step": 368,
      "train_runtime": 294.3857,
      "train_tokens_per_second": 4964.262
    },
    {
      "epoch": 28.307692307692307,
      "num_input_tokens_seen": 1461408,
      "step": 368,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 28.384615384615383,
      "grad_norm": 2.7374229431152344,
      "learning_rate": 0.01,
      "loss": 5.119778156280518,
      "num_input_tokens_seen": 1465504,
      "step": 369,
      "train_runtime": 295.1924,
      "train_tokens_per_second": 4964.572
    },
    {
      "epoch": 28.384615384615383,
      "num_input_tokens_seen": 1465504,
      "step": 369,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 28.46153846153846,
      "grad_norm": 3.7106480598449707,
      "learning_rate": 0.01,
      "loss": 5.319789886474609,
      "num_input_tokens_seen": 1469600,
      "step": 370,
      "train_runtime": 295.9993,
      "train_tokens_per_second": 4964.876
    },
    {
      "epoch": 28.46153846153846,
      "num_input_tokens_seen": 1469600,
      "step": 370,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 28.53846153846154,
      "grad_norm": 3.287628650665283,
      "learning_rate": 0.01,
      "loss": 5.204777717590332,
      "num_input_tokens_seen": 1473696,
      "step": 371,
      "train_runtime": 296.8064,
      "train_tokens_per_second": 4965.177
    },
    {
      "epoch": 28.53846153846154,
      "num_input_tokens_seen": 1473696,
      "step": 371,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 28.615384615384617,
      "grad_norm": 6.909034252166748,
      "learning_rate": 0.01,
      "loss": 5.366301536560059,
      "num_input_tokens_seen": 1477792,
      "step": 372,
      "train_runtime": 297.6133,
      "train_tokens_per_second": 4965.478
    },
    {
      "epoch": 28.615384615384617,
      "num_input_tokens_seen": 1477792,
      "step": 372,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 28.692307692307693,
      "grad_norm": 3.6321017742156982,
      "learning_rate": 0.01,
      "loss": 5.42665958404541,
      "num_input_tokens_seen": 1481888,
      "step": 373,
      "train_runtime": 298.4201,
      "train_tokens_per_second": 4965.779
    },
    {
      "epoch": 28.692307692307693,
      "num_input_tokens_seen": 1481888,
      "step": 373,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 28.76923076923077,
      "grad_norm": 4.401106834411621,
      "learning_rate": 0.01,
      "loss": 5.725541114807129,
      "num_input_tokens_seen": 1485984,
      "step": 374,
      "train_runtime": 299.227,
      "train_tokens_per_second": 4966.076
    },
    {
      "epoch": 28.76923076923077,
      "num_input_tokens_seen": 1485984,
      "step": 374,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 28.846153846153847,
      "grad_norm": 2.99000883102417,
      "learning_rate": 0.01,
      "loss": 5.187654495239258,
      "num_input_tokens_seen": 1490080,
      "step": 375,
      "train_runtime": 300.0348,
      "train_tokens_per_second": 4966.358
    },
    {
      "epoch": 28.846153846153847,
      "num_input_tokens_seen": 1490080,
      "step": 375,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 28.923076923076923,
      "grad_norm": 4.467616558074951,
      "learning_rate": 0.01,
      "loss": 5.414639472961426,
      "num_input_tokens_seen": 1494176,
      "step": 376,
      "train_runtime": 300.8386,
      "train_tokens_per_second": 4966.703
    },
    {
      "epoch": 28.923076923076923,
      "num_input_tokens_seen": 1494176,
      "step": 376,
      "train_accuracy": 0.0032573288772255182
    },
    {
      "epoch": 29.0,
      "grad_norm": 3.196613073348999,
      "learning_rate": 0.01,
      "loss": 5.394354343414307,
      "num_input_tokens_seen": 1496632,
      "step": 377,
      "train_runtime": 301.3503,
      "train_tokens_per_second": 4966.42
    },
    {
      "epoch": 29.0,
      "num_input_tokens_seen": 1496632,
      "step": 377,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 29.076923076923077,
      "grad_norm": 3.0329129695892334,
      "learning_rate": 0.01,
      "loss": 5.178316116333008,
      "num_input_tokens_seen": 1500728,
      "step": 378,
      "train_runtime": 302.1683,
      "train_tokens_per_second": 4966.529
    },
    {
      "epoch": 29.076923076923077,
      "num_input_tokens_seen": 1500728,
      "step": 378,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 29.153846153846153,
      "grad_norm": 2.940192699432373,
      "learning_rate": 0.01,
      "loss": 5.279094219207764,
      "num_input_tokens_seen": 1504824,
      "step": 379,
      "train_runtime": 302.9755,
      "train_tokens_per_second": 4966.818
    },
    {
      "epoch": 29.153846153846153,
      "num_input_tokens_seen": 1504824,
      "step": 379,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 29.23076923076923,
      "grad_norm": 3.4374678134918213,
      "learning_rate": 0.01,
      "loss": 5.204883575439453,
      "num_input_tokens_seen": 1508920,
      "step": 380,
      "train_runtime": 303.7828,
      "train_tokens_per_second": 4967.101
    },
    {
      "epoch": 29.23076923076923,
      "num_input_tokens_seen": 1508920,
      "step": 380,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 29.307692307692307,
      "grad_norm": 2.811293363571167,
      "learning_rate": 0.01,
      "loss": 5.172276973724365,
      "num_input_tokens_seen": 1513016,
      "step": 381,
      "train_runtime": 304.5893,
      "train_tokens_per_second": 4967.397
    },
    {
      "epoch": 29.307692307692307,
      "num_input_tokens_seen": 1513016,
      "step": 381,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 29.384615384615383,
      "grad_norm": 3.0140793323516846,
      "learning_rate": 0.01,
      "loss": 5.2489399909973145,
      "num_input_tokens_seen": 1517112,
      "step": 382,
      "train_runtime": 305.3955,
      "train_tokens_per_second": 4967.696
    },
    {
      "epoch": 29.384615384615383,
      "num_input_tokens_seen": 1517112,
      "step": 382,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 29.46153846153846,
      "grad_norm": 4.041913032531738,
      "learning_rate": 0.01,
      "loss": 5.229885101318359,
      "num_input_tokens_seen": 1521208,
      "step": 383,
      "train_runtime": 306.2022,
      "train_tokens_per_second": 4967.984
    },
    {
      "epoch": 29.46153846153846,
      "num_input_tokens_seen": 1521208,
      "step": 383,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 29.53846153846154,
      "grad_norm": 2.427123785018921,
      "learning_rate": 0.01,
      "loss": 4.96853494644165,
      "num_input_tokens_seen": 1525304,
      "step": 384,
      "train_runtime": 307.009,
      "train_tokens_per_second": 4968.272
    },
    {
      "epoch": 29.53846153846154,
      "num_input_tokens_seen": 1525304,
      "step": 384,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 29.615384615384617,
      "grad_norm": 2.374685525894165,
      "learning_rate": 0.01,
      "loss": 5.0407915115356445,
      "num_input_tokens_seen": 1529400,
      "step": 385,
      "train_runtime": 307.8152,
      "train_tokens_per_second": 4968.566
    },
    {
      "epoch": 29.615384615384617,
      "num_input_tokens_seen": 1529400,
      "step": 385,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 29.692307692307693,
      "grad_norm": 2.529252529144287,
      "learning_rate": 0.01,
      "loss": 5.022377014160156,
      "num_input_tokens_seen": 1533496,
      "step": 386,
      "train_runtime": 308.621,
      "train_tokens_per_second": 4968.864
    },
    {
      "epoch": 29.692307692307693,
      "num_input_tokens_seen": 1533496,
      "step": 386,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 29.76923076923077,
      "grad_norm": 2.575713872909546,
      "learning_rate": 0.01,
      "loss": 4.937716484069824,
      "num_input_tokens_seen": 1537592,
      "step": 387,
      "train_runtime": 309.4278,
      "train_tokens_per_second": 4969.146
    },
    {
      "epoch": 29.76923076923077,
      "num_input_tokens_seen": 1537592,
      "step": 387,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 29.846153846153847,
      "grad_norm": 1.810752511024475,
      "learning_rate": 0.01,
      "loss": 4.83278751373291,
      "num_input_tokens_seen": 1541688,
      "step": 388,
      "train_runtime": 310.2355,
      "train_tokens_per_second": 4969.412
    },
    {
      "epoch": 29.846153846153847,
      "num_input_tokens_seen": 1541688,
      "step": 388,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 29.923076923076923,
      "grad_norm": 2.1673622131347656,
      "learning_rate": 0.01,
      "loss": 4.854066848754883,
      "num_input_tokens_seen": 1545784,
      "step": 389,
      "train_runtime": 311.0383,
      "train_tokens_per_second": 4969.754
    },
    {
      "epoch": 29.923076923076923,
      "num_input_tokens_seen": 1545784,
      "step": 389,
      "train_accuracy": 0.013029315508902073
    },
    {
      "epoch": 30.0,
      "grad_norm": 2.0972650051116943,
      "learning_rate": 0.01,
      "loss": 4.779308795928955,
      "num_input_tokens_seen": 1548240,
      "step": 390,
      "train_runtime": 311.5498,
      "train_tokens_per_second": 4969.479
    },
    {
      "epoch": 30.0,
      "num_input_tokens_seen": 1548240,
      "step": 390,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 30.076923076923077,
      "grad_norm": 1.6017143726348877,
      "learning_rate": 0.01,
      "loss": 4.770463943481445,
      "num_input_tokens_seen": 1552336,
      "step": 391,
      "train_runtime": 312.3681,
      "train_tokens_per_second": 4969.573
    },
    {
      "epoch": 30.076923076923077,
      "num_input_tokens_seen": 1552336,
      "step": 391,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 30.153846153846153,
      "grad_norm": 1.9595004320144653,
      "learning_rate": 0.01,
      "loss": 4.839485168457031,
      "num_input_tokens_seen": 1556432,
      "step": 392,
      "train_runtime": 313.1744,
      "train_tokens_per_second": 4969.857
    },
    {
      "epoch": 30.153846153846153,
      "num_input_tokens_seen": 1556432,
      "step": 392,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 30.23076923076923,
      "grad_norm": 1.7008371353149414,
      "learning_rate": 0.01,
      "loss": 4.8051581382751465,
      "num_input_tokens_seen": 1560528,
      "step": 393,
      "train_runtime": 313.9811,
      "train_tokens_per_second": 4970.133
    },
    {
      "epoch": 30.23076923076923,
      "num_input_tokens_seen": 1560528,
      "step": 393,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 30.307692307692307,
      "grad_norm": 1.6570706367492676,
      "learning_rate": 0.01,
      "loss": 4.7867889404296875,
      "num_input_tokens_seen": 1564624,
      "step": 394,
      "train_runtime": 314.7875,
      "train_tokens_per_second": 4970.414
    },
    {
      "epoch": 30.307692307692307,
      "num_input_tokens_seen": 1564624,
      "step": 394,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 30.384615384615383,
      "grad_norm": 1.7845861911773682,
      "learning_rate": 0.01,
      "loss": 4.820751667022705,
      "num_input_tokens_seen": 1568720,
      "step": 395,
      "train_runtime": 315.5938,
      "train_tokens_per_second": 4970.693
    },
    {
      "epoch": 30.384615384615383,
      "num_input_tokens_seen": 1568720,
      "step": 395,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 30.46153846153846,
      "grad_norm": 1.7581539154052734,
      "learning_rate": 0.01,
      "loss": 4.803133964538574,
      "num_input_tokens_seen": 1572816,
      "step": 396,
      "train_runtime": 316.4007,
      "train_tokens_per_second": 4970.962
    },
    {
      "epoch": 30.46153846153846,
      "num_input_tokens_seen": 1572816,
      "step": 396,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 30.53846153846154,
      "grad_norm": 1.7303351163864136,
      "learning_rate": 0.01,
      "loss": 4.776310920715332,
      "num_input_tokens_seen": 1576912,
      "step": 397,
      "train_runtime": 317.208,
      "train_tokens_per_second": 4971.224
    },
    {
      "epoch": 30.53846153846154,
      "num_input_tokens_seen": 1576912,
      "step": 397,
      "train_accuracy": 0.0
    },
    {
      "epoch": 30.615384615384617,
      "grad_norm": 1.5727354288101196,
      "learning_rate": 0.01,
      "loss": 4.699723243713379,
      "num_input_tokens_seen": 1581008,
      "step": 398,
      "train_runtime": 318.0155,
      "train_tokens_per_second": 4971.482
    },
    {
      "epoch": 30.615384615384617,
      "num_input_tokens_seen": 1581008,
      "step": 398,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 30.692307692307693,
      "grad_norm": 1.3687564134597778,
      "learning_rate": 0.01,
      "loss": 4.686312675476074,
      "num_input_tokens_seen": 1585104,
      "step": 399,
      "train_runtime": 318.8224,
      "train_tokens_per_second": 4971.746
    },
    {
      "epoch": 30.692307692307693,
      "num_input_tokens_seen": 1585104,
      "step": 399,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 30.76923076923077,
      "grad_norm": 1.4996354579925537,
      "learning_rate": 0.01,
      "loss": 4.681853294372559,
      "num_input_tokens_seen": 1589200,
      "step": 400,
      "train_runtime": 319.6301,
      "train_tokens_per_second": 4971.997
    },
    {
      "epoch": 30.76923076923077,
      "eval_CMD_3_branch_eval_accuracy": 0.012296564195298372,
      "eval_CMD_3_branch_eval_loss": 4.700904369354248,
      "eval_CMD_3_branch_eval_runtime": 1.1322,
      "eval_CMD_3_branch_eval_samples_per_second": 2442.191,
      "eval_CMD_3_branch_eval_steps_per_second": 5.3,
      "num_input_tokens_seen": 1589200,
      "step": 400
    },
    {
      "epoch": 30.76923076923077,
      "num_input_tokens_seen": 1589200,
      "step": 400,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 30.846153846153847,
      "grad_norm": 1.5040349960327148,
      "learning_rate": 0.01,
      "loss": 4.673392295837402,
      "num_input_tokens_seen": 1593296,
      "step": 401,
      "train_runtime": 321.5745,
      "train_tokens_per_second": 4954.672
    },
    {
      "epoch": 30.846153846153847,
      "num_input_tokens_seen": 1593296,
      "step": 401,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 30.923076923076923,
      "grad_norm": 1.4539940357208252,
      "learning_rate": 0.01,
      "loss": 4.679933547973633,
      "num_input_tokens_seen": 1597392,
      "step": 402,
      "train_runtime": 322.3779,
      "train_tokens_per_second": 4955.029
    },
    {
      "epoch": 30.923076923076923,
      "num_input_tokens_seen": 1597392,
      "step": 402,
      "train_accuracy": 0.02280130237340927
    },
    {
      "epoch": 31.0,
      "grad_norm": 2.374574899673462,
      "learning_rate": 0.01,
      "loss": 4.776401042938232,
      "num_input_tokens_seen": 1599848,
      "step": 403,
      "train_runtime": 322.8894,
      "train_tokens_per_second": 4954.787
    },
    {
      "epoch": 31.0,
      "num_input_tokens_seen": 1599848,
      "step": 403,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 31.076923076923077,
      "grad_norm": 1.6019138097763062,
      "learning_rate": 0.01,
      "loss": 4.754836082458496,
      "num_input_tokens_seen": 1603944,
      "step": 404,
      "train_runtime": 323.7085,
      "train_tokens_per_second": 4954.902
    },
    {
      "epoch": 31.076923076923077,
      "num_input_tokens_seen": 1603944,
      "step": 404,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 31.153846153846153,
      "grad_norm": 1.9657342433929443,
      "learning_rate": 0.01,
      "loss": 4.811908721923828,
      "num_input_tokens_seen": 1608040,
      "step": 405,
      "train_runtime": 324.516,
      "train_tokens_per_second": 4955.194
    },
    {
      "epoch": 31.153846153846153,
      "num_input_tokens_seen": 1608040,
      "step": 405,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 31.23076923076923,
      "grad_norm": 1.391027569770813,
      "learning_rate": 0.01,
      "loss": 4.7307233810424805,
      "num_input_tokens_seen": 1612136,
      "step": 406,
      "train_runtime": 325.3232,
      "train_tokens_per_second": 4955.49
    },
    {
      "epoch": 31.23076923076923,
      "num_input_tokens_seen": 1612136,
      "step": 406,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 31.307692307692307,
      "grad_norm": 2.3124053478240967,
      "learning_rate": 0.01,
      "loss": 4.804024696350098,
      "num_input_tokens_seen": 1616232,
      "step": 407,
      "train_runtime": 326.1304,
      "train_tokens_per_second": 4955.785
    },
    {
      "epoch": 31.307692307692307,
      "num_input_tokens_seen": 1616232,
      "step": 407,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 31.384615384615383,
      "grad_norm": 2.326469898223877,
      "learning_rate": 0.01,
      "loss": 4.806546211242676,
      "num_input_tokens_seen": 1620328,
      "step": 408,
      "train_runtime": 326.937,
      "train_tokens_per_second": 4956.086
    },
    {
      "epoch": 31.384615384615383,
      "num_input_tokens_seen": 1620328,
      "step": 408,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 31.46153846153846,
      "grad_norm": 2.2424123287200928,
      "learning_rate": 0.01,
      "loss": 4.892226219177246,
      "num_input_tokens_seen": 1624424,
      "step": 409,
      "train_runtime": 327.744,
      "train_tokens_per_second": 4956.38
    },
    {
      "epoch": 31.46153846153846,
      "num_input_tokens_seen": 1624424,
      "step": 409,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 31.53846153846154,
      "grad_norm": 2.089282274246216,
      "learning_rate": 0.01,
      "loss": 4.813840389251709,
      "num_input_tokens_seen": 1628520,
      "step": 410,
      "train_runtime": 328.5515,
      "train_tokens_per_second": 4956.666
    },
    {
      "epoch": 31.53846153846154,
      "num_input_tokens_seen": 1628520,
      "step": 410,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 31.615384615384617,
      "grad_norm": 1.866958498954773,
      "learning_rate": 0.01,
      "loss": 4.844282627105713,
      "num_input_tokens_seen": 1632616,
      "step": 411,
      "train_runtime": 329.3581,
      "train_tokens_per_second": 4956.963
    },
    {
      "epoch": 31.615384615384617,
      "num_input_tokens_seen": 1632616,
      "step": 411,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 31.692307692307693,
      "grad_norm": 1.9325016736984253,
      "learning_rate": 0.01,
      "loss": 4.847882270812988,
      "num_input_tokens_seen": 1636712,
      "step": 412,
      "train_runtime": 330.1656,
      "train_tokens_per_second": 4957.245
    },
    {
      "epoch": 31.692307692307693,
      "num_input_tokens_seen": 1636712,
      "step": 412,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 31.76923076923077,
      "grad_norm": 1.4464740753173828,
      "learning_rate": 0.01,
      "loss": 4.749690055847168,
      "num_input_tokens_seen": 1640808,
      "step": 413,
      "train_runtime": 330.9725,
      "train_tokens_per_second": 4957.535
    },
    {
      "epoch": 31.76923076923077,
      "num_input_tokens_seen": 1640808,
      "step": 413,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 31.846153846153847,
      "grad_norm": 2.3502817153930664,
      "learning_rate": 0.01,
      "loss": 4.873184680938721,
      "num_input_tokens_seen": 1644904,
      "step": 414,
      "train_runtime": 331.7801,
      "train_tokens_per_second": 4957.814
    },
    {
      "epoch": 31.846153846153847,
      "num_input_tokens_seen": 1644904,
      "step": 414,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 31.923076923076923,
      "grad_norm": 2.0078091621398926,
      "learning_rate": 0.01,
      "loss": 4.7482452392578125,
      "num_input_tokens_seen": 1649000,
      "step": 415,
      "train_runtime": 332.5832,
      "train_tokens_per_second": 4958.157
    },
    {
      "epoch": 31.923076923076923,
      "num_input_tokens_seen": 1649000,
      "step": 415,
      "train_accuracy": 0.016286645084619522
    },
    {
      "epoch": 32.0,
      "grad_norm": 1.5769764184951782,
      "learning_rate": 0.01,
      "loss": 4.71989107131958,
      "num_input_tokens_seen": 1651456,
      "step": 416,
      "train_runtime": 333.0955,
      "train_tokens_per_second": 4957.906
    },
    {
      "epoch": 32.0,
      "num_input_tokens_seen": 1651456,
      "step": 416,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 32.07692307692308,
      "grad_norm": 2.7645392417907715,
      "learning_rate": 0.01,
      "loss": 4.84213924407959,
      "num_input_tokens_seen": 1655552,
      "step": 417,
      "train_runtime": 333.9141,
      "train_tokens_per_second": 4958.018
    },
    {
      "epoch": 32.07692307692308,
      "num_input_tokens_seen": 1655552,
      "step": 417,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 32.15384615384615,
      "grad_norm": 2.3278844356536865,
      "learning_rate": 0.01,
      "loss": 4.811976432800293,
      "num_input_tokens_seen": 1659648,
      "step": 418,
      "train_runtime": 334.722,
      "train_tokens_per_second": 4958.288
    },
    {
      "epoch": 32.15384615384615,
      "num_input_tokens_seen": 1659648,
      "step": 418,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 32.23076923076923,
      "grad_norm": 2.2258729934692383,
      "learning_rate": 0.01,
      "loss": 4.932866096496582,
      "num_input_tokens_seen": 1663744,
      "step": 419,
      "train_runtime": 335.5293,
      "train_tokens_per_second": 4958.565
    },
    {
      "epoch": 32.23076923076923,
      "num_input_tokens_seen": 1663744,
      "step": 419,
      "train_accuracy": 0.001953125
    },
    {
      "epoch": 32.30769230769231,
      "grad_norm": 2.404165506362915,
      "learning_rate": 0.01,
      "loss": 4.878048896789551,
      "num_input_tokens_seen": 1667840,
      "step": 420,
      "train_runtime": 336.3359,
      "train_tokens_per_second": 4958.852
    },
    {
      "epoch": 32.30769230769231,
      "num_input_tokens_seen": 1667840,
      "step": 420,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 32.38461538461539,
      "grad_norm": 2.28906512260437,
      "learning_rate": 0.01,
      "loss": 4.84922456741333,
      "num_input_tokens_seen": 1671936,
      "step": 421,
      "train_runtime": 337.1433,
      "train_tokens_per_second": 4959.125
    },
    {
      "epoch": 32.38461538461539,
      "num_input_tokens_seen": 1671936,
      "step": 421,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 32.46153846153846,
      "grad_norm": 2.0048649311065674,
      "learning_rate": 0.01,
      "loss": 4.9027628898620605,
      "num_input_tokens_seen": 1676032,
      "step": 422,
      "train_runtime": 337.9517,
      "train_tokens_per_second": 4959.384
    },
    {
      "epoch": 32.46153846153846,
      "num_input_tokens_seen": 1676032,
      "step": 422,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 32.53846153846154,
      "grad_norm": 2.3459694385528564,
      "learning_rate": 0.01,
      "loss": 4.852118492126465,
      "num_input_tokens_seen": 1680128,
      "step": 423,
      "train_runtime": 338.7586,
      "train_tokens_per_second": 4959.661
    },
    {
      "epoch": 32.53846153846154,
      "num_input_tokens_seen": 1680128,
      "step": 423,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 32.61538461538461,
      "grad_norm": 1.5540844202041626,
      "learning_rate": 0.01,
      "loss": 4.792023181915283,
      "num_input_tokens_seen": 1684224,
      "step": 424,
      "train_runtime": 339.566,
      "train_tokens_per_second": 4959.931
    },
    {
      "epoch": 32.61538461538461,
      "num_input_tokens_seen": 1684224,
      "step": 424,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 32.69230769230769,
      "grad_norm": 1.9204732179641724,
      "learning_rate": 0.01,
      "loss": 4.947464942932129,
      "num_input_tokens_seen": 1688320,
      "step": 425,
      "train_runtime": 340.3738,
      "train_tokens_per_second": 4960.194
    },
    {
      "epoch": 32.69230769230769,
      "num_input_tokens_seen": 1688320,
      "step": 425,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 32.76923076923077,
      "grad_norm": 1.775728702545166,
      "learning_rate": 0.01,
      "loss": 4.865842819213867,
      "num_input_tokens_seen": 1692416,
      "step": 426,
      "train_runtime": 341.1822,
      "train_tokens_per_second": 4960.446
    },
    {
      "epoch": 32.76923076923077,
      "num_input_tokens_seen": 1692416,
      "step": 426,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 32.84615384615385,
      "grad_norm": 1.7166199684143066,
      "learning_rate": 0.01,
      "loss": 4.792997360229492,
      "num_input_tokens_seen": 1696512,
      "step": 427,
      "train_runtime": 341.9897,
      "train_tokens_per_second": 4960.711
    },
    {
      "epoch": 32.84615384615385,
      "num_input_tokens_seen": 1696512,
      "step": 427,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 32.92307692307692,
      "grad_norm": 1.7444908618927002,
      "learning_rate": 0.01,
      "loss": 4.791705131530762,
      "num_input_tokens_seen": 1700608,
      "step": 428,
      "train_runtime": 342.7927,
      "train_tokens_per_second": 4961.04
    },
    {
      "epoch": 32.92307692307692,
      "num_input_tokens_seen": 1700608,
      "step": 428,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 33.0,
      "grad_norm": 2.254854679107666,
      "learning_rate": 0.01,
      "loss": 4.896081924438477,
      "num_input_tokens_seen": 1703064,
      "step": 429,
      "train_runtime": 343.3042,
      "train_tokens_per_second": 4960.802
    },
    {
      "epoch": 33.0,
      "num_input_tokens_seen": 1703064,
      "step": 429,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 33.07692307692308,
      "grad_norm": 2.067108392715454,
      "learning_rate": 0.01,
      "loss": 4.808799743652344,
      "num_input_tokens_seen": 1707160,
      "step": 430,
      "train_runtime": 344.1245,
      "train_tokens_per_second": 4960.879
    },
    {
      "epoch": 33.07692307692308,
      "num_input_tokens_seen": 1707160,
      "step": 430,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 33.15384615384615,
      "grad_norm": 1.6360903978347778,
      "learning_rate": 0.01,
      "loss": 4.740631103515625,
      "num_input_tokens_seen": 1711256,
      "step": 431,
      "train_runtime": 344.9322,
      "train_tokens_per_second": 4961.137
    },
    {
      "epoch": 33.15384615384615,
      "num_input_tokens_seen": 1711256,
      "step": 431,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 33.23076923076923,
      "grad_norm": 1.8674015998840332,
      "learning_rate": 0.01,
      "loss": 4.924413681030273,
      "num_input_tokens_seen": 1715352,
      "step": 432,
      "train_runtime": 345.7395,
      "train_tokens_per_second": 4961.4
    },
    {
      "epoch": 33.23076923076923,
      "num_input_tokens_seen": 1715352,
      "step": 432,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 33.30769230769231,
      "grad_norm": 2.4183735847473145,
      "learning_rate": 0.01,
      "loss": 4.848052978515625,
      "num_input_tokens_seen": 1719448,
      "step": 433,
      "train_runtime": 346.5476,
      "train_tokens_per_second": 4961.65
    },
    {
      "epoch": 33.30769230769231,
      "num_input_tokens_seen": 1719448,
      "step": 433,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 33.38461538461539,
      "grad_norm": 2.2434303760528564,
      "learning_rate": 0.01,
      "loss": 4.9350810050964355,
      "num_input_tokens_seen": 1723544,
      "step": 434,
      "train_runtime": 347.3551,
      "train_tokens_per_second": 4961.908
    },
    {
      "epoch": 33.38461538461539,
      "num_input_tokens_seen": 1723544,
      "step": 434,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 33.46153846153846,
      "grad_norm": 1.9384300708770752,
      "learning_rate": 0.01,
      "loss": 4.879378795623779,
      "num_input_tokens_seen": 1727640,
      "step": 435,
      "train_runtime": 348.1631,
      "train_tokens_per_second": 4962.157
    },
    {
      "epoch": 33.46153846153846,
      "num_input_tokens_seen": 1727640,
      "step": 435,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 33.53846153846154,
      "grad_norm": 2.2234044075012207,
      "learning_rate": 0.01,
      "loss": 4.843637466430664,
      "num_input_tokens_seen": 1731736,
      "step": 436,
      "train_runtime": 348.9708,
      "train_tokens_per_second": 4962.409
    },
    {
      "epoch": 33.53846153846154,
      "num_input_tokens_seen": 1731736,
      "step": 436,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 33.61538461538461,
      "grad_norm": 1.8920512199401855,
      "learning_rate": 0.01,
      "loss": 4.816337585449219,
      "num_input_tokens_seen": 1735832,
      "step": 437,
      "train_runtime": 349.7787,
      "train_tokens_per_second": 4962.658
    },
    {
      "epoch": 33.61538461538461,
      "num_input_tokens_seen": 1735832,
      "step": 437,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 33.69230769230769,
      "grad_norm": 1.9466722011566162,
      "learning_rate": 0.01,
      "loss": 4.822091102600098,
      "num_input_tokens_seen": 1739928,
      "step": 438,
      "train_runtime": 350.5862,
      "train_tokens_per_second": 4962.91
    },
    {
      "epoch": 33.69230769230769,
      "num_input_tokens_seen": 1739928,
      "step": 438,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 33.76923076923077,
      "grad_norm": 1.4665552377700806,
      "learning_rate": 0.01,
      "loss": 4.728626728057861,
      "num_input_tokens_seen": 1744024,
      "step": 439,
      "train_runtime": 351.3937,
      "train_tokens_per_second": 4963.163
    },
    {
      "epoch": 33.76923076923077,
      "num_input_tokens_seen": 1744024,
      "step": 439,
      "train_accuracy": 0.001953125
    },
    {
      "epoch": 33.84615384615385,
      "grad_norm": 1.747174859046936,
      "learning_rate": 0.01,
      "loss": 4.873631954193115,
      "num_input_tokens_seen": 1748120,
      "step": 440,
      "train_runtime": 352.2015,
      "train_tokens_per_second": 4963.409
    },
    {
      "epoch": 33.84615384615385,
      "num_input_tokens_seen": 1748120,
      "step": 440,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 33.92307692307692,
      "grad_norm": 1.3376929759979248,
      "learning_rate": 0.01,
      "loss": 4.754631042480469,
      "num_input_tokens_seen": 1752216,
      "step": 441,
      "train_runtime": 353.0056,
      "train_tokens_per_second": 4963.706
    },
    {
      "epoch": 33.92307692307692,
      "num_input_tokens_seen": 1752216,
      "step": 441,
      "train_accuracy": 0.0065146577544510365
    },
    {
      "epoch": 34.0,
      "grad_norm": 1.7214449644088745,
      "learning_rate": 0.01,
      "loss": 4.764516830444336,
      "num_input_tokens_seen": 1754672,
      "step": 442,
      "train_runtime": 353.5181,
      "train_tokens_per_second": 4963.457
    },
    {
      "epoch": 34.0,
      "num_input_tokens_seen": 1754672,
      "step": 442,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 34.07692307692308,
      "grad_norm": 2.1950595378875732,
      "learning_rate": 0.01,
      "loss": 4.860500812530518,
      "num_input_tokens_seen": 1758768,
      "step": 443,
      "train_runtime": 354.3391,
      "train_tokens_per_second": 4963.516
    },
    {
      "epoch": 34.07692307692308,
      "num_input_tokens_seen": 1758768,
      "step": 443,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 34.15384615384615,
      "grad_norm": 1.5873230695724487,
      "learning_rate": 0.01,
      "loss": 4.745871067047119,
      "num_input_tokens_seen": 1762864,
      "step": 444,
      "train_runtime": 355.1524,
      "train_tokens_per_second": 4963.684
    },
    {
      "epoch": 34.15384615384615,
      "num_input_tokens_seen": 1762864,
      "step": 444,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 34.23076923076923,
      "grad_norm": 1.554171085357666,
      "learning_rate": 0.01,
      "loss": 4.714781761169434,
      "num_input_tokens_seen": 1766960,
      "step": 445,
      "train_runtime": 355.9674,
      "train_tokens_per_second": 4963.826
    },
    {
      "epoch": 34.23076923076923,
      "num_input_tokens_seen": 1766960,
      "step": 445,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 34.30769230769231,
      "grad_norm": 1.4377801418304443,
      "learning_rate": 0.01,
      "loss": 4.710833549499512,
      "num_input_tokens_seen": 1771056,
      "step": 446,
      "train_runtime": 356.7789,
      "train_tokens_per_second": 4964.016
    },
    {
      "epoch": 34.30769230769231,
      "num_input_tokens_seen": 1771056,
      "step": 446,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 34.38461538461539,
      "grad_norm": 1.4280403852462769,
      "learning_rate": 0.01,
      "loss": 4.760910987854004,
      "num_input_tokens_seen": 1775152,
      "step": 447,
      "train_runtime": 357.5907,
      "train_tokens_per_second": 4964.2
    },
    {
      "epoch": 34.38461538461539,
      "num_input_tokens_seen": 1775152,
      "step": 447,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 34.46153846153846,
      "grad_norm": 1.4387681484222412,
      "learning_rate": 0.01,
      "loss": 4.687384128570557,
      "num_input_tokens_seen": 1779248,
      "step": 448,
      "train_runtime": 358.4009,
      "train_tokens_per_second": 4964.407
    },
    {
      "epoch": 34.46153846153846,
      "num_input_tokens_seen": 1779248,
      "step": 448,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 34.53846153846154,
      "grad_norm": 1.6570788621902466,
      "learning_rate": 0.01,
      "loss": 4.7208967208862305,
      "num_input_tokens_seen": 1783344,
      "step": 449,
      "train_runtime": 359.2105,
      "train_tokens_per_second": 4964.621
    },
    {
      "epoch": 34.53846153846154,
      "num_input_tokens_seen": 1783344,
      "step": 449,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 34.61538461538461,
      "grad_norm": 1.4104818105697632,
      "learning_rate": 0.01,
      "loss": 4.730859279632568,
      "num_input_tokens_seen": 1787440,
      "step": 450,
      "train_runtime": 360.0197,
      "train_tokens_per_second": 4964.839
    },
    {
      "epoch": 34.61538461538461,
      "num_input_tokens_seen": 1787440,
      "step": 450,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 34.69230769230769,
      "grad_norm": 1.1465097665786743,
      "learning_rate": 0.01,
      "loss": 4.663358688354492,
      "num_input_tokens_seen": 1791536,
      "step": 451,
      "train_runtime": 360.8301,
      "train_tokens_per_second": 4965.04
    },
    {
      "epoch": 34.69230769230769,
      "num_input_tokens_seen": 1791536,
      "step": 451,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 34.76923076923077,
      "grad_norm": 1.5145927667617798,
      "learning_rate": 0.01,
      "loss": 4.749174118041992,
      "num_input_tokens_seen": 1795632,
      "step": 452,
      "train_runtime": 361.6427,
      "train_tokens_per_second": 4965.21
    },
    {
      "epoch": 34.76923076923077,
      "num_input_tokens_seen": 1795632,
      "step": 452,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 34.84615384615385,
      "grad_norm": 1.3290702104568481,
      "learning_rate": 0.01,
      "loss": 4.6947760581970215,
      "num_input_tokens_seen": 1799728,
      "step": 453,
      "train_runtime": 362.4524,
      "train_tokens_per_second": 4965.419
    },
    {
      "epoch": 34.84615384615385,
      "num_input_tokens_seen": 1799728,
      "step": 453,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 34.92307692307692,
      "grad_norm": 1.7186955213546753,
      "learning_rate": 0.01,
      "loss": 4.8016767501831055,
      "num_input_tokens_seen": 1803824,
      "step": 454,
      "train_runtime": 363.2566,
      "train_tokens_per_second": 4965.702
    },
    {
      "epoch": 34.92307692307692,
      "num_input_tokens_seen": 1803824,
      "step": 454,
      "train_accuracy": 0.0065146577544510365
    },
    {
      "epoch": 35.0,
      "grad_norm": 2.8009908199310303,
      "learning_rate": 0.01,
      "loss": 4.814508438110352,
      "num_input_tokens_seen": 1806280,
      "step": 455,
      "train_runtime": 363.7689,
      "train_tokens_per_second": 4965.461
    },
    {
      "epoch": 35.0,
      "num_input_tokens_seen": 1806280,
      "step": 455,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 35.07692307692308,
      "grad_norm": 2.6078543663024902,
      "learning_rate": 0.01,
      "loss": 4.882625579833984,
      "num_input_tokens_seen": 1810376,
      "step": 456,
      "train_runtime": 364.5894,
      "train_tokens_per_second": 4965.521
    },
    {
      "epoch": 35.07692307692308,
      "num_input_tokens_seen": 1810376,
      "step": 456,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 35.15384615384615,
      "grad_norm": 1.9776579141616821,
      "learning_rate": 0.01,
      "loss": 4.89740514755249,
      "num_input_tokens_seen": 1814472,
      "step": 457,
      "train_runtime": 365.3969,
      "train_tokens_per_second": 4965.757
    },
    {
      "epoch": 35.15384615384615,
      "num_input_tokens_seen": 1814472,
      "step": 457,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 35.23076923076923,
      "grad_norm": 1.7558856010437012,
      "learning_rate": 0.01,
      "loss": 4.8707356452941895,
      "num_input_tokens_seen": 1818568,
      "step": 458,
      "train_runtime": 366.2051,
      "train_tokens_per_second": 4965.982
    },
    {
      "epoch": 35.23076923076923,
      "num_input_tokens_seen": 1818568,
      "step": 458,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 35.30769230769231,
      "grad_norm": 1.9807605743408203,
      "learning_rate": 0.01,
      "loss": 4.846177101135254,
      "num_input_tokens_seen": 1822664,
      "step": 459,
      "train_runtime": 367.0126,
      "train_tokens_per_second": 4966.216
    },
    {
      "epoch": 35.30769230769231,
      "num_input_tokens_seen": 1822664,
      "step": 459,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 35.38461538461539,
      "grad_norm": 1.7133845090866089,
      "learning_rate": 0.01,
      "loss": 4.851652145385742,
      "num_input_tokens_seen": 1826760,
      "step": 460,
      "train_runtime": 367.8202,
      "train_tokens_per_second": 4966.449
    },
    {
      "epoch": 35.38461538461539,
      "num_input_tokens_seen": 1826760,
      "step": 460,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 35.46153846153846,
      "grad_norm": 1.8392632007598877,
      "learning_rate": 0.01,
      "loss": 4.925780296325684,
      "num_input_tokens_seen": 1830856,
      "step": 461,
      "train_runtime": 368.6273,
      "train_tokens_per_second": 4966.686
    },
    {
      "epoch": 35.46153846153846,
      "num_input_tokens_seen": 1830856,
      "step": 461,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 35.53846153846154,
      "grad_norm": 2.557321786880493,
      "learning_rate": 0.01,
      "loss": 4.94730281829834,
      "num_input_tokens_seen": 1834952,
      "step": 462,
      "train_runtime": 369.4351,
      "train_tokens_per_second": 4966.913
    },
    {
      "epoch": 35.53846153846154,
      "num_input_tokens_seen": 1834952,
      "step": 462,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 35.61538461538461,
      "grad_norm": 1.7267029285430908,
      "learning_rate": 0.01,
      "loss": 4.850793361663818,
      "num_input_tokens_seen": 1839048,
      "step": 463,
      "train_runtime": 370.2429,
      "train_tokens_per_second": 4967.139
    },
    {
      "epoch": 35.61538461538461,
      "num_input_tokens_seen": 1839048,
      "step": 463,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 35.69230769230769,
      "grad_norm": 1.9380431175231934,
      "learning_rate": 0.01,
      "loss": 4.896787643432617,
      "num_input_tokens_seen": 1843144,
      "step": 464,
      "train_runtime": 371.05,
      "train_tokens_per_second": 4967.373
    },
    {
      "epoch": 35.69230769230769,
      "num_input_tokens_seen": 1843144,
      "step": 464,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 35.76923076923077,
      "grad_norm": 1.9114341735839844,
      "learning_rate": 0.01,
      "loss": 4.811819076538086,
      "num_input_tokens_seen": 1847240,
      "step": 465,
      "train_runtime": 371.8608,
      "train_tokens_per_second": 4967.558
    },
    {
      "epoch": 35.76923076923077,
      "num_input_tokens_seen": 1847240,
      "step": 465,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 35.84615384615385,
      "grad_norm": 1.5461879968643188,
      "learning_rate": 0.01,
      "loss": 4.788661003112793,
      "num_input_tokens_seen": 1851336,
      "step": 466,
      "train_runtime": 372.6691,
      "train_tokens_per_second": 4967.775
    },
    {
      "epoch": 35.84615384615385,
      "num_input_tokens_seen": 1851336,
      "step": 466,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 35.92307692307692,
      "grad_norm": 2.167442560195923,
      "learning_rate": 0.01,
      "loss": 4.931826114654541,
      "num_input_tokens_seen": 1855432,
      "step": 467,
      "train_runtime": 373.4723,
      "train_tokens_per_second": 4968.059
    },
    {
      "epoch": 35.92307692307692,
      "num_input_tokens_seen": 1855432,
      "step": 467,
      "train_accuracy": 0.0065146577544510365
    },
    {
      "epoch": 36.0,
      "grad_norm": 1.7010177373886108,
      "learning_rate": 0.01,
      "loss": 4.761315822601318,
      "num_input_tokens_seen": 1857888,
      "step": 468,
      "train_runtime": 373.9844,
      "train_tokens_per_second": 4967.822
    },
    {
      "epoch": 36.0,
      "num_input_tokens_seen": 1857888,
      "step": 468,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 36.07692307692308,
      "grad_norm": 1.7015635967254639,
      "learning_rate": 0.01,
      "loss": 4.743702411651611,
      "num_input_tokens_seen": 1861984,
      "step": 469,
      "train_runtime": 374.8036,
      "train_tokens_per_second": 4967.893
    },
    {
      "epoch": 36.07692307692308,
      "num_input_tokens_seen": 1861984,
      "step": 469,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 36.15384615384615,
      "grad_norm": 1.6663355827331543,
      "learning_rate": 0.01,
      "loss": 4.780470371246338,
      "num_input_tokens_seen": 1866080,
      "step": 470,
      "train_runtime": 375.6103,
      "train_tokens_per_second": 4968.128
    },
    {
      "epoch": 36.15384615384615,
      "num_input_tokens_seen": 1866080,
      "step": 470,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 36.23076923076923,
      "grad_norm": 1.3687386512756348,
      "learning_rate": 0.01,
      "loss": 4.7952775955200195,
      "num_input_tokens_seen": 1870176,
      "step": 471,
      "train_runtime": 376.4165,
      "train_tokens_per_second": 4968.369
    },
    {
      "epoch": 36.23076923076923,
      "num_input_tokens_seen": 1870176,
      "step": 471,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 36.30769230769231,
      "grad_norm": 2.32875657081604,
      "learning_rate": 0.01,
      "loss": 4.8694915771484375,
      "num_input_tokens_seen": 1874272,
      "step": 472,
      "train_runtime": 377.2236,
      "train_tokens_per_second": 4968.596
    },
    {
      "epoch": 36.30769230769231,
      "num_input_tokens_seen": 1874272,
      "step": 472,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 36.38461538461539,
      "grad_norm": 1.7562257051467896,
      "learning_rate": 0.01,
      "loss": 4.7543625831604,
      "num_input_tokens_seen": 1878368,
      "step": 473,
      "train_runtime": 378.03,
      "train_tokens_per_second": 4968.834
    },
    {
      "epoch": 36.38461538461539,
      "num_input_tokens_seen": 1878368,
      "step": 473,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 36.46153846153846,
      "grad_norm": 2.138751745223999,
      "learning_rate": 0.01,
      "loss": 4.800985336303711,
      "num_input_tokens_seen": 1882464,
      "step": 474,
      "train_runtime": 378.8384,
      "train_tokens_per_second": 4969.043
    },
    {
      "epoch": 36.46153846153846,
      "num_input_tokens_seen": 1882464,
      "step": 474,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 36.53846153846154,
      "grad_norm": 2.6692943572998047,
      "learning_rate": 0.01,
      "loss": 4.881521224975586,
      "num_input_tokens_seen": 1886560,
      "step": 475,
      "train_runtime": 379.6465,
      "train_tokens_per_second": 4969.255
    },
    {
      "epoch": 36.53846153846154,
      "num_input_tokens_seen": 1886560,
      "step": 475,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 36.61538461538461,
      "grad_norm": 1.1861162185668945,
      "learning_rate": 0.01,
      "loss": 4.710395336151123,
      "num_input_tokens_seen": 1890656,
      "step": 476,
      "train_runtime": 380.4533,
      "train_tokens_per_second": 4969.482
    },
    {
      "epoch": 36.61538461538461,
      "num_input_tokens_seen": 1890656,
      "step": 476,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 36.69230769230769,
      "grad_norm": 2.5071468353271484,
      "learning_rate": 0.01,
      "loss": 4.834889888763428,
      "num_input_tokens_seen": 1894752,
      "step": 477,
      "train_runtime": 381.2616,
      "train_tokens_per_second": 4969.69
    },
    {
      "epoch": 36.69230769230769,
      "num_input_tokens_seen": 1894752,
      "step": 477,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 36.76923076923077,
      "grad_norm": 1.7427690029144287,
      "learning_rate": 0.01,
      "loss": 4.781068325042725,
      "num_input_tokens_seen": 1898848,
      "step": 478,
      "train_runtime": 382.0678,
      "train_tokens_per_second": 4969.924
    },
    {
      "epoch": 36.76923076923077,
      "num_input_tokens_seen": 1898848,
      "step": 478,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 36.84615384615385,
      "grad_norm": 1.5795429944992065,
      "learning_rate": 0.01,
      "loss": 4.755626678466797,
      "num_input_tokens_seen": 1902944,
      "step": 479,
      "train_runtime": 382.8745,
      "train_tokens_per_second": 4970.151
    },
    {
      "epoch": 36.84615384615385,
      "num_input_tokens_seen": 1902944,
      "step": 479,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 36.92307692307692,
      "grad_norm": 1.9554100036621094,
      "learning_rate": 0.01,
      "loss": 4.753540515899658,
      "num_input_tokens_seen": 1907040,
      "step": 480,
      "train_runtime": 383.676,
      "train_tokens_per_second": 4970.444
    },
    {
      "epoch": 36.92307692307692,
      "num_input_tokens_seen": 1907040,
      "step": 480,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 37.0,
      "grad_norm": 1.7874726057052612,
      "learning_rate": 0.01,
      "loss": 4.8236589431762695,
      "num_input_tokens_seen": 1909496,
      "step": 481,
      "train_runtime": 384.1885,
      "train_tokens_per_second": 4970.207
    },
    {
      "epoch": 37.0,
      "num_input_tokens_seen": 1909496,
      "step": 481,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 37.07692307692308,
      "grad_norm": 1.234984040260315,
      "learning_rate": 0.01,
      "loss": 4.683443069458008,
      "num_input_tokens_seen": 1913592,
      "step": 482,
      "train_runtime": 385.0094,
      "train_tokens_per_second": 4970.247
    },
    {
      "epoch": 37.07692307692308,
      "num_input_tokens_seen": 1913592,
      "step": 482,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 37.15384615384615,
      "grad_norm": 1.6579338312149048,
      "learning_rate": 0.01,
      "loss": 4.802212715148926,
      "num_input_tokens_seen": 1917688,
      "step": 483,
      "train_runtime": 385.816,
      "train_tokens_per_second": 4970.473
    },
    {
      "epoch": 37.15384615384615,
      "num_input_tokens_seen": 1917688,
      "step": 483,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 37.23076923076923,
      "grad_norm": 1.476696491241455,
      "learning_rate": 0.01,
      "loss": 4.762943744659424,
      "num_input_tokens_seen": 1921784,
      "step": 484,
      "train_runtime": 386.6225,
      "train_tokens_per_second": 4970.699
    },
    {
      "epoch": 37.23076923076923,
      "num_input_tokens_seen": 1921784,
      "step": 484,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 37.30769230769231,
      "grad_norm": 1.6325160264968872,
      "learning_rate": 0.01,
      "loss": 4.768425941467285,
      "num_input_tokens_seen": 1925880,
      "step": 485,
      "train_runtime": 387.4303,
      "train_tokens_per_second": 4970.908
    },
    {
      "epoch": 37.30769230769231,
      "num_input_tokens_seen": 1925880,
      "step": 485,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 37.38461538461539,
      "grad_norm": 1.5507967472076416,
      "learning_rate": 0.01,
      "loss": 4.717211723327637,
      "num_input_tokens_seen": 1929976,
      "step": 486,
      "train_runtime": 388.2393,
      "train_tokens_per_second": 4971.099
    },
    {
      "epoch": 37.38461538461539,
      "num_input_tokens_seen": 1929976,
      "step": 486,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 37.46153846153846,
      "grad_norm": 1.1237963438034058,
      "learning_rate": 0.01,
      "loss": 4.721593856811523,
      "num_input_tokens_seen": 1934072,
      "step": 487,
      "train_runtime": 389.0462,
      "train_tokens_per_second": 4971.317
    },
    {
      "epoch": 37.46153846153846,
      "num_input_tokens_seen": 1934072,
      "step": 487,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 37.53846153846154,
      "grad_norm": 1.5940783023834229,
      "learning_rate": 0.01,
      "loss": 4.7369537353515625,
      "num_input_tokens_seen": 1938168,
      "step": 488,
      "train_runtime": 389.853,
      "train_tokens_per_second": 4971.536
    },
    {
      "epoch": 37.53846153846154,
      "num_input_tokens_seen": 1938168,
      "step": 488,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 37.61538461538461,
      "grad_norm": 1.905720591545105,
      "learning_rate": 0.01,
      "loss": 4.750986576080322,
      "num_input_tokens_seen": 1942264,
      "step": 489,
      "train_runtime": 390.6602,
      "train_tokens_per_second": 4971.748
    },
    {
      "epoch": 37.61538461538461,
      "num_input_tokens_seen": 1942264,
      "step": 489,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 37.69230769230769,
      "grad_norm": 1.314231514930725,
      "learning_rate": 0.01,
      "loss": 4.731482028961182,
      "num_input_tokens_seen": 1946360,
      "step": 490,
      "train_runtime": 391.4694,
      "train_tokens_per_second": 4971.934
    },
    {
      "epoch": 37.69230769230769,
      "num_input_tokens_seen": 1946360,
      "step": 490,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 37.76923076923077,
      "grad_norm": 1.5926564931869507,
      "learning_rate": 0.01,
      "loss": 4.735569000244141,
      "num_input_tokens_seen": 1950456,
      "step": 491,
      "train_runtime": 392.276,
      "train_tokens_per_second": 4972.152
    },
    {
      "epoch": 37.76923076923077,
      "num_input_tokens_seen": 1950456,
      "step": 491,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 37.84615384615385,
      "grad_norm": 1.6750179529190063,
      "learning_rate": 0.01,
      "loss": 4.661894798278809,
      "num_input_tokens_seen": 1954552,
      "step": 492,
      "train_runtime": 393.0829,
      "train_tokens_per_second": 4972.365
    },
    {
      "epoch": 37.84615384615385,
      "num_input_tokens_seen": 1954552,
      "step": 492,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 37.92307692307692,
      "grad_norm": 0.9824418425559998,
      "learning_rate": 0.01,
      "loss": 4.667640209197998,
      "num_input_tokens_seen": 1958648,
      "step": 493,
      "train_runtime": 393.8856,
      "train_tokens_per_second": 4972.631
    },
    {
      "epoch": 37.92307692307692,
      "num_input_tokens_seen": 1958648,
      "step": 493,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 38.0,
      "grad_norm": 1.2208882570266724,
      "learning_rate": 0.01,
      "loss": 4.651348114013672,
      "num_input_tokens_seen": 1961104,
      "step": 494,
      "train_runtime": 394.3967,
      "train_tokens_per_second": 4972.415
    },
    {
      "epoch": 38.0,
      "num_input_tokens_seen": 1961104,
      "step": 494,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 38.07692307692308,
      "grad_norm": 0.9827911257743835,
      "learning_rate": 0.01,
      "loss": 4.64877986907959,
      "num_input_tokens_seen": 1965200,
      "step": 495,
      "train_runtime": 395.2155,
      "train_tokens_per_second": 4972.477
    },
    {
      "epoch": 38.07692307692308,
      "num_input_tokens_seen": 1965200,
      "step": 495,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 38.15384615384615,
      "grad_norm": 1.0729360580444336,
      "learning_rate": 0.01,
      "loss": 4.710212230682373,
      "num_input_tokens_seen": 1969296,
      "step": 496,
      "train_runtime": 396.0229,
      "train_tokens_per_second": 4972.683
    },
    {
      "epoch": 38.15384615384615,
      "num_input_tokens_seen": 1969296,
      "step": 496,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 38.23076923076923,
      "grad_norm": 1.2178171873092651,
      "learning_rate": 0.01,
      "loss": 4.728215217590332,
      "num_input_tokens_seen": 1973392,
      "step": 497,
      "train_runtime": 396.83,
      "train_tokens_per_second": 4972.89
    },
    {
      "epoch": 38.23076923076923,
      "num_input_tokens_seen": 1973392,
      "step": 497,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 38.30769230769231,
      "grad_norm": 1.2290719747543335,
      "learning_rate": 0.01,
      "loss": 4.712850570678711,
      "num_input_tokens_seen": 1977488,
      "step": 498,
      "train_runtime": 397.6383,
      "train_tokens_per_second": 4973.082
    },
    {
      "epoch": 38.30769230769231,
      "num_input_tokens_seen": 1977488,
      "step": 498,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 38.38461538461539,
      "grad_norm": 1.0520877838134766,
      "learning_rate": 0.01,
      "loss": 4.671808242797852,
      "num_input_tokens_seen": 1981584,
      "step": 499,
      "train_runtime": 398.4459,
      "train_tokens_per_second": 4973.283
    },
    {
      "epoch": 38.38461538461539,
      "num_input_tokens_seen": 1981584,
      "step": 499,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 38.46153846153846,
      "grad_norm": 1.0742777585983276,
      "learning_rate": 0.01,
      "loss": 4.665500164031982,
      "num_input_tokens_seen": 1985680,
      "step": 500,
      "train_runtime": 399.253,
      "train_tokens_per_second": 4973.488
    },
    {
      "epoch": 38.46153846153846,
      "eval_CMD_3_branch_eval_accuracy": 0.00867992766726944,
      "eval_CMD_3_branch_eval_loss": 4.6620869636535645,
      "eval_CMD_3_branch_eval_runtime": 1.1323,
      "eval_CMD_3_branch_eval_samples_per_second": 2441.958,
      "eval_CMD_3_branch_eval_steps_per_second": 5.299,
      "num_input_tokens_seen": 1985680,
      "step": 500
    },
    {
      "epoch": 38.46153846153846,
      "num_input_tokens_seen": 1985680,
      "step": 500,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 38.53846153846154,
      "grad_norm": 0.9039981961250305,
      "learning_rate": 0.01,
      "loss": 4.649762153625488,
      "num_input_tokens_seen": 1989776,
      "step": 501,
      "train_runtime": 401.1973,
      "train_tokens_per_second": 4959.595
    },
    {
      "epoch": 38.53846153846154,
      "num_input_tokens_seen": 1989776,
      "step": 501,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 38.61538461538461,
      "grad_norm": 1.0192108154296875,
      "learning_rate": 0.01,
      "loss": 4.667194366455078,
      "num_input_tokens_seen": 1993872,
      "step": 502,
      "train_runtime": 402.005,
      "train_tokens_per_second": 4959.819
    },
    {
      "epoch": 38.61538461538461,
      "num_input_tokens_seen": 1993872,
      "step": 502,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 38.69230769230769,
      "grad_norm": 0.995784342288971,
      "learning_rate": 0.01,
      "loss": 4.644425868988037,
      "num_input_tokens_seen": 1997968,
      "step": 503,
      "train_runtime": 402.813,
      "train_tokens_per_second": 4960.039
    },
    {
      "epoch": 38.69230769230769,
      "num_input_tokens_seen": 1997968,
      "step": 503,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 38.76923076923077,
      "grad_norm": 1.5907785892486572,
      "learning_rate": 0.01,
      "loss": 4.670411109924316,
      "num_input_tokens_seen": 2002064,
      "step": 504,
      "train_runtime": 403.6209,
      "train_tokens_per_second": 4960.259
    },
    {
      "epoch": 38.76923076923077,
      "num_input_tokens_seen": 2002064,
      "step": 504,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 38.84615384615385,
      "grad_norm": 1.2107726335525513,
      "learning_rate": 0.01,
      "loss": 4.697729587554932,
      "num_input_tokens_seen": 2006160,
      "step": 505,
      "train_runtime": 404.4293,
      "train_tokens_per_second": 4960.472
    },
    {
      "epoch": 38.84615384615385,
      "num_input_tokens_seen": 2006160,
      "step": 505,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 38.92307692307692,
      "grad_norm": 1.4739183187484741,
      "learning_rate": 0.01,
      "loss": 4.691864013671875,
      "num_input_tokens_seen": 2010256,
      "step": 506,
      "train_runtime": 405.2324,
      "train_tokens_per_second": 4960.749
    },
    {
      "epoch": 38.92307692307692,
      "num_input_tokens_seen": 2010256,
      "step": 506,
      "train_accuracy": 0.0032573288772255182
    },
    {
      "epoch": 39.0,
      "grad_norm": 1.5223695039749146,
      "learning_rate": 0.01,
      "loss": 4.641530990600586,
      "num_input_tokens_seen": 2012712,
      "step": 507,
      "train_runtime": 405.7436,
      "train_tokens_per_second": 4960.551
    },
    {
      "epoch": 39.0,
      "num_input_tokens_seen": 2012712,
      "step": 507,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 39.07692307692308,
      "grad_norm": 1.0190341472625732,
      "learning_rate": 0.01,
      "loss": 4.653813362121582,
      "num_input_tokens_seen": 2016808,
      "step": 508,
      "train_runtime": 406.5634,
      "train_tokens_per_second": 4960.624
    },
    {
      "epoch": 39.07692307692308,
      "num_input_tokens_seen": 2016808,
      "step": 508,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 39.15384615384615,
      "grad_norm": 1.276046633720398,
      "learning_rate": 0.01,
      "loss": 4.758627891540527,
      "num_input_tokens_seen": 2020904,
      "step": 509,
      "train_runtime": 407.3709,
      "train_tokens_per_second": 4960.846
    },
    {
      "epoch": 39.15384615384615,
      "num_input_tokens_seen": 2020904,
      "step": 509,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 39.23076923076923,
      "grad_norm": 1.5082870721817017,
      "learning_rate": 0.01,
      "loss": 4.826314926147461,
      "num_input_tokens_seen": 2025000,
      "step": 510,
      "train_runtime": 408.1784,
      "train_tokens_per_second": 4961.066
    },
    {
      "epoch": 39.23076923076923,
      "num_input_tokens_seen": 2025000,
      "step": 510,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 39.30769230769231,
      "grad_norm": 2.431363344192505,
      "learning_rate": 0.01,
      "loss": 4.795228481292725,
      "num_input_tokens_seen": 2029096,
      "step": 511,
      "train_runtime": 408.9855,
      "train_tokens_per_second": 4961.29
    },
    {
      "epoch": 39.30769230769231,
      "num_input_tokens_seen": 2029096,
      "step": 511,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 39.38461538461539,
      "grad_norm": 1.5774593353271484,
      "learning_rate": 0.01,
      "loss": 4.711262226104736,
      "num_input_tokens_seen": 2033192,
      "step": 512,
      "train_runtime": 409.7934,
      "train_tokens_per_second": 4961.505
    },
    {
      "epoch": 39.38461538461539,
      "num_input_tokens_seen": 2033192,
      "step": 512,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 39.46153846153846,
      "grad_norm": 1.1461154222488403,
      "learning_rate": 0.01,
      "loss": 4.694036483764648,
      "num_input_tokens_seen": 2037288,
      "step": 513,
      "train_runtime": 410.6006,
      "train_tokens_per_second": 4961.726
    },
    {
      "epoch": 39.46153846153846,
      "num_input_tokens_seen": 2037288,
      "step": 513,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 39.53846153846154,
      "grad_norm": 1.1728637218475342,
      "learning_rate": 0.01,
      "loss": 4.677631378173828,
      "num_input_tokens_seen": 2041384,
      "step": 514,
      "train_runtime": 411.4092,
      "train_tokens_per_second": 4961.931
    },
    {
      "epoch": 39.53846153846154,
      "num_input_tokens_seen": 2041384,
      "step": 514,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 39.61538461538461,
      "grad_norm": 1.4426236152648926,
      "learning_rate": 0.01,
      "loss": 4.798910617828369,
      "num_input_tokens_seen": 2045480,
      "step": 515,
      "train_runtime": 412.2165,
      "train_tokens_per_second": 4962.15
    },
    {
      "epoch": 39.61538461538461,
      "num_input_tokens_seen": 2045480,
      "step": 515,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 39.69230769230769,
      "grad_norm": 1.2468035221099854,
      "learning_rate": 0.01,
      "loss": 4.738466262817383,
      "num_input_tokens_seen": 2049576,
      "step": 516,
      "train_runtime": 413.0238,
      "train_tokens_per_second": 4962.368
    },
    {
      "epoch": 39.69230769230769,
      "num_input_tokens_seen": 2049576,
      "step": 516,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 39.76923076923077,
      "grad_norm": 1.3579516410827637,
      "learning_rate": 0.01,
      "loss": 4.817930698394775,
      "num_input_tokens_seen": 2053672,
      "step": 517,
      "train_runtime": 413.8313,
      "train_tokens_per_second": 4962.582
    },
    {
      "epoch": 39.76923076923077,
      "num_input_tokens_seen": 2053672,
      "step": 517,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 39.84615384615385,
      "grad_norm": 0.9434919953346252,
      "learning_rate": 0.01,
      "loss": 4.678544998168945,
      "num_input_tokens_seen": 2057768,
      "step": 518,
      "train_runtime": 414.64,
      "train_tokens_per_second": 4962.783
    },
    {
      "epoch": 39.84615384615385,
      "num_input_tokens_seen": 2057768,
      "step": 518,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 39.92307692307692,
      "grad_norm": 2.377345323562622,
      "learning_rate": 0.01,
      "loss": 4.803607940673828,
      "num_input_tokens_seen": 2061864,
      "step": 519,
      "train_runtime": 415.4435,
      "train_tokens_per_second": 4963.043
    },
    {
      "epoch": 39.92307692307692,
      "num_input_tokens_seen": 2061864,
      "step": 519,
      "train_accuracy": 0.013029315508902073
    },
    {
      "epoch": 40.0,
      "grad_norm": 2.1644632816314697,
      "learning_rate": 0.01,
      "loss": 4.861001968383789,
      "num_input_tokens_seen": 2064320,
      "step": 520,
      "train_runtime": 415.9549,
      "train_tokens_per_second": 4962.846
    },
    {
      "epoch": 40.0,
      "num_input_tokens_seen": 2064320,
      "step": 520,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 40.07692307692308,
      "grad_norm": 1.0956887006759644,
      "learning_rate": 0.01,
      "loss": 4.702967643737793,
      "num_input_tokens_seen": 2068416,
      "step": 521,
      "train_runtime": 416.7749,
      "train_tokens_per_second": 4962.909
    },
    {
      "epoch": 40.07692307692308,
      "num_input_tokens_seen": 2068416,
      "step": 521,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 40.15384615384615,
      "grad_norm": 1.6368201971054077,
      "learning_rate": 0.01,
      "loss": 4.762077808380127,
      "num_input_tokens_seen": 2072512,
      "step": 522,
      "train_runtime": 417.5832,
      "train_tokens_per_second": 4963.111
    },
    {
      "epoch": 40.15384615384615,
      "num_input_tokens_seen": 2072512,
      "step": 522,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 40.23076923076923,
      "grad_norm": 0.9273180365562439,
      "learning_rate": 0.01,
      "loss": 4.668511390686035,
      "num_input_tokens_seen": 2076608,
      "step": 523,
      "train_runtime": 418.3917,
      "train_tokens_per_second": 4963.311
    },
    {
      "epoch": 40.23076923076923,
      "num_input_tokens_seen": 2076608,
      "step": 523,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 40.30769230769231,
      "grad_norm": 1.3700296878814697,
      "learning_rate": 0.01,
      "loss": 4.719908237457275,
      "num_input_tokens_seen": 2080704,
      "step": 524,
      "train_runtime": 419.199,
      "train_tokens_per_second": 4963.523
    },
    {
      "epoch": 40.30769230769231,
      "num_input_tokens_seen": 2080704,
      "step": 524,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 40.38461538461539,
      "grad_norm": 1.2383543252944946,
      "learning_rate": 0.01,
      "loss": 4.703868865966797,
      "num_input_tokens_seen": 2084800,
      "step": 525,
      "train_runtime": 420.0067,
      "train_tokens_per_second": 4963.73
    },
    {
      "epoch": 40.38461538461539,
      "num_input_tokens_seen": 2084800,
      "step": 525,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 40.46153846153846,
      "grad_norm": 0.9167964458465576,
      "learning_rate": 0.01,
      "loss": 4.673801422119141,
      "num_input_tokens_seen": 2088896,
      "step": 526,
      "train_runtime": 420.8142,
      "train_tokens_per_second": 4963.939
    },
    {
      "epoch": 40.46153846153846,
      "num_input_tokens_seen": 2088896,
      "step": 526,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 40.53846153846154,
      "grad_norm": 0.9295412302017212,
      "learning_rate": 0.01,
      "loss": 4.700968265533447,
      "num_input_tokens_seen": 2092992,
      "step": 527,
      "train_runtime": 421.6223,
      "train_tokens_per_second": 4964.14
    },
    {
      "epoch": 40.53846153846154,
      "num_input_tokens_seen": 2092992,
      "step": 527,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 40.61538461538461,
      "grad_norm": 1.023799180984497,
      "learning_rate": 0.01,
      "loss": 4.690126895904541,
      "num_input_tokens_seen": 2097088,
      "step": 528,
      "train_runtime": 422.4307,
      "train_tokens_per_second": 4964.336
    },
    {
      "epoch": 40.61538461538461,
      "num_input_tokens_seen": 2097088,
      "step": 528,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 40.69230769230769,
      "grad_norm": 0.9517922401428223,
      "learning_rate": 0.01,
      "loss": 4.7068867683410645,
      "num_input_tokens_seen": 2101184,
      "step": 529,
      "train_runtime": 423.2408,
      "train_tokens_per_second": 4964.512
    },
    {
      "epoch": 40.69230769230769,
      "num_input_tokens_seen": 2101184,
      "step": 529,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 40.76923076923077,
      "grad_norm": 0.964496910572052,
      "learning_rate": 0.01,
      "loss": 4.66574764251709,
      "num_input_tokens_seen": 2105280,
      "step": 530,
      "train_runtime": 424.0499,
      "train_tokens_per_second": 4964.699
    },
    {
      "epoch": 40.76923076923077,
      "num_input_tokens_seen": 2105280,
      "step": 530,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 40.84615384615385,
      "grad_norm": 0.8690102696418762,
      "learning_rate": 0.01,
      "loss": 4.646124839782715,
      "num_input_tokens_seen": 2109376,
      "step": 531,
      "train_runtime": 424.8589,
      "train_tokens_per_second": 4964.886
    },
    {
      "epoch": 40.84615384615385,
      "num_input_tokens_seen": 2109376,
      "step": 531,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 40.92307692307692,
      "grad_norm": 1.4332140684127808,
      "learning_rate": 0.01,
      "loss": 4.696255683898926,
      "num_input_tokens_seen": 2113472,
      "step": 532,
      "train_runtime": 425.6647,
      "train_tokens_per_second": 4965.11
    },
    {
      "epoch": 40.92307692307692,
      "num_input_tokens_seen": 2113472,
      "step": 532,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 41.0,
      "grad_norm": 1.0811644792556763,
      "learning_rate": 0.01,
      "loss": 4.628419399261475,
      "num_input_tokens_seen": 2115928,
      "step": 533,
      "train_runtime": 426.1769,
      "train_tokens_per_second": 4964.905
    },
    {
      "epoch": 41.0,
      "num_input_tokens_seen": 2115928,
      "step": 533,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 41.07692307692308,
      "grad_norm": 0.9979619383811951,
      "learning_rate": 0.01,
      "loss": 4.636567115783691,
      "num_input_tokens_seen": 2120024,
      "step": 534,
      "train_runtime": 426.9968,
      "train_tokens_per_second": 4964.965
    },
    {
      "epoch": 41.07692307692308,
      "num_input_tokens_seen": 2120024,
      "step": 534,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 41.15384615384615,
      "grad_norm": 1.1204509735107422,
      "learning_rate": 0.01,
      "loss": 4.685849189758301,
      "num_input_tokens_seen": 2124120,
      "step": 535,
      "train_runtime": 427.8066,
      "train_tokens_per_second": 4965.14
    },
    {
      "epoch": 41.15384615384615,
      "num_input_tokens_seen": 2124120,
      "step": 535,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 41.23076923076923,
      "grad_norm": 0.9062981605529785,
      "learning_rate": 0.01,
      "loss": 4.69474983215332,
      "num_input_tokens_seen": 2128216,
      "step": 536,
      "train_runtime": 428.6166,
      "train_tokens_per_second": 4965.314
    },
    {
      "epoch": 41.23076923076923,
      "num_input_tokens_seen": 2128216,
      "step": 536,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 41.30769230769231,
      "grad_norm": 1.2395418882369995,
      "learning_rate": 0.01,
      "loss": 4.679404258728027,
      "num_input_tokens_seen": 2132312,
      "step": 537,
      "train_runtime": 429.4268,
      "train_tokens_per_second": 4965.484
    },
    {
      "epoch": 41.30769230769231,
      "num_input_tokens_seen": 2132312,
      "step": 537,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 41.38461538461539,
      "grad_norm": 1.0222333669662476,
      "learning_rate": 0.01,
      "loss": 4.672848701477051,
      "num_input_tokens_seen": 2136408,
      "step": 538,
      "train_runtime": 430.2378,
      "train_tokens_per_second": 4965.645
    },
    {
      "epoch": 41.38461538461539,
      "num_input_tokens_seen": 2136408,
      "step": 538,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 41.46153846153846,
      "grad_norm": 0.8877430558204651,
      "learning_rate": 0.01,
      "loss": 4.6610307693481445,
      "num_input_tokens_seen": 2140504,
      "step": 539,
      "train_runtime": 431.0489,
      "train_tokens_per_second": 4965.803
    },
    {
      "epoch": 41.46153846153846,
      "num_input_tokens_seen": 2140504,
      "step": 539,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 41.53846153846154,
      "grad_norm": 1.4405112266540527,
      "learning_rate": 0.01,
      "loss": 4.744627952575684,
      "num_input_tokens_seen": 2144600,
      "step": 540,
      "train_runtime": 431.86,
      "train_tokens_per_second": 4965.961
    },
    {
      "epoch": 41.53846153846154,
      "num_input_tokens_seen": 2144600,
      "step": 540,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 41.61538461538461,
      "grad_norm": 1.2398396730422974,
      "learning_rate": 0.01,
      "loss": 4.720105171203613,
      "num_input_tokens_seen": 2148696,
      "step": 541,
      "train_runtime": 432.6712,
      "train_tokens_per_second": 4966.118
    },
    {
      "epoch": 41.61538461538461,
      "num_input_tokens_seen": 2148696,
      "step": 541,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 41.69230769230769,
      "grad_norm": 1.0383821725845337,
      "learning_rate": 0.01,
      "loss": 4.685735702514648,
      "num_input_tokens_seen": 2152792,
      "step": 542,
      "train_runtime": 433.481,
      "train_tokens_per_second": 4966.289
    },
    {
      "epoch": 41.69230769230769,
      "num_input_tokens_seen": 2152792,
      "step": 542,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 41.76923076923077,
      "grad_norm": 1.4199179410934448,
      "learning_rate": 0.01,
      "loss": 4.653835296630859,
      "num_input_tokens_seen": 2156888,
      "step": 543,
      "train_runtime": 434.2917,
      "train_tokens_per_second": 4966.45
    },
    {
      "epoch": 41.76923076923077,
      "num_input_tokens_seen": 2156888,
      "step": 543,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 41.84615384615385,
      "grad_norm": 1.0380232334136963,
      "learning_rate": 0.01,
      "loss": 4.687129974365234,
      "num_input_tokens_seen": 2160984,
      "step": 544,
      "train_runtime": 435.1018,
      "train_tokens_per_second": 4966.617
    },
    {
      "epoch": 41.84615384615385,
      "num_input_tokens_seen": 2160984,
      "step": 544,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 41.92307692307692,
      "grad_norm": 1.3905823230743408,
      "learning_rate": 0.01,
      "loss": 4.661175727844238,
      "num_input_tokens_seen": 2165080,
      "step": 545,
      "train_runtime": 435.9083,
      "train_tokens_per_second": 4966.824
    },
    {
      "epoch": 41.92307692307692,
      "num_input_tokens_seen": 2165080,
      "step": 545,
      "train_accuracy": 0.019543973729014397
    },
    {
      "epoch": 42.0,
      "grad_norm": 1.2243058681488037,
      "learning_rate": 0.01,
      "loss": 4.706570148468018,
      "num_input_tokens_seen": 2167536,
      "step": 546,
      "train_runtime": 436.4222,
      "train_tokens_per_second": 4966.603
    },
    {
      "epoch": 42.0,
      "num_input_tokens_seen": 2167536,
      "step": 546,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 42.07692307692308,
      "grad_norm": 1.379681944847107,
      "learning_rate": 0.01,
      "loss": 4.6904377937316895,
      "num_input_tokens_seen": 2171632,
      "step": 547,
      "train_runtime": 437.2445,
      "train_tokens_per_second": 4966.63
    },
    {
      "epoch": 42.07692307692308,
      "num_input_tokens_seen": 2171632,
      "step": 547,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 42.15384615384615,
      "grad_norm": 1.0322391986846924,
      "learning_rate": 0.01,
      "loss": 4.69384765625,
      "num_input_tokens_seen": 2175728,
      "step": 548,
      "train_runtime": 438.0548,
      "train_tokens_per_second": 4966.794
    },
    {
      "epoch": 42.15384615384615,
      "num_input_tokens_seen": 2175728,
      "step": 548,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 42.23076923076923,
      "grad_norm": 1.4542628526687622,
      "learning_rate": 0.01,
      "loss": 4.700136661529541,
      "num_input_tokens_seen": 2179824,
      "step": 549,
      "train_runtime": 438.8669,
      "train_tokens_per_second": 4966.937
    },
    {
      "epoch": 42.23076923076923,
      "num_input_tokens_seen": 2179824,
      "step": 549,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 42.30769230769231,
      "grad_norm": 1.1556236743927002,
      "learning_rate": 0.01,
      "loss": 4.677280426025391,
      "num_input_tokens_seen": 2183920,
      "step": 550,
      "train_runtime": 439.677,
      "train_tokens_per_second": 4967.101
    },
    {
      "epoch": 42.30769230769231,
      "num_input_tokens_seen": 2183920,
      "step": 550,
      "train_accuracy": 0.0
    },
    {
      "epoch": 42.38461538461539,
      "grad_norm": 1.9287383556365967,
      "learning_rate": 0.01,
      "loss": 4.783945083618164,
      "num_input_tokens_seen": 2188016,
      "step": 551,
      "train_runtime": 440.4864,
      "train_tokens_per_second": 4967.273
    },
    {
      "epoch": 42.38461538461539,
      "num_input_tokens_seen": 2188016,
      "step": 551,
      "train_accuracy": 0.001953125
    },
    {
      "epoch": 42.46153846153846,
      "grad_norm": 2.0787062644958496,
      "learning_rate": 0.01,
      "loss": 4.8121337890625,
      "num_input_tokens_seen": 2192112,
      "step": 552,
      "train_runtime": 441.2953,
      "train_tokens_per_second": 4967.45
    },
    {
      "epoch": 42.46153846153846,
      "num_input_tokens_seen": 2192112,
      "step": 552,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 42.53846153846154,
      "grad_norm": 0.8324519991874695,
      "learning_rate": 0.01,
      "loss": 4.6436543464660645,
      "num_input_tokens_seen": 2196208,
      "step": 553,
      "train_runtime": 442.1041,
      "train_tokens_per_second": 4967.627
    },
    {
      "epoch": 42.53846153846154,
      "num_input_tokens_seen": 2196208,
      "step": 553,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 42.61538461538461,
      "grad_norm": 1.7577073574066162,
      "learning_rate": 0.01,
      "loss": 4.760754585266113,
      "num_input_tokens_seen": 2200304,
      "step": 554,
      "train_runtime": 442.9132,
      "train_tokens_per_second": 4967.8
    },
    {
      "epoch": 42.61538461538461,
      "num_input_tokens_seen": 2200304,
      "step": 554,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 42.69230769230769,
      "grad_norm": 0.9620968699455261,
      "learning_rate": 0.01,
      "loss": 4.675868034362793,
      "num_input_tokens_seen": 2204400,
      "step": 555,
      "train_runtime": 443.7235,
      "train_tokens_per_second": 4967.958
    },
    {
      "epoch": 42.69230769230769,
      "num_input_tokens_seen": 2204400,
      "step": 555,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 42.76923076923077,
      "grad_norm": 0.7743176817893982,
      "learning_rate": 0.01,
      "loss": 4.648281097412109,
      "num_input_tokens_seen": 2208496,
      "step": 556,
      "train_runtime": 444.5325,
      "train_tokens_per_second": 4968.131
    },
    {
      "epoch": 42.76923076923077,
      "num_input_tokens_seen": 2208496,
      "step": 556,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 42.84615384615385,
      "grad_norm": 0.7625809907913208,
      "learning_rate": 0.01,
      "loss": 4.640504360198975,
      "num_input_tokens_seen": 2212592,
      "step": 557,
      "train_runtime": 445.3413,
      "train_tokens_per_second": 4968.307
    },
    {
      "epoch": 42.84615384615385,
      "num_input_tokens_seen": 2212592,
      "step": 557,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 42.92307692307692,
      "grad_norm": 0.9949125051498413,
      "learning_rate": 0.01,
      "loss": 4.657556056976318,
      "num_input_tokens_seen": 2216688,
      "step": 558,
      "train_runtime": 446.1452,
      "train_tokens_per_second": 4968.535
    },
    {
      "epoch": 42.92307692307692,
      "num_input_tokens_seen": 2216688,
      "step": 558,
      "train_accuracy": 0.016286645084619522
    },
    {
      "epoch": 43.0,
      "grad_norm": 0.904716968536377,
      "learning_rate": 0.01,
      "loss": 4.625843524932861,
      "num_input_tokens_seen": 2219144,
      "step": 559,
      "train_runtime": 446.6577,
      "train_tokens_per_second": 4968.333
    },
    {
      "epoch": 43.0,
      "num_input_tokens_seen": 2219144,
      "step": 559,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 43.07692307692308,
      "grad_norm": 1.0351663827896118,
      "learning_rate": 0.01,
      "loss": 4.719288349151611,
      "num_input_tokens_seen": 2223240,
      "step": 560,
      "train_runtime": 447.4781,
      "train_tokens_per_second": 4968.377
    },
    {
      "epoch": 43.07692307692308,
      "num_input_tokens_seen": 2223240,
      "step": 560,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 43.15384615384615,
      "grad_norm": 0.7712958455085754,
      "learning_rate": 0.01,
      "loss": 4.643596172332764,
      "num_input_tokens_seen": 2227336,
      "step": 561,
      "train_runtime": 448.2873,
      "train_tokens_per_second": 4968.546
    },
    {
      "epoch": 43.15384615384615,
      "num_input_tokens_seen": 2227336,
      "step": 561,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 43.23076923076923,
      "grad_norm": 0.703096330165863,
      "learning_rate": 0.01,
      "loss": 4.656213283538818,
      "num_input_tokens_seen": 2231432,
      "step": 562,
      "train_runtime": 449.0965,
      "train_tokens_per_second": 4968.714
    },
    {
      "epoch": 43.23076923076923,
      "num_input_tokens_seen": 2231432,
      "step": 562,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 43.30769230769231,
      "grad_norm": 0.9262785315513611,
      "learning_rate": 0.01,
      "loss": 4.732090950012207,
      "num_input_tokens_seen": 2235528,
      "step": 563,
      "train_runtime": 449.906,
      "train_tokens_per_second": 4968.878
    },
    {
      "epoch": 43.30769230769231,
      "num_input_tokens_seen": 2235528,
      "step": 563,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 43.38461538461539,
      "grad_norm": 0.8662072420120239,
      "learning_rate": 0.01,
      "loss": 4.692700386047363,
      "num_input_tokens_seen": 2239624,
      "step": 564,
      "train_runtime": 450.7149,
      "train_tokens_per_second": 4969.048
    },
    {
      "epoch": 43.38461538461539,
      "num_input_tokens_seen": 2239624,
      "step": 564,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 43.46153846153846,
      "grad_norm": 0.7943282127380371,
      "learning_rate": 0.01,
      "loss": 4.661947250366211,
      "num_input_tokens_seen": 2243720,
      "step": 565,
      "train_runtime": 451.5245,
      "train_tokens_per_second": 4969.209
    },
    {
      "epoch": 43.46153846153846,
      "num_input_tokens_seen": 2243720,
      "step": 565,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 43.53846153846154,
      "grad_norm": 0.8779962658882141,
      "learning_rate": 0.01,
      "loss": 4.677492141723633,
      "num_input_tokens_seen": 2247816,
      "step": 566,
      "train_runtime": 452.3333,
      "train_tokens_per_second": 4969.38
    },
    {
      "epoch": 43.53846153846154,
      "num_input_tokens_seen": 2247816,
      "step": 566,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 43.61538461538461,
      "grad_norm": 0.7035842537879944,
      "learning_rate": 0.01,
      "loss": 4.659738540649414,
      "num_input_tokens_seen": 2251912,
      "step": 567,
      "train_runtime": 453.1423,
      "train_tokens_per_second": 4969.547
    },
    {
      "epoch": 43.61538461538461,
      "num_input_tokens_seen": 2251912,
      "step": 567,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 43.69230769230769,
      "grad_norm": 0.7953966856002808,
      "learning_rate": 0.01,
      "loss": 4.664793968200684,
      "num_input_tokens_seen": 2256008,
      "step": 568,
      "train_runtime": 453.9513,
      "train_tokens_per_second": 4969.713
    },
    {
      "epoch": 43.69230769230769,
      "num_input_tokens_seen": 2256008,
      "step": 568,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 43.76923076923077,
      "grad_norm": 0.7259625792503357,
      "learning_rate": 0.01,
      "loss": 4.659206390380859,
      "num_input_tokens_seen": 2260104,
      "step": 569,
      "train_runtime": 454.7604,
      "train_tokens_per_second": 4969.879
    },
    {
      "epoch": 43.76923076923077,
      "num_input_tokens_seen": 2260104,
      "step": 569,
      "train_accuracy": 0.001953125
    },
    {
      "epoch": 43.84615384615385,
      "grad_norm": 0.6541624069213867,
      "learning_rate": 0.01,
      "loss": 4.595702171325684,
      "num_input_tokens_seen": 2264200,
      "step": 570,
      "train_runtime": 455.5696,
      "train_tokens_per_second": 4970.042
    },
    {
      "epoch": 43.84615384615385,
      "num_input_tokens_seen": 2264200,
      "step": 570,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 43.92307692307692,
      "grad_norm": 0.5536810755729675,
      "learning_rate": 0.01,
      "loss": 4.600469589233398,
      "num_input_tokens_seen": 2268296,
      "step": 571,
      "train_runtime": 456.3751,
      "train_tokens_per_second": 4970.245
    },
    {
      "epoch": 43.92307692307692,
      "num_input_tokens_seen": 2268296,
      "step": 571,
      "train_accuracy": 0.02280130237340927
    },
    {
      "epoch": 44.0,
      "grad_norm": 0.7504650354385376,
      "learning_rate": 0.01,
      "loss": 4.601952075958252,
      "num_input_tokens_seen": 2270752,
      "step": 572,
      "train_runtime": 456.8879,
      "train_tokens_per_second": 4970.042
    },
    {
      "epoch": 44.0,
      "num_input_tokens_seen": 2270752,
      "step": 572,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 44.07692307692308,
      "grad_norm": 0.9998853206634521,
      "learning_rate": 0.01,
      "loss": 4.685545921325684,
      "num_input_tokens_seen": 2274848,
      "step": 573,
      "train_runtime": 457.7091,
      "train_tokens_per_second": 4970.074
    },
    {
      "epoch": 44.07692307692308,
      "num_input_tokens_seen": 2274848,
      "step": 573,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 44.15384615384615,
      "grad_norm": 0.8141763806343079,
      "learning_rate": 0.01,
      "loss": 4.6422014236450195,
      "num_input_tokens_seen": 2278944,
      "step": 574,
      "train_runtime": 458.5186,
      "train_tokens_per_second": 4970.233
    },
    {
      "epoch": 44.15384615384615,
      "num_input_tokens_seen": 2278944,
      "step": 574,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 44.23076923076923,
      "grad_norm": 0.8815345764160156,
      "learning_rate": 0.01,
      "loss": 4.654372692108154,
      "num_input_tokens_seen": 2283040,
      "step": 575,
      "train_runtime": 459.3279,
      "train_tokens_per_second": 4970.393
    },
    {
      "epoch": 44.23076923076923,
      "num_input_tokens_seen": 2283040,
      "step": 575,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 44.30769230769231,
      "grad_norm": 0.6858783960342407,
      "learning_rate": 0.01,
      "loss": 4.626808166503906,
      "num_input_tokens_seen": 2287136,
      "step": 576,
      "train_runtime": 460.136,
      "train_tokens_per_second": 4970.565
    },
    {
      "epoch": 44.30769230769231,
      "num_input_tokens_seen": 2287136,
      "step": 576,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 44.38461538461539,
      "grad_norm": 0.957760214805603,
      "learning_rate": 0.01,
      "loss": 4.706740379333496,
      "num_input_tokens_seen": 2291232,
      "step": 577,
      "train_runtime": 460.9444,
      "train_tokens_per_second": 4970.734
    },
    {
      "epoch": 44.38461538461539,
      "num_input_tokens_seen": 2291232,
      "step": 577,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 44.46153846153846,
      "grad_norm": 0.8295272588729858,
      "learning_rate": 0.01,
      "loss": 4.658412933349609,
      "num_input_tokens_seen": 2295328,
      "step": 578,
      "train_runtime": 461.7536,
      "train_tokens_per_second": 4970.894
    },
    {
      "epoch": 44.46153846153846,
      "num_input_tokens_seen": 2295328,
      "step": 578,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 44.53846153846154,
      "grad_norm": 0.8796570897102356,
      "learning_rate": 0.01,
      "loss": 4.714468002319336,
      "num_input_tokens_seen": 2299424,
      "step": 579,
      "train_runtime": 462.5619,
      "train_tokens_per_second": 4971.062
    },
    {
      "epoch": 44.53846153846154,
      "num_input_tokens_seen": 2299424,
      "step": 579,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 44.61538461538461,
      "grad_norm": 0.8227230906486511,
      "learning_rate": 0.01,
      "loss": 4.655007362365723,
      "num_input_tokens_seen": 2303520,
      "step": 580,
      "train_runtime": 463.3708,
      "train_tokens_per_second": 4971.224
    },
    {
      "epoch": 44.61538461538461,
      "num_input_tokens_seen": 2303520,
      "step": 580,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 44.69230769230769,
      "grad_norm": 0.8670706748962402,
      "learning_rate": 0.01,
      "loss": 4.666321754455566,
      "num_input_tokens_seen": 2307616,
      "step": 581,
      "train_runtime": 464.1803,
      "train_tokens_per_second": 4971.378
    },
    {
      "epoch": 44.69230769230769,
      "num_input_tokens_seen": 2307616,
      "step": 581,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 44.76923076923077,
      "grad_norm": 0.955562949180603,
      "learning_rate": 0.01,
      "loss": 4.649089813232422,
      "num_input_tokens_seen": 2311712,
      "step": 582,
      "train_runtime": 464.9888,
      "train_tokens_per_second": 4971.544
    },
    {
      "epoch": 44.76923076923077,
      "num_input_tokens_seen": 2311712,
      "step": 582,
      "train_accuracy": 0.001953125
    },
    {
      "epoch": 44.84615384615385,
      "grad_norm": 0.906722903251648,
      "learning_rate": 0.01,
      "loss": 4.669605255126953,
      "num_input_tokens_seen": 2315808,
      "step": 583,
      "train_runtime": 465.7973,
      "train_tokens_per_second": 4971.708
    },
    {
      "epoch": 44.84615384615385,
      "num_input_tokens_seen": 2315808,
      "step": 583,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 44.92307692307692,
      "grad_norm": 0.9096651673316956,
      "learning_rate": 0.01,
      "loss": 4.6865081787109375,
      "num_input_tokens_seen": 2319904,
      "step": 584,
      "train_runtime": 466.6023,
      "train_tokens_per_second": 4971.909
    },
    {
      "epoch": 44.92307692307692,
      "num_input_tokens_seen": 2319904,
      "step": 584,
      "train_accuracy": 0.016286645084619522
    },
    {
      "epoch": 45.0,
      "grad_norm": 0.8906012773513794,
      "learning_rate": 0.01,
      "loss": 4.6353349685668945,
      "num_input_tokens_seen": 2322360,
      "step": 585,
      "train_runtime": 467.1148,
      "train_tokens_per_second": 4971.712
    },
    {
      "epoch": 45.0,
      "num_input_tokens_seen": 2322360,
      "step": 585,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 45.07692307692308,
      "grad_norm": 0.8796638250350952,
      "learning_rate": 0.01,
      "loss": 4.698246955871582,
      "num_input_tokens_seen": 2326456,
      "step": 586,
      "train_runtime": 467.9364,
      "train_tokens_per_second": 4971.736
    },
    {
      "epoch": 45.07692307692308,
      "num_input_tokens_seen": 2326456,
      "step": 586,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 45.15384615384615,
      "grad_norm": 1.1915932893753052,
      "learning_rate": 0.01,
      "loss": 4.661253929138184,
      "num_input_tokens_seen": 2330552,
      "step": 587,
      "train_runtime": 468.7458,
      "train_tokens_per_second": 4971.889
    },
    {
      "epoch": 45.15384615384615,
      "num_input_tokens_seen": 2330552,
      "step": 587,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 45.23076923076923,
      "grad_norm": 0.899604082107544,
      "learning_rate": 0.01,
      "loss": 4.670432090759277,
      "num_input_tokens_seen": 2334648,
      "step": 588,
      "train_runtime": 469.5557,
      "train_tokens_per_second": 4972.036
    },
    {
      "epoch": 45.23076923076923,
      "num_input_tokens_seen": 2334648,
      "step": 588,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 45.30769230769231,
      "grad_norm": 1.1505882740020752,
      "learning_rate": 0.01,
      "loss": 4.745501518249512,
      "num_input_tokens_seen": 2338744,
      "step": 589,
      "train_runtime": 470.3646,
      "train_tokens_per_second": 4972.193
    },
    {
      "epoch": 45.30769230769231,
      "num_input_tokens_seen": 2338744,
      "step": 589,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 45.38461538461539,
      "grad_norm": 1.0811315774917603,
      "learning_rate": 0.01,
      "loss": 4.690670013427734,
      "num_input_tokens_seen": 2342840,
      "step": 590,
      "train_runtime": 471.1758,
      "train_tokens_per_second": 4972.327
    },
    {
      "epoch": 45.38461538461539,
      "num_input_tokens_seen": 2342840,
      "step": 590,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 45.46153846153846,
      "grad_norm": 0.9773861169815063,
      "learning_rate": 0.01,
      "loss": 4.728013515472412,
      "num_input_tokens_seen": 2346936,
      "step": 591,
      "train_runtime": 471.9855,
      "train_tokens_per_second": 4972.474
    },
    {
      "epoch": 45.46153846153846,
      "num_input_tokens_seen": 2346936,
      "step": 591,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 45.53846153846154,
      "grad_norm": 1.1385358572006226,
      "learning_rate": 0.01,
      "loss": 4.697521209716797,
      "num_input_tokens_seen": 2351032,
      "step": 592,
      "train_runtime": 472.7948,
      "train_tokens_per_second": 4972.626
    },
    {
      "epoch": 45.53846153846154,
      "num_input_tokens_seen": 2351032,
      "step": 592,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 45.61538461538461,
      "grad_norm": 0.7973088026046753,
      "learning_rate": 0.01,
      "loss": 4.714956283569336,
      "num_input_tokens_seen": 2355128,
      "step": 593,
      "train_runtime": 473.604,
      "train_tokens_per_second": 4972.779
    },
    {
      "epoch": 45.61538461538461,
      "num_input_tokens_seen": 2355128,
      "step": 593,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 45.69230769230769,
      "grad_norm": 0.8927688598632812,
      "learning_rate": 0.01,
      "loss": 4.65383243560791,
      "num_input_tokens_seen": 2359224,
      "step": 594,
      "train_runtime": 474.4134,
      "train_tokens_per_second": 4972.929
    },
    {
      "epoch": 45.69230769230769,
      "num_input_tokens_seen": 2359224,
      "step": 594,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 45.76923076923077,
      "grad_norm": 0.79509037733078,
      "learning_rate": 0.01,
      "loss": 4.636276721954346,
      "num_input_tokens_seen": 2363320,
      "step": 595,
      "train_runtime": 475.2229,
      "train_tokens_per_second": 4973.077
    },
    {
      "epoch": 45.76923076923077,
      "num_input_tokens_seen": 2363320,
      "step": 595,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 45.84615384615385,
      "grad_norm": 0.9098619222640991,
      "learning_rate": 0.01,
      "loss": 4.748471260070801,
      "num_input_tokens_seen": 2367416,
      "step": 596,
      "train_runtime": 476.032,
      "train_tokens_per_second": 4973.229
    },
    {
      "epoch": 45.84615384615385,
      "num_input_tokens_seen": 2367416,
      "step": 596,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 45.92307692307692,
      "grad_norm": 0.6853192448616028,
      "learning_rate": 0.01,
      "loss": 4.684442520141602,
      "num_input_tokens_seen": 2371512,
      "step": 597,
      "train_runtime": 476.838,
      "train_tokens_per_second": 4973.413
    },
    {
      "epoch": 45.92307692307692,
      "num_input_tokens_seen": 2371512,
      "step": 597,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 46.0,
      "grad_norm": 1.149055004119873,
      "learning_rate": 0.01,
      "loss": 4.697789192199707,
      "num_input_tokens_seen": 2373968,
      "step": 598,
      "train_runtime": 477.3504,
      "train_tokens_per_second": 4973.218
    },
    {
      "epoch": 46.0,
      "num_input_tokens_seen": 2373968,
      "step": 598,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 46.07692307692308,
      "grad_norm": 0.8396105766296387,
      "learning_rate": 0.01,
      "loss": 4.683511734008789,
      "num_input_tokens_seen": 2378064,
      "step": 599,
      "train_runtime": 478.1716,
      "train_tokens_per_second": 4973.244
    },
    {
      "epoch": 46.07692307692308,
      "num_input_tokens_seen": 2378064,
      "step": 599,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 46.15384615384615,
      "grad_norm": 0.7613953948020935,
      "learning_rate": 0.01,
      "loss": 4.631681442260742,
      "num_input_tokens_seen": 2382160,
      "step": 600,
      "train_runtime": 478.9815,
      "train_tokens_per_second": 4973.387
    },
    {
      "epoch": 46.15384615384615,
      "eval_CMD_3_branch_eval_accuracy": 0.007956600361663653,
      "eval_CMD_3_branch_eval_loss": 4.713253021240234,
      "eval_CMD_3_branch_eval_runtime": 1.1395,
      "eval_CMD_3_branch_eval_samples_per_second": 2426.455,
      "eval_CMD_3_branch_eval_steps_per_second": 5.265,
      "num_input_tokens_seen": 2382160,
      "step": 600
    },
    {
      "epoch": 46.15384615384615,
      "num_input_tokens_seen": 2382160,
      "step": 600,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 46.23076923076923,
      "grad_norm": 0.7381619811058044,
      "learning_rate": 0.01,
      "loss": 4.686368942260742,
      "num_input_tokens_seen": 2386256,
      "step": 601,
      "train_runtime": 480.9359,
      "train_tokens_per_second": 4961.693
    },
    {
      "epoch": 46.23076923076923,
      "num_input_tokens_seen": 2386256,
      "step": 601,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 46.30769230769231,
      "grad_norm": 0.5763469934463501,
      "learning_rate": 0.01,
      "loss": 4.618984222412109,
      "num_input_tokens_seen": 2390352,
      "step": 602,
      "train_runtime": 481.7451,
      "train_tokens_per_second": 4961.861
    },
    {
      "epoch": 46.30769230769231,
      "num_input_tokens_seen": 2390352,
      "step": 602,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 46.38461538461539,
      "grad_norm": 0.6603495478630066,
      "learning_rate": 0.01,
      "loss": 4.622781753540039,
      "num_input_tokens_seen": 2394448,
      "step": 603,
      "train_runtime": 482.5554,
      "train_tokens_per_second": 4962.017
    },
    {
      "epoch": 46.38461538461539,
      "num_input_tokens_seen": 2394448,
      "step": 603,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 46.46153846153846,
      "grad_norm": 1.2545437812805176,
      "learning_rate": 0.01,
      "loss": 4.671928405761719,
      "num_input_tokens_seen": 2398544,
      "step": 604,
      "train_runtime": 483.3655,
      "train_tokens_per_second": 4962.174
    },
    {
      "epoch": 46.46153846153846,
      "num_input_tokens_seen": 2398544,
      "step": 604,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 46.53846153846154,
      "grad_norm": 1.3104106187820435,
      "learning_rate": 0.01,
      "loss": 4.700450897216797,
      "num_input_tokens_seen": 2402640,
      "step": 605,
      "train_runtime": 484.1767,
      "train_tokens_per_second": 4962.32
    },
    {
      "epoch": 46.53846153846154,
      "num_input_tokens_seen": 2402640,
      "step": 605,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 46.61538461538461,
      "grad_norm": 0.7127174139022827,
      "learning_rate": 0.01,
      "loss": 4.673466205596924,
      "num_input_tokens_seen": 2406736,
      "step": 606,
      "train_runtime": 484.9871,
      "train_tokens_per_second": 4962.475
    },
    {
      "epoch": 46.61538461538461,
      "num_input_tokens_seen": 2406736,
      "step": 606,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 46.69230769230769,
      "grad_norm": 0.7060525417327881,
      "learning_rate": 0.01,
      "loss": 4.623709678649902,
      "num_input_tokens_seen": 2410832,
      "step": 607,
      "train_runtime": 485.797,
      "train_tokens_per_second": 4962.633
    },
    {
      "epoch": 46.69230769230769,
      "num_input_tokens_seen": 2410832,
      "step": 607,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 46.76923076923077,
      "grad_norm": 0.6495184302330017,
      "learning_rate": 0.01,
      "loss": 4.624743938446045,
      "num_input_tokens_seen": 2414928,
      "step": 608,
      "train_runtime": 486.6082,
      "train_tokens_per_second": 4962.777
    },
    {
      "epoch": 46.76923076923077,
      "num_input_tokens_seen": 2414928,
      "step": 608,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 46.84615384615385,
      "grad_norm": 0.6529079675674438,
      "learning_rate": 0.01,
      "loss": 4.624449729919434,
      "num_input_tokens_seen": 2419024,
      "step": 609,
      "train_runtime": 487.418,
      "train_tokens_per_second": 4962.935
    },
    {
      "epoch": 46.84615384615385,
      "num_input_tokens_seen": 2419024,
      "step": 609,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 46.92307692307692,
      "grad_norm": 0.888717532157898,
      "learning_rate": 0.01,
      "loss": 4.675288200378418,
      "num_input_tokens_seen": 2423120,
      "step": 610,
      "train_runtime": 488.2232,
      "train_tokens_per_second": 4963.139
    },
    {
      "epoch": 46.92307692307692,
      "num_input_tokens_seen": 2423120,
      "step": 610,
      "train_accuracy": 0.019543973729014397
    },
    {
      "epoch": 47.0,
      "grad_norm": 1.031438946723938,
      "learning_rate": 0.01,
      "loss": 4.670546054840088,
      "num_input_tokens_seen": 2425576,
      "step": 611,
      "train_runtime": 488.7364,
      "train_tokens_per_second": 4962.953
    },
    {
      "epoch": 47.0,
      "num_input_tokens_seen": 2425576,
      "step": 611,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 47.07692307692308,
      "grad_norm": 0.6399885416030884,
      "learning_rate": 0.01,
      "loss": 4.624326705932617,
      "num_input_tokens_seen": 2429672,
      "step": 612,
      "train_runtime": 489.558,
      "train_tokens_per_second": 4962.991
    },
    {
      "epoch": 47.07692307692308,
      "num_input_tokens_seen": 2429672,
      "step": 612,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 47.15384615384615,
      "grad_norm": 0.8436721563339233,
      "learning_rate": 0.01,
      "loss": 4.636343479156494,
      "num_input_tokens_seen": 2433768,
      "step": 613,
      "train_runtime": 490.3673,
      "train_tokens_per_second": 4963.154
    },
    {
      "epoch": 47.15384615384615,
      "num_input_tokens_seen": 2433768,
      "step": 613,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 47.23076923076923,
      "grad_norm": 0.6118916273117065,
      "learning_rate": 0.01,
      "loss": 4.632634162902832,
      "num_input_tokens_seen": 2437864,
      "step": 614,
      "train_runtime": 491.1772,
      "train_tokens_per_second": 4963.309
    },
    {
      "epoch": 47.23076923076923,
      "num_input_tokens_seen": 2437864,
      "step": 614,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 47.30769230769231,
      "grad_norm": 0.922429084777832,
      "learning_rate": 0.01,
      "loss": 4.65765380859375,
      "num_input_tokens_seen": 2441960,
      "step": 615,
      "train_runtime": 491.9875,
      "train_tokens_per_second": 4963.46
    },
    {
      "epoch": 47.30769230769231,
      "num_input_tokens_seen": 2441960,
      "step": 615,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 47.38461538461539,
      "grad_norm": 0.5347769260406494,
      "learning_rate": 0.01,
      "loss": 4.66064453125,
      "num_input_tokens_seen": 2446056,
      "step": 616,
      "train_runtime": 492.7971,
      "train_tokens_per_second": 4963.617
    },
    {
      "epoch": 47.38461538461539,
      "num_input_tokens_seen": 2446056,
      "step": 616,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 47.46153846153846,
      "grad_norm": 1.1451373100280762,
      "learning_rate": 0.01,
      "loss": 4.636528968811035,
      "num_input_tokens_seen": 2450152,
      "step": 617,
      "train_runtime": 493.6084,
      "train_tokens_per_second": 4963.756
    },
    {
      "epoch": 47.46153846153846,
      "num_input_tokens_seen": 2450152,
      "step": 617,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 47.53846153846154,
      "grad_norm": 0.7594324350357056,
      "learning_rate": 0.01,
      "loss": 4.647350311279297,
      "num_input_tokens_seen": 2454248,
      "step": 618,
      "train_runtime": 494.418,
      "train_tokens_per_second": 4963.913
    },
    {
      "epoch": 47.53846153846154,
      "num_input_tokens_seen": 2454248,
      "step": 618,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 47.61538461538461,
      "grad_norm": 0.7688575387001038,
      "learning_rate": 0.01,
      "loss": 4.643795013427734,
      "num_input_tokens_seen": 2458344,
      "step": 619,
      "train_runtime": 495.2287,
      "train_tokens_per_second": 4964.058
    },
    {
      "epoch": 47.61538461538461,
      "num_input_tokens_seen": 2458344,
      "step": 619,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 47.69230769230769,
      "grad_norm": 0.571732759475708,
      "learning_rate": 0.01,
      "loss": 4.648545265197754,
      "num_input_tokens_seen": 2462440,
      "step": 620,
      "train_runtime": 496.0391,
      "train_tokens_per_second": 4964.205
    },
    {
      "epoch": 47.69230769230769,
      "num_input_tokens_seen": 2462440,
      "step": 620,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 47.76923076923077,
      "grad_norm": 0.9158742427825928,
      "learning_rate": 0.01,
      "loss": 4.65397834777832,
      "num_input_tokens_seen": 2466536,
      "step": 621,
      "train_runtime": 496.8495,
      "train_tokens_per_second": 4964.352
    },
    {
      "epoch": 47.76923076923077,
      "num_input_tokens_seen": 2466536,
      "step": 621,
      "train_accuracy": 0.001953125
    },
    {
      "epoch": 47.84615384615385,
      "grad_norm": 0.6186763048171997,
      "learning_rate": 0.01,
      "loss": 4.6529364585876465,
      "num_input_tokens_seen": 2470632,
      "step": 622,
      "train_runtime": 497.6592,
      "train_tokens_per_second": 4964.505
    },
    {
      "epoch": 47.84615384615385,
      "num_input_tokens_seen": 2470632,
      "step": 622,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 47.92307692307692,
      "grad_norm": 0.6545286774635315,
      "learning_rate": 0.01,
      "loss": 4.616402626037598,
      "num_input_tokens_seen": 2474728,
      "step": 623,
      "train_runtime": 498.4647,
      "train_tokens_per_second": 4964.701
    },
    {
      "epoch": 47.92307692307692,
      "num_input_tokens_seen": 2474728,
      "step": 623,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 48.0,
      "grad_norm": 0.7522267699241638,
      "learning_rate": 0.01,
      "loss": 4.644376754760742,
      "num_input_tokens_seen": 2477184,
      "step": 624,
      "train_runtime": 498.9772,
      "train_tokens_per_second": 4964.523
    },
    {
      "epoch": 48.0,
      "num_input_tokens_seen": 2477184,
      "step": 624,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 48.07692307692308,
      "grad_norm": 0.6838968992233276,
      "learning_rate": 0.01,
      "loss": 4.639554023742676,
      "num_input_tokens_seen": 2481280,
      "step": 625,
      "train_runtime": 499.7983,
      "train_tokens_per_second": 4964.563
    },
    {
      "epoch": 48.07692307692308,
      "num_input_tokens_seen": 2481280,
      "step": 625,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 48.15384615384615,
      "grad_norm": 0.8111297488212585,
      "learning_rate": 0.01,
      "loss": 4.663860321044922,
      "num_input_tokens_seen": 2485376,
      "step": 626,
      "train_runtime": 500.6081,
      "train_tokens_per_second": 4964.714
    },
    {
      "epoch": 48.15384615384615,
      "num_input_tokens_seen": 2485376,
      "step": 626,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 48.23076923076923,
      "grad_norm": 0.7217122912406921,
      "learning_rate": 0.01,
      "loss": 4.6131815910339355,
      "num_input_tokens_seen": 2489472,
      "step": 627,
      "train_runtime": 501.4182,
      "train_tokens_per_second": 4964.862
    },
    {
      "epoch": 48.23076923076923,
      "num_input_tokens_seen": 2489472,
      "step": 627,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 48.30769230769231,
      "grad_norm": 0.6450233459472656,
      "learning_rate": 0.01,
      "loss": 4.653800964355469,
      "num_input_tokens_seen": 2493568,
      "step": 628,
      "train_runtime": 502.8205,
      "train_tokens_per_second": 4959.161
    },
    {
      "epoch": 48.30769230769231,
      "num_input_tokens_seen": 2493568,
      "step": 628,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 48.38461538461539,
      "grad_norm": 0.6552592515945435,
      "learning_rate": 0.01,
      "loss": 4.636212348937988,
      "num_input_tokens_seen": 2497664,
      "step": 629,
      "train_runtime": 503.6295,
      "train_tokens_per_second": 4959.328
    },
    {
      "epoch": 48.38461538461539,
      "num_input_tokens_seen": 2497664,
      "step": 629,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 48.46153846153846,
      "grad_norm": 0.6383926868438721,
      "learning_rate": 0.01,
      "loss": 4.636415958404541,
      "num_input_tokens_seen": 2501760,
      "step": 630,
      "train_runtime": 504.4393,
      "train_tokens_per_second": 4959.487
    },
    {
      "epoch": 48.46153846153846,
      "num_input_tokens_seen": 2501760,
      "step": 630,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 48.53846153846154,
      "grad_norm": 0.7733830213546753,
      "learning_rate": 0.01,
      "loss": 4.661928176879883,
      "num_input_tokens_seen": 2505856,
      "step": 631,
      "train_runtime": 505.2491,
      "train_tokens_per_second": 4959.644
    },
    {
      "epoch": 48.53846153846154,
      "num_input_tokens_seen": 2505856,
      "step": 631,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 48.61538461538461,
      "grad_norm": 0.5787567496299744,
      "learning_rate": 0.01,
      "loss": 4.6234893798828125,
      "num_input_tokens_seen": 2509952,
      "step": 632,
      "train_runtime": 506.0584,
      "train_tokens_per_second": 4959.807
    },
    {
      "epoch": 48.61538461538461,
      "num_input_tokens_seen": 2509952,
      "step": 632,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 48.69230769230769,
      "grad_norm": 1.2400026321411133,
      "learning_rate": 0.01,
      "loss": 4.647170066833496,
      "num_input_tokens_seen": 2514048,
      "step": 633,
      "train_runtime": 506.8688,
      "train_tokens_per_second": 4959.958
    },
    {
      "epoch": 48.69230769230769,
      "num_input_tokens_seen": 2514048,
      "step": 633,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 48.76923076923077,
      "grad_norm": 1.0741299390792847,
      "learning_rate": 0.01,
      "loss": 4.729347229003906,
      "num_input_tokens_seen": 2518144,
      "step": 634,
      "train_runtime": 507.6777,
      "train_tokens_per_second": 4960.123
    },
    {
      "epoch": 48.76923076923077,
      "num_input_tokens_seen": 2518144,
      "step": 634,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 48.84615384615385,
      "grad_norm": 1.0568114519119263,
      "learning_rate": 0.01,
      "loss": 4.669841766357422,
      "num_input_tokens_seen": 2522240,
      "step": 635,
      "train_runtime": 508.486,
      "train_tokens_per_second": 4960.293
    },
    {
      "epoch": 48.84615384615385,
      "num_input_tokens_seen": 2522240,
      "step": 635,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 48.92307692307692,
      "grad_norm": 0.9858353734016418,
      "learning_rate": 0.01,
      "loss": 4.701937198638916,
      "num_input_tokens_seen": 2526336,
      "step": 636,
      "train_runtime": 509.2909,
      "train_tokens_per_second": 4960.497
    },
    {
      "epoch": 48.92307692307692,
      "num_input_tokens_seen": 2526336,
      "step": 636,
      "train_accuracy": 0.026058631017804146
    },
    {
      "epoch": 49.0,
      "grad_norm": 1.1662908792495728,
      "learning_rate": 0.01,
      "loss": 4.6463212966918945,
      "num_input_tokens_seen": 2528792,
      "step": 637,
      "train_runtime": 509.8034,
      "train_tokens_per_second": 4960.328
    },
    {
      "epoch": 49.0,
      "num_input_tokens_seen": 2528792,
      "step": 637,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 49.07692307692308,
      "grad_norm": 0.9272441864013672,
      "learning_rate": 0.01,
      "loss": 4.6829328536987305,
      "num_input_tokens_seen": 2532888,
      "step": 638,
      "train_runtime": 510.625,
      "train_tokens_per_second": 4960.368
    },
    {
      "epoch": 49.07692307692308,
      "num_input_tokens_seen": 2532888,
      "step": 638,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 49.15384615384615,
      "grad_norm": 1.0570746660232544,
      "learning_rate": 0.01,
      "loss": 4.69055700302124,
      "num_input_tokens_seen": 2536984,
      "step": 639,
      "train_runtime": 511.4358,
      "train_tokens_per_second": 4960.513
    },
    {
      "epoch": 49.15384615384615,
      "num_input_tokens_seen": 2536984,
      "step": 639,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 49.23076923076923,
      "grad_norm": 1.1872737407684326,
      "learning_rate": 0.01,
      "loss": 4.695536136627197,
      "num_input_tokens_seen": 2541080,
      "step": 640,
      "train_runtime": 512.246,
      "train_tokens_per_second": 4960.664
    },
    {
      "epoch": 49.23076923076923,
      "num_input_tokens_seen": 2541080,
      "step": 640,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 49.30769230769231,
      "grad_norm": 0.7186793088912964,
      "learning_rate": 0.01,
      "loss": 4.674342155456543,
      "num_input_tokens_seen": 2545176,
      "step": 641,
      "train_runtime": 513.0556,
      "train_tokens_per_second": 4960.819
    },
    {
      "epoch": 49.30769230769231,
      "num_input_tokens_seen": 2545176,
      "step": 641,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 49.38461538461539,
      "grad_norm": 1.009905219078064,
      "learning_rate": 0.01,
      "loss": 4.646609306335449,
      "num_input_tokens_seen": 2549272,
      "step": 642,
      "train_runtime": 513.8658,
      "train_tokens_per_second": 4960.968
    },
    {
      "epoch": 49.38461538461539,
      "num_input_tokens_seen": 2549272,
      "step": 642,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 49.46153846153846,
      "grad_norm": 0.5971208214759827,
      "learning_rate": 0.01,
      "loss": 4.600564956665039,
      "num_input_tokens_seen": 2553368,
      "step": 643,
      "train_runtime": 514.6756,
      "train_tokens_per_second": 4961.121
    },
    {
      "epoch": 49.46153846153846,
      "num_input_tokens_seen": 2553368,
      "step": 643,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 49.53846153846154,
      "grad_norm": 0.8048012256622314,
      "learning_rate": 0.01,
      "loss": 4.6409101486206055,
      "num_input_tokens_seen": 2557464,
      "step": 644,
      "train_runtime": 515.4846,
      "train_tokens_per_second": 4961.281
    },
    {
      "epoch": 49.53846153846154,
      "num_input_tokens_seen": 2557464,
      "step": 644,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 49.61538461538461,
      "grad_norm": 0.7245068550109863,
      "learning_rate": 0.01,
      "loss": 4.6511125564575195,
      "num_input_tokens_seen": 2561560,
      "step": 645,
      "train_runtime": 516.2954,
      "train_tokens_per_second": 4961.424
    },
    {
      "epoch": 49.61538461538461,
      "num_input_tokens_seen": 2561560,
      "step": 645,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 49.69230769230769,
      "grad_norm": 0.6752263307571411,
      "learning_rate": 0.01,
      "loss": 4.658726692199707,
      "num_input_tokens_seen": 2565656,
      "step": 646,
      "train_runtime": 517.1067,
      "train_tokens_per_second": 4961.56
    },
    {
      "epoch": 49.69230769230769,
      "num_input_tokens_seen": 2565656,
      "step": 646,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 49.76923076923077,
      "grad_norm": 0.7821914553642273,
      "learning_rate": 0.01,
      "loss": 4.655362129211426,
      "num_input_tokens_seen": 2569752,
      "step": 647,
      "train_runtime": 517.9168,
      "train_tokens_per_second": 4961.708
    },
    {
      "epoch": 49.76923076923077,
      "num_input_tokens_seen": 2569752,
      "step": 647,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 49.84615384615385,
      "grad_norm": 0.6703137755393982,
      "learning_rate": 0.01,
      "loss": 4.617711067199707,
      "num_input_tokens_seen": 2573848,
      "step": 648,
      "train_runtime": 518.7273,
      "train_tokens_per_second": 4961.852
    },
    {
      "epoch": 49.84615384615385,
      "num_input_tokens_seen": 2573848,
      "step": 648,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 49.92307692307692,
      "grad_norm": 0.7096541523933411,
      "learning_rate": 0.01,
      "loss": 4.640897750854492,
      "num_input_tokens_seen": 2577944,
      "step": 649,
      "train_runtime": 519.5333,
      "train_tokens_per_second": 4962.038
    },
    {
      "epoch": 49.92307692307692,
      "num_input_tokens_seen": 2577944,
      "step": 649,
      "train_accuracy": 0.013029315508902073
    },
    {
      "epoch": 50.0,
      "grad_norm": 0.8702600598335266,
      "learning_rate": 0.01,
      "loss": 4.709993362426758,
      "num_input_tokens_seen": 2580400,
      "step": 650,
      "train_runtime": 520.0465,
      "train_tokens_per_second": 4961.864
    },
    {
      "epoch": 50.0,
      "num_input_tokens_seen": 2580400,
      "step": 650,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 50.07692307692308,
      "grad_norm": 0.42548519372940063,
      "learning_rate": 0.01,
      "loss": 4.5709357261657715,
      "num_input_tokens_seen": 2584496,
      "step": 651,
      "train_runtime": 520.8691,
      "train_tokens_per_second": 4961.892
    },
    {
      "epoch": 50.07692307692308,
      "num_input_tokens_seen": 2584496,
      "step": 651,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 50.15384615384615,
      "grad_norm": 0.5908008217811584,
      "learning_rate": 0.01,
      "loss": 4.604125499725342,
      "num_input_tokens_seen": 2588592,
      "step": 652,
      "train_runtime": 521.679,
      "train_tokens_per_second": 4962.04
    },
    {
      "epoch": 50.15384615384615,
      "num_input_tokens_seen": 2588592,
      "step": 652,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 50.23076923076923,
      "grad_norm": 0.4799763262271881,
      "learning_rate": 0.01,
      "loss": 4.620506286621094,
      "num_input_tokens_seen": 2592688,
      "step": 653,
      "train_runtime": 522.489,
      "train_tokens_per_second": 4962.187
    },
    {
      "epoch": 50.23076923076923,
      "num_input_tokens_seen": 2592688,
      "step": 653,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 50.30769230769231,
      "grad_norm": 0.5626305937767029,
      "learning_rate": 0.01,
      "loss": 4.628401756286621,
      "num_input_tokens_seen": 2596784,
      "step": 654,
      "train_runtime": 523.2995,
      "train_tokens_per_second": 4962.328
    },
    {
      "epoch": 50.30769230769231,
      "num_input_tokens_seen": 2596784,
      "step": 654,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 50.38461538461539,
      "grad_norm": 0.5553593635559082,
      "learning_rate": 0.01,
      "loss": 4.606263160705566,
      "num_input_tokens_seen": 2600880,
      "step": 655,
      "train_runtime": 524.1095,
      "train_tokens_per_second": 4962.475
    },
    {
      "epoch": 50.38461538461539,
      "num_input_tokens_seen": 2600880,
      "step": 655,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 50.46153846153846,
      "grad_norm": 0.4625200629234314,
      "learning_rate": 0.01,
      "loss": 4.608486175537109,
      "num_input_tokens_seen": 2604976,
      "step": 656,
      "train_runtime": 524.9193,
      "train_tokens_per_second": 4962.622
    },
    {
      "epoch": 50.46153846153846,
      "num_input_tokens_seen": 2604976,
      "step": 656,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 50.53846153846154,
      "grad_norm": 0.5337545275688171,
      "learning_rate": 0.01,
      "loss": 4.64743185043335,
      "num_input_tokens_seen": 2609072,
      "step": 657,
      "train_runtime": 525.7294,
      "train_tokens_per_second": 4962.766
    },
    {
      "epoch": 50.53846153846154,
      "num_input_tokens_seen": 2609072,
      "step": 657,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 50.61538461538461,
      "grad_norm": 0.4707218110561371,
      "learning_rate": 0.01,
      "loss": 4.617958068847656,
      "num_input_tokens_seen": 2613168,
      "step": 658,
      "train_runtime": 526.54,
      "train_tokens_per_second": 4962.905
    },
    {
      "epoch": 50.61538461538461,
      "num_input_tokens_seen": 2613168,
      "step": 658,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 50.69230769230769,
      "grad_norm": 0.5861688852310181,
      "learning_rate": 0.01,
      "loss": 4.668414115905762,
      "num_input_tokens_seen": 2617264,
      "step": 659,
      "train_runtime": 527.3512,
      "train_tokens_per_second": 4963.038
    },
    {
      "epoch": 50.69230769230769,
      "num_input_tokens_seen": 2617264,
      "step": 659,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 50.76923076923077,
      "grad_norm": 0.4923286736011505,
      "learning_rate": 0.01,
      "loss": 4.589829444885254,
      "num_input_tokens_seen": 2621360,
      "step": 660,
      "train_runtime": 528.1611,
      "train_tokens_per_second": 4963.182
    },
    {
      "epoch": 50.76923076923077,
      "num_input_tokens_seen": 2621360,
      "step": 660,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 50.84615384615385,
      "grad_norm": 0.4617142379283905,
      "learning_rate": 0.01,
      "loss": 4.619271278381348,
      "num_input_tokens_seen": 2625456,
      "step": 661,
      "train_runtime": 528.9712,
      "train_tokens_per_second": 4963.325
    },
    {
      "epoch": 50.84615384615385,
      "num_input_tokens_seen": 2625456,
      "step": 661,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 50.92307692307692,
      "grad_norm": 0.6571746468544006,
      "learning_rate": 0.01,
      "loss": 4.60849666595459,
      "num_input_tokens_seen": 2629552,
      "step": 662,
      "train_runtime": 529.7773,
      "train_tokens_per_second": 4963.505
    },
    {
      "epoch": 50.92307692307692,
      "num_input_tokens_seen": 2629552,
      "step": 662,
      "train_accuracy": 0.0065146577544510365
    },
    {
      "epoch": 51.0,
      "grad_norm": 0.5296902656555176,
      "learning_rate": 0.01,
      "loss": 4.624043941497803,
      "num_input_tokens_seen": 2632008,
      "step": 663,
      "train_runtime": 530.2911,
      "train_tokens_per_second": 4963.326
    },
    {
      "epoch": 51.0,
      "num_input_tokens_seen": 2632008,
      "step": 663,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 51.07692307692308,
      "grad_norm": 0.4619719088077545,
      "learning_rate": 0.01,
      "loss": 4.613448143005371,
      "num_input_tokens_seen": 2636104,
      "step": 664,
      "train_runtime": 531.1132,
      "train_tokens_per_second": 4963.356
    },
    {
      "epoch": 51.07692307692308,
      "num_input_tokens_seen": 2636104,
      "step": 664,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 51.15384615384615,
      "grad_norm": 0.4793144166469574,
      "learning_rate": 0.01,
      "loss": 4.639217853546143,
      "num_input_tokens_seen": 2640200,
      "step": 665,
      "train_runtime": 531.9243,
      "train_tokens_per_second": 4963.488
    },
    {
      "epoch": 51.15384615384615,
      "num_input_tokens_seen": 2640200,
      "step": 665,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 51.23076923076923,
      "grad_norm": 0.43797430396080017,
      "learning_rate": 0.01,
      "loss": 4.610978126525879,
      "num_input_tokens_seen": 2644296,
      "step": 666,
      "train_runtime": 532.7351,
      "train_tokens_per_second": 4963.622
    },
    {
      "epoch": 51.23076923076923,
      "num_input_tokens_seen": 2644296,
      "step": 666,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 51.30769230769231,
      "grad_norm": 0.41927510499954224,
      "learning_rate": 0.01,
      "loss": 4.595067024230957,
      "num_input_tokens_seen": 2648392,
      "step": 667,
      "train_runtime": 533.5461,
      "train_tokens_per_second": 4963.754
    },
    {
      "epoch": 51.30769230769231,
      "num_input_tokens_seen": 2648392,
      "step": 667,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 51.38461538461539,
      "grad_norm": 0.532819390296936,
      "learning_rate": 0.01,
      "loss": 4.598370552062988,
      "num_input_tokens_seen": 2652488,
      "step": 668,
      "train_runtime": 534.3567,
      "train_tokens_per_second": 4963.89
    },
    {
      "epoch": 51.38461538461539,
      "num_input_tokens_seen": 2652488,
      "step": 668,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 51.46153846153846,
      "grad_norm": 0.6052037477493286,
      "learning_rate": 0.01,
      "loss": 4.617681503295898,
      "num_input_tokens_seen": 2656584,
      "step": 669,
      "train_runtime": 535.1672,
      "train_tokens_per_second": 4964.026
    },
    {
      "epoch": 51.46153846153846,
      "num_input_tokens_seen": 2656584,
      "step": 669,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 51.53846153846154,
      "grad_norm": 0.4389113485813141,
      "learning_rate": 0.01,
      "loss": 4.626598358154297,
      "num_input_tokens_seen": 2660680,
      "step": 670,
      "train_runtime": 535.9782,
      "train_tokens_per_second": 4964.157
    },
    {
      "epoch": 51.53846153846154,
      "num_input_tokens_seen": 2660680,
      "step": 670,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 51.61538461538461,
      "grad_norm": 0.639845609664917,
      "learning_rate": 0.01,
      "loss": 4.6394243240356445,
      "num_input_tokens_seen": 2664776,
      "step": 671,
      "train_runtime": 536.788,
      "train_tokens_per_second": 4964.298
    },
    {
      "epoch": 51.61538461538461,
      "num_input_tokens_seen": 2664776,
      "step": 671,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 51.69230769230769,
      "grad_norm": 0.560914933681488,
      "learning_rate": 0.01,
      "loss": 4.6167473793029785,
      "num_input_tokens_seen": 2668872,
      "step": 672,
      "train_runtime": 537.5985,
      "train_tokens_per_second": 4964.433
    },
    {
      "epoch": 51.69230769230769,
      "num_input_tokens_seen": 2668872,
      "step": 672,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 51.76923076923077,
      "grad_norm": 0.4604809880256653,
      "learning_rate": 0.01,
      "loss": 4.6317315101623535,
      "num_input_tokens_seen": 2672968,
      "step": 673,
      "train_runtime": 538.4093,
      "train_tokens_per_second": 4964.565
    },
    {
      "epoch": 51.76923076923077,
      "num_input_tokens_seen": 2672968,
      "step": 673,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 51.84615384615385,
      "grad_norm": 0.518646240234375,
      "learning_rate": 0.01,
      "loss": 4.621442794799805,
      "num_input_tokens_seen": 2677064,
      "step": 674,
      "train_runtime": 539.2199,
      "train_tokens_per_second": 4964.698
    },
    {
      "epoch": 51.84615384615385,
      "num_input_tokens_seen": 2677064,
      "step": 674,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 51.92307692307692,
      "grad_norm": 0.4244747459888458,
      "learning_rate": 0.01,
      "loss": 4.609426498413086,
      "num_input_tokens_seen": 2681160,
      "step": 675,
      "train_runtime": 540.0266,
      "train_tokens_per_second": 4964.867
    },
    {
      "epoch": 51.92307692307692,
      "num_input_tokens_seen": 2681160,
      "step": 675,
      "train_accuracy": 0.016286645084619522
    },
    {
      "epoch": 52.0,
      "grad_norm": 0.7208302021026611,
      "learning_rate": 0.01,
      "loss": 4.620415210723877,
      "num_input_tokens_seen": 2683616,
      "step": 676,
      "train_runtime": 540.5399,
      "train_tokens_per_second": 4964.696
    },
    {
      "epoch": 52.0,
      "num_input_tokens_seen": 2683616,
      "step": 676,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 52.07692307692308,
      "grad_norm": 0.6297442317008972,
      "learning_rate": 0.01,
      "loss": 4.6504974365234375,
      "num_input_tokens_seen": 2687712,
      "step": 677,
      "train_runtime": 541.3618,
      "train_tokens_per_second": 4964.724
    },
    {
      "epoch": 52.07692307692308,
      "num_input_tokens_seen": 2687712,
      "step": 677,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 52.15384615384615,
      "grad_norm": 0.5211429595947266,
      "learning_rate": 0.01,
      "loss": 4.601546287536621,
      "num_input_tokens_seen": 2691808,
      "step": 678,
      "train_runtime": 542.1729,
      "train_tokens_per_second": 4964.851
    },
    {
      "epoch": 52.15384615384615,
      "num_input_tokens_seen": 2691808,
      "step": 678,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 52.23076923076923,
      "grad_norm": 0.7010127902030945,
      "learning_rate": 0.01,
      "loss": 4.647871971130371,
      "num_input_tokens_seen": 2695904,
      "step": 679,
      "train_runtime": 542.9839,
      "train_tokens_per_second": 4964.98
    },
    {
      "epoch": 52.23076923076923,
      "num_input_tokens_seen": 2695904,
      "step": 679,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 52.30769230769231,
      "grad_norm": 0.6510889530181885,
      "learning_rate": 0.01,
      "loss": 4.646492958068848,
      "num_input_tokens_seen": 2700000,
      "step": 680,
      "train_runtime": 543.795,
      "train_tokens_per_second": 4965.106
    },
    {
      "epoch": 52.30769230769231,
      "num_input_tokens_seen": 2700000,
      "step": 680,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 52.38461538461539,
      "grad_norm": 0.616940438747406,
      "learning_rate": 0.01,
      "loss": 4.585207462310791,
      "num_input_tokens_seen": 2704096,
      "step": 681,
      "train_runtime": 544.6071,
      "train_tokens_per_second": 4965.224
    },
    {
      "epoch": 52.38461538461539,
      "num_input_tokens_seen": 2704096,
      "step": 681,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 52.46153846153846,
      "grad_norm": 0.4518240988254547,
      "learning_rate": 0.01,
      "loss": 4.583968639373779,
      "num_input_tokens_seen": 2708192,
      "step": 682,
      "train_runtime": 545.4181,
      "train_tokens_per_second": 4965.35
    },
    {
      "epoch": 52.46153846153846,
      "num_input_tokens_seen": 2708192,
      "step": 682,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 52.53846153846154,
      "grad_norm": 0.8632458448410034,
      "learning_rate": 0.01,
      "loss": 4.667637825012207,
      "num_input_tokens_seen": 2712288,
      "step": 683,
      "train_runtime": 546.2306,
      "train_tokens_per_second": 4965.464
    },
    {
      "epoch": 52.53846153846154,
      "num_input_tokens_seen": 2712288,
      "step": 683,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 52.61538461538461,
      "grad_norm": 0.9562072157859802,
      "learning_rate": 0.01,
      "loss": 4.669644355773926,
      "num_input_tokens_seen": 2716384,
      "step": 684,
      "train_runtime": 547.0419,
      "train_tokens_per_second": 4965.587
    },
    {
      "epoch": 52.61538461538461,
      "num_input_tokens_seen": 2716384,
      "step": 684,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 52.69230769230769,
      "grad_norm": 0.48732319474220276,
      "learning_rate": 0.01,
      "loss": 4.63418436050415,
      "num_input_tokens_seen": 2720480,
      "step": 685,
      "train_runtime": 547.8534,
      "train_tokens_per_second": 4965.708
    },
    {
      "epoch": 52.69230769230769,
      "num_input_tokens_seen": 2720480,
      "step": 685,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 52.76923076923077,
      "grad_norm": 0.8183702826499939,
      "learning_rate": 0.01,
      "loss": 4.641991138458252,
      "num_input_tokens_seen": 2724576,
      "step": 686,
      "train_runtime": 548.6662,
      "train_tokens_per_second": 4965.817
    },
    {
      "epoch": 52.76923076923077,
      "num_input_tokens_seen": 2724576,
      "step": 686,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 52.84615384615385,
      "grad_norm": 0.4833761751651764,
      "learning_rate": 0.01,
      "loss": 4.62351131439209,
      "num_input_tokens_seen": 2728672,
      "step": 687,
      "train_runtime": 549.477,
      "train_tokens_per_second": 4965.944
    },
    {
      "epoch": 52.84615384615385,
      "num_input_tokens_seen": 2728672,
      "step": 687,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 52.92307692307692,
      "grad_norm": 0.9291377067565918,
      "learning_rate": 0.01,
      "loss": 4.693155288696289,
      "num_input_tokens_seen": 2732768,
      "step": 688,
      "train_runtime": 550.2834,
      "train_tokens_per_second": 4966.11
    },
    {
      "epoch": 52.92307692307692,
      "num_input_tokens_seen": 2732768,
      "step": 688,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 53.0,
      "grad_norm": 0.6080514788627625,
      "learning_rate": 0.01,
      "loss": 4.683415412902832,
      "num_input_tokens_seen": 2735224,
      "step": 689,
      "train_runtime": 550.7971,
      "train_tokens_per_second": 4965.937
    },
    {
      "epoch": 53.0,
      "num_input_tokens_seen": 2735224,
      "step": 689,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 53.07692307692308,
      "grad_norm": 1.0228312015533447,
      "learning_rate": 0.01,
      "loss": 4.675637245178223,
      "num_input_tokens_seen": 2739320,
      "step": 690,
      "train_runtime": 551.6232,
      "train_tokens_per_second": 4965.926
    },
    {
      "epoch": 53.07692307692308,
      "num_input_tokens_seen": 2739320,
      "step": 690,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 53.15384615384615,
      "grad_norm": 0.5086424946784973,
      "learning_rate": 0.01,
      "loss": 4.654016971588135,
      "num_input_tokens_seen": 2743416,
      "step": 691,
      "train_runtime": 552.4341,
      "train_tokens_per_second": 4966.051
    },
    {
      "epoch": 53.15384615384615,
      "num_input_tokens_seen": 2743416,
      "step": 691,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 53.23076923076923,
      "grad_norm": 0.929985523223877,
      "learning_rate": 0.01,
      "loss": 4.672959327697754,
      "num_input_tokens_seen": 2747512,
      "step": 692,
      "train_runtime": 553.2448,
      "train_tokens_per_second": 4966.178
    },
    {
      "epoch": 53.23076923076923,
      "num_input_tokens_seen": 2747512,
      "step": 692,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 53.30769230769231,
      "grad_norm": 0.49006387591362,
      "learning_rate": 0.01,
      "loss": 4.641401767730713,
      "num_input_tokens_seen": 2751608,
      "step": 693,
      "train_runtime": 554.0558,
      "train_tokens_per_second": 4966.301
    },
    {
      "epoch": 53.30769230769231,
      "num_input_tokens_seen": 2751608,
      "step": 693,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 53.38461538461539,
      "grad_norm": 1.072021245956421,
      "learning_rate": 0.01,
      "loss": 4.657723426818848,
      "num_input_tokens_seen": 2755704,
      "step": 694,
      "train_runtime": 554.8672,
      "train_tokens_per_second": 4966.421
    },
    {
      "epoch": 53.38461538461539,
      "num_input_tokens_seen": 2755704,
      "step": 694,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 53.46153846153846,
      "grad_norm": 0.491686075925827,
      "learning_rate": 0.01,
      "loss": 4.653244972229004,
      "num_input_tokens_seen": 2759800,
      "step": 695,
      "train_runtime": 555.6775,
      "train_tokens_per_second": 4966.55
    },
    {
      "epoch": 53.46153846153846,
      "num_input_tokens_seen": 2759800,
      "step": 695,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 53.53846153846154,
      "grad_norm": 1.2374560832977295,
      "learning_rate": 0.01,
      "loss": 4.721195220947266,
      "num_input_tokens_seen": 2763896,
      "step": 696,
      "train_runtime": 556.4879,
      "train_tokens_per_second": 4966.677
    },
    {
      "epoch": 53.53846153846154,
      "num_input_tokens_seen": 2763896,
      "step": 696,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 53.61538461538461,
      "grad_norm": 0.61308753490448,
      "learning_rate": 0.01,
      "loss": 4.64759635925293,
      "num_input_tokens_seen": 2767992,
      "step": 697,
      "train_runtime": 557.2994,
      "train_tokens_per_second": 4966.795
    },
    {
      "epoch": 53.61538461538461,
      "num_input_tokens_seen": 2767992,
      "step": 697,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 53.69230769230769,
      "grad_norm": 0.5914096236228943,
      "learning_rate": 0.01,
      "loss": 4.6311354637146,
      "num_input_tokens_seen": 2772088,
      "step": 698,
      "train_runtime": 558.1101,
      "train_tokens_per_second": 4966.92
    },
    {
      "epoch": 53.69230769230769,
      "num_input_tokens_seen": 2772088,
      "step": 698,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 53.76923076923077,
      "grad_norm": 0.5452708601951599,
      "learning_rate": 0.01,
      "loss": 4.661554336547852,
      "num_input_tokens_seen": 2776184,
      "step": 699,
      "train_runtime": 558.9212,
      "train_tokens_per_second": 4967.04
    },
    {
      "epoch": 53.76923076923077,
      "num_input_tokens_seen": 2776184,
      "step": 699,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 53.84615384615385,
      "grad_norm": 0.8172363638877869,
      "learning_rate": 0.01,
      "loss": 4.6825737953186035,
      "num_input_tokens_seen": 2780280,
      "step": 700,
      "train_runtime": 559.7322,
      "train_tokens_per_second": 4967.161
    },
    {
      "epoch": 53.84615384615385,
      "eval_CMD_3_branch_eval_accuracy": 0.012658227848101266,
      "eval_CMD_3_branch_eval_loss": 4.668797016143799,
      "eval_CMD_3_branch_eval_runtime": 1.1401,
      "eval_CMD_3_branch_eval_samples_per_second": 2425.183,
      "eval_CMD_3_branch_eval_steps_per_second": 5.263,
      "num_input_tokens_seen": 2780280,
      "step": 700
    },
    {
      "epoch": 53.84615384615385,
      "num_input_tokens_seen": 2780280,
      "step": 700,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 53.92307692307692,
      "grad_norm": 0.7303228974342346,
      "learning_rate": 0.01,
      "loss": 4.647892951965332,
      "num_input_tokens_seen": 2784376,
      "step": 701,
      "train_runtime": 561.6836,
      "train_tokens_per_second": 4957.197
    },
    {
      "epoch": 53.92307692307692,
      "num_input_tokens_seen": 2784376,
      "step": 701,
      "train_accuracy": 0.013029315508902073
    },
    {
      "epoch": 54.0,
      "grad_norm": 0.6225734353065491,
      "learning_rate": 0.01,
      "loss": 4.695523738861084,
      "num_input_tokens_seen": 2786832,
      "step": 702,
      "train_runtime": 562.197,
      "train_tokens_per_second": 4957.038
    },
    {
      "epoch": 54.0,
      "num_input_tokens_seen": 2786832,
      "step": 702,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 54.07692307692308,
      "grad_norm": 0.3752252459526062,
      "learning_rate": 0.01,
      "loss": 4.623847007751465,
      "num_input_tokens_seen": 2790928,
      "step": 703,
      "train_runtime": 563.0201,
      "train_tokens_per_second": 4957.066
    },
    {
      "epoch": 54.07692307692308,
      "num_input_tokens_seen": 2790928,
      "step": 703,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 54.15384615384615,
      "grad_norm": 0.5645386576652527,
      "learning_rate": 0.01,
      "loss": 4.653515815734863,
      "num_input_tokens_seen": 2795024,
      "step": 704,
      "train_runtime": 563.8314,
      "train_tokens_per_second": 4957.198
    },
    {
      "epoch": 54.15384615384615,
      "num_input_tokens_seen": 2795024,
      "step": 704,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 54.23076923076923,
      "grad_norm": 0.6331173181533813,
      "learning_rate": 0.01,
      "loss": 4.634461402893066,
      "num_input_tokens_seen": 2799120,
      "step": 705,
      "train_runtime": 564.6431,
      "train_tokens_per_second": 4957.326
    },
    {
      "epoch": 54.23076923076923,
      "num_input_tokens_seen": 2799120,
      "step": 705,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 54.30769230769231,
      "grad_norm": 0.655329704284668,
      "learning_rate": 0.01,
      "loss": 4.607893466949463,
      "num_input_tokens_seen": 2803216,
      "step": 706,
      "train_runtime": 565.4538,
      "train_tokens_per_second": 4957.462
    },
    {
      "epoch": 54.30769230769231,
      "num_input_tokens_seen": 2803216,
      "step": 706,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 54.38461538461539,
      "grad_norm": 0.5159600973129272,
      "learning_rate": 0.01,
      "loss": 4.5496978759765625,
      "num_input_tokens_seen": 2807312,
      "step": 707,
      "train_runtime": 566.2658,
      "train_tokens_per_second": 4957.587
    },
    {
      "epoch": 54.38461538461539,
      "num_input_tokens_seen": 2807312,
      "step": 707,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 54.46153846153846,
      "grad_norm": 0.529163122177124,
      "learning_rate": 0.01,
      "loss": 4.6628851890563965,
      "num_input_tokens_seen": 2811408,
      "step": 708,
      "train_runtime": 567.0773,
      "train_tokens_per_second": 4957.716
    },
    {
      "epoch": 54.46153846153846,
      "num_input_tokens_seen": 2811408,
      "step": 708,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 54.53846153846154,
      "grad_norm": 0.7172780632972717,
      "learning_rate": 0.01,
      "loss": 4.659363746643066,
      "num_input_tokens_seen": 2815504,
      "step": 709,
      "train_runtime": 567.8884,
      "train_tokens_per_second": 4957.847
    },
    {
      "epoch": 54.53846153846154,
      "num_input_tokens_seen": 2815504,
      "step": 709,
      "train_accuracy": 0.001953125
    },
    {
      "epoch": 54.61538461538461,
      "grad_norm": 0.5487365126609802,
      "learning_rate": 0.01,
      "loss": 4.644710540771484,
      "num_input_tokens_seen": 2819600,
      "step": 710,
      "train_runtime": 568.7011,
      "train_tokens_per_second": 4957.964
    },
    {
      "epoch": 54.61538461538461,
      "num_input_tokens_seen": 2819600,
      "step": 710,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 54.69230769230769,
      "grad_norm": 0.5705341100692749,
      "learning_rate": 0.01,
      "loss": 4.6619110107421875,
      "num_input_tokens_seen": 2823696,
      "step": 711,
      "train_runtime": 569.5132,
      "train_tokens_per_second": 4958.087
    },
    {
      "epoch": 54.69230769230769,
      "num_input_tokens_seen": 2823696,
      "step": 711,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 54.76923076923077,
      "grad_norm": 0.37892526388168335,
      "learning_rate": 0.01,
      "loss": 4.606176376342773,
      "num_input_tokens_seen": 2827792,
      "step": 712,
      "train_runtime": 570.3267,
      "train_tokens_per_second": 4958.197
    },
    {
      "epoch": 54.76923076923077,
      "num_input_tokens_seen": 2827792,
      "step": 712,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 54.84615384615385,
      "grad_norm": 0.8703420758247375,
      "learning_rate": 0.01,
      "loss": 4.640538692474365,
      "num_input_tokens_seen": 2831888,
      "step": 713,
      "train_runtime": 571.1386,
      "train_tokens_per_second": 4958.32
    },
    {
      "epoch": 54.84615384615385,
      "num_input_tokens_seen": 2831888,
      "step": 713,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 54.92307692307692,
      "grad_norm": 0.3756444752216339,
      "learning_rate": 0.01,
      "loss": 4.595385551452637,
      "num_input_tokens_seen": 2835984,
      "step": 714,
      "train_runtime": 571.9462,
      "train_tokens_per_second": 4958.48
    },
    {
      "epoch": 54.92307692307692,
      "num_input_tokens_seen": 2835984,
      "step": 714,
      "train_accuracy": 0.0032573288772255182
    },
    {
      "epoch": 55.0,
      "grad_norm": 0.9132266044616699,
      "learning_rate": 0.01,
      "loss": 4.677668571472168,
      "num_input_tokens_seen": 2838440,
      "step": 715,
      "train_runtime": 572.461,
      "train_tokens_per_second": 4958.311
    },
    {
      "epoch": 55.0,
      "num_input_tokens_seen": 2838440,
      "step": 715,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 55.07692307692308,
      "grad_norm": 0.4334953725337982,
      "learning_rate": 0.01,
      "loss": 4.614554405212402,
      "num_input_tokens_seen": 2842536,
      "step": 716,
      "train_runtime": 573.2851,
      "train_tokens_per_second": 4958.328
    },
    {
      "epoch": 55.07692307692308,
      "num_input_tokens_seen": 2842536,
      "step": 716,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 55.15384615384615,
      "grad_norm": 0.4798298478126526,
      "learning_rate": 0.01,
      "loss": 4.58780574798584,
      "num_input_tokens_seen": 2846632,
      "step": 717,
      "train_runtime": 574.0977,
      "train_tokens_per_second": 4958.445
    },
    {
      "epoch": 55.15384615384615,
      "num_input_tokens_seen": 2846632,
      "step": 717,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 55.23076923076923,
      "grad_norm": 0.592604398727417,
      "learning_rate": 0.01,
      "loss": 4.646493911743164,
      "num_input_tokens_seen": 2850728,
      "step": 718,
      "train_runtime": 574.909,
      "train_tokens_per_second": 4958.573
    },
    {
      "epoch": 55.23076923076923,
      "num_input_tokens_seen": 2850728,
      "step": 718,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 55.30769230769231,
      "grad_norm": 0.4955499470233917,
      "learning_rate": 0.01,
      "loss": 4.645220756530762,
      "num_input_tokens_seen": 2854824,
      "step": 719,
      "train_runtime": 575.7206,
      "train_tokens_per_second": 4958.697
    },
    {
      "epoch": 55.30769230769231,
      "num_input_tokens_seen": 2854824,
      "step": 719,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 55.38461538461539,
      "grad_norm": 0.785028874874115,
      "learning_rate": 0.01,
      "loss": 4.696919918060303,
      "num_input_tokens_seen": 2858920,
      "step": 720,
      "train_runtime": 576.5321,
      "train_tokens_per_second": 4958.822
    },
    {
      "epoch": 55.38461538461539,
      "num_input_tokens_seen": 2858920,
      "step": 720,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 55.46153846153846,
      "grad_norm": 0.42734456062316895,
      "learning_rate": 0.01,
      "loss": 4.600754737854004,
      "num_input_tokens_seen": 2863016,
      "step": 721,
      "train_runtime": 577.3449,
      "train_tokens_per_second": 4958.935
    },
    {
      "epoch": 55.46153846153846,
      "num_input_tokens_seen": 2863016,
      "step": 721,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 55.53846153846154,
      "grad_norm": 0.6232092976570129,
      "learning_rate": 0.01,
      "loss": 4.644918441772461,
      "num_input_tokens_seen": 2867112,
      "step": 722,
      "train_runtime": 578.1562,
      "train_tokens_per_second": 4959.061
    },
    {
      "epoch": 55.53846153846154,
      "num_input_tokens_seen": 2867112,
      "step": 722,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 55.61538461538461,
      "grad_norm": 0.5796198844909668,
      "learning_rate": 0.01,
      "loss": 4.643170356750488,
      "num_input_tokens_seen": 2871208,
      "step": 723,
      "train_runtime": 578.9673,
      "train_tokens_per_second": 4959.189
    },
    {
      "epoch": 55.61538461538461,
      "num_input_tokens_seen": 2871208,
      "step": 723,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 55.69230769230769,
      "grad_norm": 0.5011171698570251,
      "learning_rate": 0.01,
      "loss": 4.630743026733398,
      "num_input_tokens_seen": 2875304,
      "step": 724,
      "train_runtime": 579.779,
      "train_tokens_per_second": 4959.31
    },
    {
      "epoch": 55.69230769230769,
      "num_input_tokens_seen": 2875304,
      "step": 724,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 55.76923076923077,
      "grad_norm": 0.4246017038822174,
      "learning_rate": 0.01,
      "loss": 4.622054100036621,
      "num_input_tokens_seen": 2879400,
      "step": 725,
      "train_runtime": 580.5921,
      "train_tokens_per_second": 4959.42
    },
    {
      "epoch": 55.76923076923077,
      "num_input_tokens_seen": 2879400,
      "step": 725,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 55.84615384615385,
      "grad_norm": 0.3628958761692047,
      "learning_rate": 0.01,
      "loss": 4.581844329833984,
      "num_input_tokens_seen": 2883496,
      "step": 726,
      "train_runtime": 581.4058,
      "train_tokens_per_second": 4959.524
    },
    {
      "epoch": 55.84615384615385,
      "num_input_tokens_seen": 2883496,
      "step": 726,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 55.92307692307692,
      "grad_norm": 0.40375009179115295,
      "learning_rate": 0.01,
      "loss": 4.625261306762695,
      "num_input_tokens_seen": 2887592,
      "step": 727,
      "train_runtime": 582.2139,
      "train_tokens_per_second": 4959.676
    },
    {
      "epoch": 55.92307692307692,
      "num_input_tokens_seen": 2887592,
      "step": 727,
      "train_accuracy": 0.016286645084619522
    },
    {
      "epoch": 56.0,
      "grad_norm": 0.4562211036682129,
      "learning_rate": 0.01,
      "loss": 4.64891242980957,
      "num_input_tokens_seen": 2890048,
      "step": 728,
      "train_runtime": 582.7281,
      "train_tokens_per_second": 4959.514
    },
    {
      "epoch": 56.0,
      "num_input_tokens_seen": 2890048,
      "step": 728,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 56.07692307692308,
      "grad_norm": 0.4938841164112091,
      "learning_rate": 0.01,
      "loss": 4.630170822143555,
      "num_input_tokens_seen": 2894144,
      "step": 729,
      "train_runtime": 583.5544,
      "train_tokens_per_second": 4959.51
    },
    {
      "epoch": 56.07692307692308,
      "num_input_tokens_seen": 2894144,
      "step": 729,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 56.15384615384615,
      "grad_norm": 0.43323031067848206,
      "learning_rate": 0.01,
      "loss": 4.594569206237793,
      "num_input_tokens_seen": 2898240,
      "step": 730,
      "train_runtime": 584.3662,
      "train_tokens_per_second": 4959.63
    },
    {
      "epoch": 56.15384615384615,
      "num_input_tokens_seen": 2898240,
      "step": 730,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 56.23076923076923,
      "grad_norm": 0.41912975907325745,
      "learning_rate": 0.01,
      "loss": 4.605883598327637,
      "num_input_tokens_seen": 2902336,
      "step": 731,
      "train_runtime": 585.1772,
      "train_tokens_per_second": 4959.756
    },
    {
      "epoch": 56.23076923076923,
      "num_input_tokens_seen": 2902336,
      "step": 731,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 56.30769230769231,
      "grad_norm": 0.31143659353256226,
      "learning_rate": 0.01,
      "loss": 4.563133239746094,
      "num_input_tokens_seen": 2906432,
      "step": 732,
      "train_runtime": 585.9885,
      "train_tokens_per_second": 4959.879
    },
    {
      "epoch": 56.30769230769231,
      "num_input_tokens_seen": 2906432,
      "step": 732,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 56.38461538461539,
      "grad_norm": 0.6525051593780518,
      "learning_rate": 0.01,
      "loss": 4.656364917755127,
      "num_input_tokens_seen": 2910528,
      "step": 733,
      "train_runtime": 586.7999,
      "train_tokens_per_second": 4960.0
    },
    {
      "epoch": 56.38461538461539,
      "num_input_tokens_seen": 2910528,
      "step": 733,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 56.46153846153846,
      "grad_norm": 0.3329682946205139,
      "learning_rate": 0.01,
      "loss": 4.611915588378906,
      "num_input_tokens_seen": 2914624,
      "step": 734,
      "train_runtime": 587.6135,
      "train_tokens_per_second": 4960.104
    },
    {
      "epoch": 56.46153846153846,
      "num_input_tokens_seen": 2914624,
      "step": 734,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 56.53846153846154,
      "grad_norm": 0.640602707862854,
      "learning_rate": 0.01,
      "loss": 4.685609340667725,
      "num_input_tokens_seen": 2918720,
      "step": 735,
      "train_runtime": 588.4257,
      "train_tokens_per_second": 4960.219
    },
    {
      "epoch": 56.53846153846154,
      "num_input_tokens_seen": 2918720,
      "step": 735,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 56.61538461538461,
      "grad_norm": 0.47230076789855957,
      "learning_rate": 0.01,
      "loss": 4.608022689819336,
      "num_input_tokens_seen": 2922816,
      "step": 736,
      "train_runtime": 589.2396,
      "train_tokens_per_second": 4960.318
    },
    {
      "epoch": 56.61538461538461,
      "num_input_tokens_seen": 2922816,
      "step": 736,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 56.69230769230769,
      "grad_norm": 0.452637642621994,
      "learning_rate": 0.01,
      "loss": 4.626341819763184,
      "num_input_tokens_seen": 2926912,
      "step": 737,
      "train_runtime": 590.0542,
      "train_tokens_per_second": 4960.412
    },
    {
      "epoch": 56.69230769230769,
      "num_input_tokens_seen": 2926912,
      "step": 737,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 56.76923076923077,
      "grad_norm": 0.37576910853385925,
      "learning_rate": 0.01,
      "loss": 4.564453125,
      "num_input_tokens_seen": 2931008,
      "step": 738,
      "train_runtime": 590.8688,
      "train_tokens_per_second": 4960.506
    },
    {
      "epoch": 56.76923076923077,
      "num_input_tokens_seen": 2931008,
      "step": 738,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 56.84615384615385,
      "grad_norm": 0.4727272093296051,
      "learning_rate": 0.01,
      "loss": 4.6210150718688965,
      "num_input_tokens_seen": 2935104,
      "step": 739,
      "train_runtime": 591.6807,
      "train_tokens_per_second": 4960.622
    },
    {
      "epoch": 56.84615384615385,
      "num_input_tokens_seen": 2935104,
      "step": 739,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 56.92307692307692,
      "grad_norm": 1.8359097242355347,
      "learning_rate": 0.01,
      "loss": 4.627798080444336,
      "num_input_tokens_seen": 2939200,
      "step": 740,
      "train_runtime": 592.4882,
      "train_tokens_per_second": 4960.774
    },
    {
      "epoch": 56.92307692307692,
      "num_input_tokens_seen": 2939200,
      "step": 740,
      "train_accuracy": 0.013029315508902073
    },
    {
      "epoch": 57.0,
      "grad_norm": 0.823265552520752,
      "learning_rate": 0.01,
      "loss": 4.645924091339111,
      "num_input_tokens_seen": 2941656,
      "step": 741,
      "train_runtime": 593.0026,
      "train_tokens_per_second": 4960.612
    },
    {
      "epoch": 57.0,
      "num_input_tokens_seen": 2941656,
      "step": 741,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 57.07692307692308,
      "grad_norm": 0.4464469850063324,
      "learning_rate": 0.01,
      "loss": 4.5953593254089355,
      "num_input_tokens_seen": 2945752,
      "step": 742,
      "train_runtime": 593.8268,
      "train_tokens_per_second": 4960.625
    },
    {
      "epoch": 57.07692307692308,
      "num_input_tokens_seen": 2945752,
      "step": 742,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 57.15384615384615,
      "grad_norm": 0.5173393487930298,
      "learning_rate": 0.01,
      "loss": 4.618371486663818,
      "num_input_tokens_seen": 2949848,
      "step": 743,
      "train_runtime": 594.6411,
      "train_tokens_per_second": 4960.72
    },
    {
      "epoch": 57.15384615384615,
      "num_input_tokens_seen": 2949848,
      "step": 743,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 57.23076923076923,
      "grad_norm": 0.6752522587776184,
      "learning_rate": 0.01,
      "loss": 4.671770095825195,
      "num_input_tokens_seen": 2953944,
      "step": 744,
      "train_runtime": 595.4542,
      "train_tokens_per_second": 4960.825
    },
    {
      "epoch": 57.23076923076923,
      "num_input_tokens_seen": 2953944,
      "step": 744,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 57.30769230769231,
      "grad_norm": 0.41735801100730896,
      "learning_rate": 0.01,
      "loss": 4.670845985412598,
      "num_input_tokens_seen": 2958040,
      "step": 745,
      "train_runtime": 596.2691,
      "train_tokens_per_second": 4960.915
    },
    {
      "epoch": 57.30769230769231,
      "num_input_tokens_seen": 2958040,
      "step": 745,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 57.38461538461539,
      "grad_norm": 3.265552520751953,
      "learning_rate": 0.01,
      "loss": 4.609025001525879,
      "num_input_tokens_seen": 2962136,
      "step": 746,
      "train_runtime": 597.0835,
      "train_tokens_per_second": 4961.008
    },
    {
      "epoch": 57.38461538461539,
      "num_input_tokens_seen": 2962136,
      "step": 746,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 57.46153846153846,
      "grad_norm": 0.4434090852737427,
      "learning_rate": 0.01,
      "loss": 4.605268955230713,
      "num_input_tokens_seen": 2966232,
      "step": 747,
      "train_runtime": 597.8969,
      "train_tokens_per_second": 4961.11
    },
    {
      "epoch": 57.46153846153846,
      "num_input_tokens_seen": 2966232,
      "step": 747,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 57.53846153846154,
      "grad_norm": 0.4303193986415863,
      "learning_rate": 0.01,
      "loss": 4.568668365478516,
      "num_input_tokens_seen": 2970328,
      "step": 748,
      "train_runtime": 598.7102,
      "train_tokens_per_second": 4961.212
    },
    {
      "epoch": 57.53846153846154,
      "num_input_tokens_seen": 2970328,
      "step": 748,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 57.61538461538461,
      "grad_norm": 0.5301846861839294,
      "learning_rate": 0.01,
      "loss": 4.630484580993652,
      "num_input_tokens_seen": 2974424,
      "step": 749,
      "train_runtime": 599.5228,
      "train_tokens_per_second": 4961.319
    },
    {
      "epoch": 57.61538461538461,
      "num_input_tokens_seen": 2974424,
      "step": 749,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 57.69230769230769,
      "grad_norm": 0.5873321890830994,
      "learning_rate": 0.01,
      "loss": 4.623235702514648,
      "num_input_tokens_seen": 2978520,
      "step": 750,
      "train_runtime": 600.3348,
      "train_tokens_per_second": 4961.431
    },
    {
      "epoch": 57.69230769230769,
      "num_input_tokens_seen": 2978520,
      "step": 750,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 57.76923076923077,
      "grad_norm": 0.4589889943599701,
      "learning_rate": 0.01,
      "loss": 4.675217628479004,
      "num_input_tokens_seen": 2982616,
      "step": 751,
      "train_runtime": 601.1469,
      "train_tokens_per_second": 4961.543
    },
    {
      "epoch": 57.76923076923077,
      "num_input_tokens_seen": 2982616,
      "step": 751,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 57.84615384615385,
      "grad_norm": 0.41689085960388184,
      "learning_rate": 0.01,
      "loss": 4.664083957672119,
      "num_input_tokens_seen": 2986712,
      "step": 752,
      "train_runtime": 601.9606,
      "train_tokens_per_second": 4961.641
    },
    {
      "epoch": 57.84615384615385,
      "num_input_tokens_seen": 2986712,
      "step": 752,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 57.92307692307692,
      "grad_norm": 0.38346293568611145,
      "learning_rate": 0.01,
      "loss": 4.640788555145264,
      "num_input_tokens_seen": 2990808,
      "step": 753,
      "train_runtime": 602.7696,
      "train_tokens_per_second": 4961.777
    },
    {
      "epoch": 57.92307692307692,
      "num_input_tokens_seen": 2990808,
      "step": 753,
      "train_accuracy": 0.016286645084619522
    },
    {
      "epoch": 58.0,
      "grad_norm": 0.5533782839775085,
      "learning_rate": 0.01,
      "loss": 4.685469627380371,
      "num_input_tokens_seen": 2993264,
      "step": 754,
      "train_runtime": 603.2851,
      "train_tokens_per_second": 4961.608
    },
    {
      "epoch": 58.0,
      "num_input_tokens_seen": 2993264,
      "step": 754,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 58.07692307692308,
      "grad_norm": 0.5619157552719116,
      "learning_rate": 0.01,
      "loss": 4.628260135650635,
      "num_input_tokens_seen": 2997360,
      "step": 755,
      "train_runtime": 604.1118,
      "train_tokens_per_second": 4961.598
    },
    {
      "epoch": 58.07692307692308,
      "num_input_tokens_seen": 2997360,
      "step": 755,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 58.15384615384615,
      "grad_norm": 0.5559141039848328,
      "learning_rate": 0.01,
      "loss": 4.6344099044799805,
      "num_input_tokens_seen": 3001456,
      "step": 756,
      "train_runtime": 604.9264,
      "train_tokens_per_second": 4961.688
    },
    {
      "epoch": 58.15384615384615,
      "num_input_tokens_seen": 3001456,
      "step": 756,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 58.23076923076923,
      "grad_norm": 0.4944635033607483,
      "learning_rate": 0.01,
      "loss": 4.611584186553955,
      "num_input_tokens_seen": 3005552,
      "step": 757,
      "train_runtime": 605.7408,
      "train_tokens_per_second": 4961.779
    },
    {
      "epoch": 58.23076923076923,
      "num_input_tokens_seen": 3005552,
      "step": 757,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 58.30769230769231,
      "grad_norm": 0.42360833287239075,
      "learning_rate": 0.01,
      "loss": 4.624504566192627,
      "num_input_tokens_seen": 3009648,
      "step": 758,
      "train_runtime": 606.5549,
      "train_tokens_per_second": 4961.873
    },
    {
      "epoch": 58.30769230769231,
      "num_input_tokens_seen": 3009648,
      "step": 758,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 58.38461538461539,
      "grad_norm": 0.4611240327358246,
      "learning_rate": 0.01,
      "loss": 4.658176898956299,
      "num_input_tokens_seen": 3013744,
      "step": 759,
      "train_runtime": 607.3699,
      "train_tokens_per_second": 4961.958
    },
    {
      "epoch": 58.38461538461539,
      "num_input_tokens_seen": 3013744,
      "step": 759,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 58.46153846153846,
      "grad_norm": 0.3276404142379761,
      "learning_rate": 0.01,
      "loss": 4.597156524658203,
      "num_input_tokens_seen": 3017840,
      "step": 760,
      "train_runtime": 608.1852,
      "train_tokens_per_second": 4962.041
    },
    {
      "epoch": 58.46153846153846,
      "num_input_tokens_seen": 3017840,
      "step": 760,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 58.53846153846154,
      "grad_norm": 0.5247175097465515,
      "learning_rate": 0.01,
      "loss": 4.647055625915527,
      "num_input_tokens_seen": 3021936,
      "step": 761,
      "train_runtime": 608.9994,
      "train_tokens_per_second": 4962.133
    },
    {
      "epoch": 58.53846153846154,
      "num_input_tokens_seen": 3021936,
      "step": 761,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 58.61538461538461,
      "grad_norm": 0.9190148711204529,
      "learning_rate": 0.01,
      "loss": 4.61857271194458,
      "num_input_tokens_seen": 3026032,
      "step": 762,
      "train_runtime": 609.8158,
      "train_tokens_per_second": 4962.207
    },
    {
      "epoch": 58.61538461538461,
      "num_input_tokens_seen": 3026032,
      "step": 762,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 58.69230769230769,
      "grad_norm": 0.41588884592056274,
      "learning_rate": 0.01,
      "loss": 4.630764007568359,
      "num_input_tokens_seen": 3030128,
      "step": 763,
      "train_runtime": 610.6304,
      "train_tokens_per_second": 4962.295
    },
    {
      "epoch": 58.69230769230769,
      "num_input_tokens_seen": 3030128,
      "step": 763,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 58.76923076923077,
      "grad_norm": 0.4531322717666626,
      "learning_rate": 0.01,
      "loss": 4.6122589111328125,
      "num_input_tokens_seen": 3034224,
      "step": 764,
      "train_runtime": 611.4467,
      "train_tokens_per_second": 4962.369
    },
    {
      "epoch": 58.76923076923077,
      "num_input_tokens_seen": 3034224,
      "step": 764,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 58.84615384615385,
      "grad_norm": 0.5354483723640442,
      "learning_rate": 0.01,
      "loss": 4.611150741577148,
      "num_input_tokens_seen": 3038320,
      "step": 765,
      "train_runtime": 612.2621,
      "train_tokens_per_second": 4962.45
    },
    {
      "epoch": 58.84615384615385,
      "num_input_tokens_seen": 3038320,
      "step": 765,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 58.92307692307692,
      "grad_norm": 0.35458430647850037,
      "learning_rate": 0.01,
      "loss": 4.661350727081299,
      "num_input_tokens_seen": 3042416,
      "step": 766,
      "train_runtime": 613.0735,
      "train_tokens_per_second": 4962.563
    },
    {
      "epoch": 58.92307692307692,
      "num_input_tokens_seen": 3042416,
      "step": 766,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 59.0,
      "grad_norm": 0.5755992531776428,
      "learning_rate": 0.01,
      "loss": 4.634483337402344,
      "num_input_tokens_seen": 3044872,
      "step": 767,
      "train_runtime": 613.5895,
      "train_tokens_per_second": 4962.392
    },
    {
      "epoch": 59.0,
      "num_input_tokens_seen": 3044872,
      "step": 767,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 59.07692307692308,
      "grad_norm": 0.6252405047416687,
      "learning_rate": 0.01,
      "loss": 4.6295247077941895,
      "num_input_tokens_seen": 3048968,
      "step": 768,
      "train_runtime": 614.4156,
      "train_tokens_per_second": 4962.387
    },
    {
      "epoch": 59.07692307692308,
      "num_input_tokens_seen": 3048968,
      "step": 768,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 59.15384615384615,
      "grad_norm": 0.6782728433609009,
      "learning_rate": 0.01,
      "loss": 4.613170623779297,
      "num_input_tokens_seen": 3053064,
      "step": 769,
      "train_runtime": 615.2305,
      "train_tokens_per_second": 4962.472
    },
    {
      "epoch": 59.15384615384615,
      "num_input_tokens_seen": 3053064,
      "step": 769,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 59.23076923076923,
      "grad_norm": 0.5422839522361755,
      "learning_rate": 0.01,
      "loss": 4.658756256103516,
      "num_input_tokens_seen": 3057160,
      "step": 770,
      "train_runtime": 616.0472,
      "train_tokens_per_second": 4962.542
    },
    {
      "epoch": 59.23076923076923,
      "num_input_tokens_seen": 3057160,
      "step": 770,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 59.30769230769231,
      "grad_norm": 0.7251322865486145,
      "learning_rate": 0.01,
      "loss": 4.618936538696289,
      "num_input_tokens_seen": 3061256,
      "step": 771,
      "train_runtime": 616.8623,
      "train_tokens_per_second": 4962.625
    },
    {
      "epoch": 59.30769230769231,
      "num_input_tokens_seen": 3061256,
      "step": 771,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 59.38461538461539,
      "grad_norm": 0.47734972834587097,
      "learning_rate": 0.01,
      "loss": 4.614592552185059,
      "num_input_tokens_seen": 3065352,
      "step": 772,
      "train_runtime": 617.6782,
      "train_tokens_per_second": 4962.701
    },
    {
      "epoch": 59.38461538461539,
      "num_input_tokens_seen": 3065352,
      "step": 772,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 59.46153846153846,
      "grad_norm": 0.5645265579223633,
      "learning_rate": 0.01,
      "loss": 4.658867359161377,
      "num_input_tokens_seen": 3069448,
      "step": 773,
      "train_runtime": 618.4939,
      "train_tokens_per_second": 4962.778
    },
    {
      "epoch": 59.46153846153846,
      "num_input_tokens_seen": 3069448,
      "step": 773,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 59.53846153846154,
      "grad_norm": 0.5353477001190186,
      "learning_rate": 0.01,
      "loss": 4.621963024139404,
      "num_input_tokens_seen": 3073544,
      "step": 774,
      "train_runtime": 619.3092,
      "train_tokens_per_second": 4962.858
    },
    {
      "epoch": 59.53846153846154,
      "num_input_tokens_seen": 3073544,
      "step": 774,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 59.61538461538461,
      "grad_norm": 0.3467482924461365,
      "learning_rate": 0.01,
      "loss": 4.559604644775391,
      "num_input_tokens_seen": 3077640,
      "step": 775,
      "train_runtime": 620.1247,
      "train_tokens_per_second": 4962.937
    },
    {
      "epoch": 59.61538461538461,
      "num_input_tokens_seen": 3077640,
      "step": 775,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 59.69230769230769,
      "grad_norm": 0.6668217778205872,
      "learning_rate": 0.01,
      "loss": 4.629401206970215,
      "num_input_tokens_seen": 3081736,
      "step": 776,
      "train_runtime": 620.9392,
      "train_tokens_per_second": 4963.023
    },
    {
      "epoch": 59.69230769230769,
      "num_input_tokens_seen": 3081736,
      "step": 776,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 59.76923076923077,
      "grad_norm": 0.5376278758049011,
      "learning_rate": 0.01,
      "loss": 4.66391658782959,
      "num_input_tokens_seen": 3085832,
      "step": 777,
      "train_runtime": 621.7551,
      "train_tokens_per_second": 4963.099
    },
    {
      "epoch": 59.76923076923077,
      "num_input_tokens_seen": 3085832,
      "step": 777,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 59.84615384615385,
      "grad_norm": 0.6031669974327087,
      "learning_rate": 0.01,
      "loss": 4.584033012390137,
      "num_input_tokens_seen": 3089928,
      "step": 778,
      "train_runtime": 622.5694,
      "train_tokens_per_second": 4963.187
    },
    {
      "epoch": 59.84615384615385,
      "num_input_tokens_seen": 3089928,
      "step": 778,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 59.92307692307692,
      "grad_norm": 0.4091344475746155,
      "learning_rate": 0.01,
      "loss": 4.601544380187988,
      "num_input_tokens_seen": 3094024,
      "step": 779,
      "train_runtime": 623.3806,
      "train_tokens_per_second": 4963.299
    },
    {
      "epoch": 59.92307692307692,
      "num_input_tokens_seen": 3094024,
      "step": 779,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 60.0,
      "grad_norm": 0.769387423992157,
      "learning_rate": 0.01,
      "loss": 4.655564308166504,
      "num_input_tokens_seen": 3096480,
      "step": 780,
      "train_runtime": 623.8971,
      "train_tokens_per_second": 4963.126
    },
    {
      "epoch": 60.0,
      "num_input_tokens_seen": 3096480,
      "step": 780,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 60.07692307692308,
      "grad_norm": 1.0533069372177124,
      "learning_rate": 0.01,
      "loss": 4.592347621917725,
      "num_input_tokens_seen": 3100576,
      "step": 781,
      "train_runtime": 624.7232,
      "train_tokens_per_second": 4963.12
    },
    {
      "epoch": 60.07692307692308,
      "num_input_tokens_seen": 3100576,
      "step": 781,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 60.15384615384615,
      "grad_norm": 0.5696276426315308,
      "learning_rate": 0.01,
      "loss": 4.577451705932617,
      "num_input_tokens_seen": 3104672,
      "step": 782,
      "train_runtime": 625.5381,
      "train_tokens_per_second": 4963.202
    },
    {
      "epoch": 60.15384615384615,
      "num_input_tokens_seen": 3104672,
      "step": 782,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 60.23076923076923,
      "grad_norm": 0.4790893793106079,
      "learning_rate": 0.01,
      "loss": 4.570064544677734,
      "num_input_tokens_seen": 3108768,
      "step": 783,
      "train_runtime": 626.3531,
      "train_tokens_per_second": 4963.283
    },
    {
      "epoch": 60.23076923076923,
      "num_input_tokens_seen": 3108768,
      "step": 783,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 60.30769230769231,
      "grad_norm": 0.3884826600551605,
      "learning_rate": 0.01,
      "loss": 4.581583023071289,
      "num_input_tokens_seen": 3112864,
      "step": 784,
      "train_runtime": 627.1676,
      "train_tokens_per_second": 4963.369
    },
    {
      "epoch": 60.30769230769231,
      "num_input_tokens_seen": 3112864,
      "step": 784,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 60.38461538461539,
      "grad_norm": 0.5839759707450867,
      "learning_rate": 0.01,
      "loss": 4.604778289794922,
      "num_input_tokens_seen": 3116960,
      "step": 785,
      "train_runtime": 627.9834,
      "train_tokens_per_second": 4963.444
    },
    {
      "epoch": 60.38461538461539,
      "num_input_tokens_seen": 3116960,
      "step": 785,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 60.46153846153846,
      "grad_norm": 0.337950736284256,
      "learning_rate": 0.01,
      "loss": 4.588187217712402,
      "num_input_tokens_seen": 3121056,
      "step": 786,
      "train_runtime": 628.7993,
      "train_tokens_per_second": 4963.517
    },
    {
      "epoch": 60.46153846153846,
      "num_input_tokens_seen": 3121056,
      "step": 786,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 60.53846153846154,
      "grad_norm": 0.7020766735076904,
      "learning_rate": 0.01,
      "loss": 4.6410064697265625,
      "num_input_tokens_seen": 3125152,
      "step": 787,
      "train_runtime": 629.6208,
      "train_tokens_per_second": 4963.546
    },
    {
      "epoch": 60.53846153846154,
      "num_input_tokens_seen": 3125152,
      "step": 787,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 60.61538461538461,
      "grad_norm": 0.5368402600288391,
      "learning_rate": 0.01,
      "loss": 4.639837265014648,
      "num_input_tokens_seen": 3129248,
      "step": 788,
      "train_runtime": 630.4378,
      "train_tokens_per_second": 4963.611
    },
    {
      "epoch": 60.61538461538461,
      "num_input_tokens_seen": 3129248,
      "step": 788,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 60.69230769230769,
      "grad_norm": 0.47061866521835327,
      "learning_rate": 0.01,
      "loss": 4.574796676635742,
      "num_input_tokens_seen": 3133344,
      "step": 789,
      "train_runtime": 631.2552,
      "train_tokens_per_second": 4963.672
    },
    {
      "epoch": 60.69230769230769,
      "num_input_tokens_seen": 3133344,
      "step": 789,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 60.76923076923077,
      "grad_norm": 0.4266293942928314,
      "learning_rate": 0.01,
      "loss": 4.553921699523926,
      "num_input_tokens_seen": 3137440,
      "step": 790,
      "train_runtime": 632.0714,
      "train_tokens_per_second": 4963.743
    },
    {
      "epoch": 60.76923076923077,
      "num_input_tokens_seen": 3137440,
      "step": 790,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 60.84615384615385,
      "grad_norm": 0.38818246126174927,
      "learning_rate": 0.01,
      "loss": 4.5928239822387695,
      "num_input_tokens_seen": 3141536,
      "step": 791,
      "train_runtime": 632.8886,
      "train_tokens_per_second": 4963.806
    },
    {
      "epoch": 60.84615384615385,
      "num_input_tokens_seen": 3141536,
      "step": 791,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 60.92307692307692,
      "grad_norm": 0.5804455280303955,
      "learning_rate": 0.01,
      "loss": 4.585761070251465,
      "num_input_tokens_seen": 3145632,
      "step": 792,
      "train_runtime": 633.7005,
      "train_tokens_per_second": 4963.91
    },
    {
      "epoch": 60.92307692307692,
      "num_input_tokens_seen": 3145632,
      "step": 792,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 61.0,
      "grad_norm": 0.480523943901062,
      "learning_rate": 0.01,
      "loss": 4.664402961730957,
      "num_input_tokens_seen": 3148088,
      "step": 793,
      "train_runtime": 634.2165,
      "train_tokens_per_second": 4963.743
    },
    {
      "epoch": 61.0,
      "num_input_tokens_seen": 3148088,
      "step": 793,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 61.07692307692308,
      "grad_norm": 0.3555164933204651,
      "learning_rate": 0.01,
      "loss": 4.57891845703125,
      "num_input_tokens_seen": 3152184,
      "step": 794,
      "train_runtime": 635.0434,
      "train_tokens_per_second": 4963.73
    },
    {
      "epoch": 61.07692307692308,
      "num_input_tokens_seen": 3152184,
      "step": 794,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 61.15384615384615,
      "grad_norm": 0.4352375864982605,
      "learning_rate": 0.01,
      "loss": 4.668402194976807,
      "num_input_tokens_seen": 3156280,
      "step": 795,
      "train_runtime": 635.8599,
      "train_tokens_per_second": 4963.798
    },
    {
      "epoch": 61.15384615384615,
      "num_input_tokens_seen": 3156280,
      "step": 795,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 61.23076923076923,
      "grad_norm": 0.6781975626945496,
      "learning_rate": 0.01,
      "loss": 4.623663902282715,
      "num_input_tokens_seen": 3160376,
      "step": 796,
      "train_runtime": 636.6749,
      "train_tokens_per_second": 4963.877
    },
    {
      "epoch": 61.23076923076923,
      "num_input_tokens_seen": 3160376,
      "step": 796,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 61.30769230769231,
      "grad_norm": 0.47450610995292664,
      "learning_rate": 0.01,
      "loss": 4.609356880187988,
      "num_input_tokens_seen": 3164472,
      "step": 797,
      "train_runtime": 637.4895,
      "train_tokens_per_second": 4963.959
    },
    {
      "epoch": 61.30769230769231,
      "num_input_tokens_seen": 3164472,
      "step": 797,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 61.38461538461539,
      "grad_norm": 0.63689124584198,
      "learning_rate": 0.01,
      "loss": 4.566770553588867,
      "num_input_tokens_seen": 3168568,
      "step": 798,
      "train_runtime": 638.3052,
      "train_tokens_per_second": 4964.033
    },
    {
      "epoch": 61.38461538461539,
      "num_input_tokens_seen": 3168568,
      "step": 798,
      "train_accuracy": 0.03125
    },
    {
      "epoch": 61.46153846153846,
      "grad_norm": 0.2824016511440277,
      "learning_rate": 0.01,
      "loss": 4.541934013366699,
      "num_input_tokens_seen": 3172664,
      "step": 799,
      "train_runtime": 639.1201,
      "train_tokens_per_second": 4964.113
    },
    {
      "epoch": 61.46153846153846,
      "num_input_tokens_seen": 3172664,
      "step": 799,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 61.53846153846154,
      "grad_norm": 0.5654345154762268,
      "learning_rate": 0.01,
      "loss": 4.638147354125977,
      "num_input_tokens_seen": 3176760,
      "step": 800,
      "train_runtime": 639.9355,
      "train_tokens_per_second": 4964.188
    },
    {
      "epoch": 61.53846153846154,
      "eval_CMD_3_branch_eval_accuracy": 0.009403254972875226,
      "eval_CMD_3_branch_eval_loss": 4.673147678375244,
      "eval_CMD_3_branch_eval_runtime": 1.1525,
      "eval_CMD_3_branch_eval_samples_per_second": 2399.235,
      "eval_CMD_3_branch_eval_steps_per_second": 5.206,
      "num_input_tokens_seen": 3176760,
      "step": 800
    },
    {
      "epoch": 61.53846153846154,
      "num_input_tokens_seen": 3176760,
      "step": 800,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 61.61538461538461,
      "grad_norm": 0.4775870740413666,
      "learning_rate": 0.01,
      "loss": 4.564815521240234,
      "num_input_tokens_seen": 3180856,
      "step": 801,
      "train_runtime": 641.9086,
      "train_tokens_per_second": 4955.31
    },
    {
      "epoch": 61.61538461538461,
      "num_input_tokens_seen": 3180856,
      "step": 801,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 61.69230769230769,
      "grad_norm": 0.4272870421409607,
      "learning_rate": 0.01,
      "loss": 4.57087516784668,
      "num_input_tokens_seen": 3184952,
      "step": 802,
      "train_runtime": 642.724,
      "train_tokens_per_second": 4955.396
    },
    {
      "epoch": 61.69230769230769,
      "num_input_tokens_seen": 3184952,
      "step": 802,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 61.76923076923077,
      "grad_norm": 0.44081181287765503,
      "learning_rate": 0.01,
      "loss": 4.596053123474121,
      "num_input_tokens_seen": 3189048,
      "step": 803,
      "train_runtime": 643.5392,
      "train_tokens_per_second": 4955.484
    },
    {
      "epoch": 61.76923076923077,
      "num_input_tokens_seen": 3189048,
      "step": 803,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 61.84615384615385,
      "grad_norm": 0.8313249945640564,
      "learning_rate": 0.01,
      "loss": 4.604004859924316,
      "num_input_tokens_seen": 3193144,
      "step": 804,
      "train_runtime": 644.355,
      "train_tokens_per_second": 4955.566
    },
    {
      "epoch": 61.84615384615385,
      "num_input_tokens_seen": 3193144,
      "step": 804,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 61.92307692307692,
      "grad_norm": 0.40859705209732056,
      "learning_rate": 0.01,
      "loss": 4.617779731750488,
      "num_input_tokens_seen": 3197240,
      "step": 805,
      "train_runtime": 645.1682,
      "train_tokens_per_second": 4955.669
    },
    {
      "epoch": 61.92307692307692,
      "num_input_tokens_seen": 3197240,
      "step": 805,
      "train_accuracy": 0.013029315508902073
    },
    {
      "epoch": 62.0,
      "grad_norm": 0.8009452223777771,
      "learning_rate": 0.01,
      "loss": 4.670826435089111,
      "num_input_tokens_seen": 3199696,
      "step": 806,
      "train_runtime": 645.6839,
      "train_tokens_per_second": 4955.515
    },
    {
      "epoch": 62.0,
      "num_input_tokens_seen": 3199696,
      "step": 806,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 62.07692307692308,
      "grad_norm": 0.4048238694667816,
      "learning_rate": 0.01,
      "loss": 4.574530601501465,
      "num_input_tokens_seen": 3203792,
      "step": 807,
      "train_runtime": 646.5116,
      "train_tokens_per_second": 4955.506
    },
    {
      "epoch": 62.07692307692308,
      "num_input_tokens_seen": 3203792,
      "step": 807,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 62.15384615384615,
      "grad_norm": 0.5081648230552673,
      "learning_rate": 0.01,
      "loss": 4.61641263961792,
      "num_input_tokens_seen": 3207888,
      "step": 808,
      "train_runtime": 647.3274,
      "train_tokens_per_second": 4955.588
    },
    {
      "epoch": 62.15384615384615,
      "num_input_tokens_seen": 3207888,
      "step": 808,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 62.23076923076923,
      "grad_norm": 0.3222755491733551,
      "learning_rate": 0.01,
      "loss": 4.564643859863281,
      "num_input_tokens_seen": 3211984,
      "step": 809,
      "train_runtime": 648.144,
      "train_tokens_per_second": 4955.664
    },
    {
      "epoch": 62.23076923076923,
      "num_input_tokens_seen": 3211984,
      "step": 809,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 62.30769230769231,
      "grad_norm": 0.41443246603012085,
      "learning_rate": 0.01,
      "loss": 4.546332359313965,
      "num_input_tokens_seen": 3216080,
      "step": 810,
      "train_runtime": 648.9592,
      "train_tokens_per_second": 4955.75
    },
    {
      "epoch": 62.30769230769231,
      "num_input_tokens_seen": 3216080,
      "step": 810,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 62.38461538461539,
      "grad_norm": 0.39902424812316895,
      "learning_rate": 0.01,
      "loss": 4.5842790603637695,
      "num_input_tokens_seen": 3220176,
      "step": 811,
      "train_runtime": 649.7759,
      "train_tokens_per_second": 4955.826
    },
    {
      "epoch": 62.38461538461539,
      "num_input_tokens_seen": 3220176,
      "step": 811,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 62.46153846153846,
      "grad_norm": 0.7111110091209412,
      "learning_rate": 0.01,
      "loss": 4.599859237670898,
      "num_input_tokens_seen": 3224272,
      "step": 812,
      "train_runtime": 650.5915,
      "train_tokens_per_second": 4955.909
    },
    {
      "epoch": 62.46153846153846,
      "num_input_tokens_seen": 3224272,
      "step": 812,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 62.53846153846154,
      "grad_norm": 0.34514543414115906,
      "learning_rate": 0.01,
      "loss": 4.589466094970703,
      "num_input_tokens_seen": 3228368,
      "step": 813,
      "train_runtime": 651.4081,
      "train_tokens_per_second": 4955.984
    },
    {
      "epoch": 62.53846153846154,
      "num_input_tokens_seen": 3228368,
      "step": 813,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 62.61538461538461,
      "grad_norm": 0.36978796124458313,
      "learning_rate": 0.01,
      "loss": 4.593320846557617,
      "num_input_tokens_seen": 3232464,
      "step": 814,
      "train_runtime": 652.2257,
      "train_tokens_per_second": 4956.051
    },
    {
      "epoch": 62.61538461538461,
      "num_input_tokens_seen": 3232464,
      "step": 814,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 62.69230769230769,
      "grad_norm": 0.5854072570800781,
      "learning_rate": 0.01,
      "loss": 4.576139450073242,
      "num_input_tokens_seen": 3236560,
      "step": 815,
      "train_runtime": 653.044,
      "train_tokens_per_second": 4956.113
    },
    {
      "epoch": 62.69230769230769,
      "num_input_tokens_seen": 3236560,
      "step": 815,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 62.76923076923077,
      "grad_norm": 0.3284805417060852,
      "learning_rate": 0.01,
      "loss": 4.58418083190918,
      "num_input_tokens_seen": 3240656,
      "step": 816,
      "train_runtime": 653.8605,
      "train_tokens_per_second": 4956.189
    },
    {
      "epoch": 62.76923076923077,
      "num_input_tokens_seen": 3240656,
      "step": 816,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 62.84615384615385,
      "grad_norm": 0.5775055885314941,
      "learning_rate": 0.01,
      "loss": 4.60098934173584,
      "num_input_tokens_seen": 3244752,
      "step": 817,
      "train_runtime": 654.6787,
      "train_tokens_per_second": 4956.251
    },
    {
      "epoch": 62.84615384615385,
      "num_input_tokens_seen": 3244752,
      "step": 817,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 62.92307692307692,
      "grad_norm": 0.4311469495296478,
      "learning_rate": 0.01,
      "loss": 4.5810322761535645,
      "num_input_tokens_seen": 3248848,
      "step": 818,
      "train_runtime": 655.49,
      "train_tokens_per_second": 4956.365
    },
    {
      "epoch": 62.92307692307692,
      "num_input_tokens_seen": 3248848,
      "step": 818,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 63.0,
      "grad_norm": 0.41363051533699036,
      "learning_rate": 0.01,
      "loss": 4.620242595672607,
      "num_input_tokens_seen": 3251304,
      "step": 819,
      "train_runtime": 656.007,
      "train_tokens_per_second": 4956.203
    },
    {
      "epoch": 63.0,
      "num_input_tokens_seen": 3251304,
      "step": 819,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 63.07692307692308,
      "grad_norm": 0.46569857001304626,
      "learning_rate": 0.01,
      "loss": 4.544882774353027,
      "num_input_tokens_seen": 3255400,
      "step": 820,
      "train_runtime": 656.8351,
      "train_tokens_per_second": 4956.191
    },
    {
      "epoch": 63.07692307692308,
      "num_input_tokens_seen": 3255400,
      "step": 820,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 63.15384615384615,
      "grad_norm": 0.5212401151657104,
      "learning_rate": 0.01,
      "loss": 4.550348281860352,
      "num_input_tokens_seen": 3259496,
      "step": 821,
      "train_runtime": 657.653,
      "train_tokens_per_second": 4956.255
    },
    {
      "epoch": 63.15384615384615,
      "num_input_tokens_seen": 3259496,
      "step": 821,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 63.23076923076923,
      "grad_norm": 0.3535076975822449,
      "learning_rate": 0.01,
      "loss": 4.574855804443359,
      "num_input_tokens_seen": 3263592,
      "step": 822,
      "train_runtime": 658.4702,
      "train_tokens_per_second": 4956.325
    },
    {
      "epoch": 63.23076923076923,
      "num_input_tokens_seen": 3263592,
      "step": 822,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 63.30769230769231,
      "grad_norm": 0.31359365582466125,
      "learning_rate": 0.01,
      "loss": 4.522514820098877,
      "num_input_tokens_seen": 3267688,
      "step": 823,
      "train_runtime": 659.2856,
      "train_tokens_per_second": 4956.407
    },
    {
      "epoch": 63.30769230769231,
      "num_input_tokens_seen": 3267688,
      "step": 823,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 63.38461538461539,
      "grad_norm": 2.1746299266815186,
      "learning_rate": 0.01,
      "loss": 4.580921173095703,
      "num_input_tokens_seen": 3271784,
      "step": 824,
      "train_runtime": 660.1022,
      "train_tokens_per_second": 4956.481
    },
    {
      "epoch": 63.38461538461539,
      "num_input_tokens_seen": 3271784,
      "step": 824,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 63.46153846153846,
      "grad_norm": 0.36821866035461426,
      "learning_rate": 0.01,
      "loss": 4.554903984069824,
      "num_input_tokens_seen": 3275880,
      "step": 825,
      "train_runtime": 660.9182,
      "train_tokens_per_second": 4956.559
    },
    {
      "epoch": 63.46153846153846,
      "num_input_tokens_seen": 3275880,
      "step": 825,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 63.53846153846154,
      "grad_norm": 0.8334118127822876,
      "learning_rate": 0.01,
      "loss": 4.595179557800293,
      "num_input_tokens_seen": 3279976,
      "step": 826,
      "train_runtime": 661.7356,
      "train_tokens_per_second": 4956.626
    },
    {
      "epoch": 63.53846153846154,
      "num_input_tokens_seen": 3279976,
      "step": 826,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 63.61538461538461,
      "grad_norm": 0.459299236536026,
      "learning_rate": 0.01,
      "loss": 4.620095252990723,
      "num_input_tokens_seen": 3284072,
      "step": 827,
      "train_runtime": 662.554,
      "train_tokens_per_second": 4956.686
    },
    {
      "epoch": 63.61538461538461,
      "num_input_tokens_seen": 3284072,
      "step": 827,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 63.69230769230769,
      "grad_norm": 0.2941116392612457,
      "learning_rate": 0.01,
      "loss": 4.52593994140625,
      "num_input_tokens_seen": 3288168,
      "step": 828,
      "train_runtime": 663.3716,
      "train_tokens_per_second": 4956.751
    },
    {
      "epoch": 63.69230769230769,
      "num_input_tokens_seen": 3288168,
      "step": 828,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 63.76923076923077,
      "grad_norm": 0.42291563749313354,
      "learning_rate": 0.01,
      "loss": 4.538573265075684,
      "num_input_tokens_seen": 3292264,
      "step": 829,
      "train_runtime": 664.188,
      "train_tokens_per_second": 4956.826
    },
    {
      "epoch": 63.76923076923077,
      "num_input_tokens_seen": 3292264,
      "step": 829,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 63.84615384615385,
      "grad_norm": 0.868320643901825,
      "learning_rate": 0.01,
      "loss": 4.574146747589111,
      "num_input_tokens_seen": 3296360,
      "step": 830,
      "train_runtime": 665.0058,
      "train_tokens_per_second": 4956.889
    },
    {
      "epoch": 63.84615384615385,
      "num_input_tokens_seen": 3296360,
      "step": 830,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 63.92307692307692,
      "grad_norm": 0.25749003887176514,
      "learning_rate": 0.01,
      "loss": 4.5796589851379395,
      "num_input_tokens_seen": 3300456,
      "step": 831,
      "train_runtime": 665.8176,
      "train_tokens_per_second": 4956.997
    },
    {
      "epoch": 63.92307692307692,
      "num_input_tokens_seen": 3300456,
      "step": 831,
      "train_accuracy": 0.013029315508902073
    },
    {
      "epoch": 64.0,
      "grad_norm": 0.52784663438797,
      "learning_rate": 0.01,
      "loss": 4.594731330871582,
      "num_input_tokens_seen": 3302912,
      "step": 832,
      "train_runtime": 666.3346,
      "train_tokens_per_second": 4956.837
    },
    {
      "epoch": 64.0,
      "num_input_tokens_seen": 3302912,
      "step": 832,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 64.07692307692308,
      "grad_norm": 0.3691712021827698,
      "learning_rate": 0.01,
      "loss": 4.568826675415039,
      "num_input_tokens_seen": 3307008,
      "step": 833,
      "train_runtime": 667.1629,
      "train_tokens_per_second": 4956.823
    },
    {
      "epoch": 64.07692307692308,
      "num_input_tokens_seen": 3307008,
      "step": 833,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 64.15384615384616,
      "grad_norm": 0.35906437039375305,
      "learning_rate": 0.01,
      "loss": 4.541396141052246,
      "num_input_tokens_seen": 3311104,
      "step": 834,
      "train_runtime": 667.9803,
      "train_tokens_per_second": 4956.889
    },
    {
      "epoch": 64.15384615384616,
      "num_input_tokens_seen": 3311104,
      "step": 834,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 64.23076923076923,
      "grad_norm": 0.446509450674057,
      "learning_rate": 0.01,
      "loss": 4.571511268615723,
      "num_input_tokens_seen": 3315200,
      "step": 835,
      "train_runtime": 668.7963,
      "train_tokens_per_second": 4956.965
    },
    {
      "epoch": 64.23076923076923,
      "num_input_tokens_seen": 3315200,
      "step": 835,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 64.3076923076923,
      "grad_norm": 0.39407438039779663,
      "learning_rate": 0.01,
      "loss": 4.596895217895508,
      "num_input_tokens_seen": 3319296,
      "step": 836,
      "train_runtime": 669.6128,
      "train_tokens_per_second": 4957.038
    },
    {
      "epoch": 64.3076923076923,
      "num_input_tokens_seen": 3319296,
      "step": 836,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 64.38461538461539,
      "grad_norm": 0.29259923100471497,
      "learning_rate": 0.01,
      "loss": 4.561599254608154,
      "num_input_tokens_seen": 3323392,
      "step": 837,
      "train_runtime": 670.43,
      "train_tokens_per_second": 4957.105
    },
    {
      "epoch": 64.38461538461539,
      "num_input_tokens_seen": 3323392,
      "step": 837,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 64.46153846153847,
      "grad_norm": 0.3567143380641937,
      "learning_rate": 0.01,
      "loss": 4.5362067222595215,
      "num_input_tokens_seen": 3327488,
      "step": 838,
      "train_runtime": 671.2476,
      "train_tokens_per_second": 4957.169
    },
    {
      "epoch": 64.46153846153847,
      "num_input_tokens_seen": 3327488,
      "step": 838,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 64.53846153846153,
      "grad_norm": 0.5534819960594177,
      "learning_rate": 0.01,
      "loss": 4.5910515785217285,
      "num_input_tokens_seen": 3331584,
      "step": 839,
      "train_runtime": 672.0656,
      "train_tokens_per_second": 4957.23
    },
    {
      "epoch": 64.53846153846153,
      "num_input_tokens_seen": 3331584,
      "step": 839,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 64.61538461538461,
      "grad_norm": 0.3455497920513153,
      "learning_rate": 0.01,
      "loss": 4.5095014572143555,
      "num_input_tokens_seen": 3335680,
      "step": 840,
      "train_runtime": 672.883,
      "train_tokens_per_second": 4957.295
    },
    {
      "epoch": 64.61538461538461,
      "num_input_tokens_seen": 3335680,
      "step": 840,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 64.6923076923077,
      "grad_norm": 0.6076437830924988,
      "learning_rate": 0.01,
      "loss": 4.570412635803223,
      "num_input_tokens_seen": 3339776,
      "step": 841,
      "train_runtime": 673.7006,
      "train_tokens_per_second": 4957.359
    },
    {
      "epoch": 64.6923076923077,
      "num_input_tokens_seen": 3339776,
      "step": 841,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 64.76923076923077,
      "grad_norm": 0.6800243258476257,
      "learning_rate": 0.01,
      "loss": 4.616314888000488,
      "num_input_tokens_seen": 3343872,
      "step": 842,
      "train_runtime": 674.518,
      "train_tokens_per_second": 4957.425
    },
    {
      "epoch": 64.76923076923077,
      "num_input_tokens_seen": 3343872,
      "step": 842,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 64.84615384615384,
      "grad_norm": 0.741438627243042,
      "learning_rate": 0.01,
      "loss": 4.646421432495117,
      "num_input_tokens_seen": 3347968,
      "step": 843,
      "train_runtime": 675.3358,
      "train_tokens_per_second": 4957.486
    },
    {
      "epoch": 64.84615384615384,
      "num_input_tokens_seen": 3347968,
      "step": 843,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 64.92307692307692,
      "grad_norm": 1.0583053827285767,
      "learning_rate": 0.01,
      "loss": 4.638894081115723,
      "num_input_tokens_seen": 3352064,
      "step": 844,
      "train_runtime": 676.153,
      "train_tokens_per_second": 4957.553
    },
    {
      "epoch": 64.92307692307692,
      "num_input_tokens_seen": 3352064,
      "step": 844,
      "train_accuracy": 0.013029315508902073
    },
    {
      "epoch": 65.0,
      "grad_norm": 0.32858121395111084,
      "learning_rate": 0.01,
      "loss": 4.560145378112793,
      "num_input_tokens_seen": 3354520,
      "step": 845,
      "train_runtime": 676.6712,
      "train_tokens_per_second": 4957.385
    },
    {
      "epoch": 65.0,
      "num_input_tokens_seen": 3354520,
      "step": 845,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 65.07692307692308,
      "grad_norm": 0.6705158948898315,
      "learning_rate": 0.01,
      "loss": 4.5458149909973145,
      "num_input_tokens_seen": 3358616,
      "step": 846,
      "train_runtime": 677.5014,
      "train_tokens_per_second": 4957.356
    },
    {
      "epoch": 65.07692307692308,
      "num_input_tokens_seen": 3358616,
      "step": 846,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 65.15384615384616,
      "grad_norm": 0.3764450252056122,
      "learning_rate": 0.01,
      "loss": 4.460278511047363,
      "num_input_tokens_seen": 3362712,
      "step": 847,
      "train_runtime": 678.3184,
      "train_tokens_per_second": 4957.424
    },
    {
      "epoch": 65.15384615384616,
      "num_input_tokens_seen": 3362712,
      "step": 847,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 65.23076923076923,
      "grad_norm": 0.5173666477203369,
      "learning_rate": 0.01,
      "loss": 4.573984146118164,
      "num_input_tokens_seen": 3366808,
      "step": 848,
      "train_runtime": 679.1352,
      "train_tokens_per_second": 4957.493
    },
    {
      "epoch": 65.23076923076923,
      "num_input_tokens_seen": 3366808,
      "step": 848,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 65.3076923076923,
      "grad_norm": 0.3733808696269989,
      "learning_rate": 0.01,
      "loss": 4.59464168548584,
      "num_input_tokens_seen": 3370904,
      "step": 849,
      "train_runtime": 679.9537,
      "train_tokens_per_second": 4957.549
    },
    {
      "epoch": 65.3076923076923,
      "num_input_tokens_seen": 3370904,
      "step": 849,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 65.38461538461539,
      "grad_norm": 0.35254454612731934,
      "learning_rate": 0.01,
      "loss": 4.5532121658325195,
      "num_input_tokens_seen": 3375000,
      "step": 850,
      "train_runtime": 680.7718,
      "train_tokens_per_second": 4957.609
    },
    {
      "epoch": 65.38461538461539,
      "num_input_tokens_seen": 3375000,
      "step": 850,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 65.46153846153847,
      "grad_norm": 0.3959769010543823,
      "learning_rate": 0.01,
      "loss": 4.507648468017578,
      "num_input_tokens_seen": 3379096,
      "step": 851,
      "train_runtime": 681.589,
      "train_tokens_per_second": 4957.674
    },
    {
      "epoch": 65.46153846153847,
      "num_input_tokens_seen": 3379096,
      "step": 851,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 65.53846153846153,
      "grad_norm": 0.5590829849243164,
      "learning_rate": 0.01,
      "loss": 4.564872741699219,
      "num_input_tokens_seen": 3383192,
      "step": 852,
      "train_runtime": 682.4069,
      "train_tokens_per_second": 4957.734
    },
    {
      "epoch": 65.53846153846153,
      "num_input_tokens_seen": 3383192,
      "step": 852,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 65.61538461538461,
      "grad_norm": 0.3259311318397522,
      "learning_rate": 0.01,
      "loss": 4.633062362670898,
      "num_input_tokens_seen": 3387288,
      "step": 853,
      "train_runtime": 683.2256,
      "train_tokens_per_second": 4957.788
    },
    {
      "epoch": 65.61538461538461,
      "num_input_tokens_seen": 3387288,
      "step": 853,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 65.6923076923077,
      "grad_norm": 0.46478477120399475,
      "learning_rate": 0.01,
      "loss": 4.621218681335449,
      "num_input_tokens_seen": 3391384,
      "step": 854,
      "train_runtime": 684.0439,
      "train_tokens_per_second": 4957.845
    },
    {
      "epoch": 65.6923076923077,
      "num_input_tokens_seen": 3391384,
      "step": 854,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 65.76923076923077,
      "grad_norm": 0.31208330392837524,
      "learning_rate": 0.01,
      "loss": 4.556248664855957,
      "num_input_tokens_seen": 3395480,
      "step": 855,
      "train_runtime": 684.8618,
      "train_tokens_per_second": 4957.905
    },
    {
      "epoch": 65.76923076923077,
      "num_input_tokens_seen": 3395480,
      "step": 855,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 65.84615384615384,
      "grad_norm": 0.3860895037651062,
      "learning_rate": 0.01,
      "loss": 4.599390506744385,
      "num_input_tokens_seen": 3399576,
      "step": 856,
      "train_runtime": 685.6793,
      "train_tokens_per_second": 4957.968
    },
    {
      "epoch": 65.84615384615384,
      "num_input_tokens_seen": 3399576,
      "step": 856,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 65.92307692307692,
      "grad_norm": 0.37738537788391113,
      "learning_rate": 0.01,
      "loss": 4.5208048820495605,
      "num_input_tokens_seen": 3403672,
      "step": 857,
      "train_runtime": 686.4916,
      "train_tokens_per_second": 4958.068
    },
    {
      "epoch": 65.92307692307692,
      "num_input_tokens_seen": 3403672,
      "step": 857,
      "train_accuracy": 0.0032573288772255182
    },
    {
      "epoch": 66.0,
      "grad_norm": 0.8981891870498657,
      "learning_rate": 0.01,
      "loss": 4.5489935874938965,
      "num_input_tokens_seen": 3406128,
      "step": 858,
      "train_runtime": 687.0093,
      "train_tokens_per_second": 4957.907
    },
    {
      "epoch": 66.0,
      "num_input_tokens_seen": 3406128,
      "step": 858,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 66.07692307692308,
      "grad_norm": 0.6242077946662903,
      "learning_rate": 0.01,
      "loss": 4.539294242858887,
      "num_input_tokens_seen": 3410224,
      "step": 859,
      "train_runtime": 687.8387,
      "train_tokens_per_second": 4957.883
    },
    {
      "epoch": 66.07692307692308,
      "num_input_tokens_seen": 3410224,
      "step": 859,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 66.15384615384616,
      "grad_norm": 0.7122806310653687,
      "learning_rate": 0.01,
      "loss": 4.606278419494629,
      "num_input_tokens_seen": 3414320,
      "step": 860,
      "train_runtime": 688.6663,
      "train_tokens_per_second": 4957.873
    },
    {
      "epoch": 66.15384615384616,
      "num_input_tokens_seen": 3414320,
      "step": 860,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 66.23076923076923,
      "grad_norm": 0.708748459815979,
      "learning_rate": 0.01,
      "loss": 4.507891654968262,
      "num_input_tokens_seen": 3418416,
      "step": 861,
      "train_runtime": 689.4846,
      "train_tokens_per_second": 4957.929
    },
    {
      "epoch": 66.23076923076923,
      "num_input_tokens_seen": 3418416,
      "step": 861,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 66.3076923076923,
      "grad_norm": 0.45489710569381714,
      "learning_rate": 0.01,
      "loss": 4.558891296386719,
      "num_input_tokens_seen": 3422512,
      "step": 862,
      "train_runtime": 690.3039,
      "train_tokens_per_second": 4957.979
    },
    {
      "epoch": 66.3076923076923,
      "num_input_tokens_seen": 3422512,
      "step": 862,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 66.38461538461539,
      "grad_norm": 0.7404343485832214,
      "learning_rate": 0.01,
      "loss": 4.598685264587402,
      "num_input_tokens_seen": 3426608,
      "step": 863,
      "train_runtime": 691.121,
      "train_tokens_per_second": 4958.043
    },
    {
      "epoch": 66.38461538461539,
      "num_input_tokens_seen": 3426608,
      "step": 863,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 66.46153846153847,
      "grad_norm": 0.5132432579994202,
      "learning_rate": 0.01,
      "loss": 4.531277656555176,
      "num_input_tokens_seen": 3430704,
      "step": 864,
      "train_runtime": 691.9398,
      "train_tokens_per_second": 4958.096
    },
    {
      "epoch": 66.46153846153847,
      "num_input_tokens_seen": 3430704,
      "step": 864,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 66.53846153846153,
      "grad_norm": 0.45672670006752014,
      "learning_rate": 0.01,
      "loss": 4.500669479370117,
      "num_input_tokens_seen": 3434800,
      "step": 865,
      "train_runtime": 692.7587,
      "train_tokens_per_second": 4958.148
    },
    {
      "epoch": 66.53846153846153,
      "num_input_tokens_seen": 3434800,
      "step": 865,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 66.61538461538461,
      "grad_norm": 0.8451108336448669,
      "learning_rate": 0.01,
      "loss": 4.559305191040039,
      "num_input_tokens_seen": 3438896,
      "step": 866,
      "train_runtime": 693.5777,
      "train_tokens_per_second": 4958.199
    },
    {
      "epoch": 66.61538461538461,
      "num_input_tokens_seen": 3438896,
      "step": 866,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 66.6923076923077,
      "grad_norm": 0.37316787242889404,
      "learning_rate": 0.01,
      "loss": 4.522036552429199,
      "num_input_tokens_seen": 3442992,
      "step": 867,
      "train_runtime": 694.395,
      "train_tokens_per_second": 4958.261
    },
    {
      "epoch": 66.6923076923077,
      "num_input_tokens_seen": 3442992,
      "step": 867,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 66.76923076923077,
      "grad_norm": 0.46986570954322815,
      "learning_rate": 0.01,
      "loss": 4.549263954162598,
      "num_input_tokens_seen": 3447088,
      "step": 868,
      "train_runtime": 695.2145,
      "train_tokens_per_second": 4958.309
    },
    {
      "epoch": 66.76923076923077,
      "num_input_tokens_seen": 3447088,
      "step": 868,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 66.84615384615384,
      "grad_norm": 0.42503783106803894,
      "learning_rate": 0.01,
      "loss": 4.543720245361328,
      "num_input_tokens_seen": 3451184,
      "step": 869,
      "train_runtime": 696.0316,
      "train_tokens_per_second": 4958.372
    },
    {
      "epoch": 66.84615384615384,
      "num_input_tokens_seen": 3451184,
      "step": 869,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 66.92307692307692,
      "grad_norm": 0.7722694873809814,
      "learning_rate": 0.01,
      "loss": 4.536430358886719,
      "num_input_tokens_seen": 3455280,
      "step": 870,
      "train_runtime": 696.8446,
      "train_tokens_per_second": 4958.465
    },
    {
      "epoch": 66.92307692307692,
      "num_input_tokens_seen": 3455280,
      "step": 870,
      "train_accuracy": 0.0065146577544510365
    },
    {
      "epoch": 67.0,
      "grad_norm": 0.5014510750770569,
      "learning_rate": 0.01,
      "loss": 4.531091690063477,
      "num_input_tokens_seen": 3457736,
      "step": 871,
      "train_runtime": 697.3616,
      "train_tokens_per_second": 4958.311
    },
    {
      "epoch": 67.0,
      "num_input_tokens_seen": 3457736,
      "step": 871,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 67.07692307692308,
      "grad_norm": 0.32843247056007385,
      "learning_rate": 0.01,
      "loss": 4.4693732261657715,
      "num_input_tokens_seen": 3461832,
      "step": 872,
      "train_runtime": 698.1915,
      "train_tokens_per_second": 4958.284
    },
    {
      "epoch": 67.07692307692308,
      "num_input_tokens_seen": 3461832,
      "step": 872,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 67.15384615384616,
      "grad_norm": 0.5785554051399231,
      "learning_rate": 0.01,
      "loss": 4.559943199157715,
      "num_input_tokens_seen": 3465928,
      "step": 873,
      "train_runtime": 699.01,
      "train_tokens_per_second": 4958.338
    },
    {
      "epoch": 67.15384615384616,
      "num_input_tokens_seen": 3465928,
      "step": 873,
      "train_accuracy": 0.0390625
    },
    {
      "epoch": 67.23076923076923,
      "grad_norm": 0.8781976699829102,
      "learning_rate": 0.01,
      "loss": 4.5055108070373535,
      "num_input_tokens_seen": 3470024,
      "step": 874,
      "train_runtime": 699.827,
      "train_tokens_per_second": 4958.403
    },
    {
      "epoch": 67.23076923076923,
      "num_input_tokens_seen": 3470024,
      "step": 874,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 67.3076923076923,
      "grad_norm": 0.439237117767334,
      "learning_rate": 0.01,
      "loss": 4.49114465713501,
      "num_input_tokens_seen": 3474120,
      "step": 875,
      "train_runtime": 700.6457,
      "train_tokens_per_second": 4958.455
    },
    {
      "epoch": 67.3076923076923,
      "num_input_tokens_seen": 3474120,
      "step": 875,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 67.38461538461539,
      "grad_norm": 0.4294290840625763,
      "learning_rate": 0.01,
      "loss": 4.490567207336426,
      "num_input_tokens_seen": 3478216,
      "step": 876,
      "train_runtime": 701.4635,
      "train_tokens_per_second": 4958.513
    },
    {
      "epoch": 67.38461538461539,
      "num_input_tokens_seen": 3478216,
      "step": 876,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 67.46153846153847,
      "grad_norm": 0.6100071668624878,
      "learning_rate": 0.01,
      "loss": 4.5203142166137695,
      "num_input_tokens_seen": 3482312,
      "step": 877,
      "train_runtime": 702.2855,
      "train_tokens_per_second": 4958.542
    },
    {
      "epoch": 67.46153846153847,
      "num_input_tokens_seen": 3482312,
      "step": 877,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 67.53846153846153,
      "grad_norm": 0.7818669676780701,
      "learning_rate": 0.01,
      "loss": 4.531253814697266,
      "num_input_tokens_seen": 3486408,
      "step": 878,
      "train_runtime": 703.1032,
      "train_tokens_per_second": 4958.601
    },
    {
      "epoch": 67.53846153846153,
      "num_input_tokens_seen": 3486408,
      "step": 878,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 67.61538461538461,
      "grad_norm": 0.3752618730068207,
      "learning_rate": 0.01,
      "loss": 4.493743896484375,
      "num_input_tokens_seen": 3490504,
      "step": 879,
      "train_runtime": 703.9239,
      "train_tokens_per_second": 4958.638
    },
    {
      "epoch": 67.61538461538461,
      "num_input_tokens_seen": 3490504,
      "step": 879,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 67.6923076923077,
      "grad_norm": 0.5436190366744995,
      "learning_rate": 0.01,
      "loss": 4.5386552810668945,
      "num_input_tokens_seen": 3494600,
      "step": 880,
      "train_runtime": 704.7419,
      "train_tokens_per_second": 4958.695
    },
    {
      "epoch": 67.6923076923077,
      "num_input_tokens_seen": 3494600,
      "step": 880,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 67.76923076923077,
      "grad_norm": 0.5620259642601013,
      "learning_rate": 0.01,
      "loss": 4.5473809242248535,
      "num_input_tokens_seen": 3498696,
      "step": 881,
      "train_runtime": 705.5662,
      "train_tokens_per_second": 4958.707
    },
    {
      "epoch": 67.76923076923077,
      "num_input_tokens_seen": 3498696,
      "step": 881,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 67.84615384615384,
      "grad_norm": 0.3890678584575653,
      "learning_rate": 0.01,
      "loss": 4.502055644989014,
      "num_input_tokens_seen": 3502792,
      "step": 882,
      "train_runtime": 706.385,
      "train_tokens_per_second": 4958.757
    },
    {
      "epoch": 67.84615384615384,
      "num_input_tokens_seen": 3502792,
      "step": 882,
      "train_accuracy": 0.03125
    },
    {
      "epoch": 67.92307692307692,
      "grad_norm": 0.3027574121952057,
      "learning_rate": 0.01,
      "loss": 4.52012300491333,
      "num_input_tokens_seen": 3506888,
      "step": 883,
      "train_runtime": 707.1973,
      "train_tokens_per_second": 4958.853
    },
    {
      "epoch": 67.92307692307692,
      "num_input_tokens_seen": 3506888,
      "step": 883,
      "train_accuracy": 0.0065146577544510365
    },
    {
      "epoch": 68.0,
      "grad_norm": 0.3700380325317383,
      "learning_rate": 0.01,
      "loss": 4.579263210296631,
      "num_input_tokens_seen": 3509344,
      "step": 884,
      "train_runtime": 707.7159,
      "train_tokens_per_second": 4958.69
    },
    {
      "epoch": 68.0,
      "num_input_tokens_seen": 3509344,
      "step": 884,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 68.07692307692308,
      "grad_norm": 0.48986291885375977,
      "learning_rate": 0.01,
      "loss": 4.526958465576172,
      "num_input_tokens_seen": 3513440,
      "step": 885,
      "train_runtime": 708.5483,
      "train_tokens_per_second": 4958.646
    },
    {
      "epoch": 68.07692307692308,
      "num_input_tokens_seen": 3513440,
      "step": 885,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 68.15384615384616,
      "grad_norm": 0.3456747233867645,
      "learning_rate": 0.01,
      "loss": 4.496724605560303,
      "num_input_tokens_seen": 3517536,
      "step": 886,
      "train_runtime": 709.3656,
      "train_tokens_per_second": 4958.707
    },
    {
      "epoch": 68.15384615384616,
      "num_input_tokens_seen": 3517536,
      "step": 886,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 68.23076923076923,
      "grad_norm": 0.8218653798103333,
      "learning_rate": 0.01,
      "loss": 4.564673900604248,
      "num_input_tokens_seen": 3521632,
      "step": 887,
      "train_runtime": 710.1841,
      "train_tokens_per_second": 4958.76
    },
    {
      "epoch": 68.23076923076923,
      "num_input_tokens_seen": 3521632,
      "step": 887,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 68.3076923076923,
      "grad_norm": 0.40607571601867676,
      "learning_rate": 0.01,
      "loss": 4.539470672607422,
      "num_input_tokens_seen": 3525728,
      "step": 888,
      "train_runtime": 711.0016,
      "train_tokens_per_second": 4958.819
    },
    {
      "epoch": 68.3076923076923,
      "num_input_tokens_seen": 3525728,
      "step": 888,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 68.38461538461539,
      "grad_norm": 1.5847623348236084,
      "learning_rate": 0.01,
      "loss": 4.538667678833008,
      "num_input_tokens_seen": 3529824,
      "step": 889,
      "train_runtime": 711.8206,
      "train_tokens_per_second": 4958.867
    },
    {
      "epoch": 68.38461538461539,
      "num_input_tokens_seen": 3529824,
      "step": 889,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 68.46153846153847,
      "grad_norm": 0.5893070697784424,
      "learning_rate": 0.01,
      "loss": 4.502369403839111,
      "num_input_tokens_seen": 3533920,
      "step": 890,
      "train_runtime": 712.639,
      "train_tokens_per_second": 4958.92
    },
    {
      "epoch": 68.46153846153847,
      "num_input_tokens_seen": 3533920,
      "step": 890,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 68.53846153846153,
      "grad_norm": 0.6661847233772278,
      "learning_rate": 0.01,
      "loss": 4.539044380187988,
      "num_input_tokens_seen": 3538016,
      "step": 891,
      "train_runtime": 713.4595,
      "train_tokens_per_second": 4958.958
    },
    {
      "epoch": 68.53846153846153,
      "num_input_tokens_seen": 3538016,
      "step": 891,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 68.61538461538461,
      "grad_norm": 0.5200782418251038,
      "learning_rate": 0.01,
      "loss": 4.511730194091797,
      "num_input_tokens_seen": 3542112,
      "step": 892,
      "train_runtime": 714.2787,
      "train_tokens_per_second": 4959.006
    },
    {
      "epoch": 68.61538461538461,
      "num_input_tokens_seen": 3542112,
      "step": 892,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 68.6923076923077,
      "grad_norm": 0.5108278393745422,
      "learning_rate": 0.01,
      "loss": 4.490791320800781,
      "num_input_tokens_seen": 3546208,
      "step": 893,
      "train_runtime": 715.0995,
      "train_tokens_per_second": 4959.041
    },
    {
      "epoch": 68.6923076923077,
      "num_input_tokens_seen": 3546208,
      "step": 893,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 68.76923076923077,
      "grad_norm": 0.7380408644676208,
      "learning_rate": 0.01,
      "loss": 4.563013553619385,
      "num_input_tokens_seen": 3550304,
      "step": 894,
      "train_runtime": 715.9188,
      "train_tokens_per_second": 4959.088
    },
    {
      "epoch": 68.76923076923077,
      "num_input_tokens_seen": 3550304,
      "step": 894,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 68.84615384615384,
      "grad_norm": 0.33585381507873535,
      "learning_rate": 0.01,
      "loss": 4.481448173522949,
      "num_input_tokens_seen": 3554400,
      "step": 895,
      "train_runtime": 716.7379,
      "train_tokens_per_second": 4959.135
    },
    {
      "epoch": 68.84615384615384,
      "num_input_tokens_seen": 3554400,
      "step": 895,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 68.92307692307692,
      "grad_norm": 0.43700939416885376,
      "learning_rate": 0.01,
      "loss": 4.583843231201172,
      "num_input_tokens_seen": 3558496,
      "step": 896,
      "train_runtime": 717.5513,
      "train_tokens_per_second": 4959.221
    },
    {
      "epoch": 68.92307692307692,
      "num_input_tokens_seen": 3558496,
      "step": 896,
      "train_accuracy": 0.019543973729014397
    },
    {
      "epoch": 69.0,
      "grad_norm": 0.6429909467697144,
      "learning_rate": 0.01,
      "loss": 4.469284534454346,
      "num_input_tokens_seen": 3560952,
      "step": 897,
      "train_runtime": 718.0693,
      "train_tokens_per_second": 4959.064
    },
    {
      "epoch": 69.0,
      "num_input_tokens_seen": 3560952,
      "step": 897,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 69.07692307692308,
      "grad_norm": 0.44672366976737976,
      "learning_rate": 0.01,
      "loss": 4.423923969268799,
      "num_input_tokens_seen": 3565048,
      "step": 898,
      "train_runtime": 718.8988,
      "train_tokens_per_second": 4959.04
    },
    {
      "epoch": 69.07692307692308,
      "num_input_tokens_seen": 3565048,
      "step": 898,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 69.15384615384616,
      "grad_norm": 0.4783729612827301,
      "learning_rate": 0.01,
      "loss": 4.454030990600586,
      "num_input_tokens_seen": 3569144,
      "step": 899,
      "train_runtime": 719.7174,
      "train_tokens_per_second": 4959.091
    },
    {
      "epoch": 69.15384615384616,
      "num_input_tokens_seen": 3569144,
      "step": 899,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 69.23076923076923,
      "grad_norm": 0.40234968066215515,
      "learning_rate": 0.01,
      "loss": 4.422664642333984,
      "num_input_tokens_seen": 3573240,
      "step": 900,
      "train_runtime": 720.536,
      "train_tokens_per_second": 4959.142
    },
    {
      "epoch": 69.23076923076923,
      "eval_CMD_3_branch_eval_accuracy": 0.010126582278481013,
      "eval_CMD_3_branch_eval_loss": 4.629105091094971,
      "eval_CMD_3_branch_eval_runtime": 1.1629,
      "eval_CMD_3_branch_eval_samples_per_second": 2377.736,
      "eval_CMD_3_branch_eval_steps_per_second": 5.16,
      "num_input_tokens_seen": 3573240,
      "step": 900
    },
    {
      "epoch": 69.23076923076923,
      "num_input_tokens_seen": 3573240,
      "step": 900,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 69.3076923076923,
      "grad_norm": 0.45272207260131836,
      "learning_rate": 0.01,
      "loss": 4.406930446624756,
      "num_input_tokens_seen": 3577336,
      "step": 901,
      "train_runtime": 722.5235,
      "train_tokens_per_second": 4951.169
    },
    {
      "epoch": 69.3076923076923,
      "num_input_tokens_seen": 3577336,
      "step": 901,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 69.38461538461539,
      "grad_norm": 0.4514341652393341,
      "learning_rate": 0.01,
      "loss": 4.4331464767456055,
      "num_input_tokens_seen": 3581432,
      "step": 902,
      "train_runtime": 723.3415,
      "train_tokens_per_second": 4951.233
    },
    {
      "epoch": 69.38461538461539,
      "num_input_tokens_seen": 3581432,
      "step": 902,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 69.46153846153847,
      "grad_norm": 0.3631833493709564,
      "learning_rate": 0.01,
      "loss": 4.447068691253662,
      "num_input_tokens_seen": 3585528,
      "step": 903,
      "train_runtime": 724.1645,
      "train_tokens_per_second": 4951.262
    },
    {
      "epoch": 69.46153846153847,
      "num_input_tokens_seen": 3585528,
      "step": 903,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 69.53846153846153,
      "grad_norm": 0.2641160786151886,
      "learning_rate": 0.01,
      "loss": 4.469538688659668,
      "num_input_tokens_seen": 3589624,
      "step": 904,
      "train_runtime": 724.9842,
      "train_tokens_per_second": 4951.313
    },
    {
      "epoch": 69.53846153846153,
      "num_input_tokens_seen": 3589624,
      "step": 904,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 69.61538461538461,
      "grad_norm": 1.2099906206130981,
      "learning_rate": 0.01,
      "loss": 4.496251106262207,
      "num_input_tokens_seen": 3593720,
      "step": 905,
      "train_runtime": 725.8041,
      "train_tokens_per_second": 4951.364
    },
    {
      "epoch": 69.61538461538461,
      "num_input_tokens_seen": 3593720,
      "step": 905,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 69.6923076923077,
      "grad_norm": 0.33517658710479736,
      "learning_rate": 0.01,
      "loss": 4.484493732452393,
      "num_input_tokens_seen": 3597816,
      "step": 906,
      "train_runtime": 726.6217,
      "train_tokens_per_second": 4951.43
    },
    {
      "epoch": 69.6923076923077,
      "num_input_tokens_seen": 3597816,
      "step": 906,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 69.76923076923077,
      "grad_norm": 0.3322330415248871,
      "learning_rate": 0.01,
      "loss": 4.471845626831055,
      "num_input_tokens_seen": 3601912,
      "step": 907,
      "train_runtime": 727.4411,
      "train_tokens_per_second": 4951.483
    },
    {
      "epoch": 69.76923076923077,
      "num_input_tokens_seen": 3601912,
      "step": 907,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 69.84615384615384,
      "grad_norm": 0.29685264825820923,
      "learning_rate": 0.01,
      "loss": 4.48011589050293,
      "num_input_tokens_seen": 3606008,
      "step": 908,
      "train_runtime": 728.2599,
      "train_tokens_per_second": 4951.54
    },
    {
      "epoch": 69.84615384615384,
      "num_input_tokens_seen": 3606008,
      "step": 908,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 69.92307692307692,
      "grad_norm": 0.5841346979141235,
      "learning_rate": 0.01,
      "loss": 4.4899749755859375,
      "num_input_tokens_seen": 3610104,
      "step": 909,
      "train_runtime": 729.0733,
      "train_tokens_per_second": 4951.634
    },
    {
      "epoch": 69.92307692307692,
      "num_input_tokens_seen": 3610104,
      "step": 909,
      "train_accuracy": 0.02280130237340927
    },
    {
      "epoch": 70.0,
      "grad_norm": 0.9717252254486084,
      "learning_rate": 0.01,
      "loss": 4.529669284820557,
      "num_input_tokens_seen": 3612560,
      "step": 910,
      "train_runtime": 729.5916,
      "train_tokens_per_second": 4951.483
    },
    {
      "epoch": 70.0,
      "num_input_tokens_seen": 3612560,
      "step": 910,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 70.07692307692308,
      "grad_norm": 0.361969918012619,
      "learning_rate": 0.01,
      "loss": 4.41936731338501,
      "num_input_tokens_seen": 3616656,
      "step": 911,
      "train_runtime": 730.4236,
      "train_tokens_per_second": 4951.45
    },
    {
      "epoch": 70.07692307692308,
      "num_input_tokens_seen": 3616656,
      "step": 911,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 70.15384615384616,
      "grad_norm": 0.9180896878242493,
      "learning_rate": 0.01,
      "loss": 4.470214366912842,
      "num_input_tokens_seen": 3620752,
      "step": 912,
      "train_runtime": 731.2415,
      "train_tokens_per_second": 4951.513
    },
    {
      "epoch": 70.15384615384616,
      "num_input_tokens_seen": 3620752,
      "step": 912,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 70.23076923076923,
      "grad_norm": 0.5340433716773987,
      "learning_rate": 0.01,
      "loss": 4.378235816955566,
      "num_input_tokens_seen": 3624848,
      "step": 913,
      "train_runtime": 732.0595,
      "train_tokens_per_second": 4951.575
    },
    {
      "epoch": 70.23076923076923,
      "num_input_tokens_seen": 3624848,
      "step": 913,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 70.3076923076923,
      "grad_norm": 1.6778451204299927,
      "learning_rate": 0.01,
      "loss": 4.447396755218506,
      "num_input_tokens_seen": 3628944,
      "step": 914,
      "train_runtime": 732.8801,
      "train_tokens_per_second": 4951.62
    },
    {
      "epoch": 70.3076923076923,
      "num_input_tokens_seen": 3628944,
      "step": 914,
      "train_accuracy": 0.03125
    },
    {
      "epoch": 70.38461538461539,
      "grad_norm": 0.616068422794342,
      "learning_rate": 0.01,
      "loss": 4.4343061447143555,
      "num_input_tokens_seen": 3633040,
      "step": 915,
      "train_runtime": 733.7009,
      "train_tokens_per_second": 4951.664
    },
    {
      "epoch": 70.38461538461539,
      "num_input_tokens_seen": 3633040,
      "step": 915,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 70.46153846153847,
      "grad_norm": 0.3702734410762787,
      "learning_rate": 0.01,
      "loss": 4.424246788024902,
      "num_input_tokens_seen": 3637136,
      "step": 916,
      "train_runtime": 734.5218,
      "train_tokens_per_second": 4951.706
    },
    {
      "epoch": 70.46153846153847,
      "num_input_tokens_seen": 3637136,
      "step": 916,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 70.53846153846153,
      "grad_norm": 0.5244331955909729,
      "learning_rate": 0.01,
      "loss": 4.418830871582031,
      "num_input_tokens_seen": 3641232,
      "step": 917,
      "train_runtime": 735.3397,
      "train_tokens_per_second": 4951.768
    },
    {
      "epoch": 70.53846153846153,
      "num_input_tokens_seen": 3641232,
      "step": 917,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 70.61538461538461,
      "grad_norm": 0.5785495042800903,
      "learning_rate": 0.01,
      "loss": 4.386415958404541,
      "num_input_tokens_seen": 3645328,
      "step": 918,
      "train_runtime": 736.1584,
      "train_tokens_per_second": 4951.826
    },
    {
      "epoch": 70.61538461538461,
      "num_input_tokens_seen": 3645328,
      "step": 918,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 70.6923076923077,
      "grad_norm": 0.44390246272087097,
      "learning_rate": 0.01,
      "loss": 4.42064094543457,
      "num_input_tokens_seen": 3649424,
      "step": 919,
      "train_runtime": 736.9773,
      "train_tokens_per_second": 4951.881
    },
    {
      "epoch": 70.6923076923077,
      "num_input_tokens_seen": 3649424,
      "step": 919,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 70.76923076923077,
      "grad_norm": 0.48718634247779846,
      "learning_rate": 0.01,
      "loss": 4.4073896408081055,
      "num_input_tokens_seen": 3653520,
      "step": 920,
      "train_runtime": 737.7959,
      "train_tokens_per_second": 4951.939
    },
    {
      "epoch": 70.76923076923077,
      "num_input_tokens_seen": 3653520,
      "step": 920,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 70.84615384615384,
      "grad_norm": 0.39334720373153687,
      "learning_rate": 0.01,
      "loss": 4.4013895988464355,
      "num_input_tokens_seen": 3657616,
      "step": 921,
      "train_runtime": 738.6148,
      "train_tokens_per_second": 4951.994
    },
    {
      "epoch": 70.84615384615384,
      "num_input_tokens_seen": 3657616,
      "step": 921,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 70.92307692307692,
      "grad_norm": 0.4572714865207672,
      "learning_rate": 0.01,
      "loss": 4.39383602142334,
      "num_input_tokens_seen": 3661712,
      "step": 922,
      "train_runtime": 739.4301,
      "train_tokens_per_second": 4952.073
    },
    {
      "epoch": 70.92307692307692,
      "num_input_tokens_seen": 3661712,
      "step": 922,
      "train_accuracy": 0.02280130237340927
    },
    {
      "epoch": 71.0,
      "grad_norm": 0.5256152749061584,
      "learning_rate": 0.01,
      "loss": 4.438015460968018,
      "num_input_tokens_seen": 3664168,
      "step": 923,
      "train_runtime": 739.948,
      "train_tokens_per_second": 4951.926
    },
    {
      "epoch": 71.0,
      "num_input_tokens_seen": 3664168,
      "step": 923,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 71.07692307692308,
      "grad_norm": 0.7388529181480408,
      "learning_rate": 0.01,
      "loss": 4.342407703399658,
      "num_input_tokens_seen": 3668264,
      "step": 924,
      "train_runtime": 740.7783,
      "train_tokens_per_second": 4951.905
    },
    {
      "epoch": 71.07692307692308,
      "num_input_tokens_seen": 3668264,
      "step": 924,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 71.15384615384616,
      "grad_norm": 0.7544814944267273,
      "learning_rate": 0.01,
      "loss": 4.366703510284424,
      "num_input_tokens_seen": 3672360,
      "step": 925,
      "train_runtime": 741.5977,
      "train_tokens_per_second": 4951.957
    },
    {
      "epoch": 71.15384615384616,
      "num_input_tokens_seen": 3672360,
      "step": 925,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 71.23076923076923,
      "grad_norm": 0.47264015674591064,
      "learning_rate": 0.01,
      "loss": 4.39190149307251,
      "num_input_tokens_seen": 3676456,
      "step": 926,
      "train_runtime": 742.4173,
      "train_tokens_per_second": 4952.007
    },
    {
      "epoch": 71.23076923076923,
      "num_input_tokens_seen": 3676456,
      "step": 926,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 71.3076923076923,
      "grad_norm": 0.7658267021179199,
      "learning_rate": 0.01,
      "loss": 4.420034408569336,
      "num_input_tokens_seen": 3680552,
      "step": 927,
      "train_runtime": 743.2371,
      "train_tokens_per_second": 4952.056
    },
    {
      "epoch": 71.3076923076923,
      "num_input_tokens_seen": 3680552,
      "step": 927,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 71.38461538461539,
      "grad_norm": 0.44571179151535034,
      "learning_rate": 0.01,
      "loss": 4.437005519866943,
      "num_input_tokens_seen": 3684648,
      "step": 928,
      "train_runtime": 744.063,
      "train_tokens_per_second": 4952.064
    },
    {
      "epoch": 71.38461538461539,
      "num_input_tokens_seen": 3684648,
      "step": 928,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 71.46153846153847,
      "grad_norm": 1.2002772092819214,
      "learning_rate": 0.01,
      "loss": 4.350293159484863,
      "num_input_tokens_seen": 3688744,
      "step": 929,
      "train_runtime": 744.8838,
      "train_tokens_per_second": 4952.106
    },
    {
      "epoch": 71.46153846153847,
      "num_input_tokens_seen": 3688744,
      "step": 929,
      "train_accuracy": 0.03515625
    },
    {
      "epoch": 71.53846153846153,
      "grad_norm": 0.8120535612106323,
      "learning_rate": 0.01,
      "loss": 4.378268241882324,
      "num_input_tokens_seen": 3692840,
      "step": 930,
      "train_runtime": 745.7028,
      "train_tokens_per_second": 4952.16
    },
    {
      "epoch": 71.53846153846153,
      "num_input_tokens_seen": 3692840,
      "step": 930,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 71.61538461538461,
      "grad_norm": 0.8198422193527222,
      "learning_rate": 0.01,
      "loss": 4.386263370513916,
      "num_input_tokens_seen": 3696936,
      "step": 931,
      "train_runtime": 746.5241,
      "train_tokens_per_second": 4952.199
    },
    {
      "epoch": 71.61538461538461,
      "num_input_tokens_seen": 3696936,
      "step": 931,
      "train_accuracy": 0.03515625
    },
    {
      "epoch": 71.6923076923077,
      "grad_norm": 0.4060828685760498,
      "learning_rate": 0.01,
      "loss": 4.37192964553833,
      "num_input_tokens_seen": 3701032,
      "step": 932,
      "train_runtime": 747.3444,
      "train_tokens_per_second": 4952.244
    },
    {
      "epoch": 71.6923076923077,
      "num_input_tokens_seen": 3701032,
      "step": 932,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 71.76923076923077,
      "grad_norm": 0.7270469665527344,
      "learning_rate": 0.01,
      "loss": 4.358577728271484,
      "num_input_tokens_seen": 3705128,
      "step": 933,
      "train_runtime": 748.1642,
      "train_tokens_per_second": 4952.293
    },
    {
      "epoch": 71.76923076923077,
      "num_input_tokens_seen": 3705128,
      "step": 933,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 71.84615384615384,
      "grad_norm": 0.4120619595050812,
      "learning_rate": 0.01,
      "loss": 4.354004383087158,
      "num_input_tokens_seen": 3709224,
      "step": 934,
      "train_runtime": 748.9849,
      "train_tokens_per_second": 4952.335
    },
    {
      "epoch": 71.84615384615384,
      "num_input_tokens_seen": 3709224,
      "step": 934,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 71.92307692307692,
      "grad_norm": 0.6477432250976562,
      "learning_rate": 0.01,
      "loss": 4.411651611328125,
      "num_input_tokens_seen": 3713320,
      "step": 935,
      "train_runtime": 749.7987,
      "train_tokens_per_second": 4952.423
    },
    {
      "epoch": 71.92307692307692,
      "num_input_tokens_seen": 3713320,
      "step": 935,
      "train_accuracy": 0.032573290169239044
    },
    {
      "epoch": 72.0,
      "grad_norm": 1.7840039730072021,
      "learning_rate": 0.01,
      "loss": 4.346638202667236,
      "num_input_tokens_seen": 3715776,
      "step": 936,
      "train_runtime": 750.3172,
      "train_tokens_per_second": 4952.274
    },
    {
      "epoch": 72.0,
      "num_input_tokens_seen": 3715776,
      "step": 936,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 72.07692307692308,
      "grad_norm": 1.114080786705017,
      "learning_rate": 0.01,
      "loss": 4.222871780395508,
      "num_input_tokens_seen": 3719872,
      "step": 937,
      "train_runtime": 751.1472,
      "train_tokens_per_second": 4952.254
    },
    {
      "epoch": 72.07692307692308,
      "num_input_tokens_seen": 3719872,
      "step": 937,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 72.15384615384616,
      "grad_norm": 0.4295608103275299,
      "learning_rate": 0.01,
      "loss": 4.352922439575195,
      "num_input_tokens_seen": 3723968,
      "step": 938,
      "train_runtime": 751.97,
      "train_tokens_per_second": 4952.283
    },
    {
      "epoch": 72.15384615384616,
      "num_input_tokens_seen": 3723968,
      "step": 938,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 72.23076923076923,
      "grad_norm": 0.5601105690002441,
      "learning_rate": 0.01,
      "loss": 4.34294319152832,
      "num_input_tokens_seen": 3728064,
      "step": 939,
      "train_runtime": 752.789,
      "train_tokens_per_second": 4952.336
    },
    {
      "epoch": 72.23076923076923,
      "num_input_tokens_seen": 3728064,
      "step": 939,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 72.3076923076923,
      "grad_norm": 2.131270170211792,
      "learning_rate": 0.01,
      "loss": 4.369997978210449,
      "num_input_tokens_seen": 3732160,
      "step": 940,
      "train_runtime": 753.608,
      "train_tokens_per_second": 4952.389
    },
    {
      "epoch": 72.3076923076923,
      "num_input_tokens_seen": 3732160,
      "step": 940,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 72.38461538461539,
      "grad_norm": 0.6599200367927551,
      "learning_rate": 0.01,
      "loss": 4.3323774337768555,
      "num_input_tokens_seen": 3736256,
      "step": 941,
      "train_runtime": 754.4262,
      "train_tokens_per_second": 4952.447
    },
    {
      "epoch": 72.38461538461539,
      "num_input_tokens_seen": 3736256,
      "step": 941,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 72.46153846153847,
      "grad_norm": 0.6697801947593689,
      "learning_rate": 0.01,
      "loss": 4.381878852844238,
      "num_input_tokens_seen": 3740352,
      "step": 942,
      "train_runtime": 755.2452,
      "train_tokens_per_second": 4952.5
    },
    {
      "epoch": 72.46153846153847,
      "num_input_tokens_seen": 3740352,
      "step": 942,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 72.53846153846153,
      "grad_norm": 2.5567116737365723,
      "learning_rate": 0.01,
      "loss": 4.386260986328125,
      "num_input_tokens_seen": 3744448,
      "step": 943,
      "train_runtime": 756.0642,
      "train_tokens_per_second": 4952.553
    },
    {
      "epoch": 72.53846153846153,
      "num_input_tokens_seen": 3744448,
      "step": 943,
      "train_accuracy": 0.03515625
    },
    {
      "epoch": 72.61538461538461,
      "grad_norm": 0.4510970413684845,
      "learning_rate": 0.01,
      "loss": 4.402202606201172,
      "num_input_tokens_seen": 3748544,
      "step": 944,
      "train_runtime": 756.8828,
      "train_tokens_per_second": 4952.608
    },
    {
      "epoch": 72.61538461538461,
      "num_input_tokens_seen": 3748544,
      "step": 944,
      "train_accuracy": 0.03125
    },
    {
      "epoch": 72.6923076923077,
      "grad_norm": 0.47819021344184875,
      "learning_rate": 0.01,
      "loss": 4.406796455383301,
      "num_input_tokens_seen": 3752640,
      "step": 945,
      "train_runtime": 757.7012,
      "train_tokens_per_second": 4952.665
    },
    {
      "epoch": 72.6923076923077,
      "num_input_tokens_seen": 3752640,
      "step": 945,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 72.76923076923077,
      "grad_norm": 0.8835750818252563,
      "learning_rate": 0.01,
      "loss": 4.430482387542725,
      "num_input_tokens_seen": 3756736,
      "step": 946,
      "train_runtime": 758.521,
      "train_tokens_per_second": 4952.712
    },
    {
      "epoch": 72.76923076923077,
      "num_input_tokens_seen": 3756736,
      "step": 946,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 72.84615384615384,
      "grad_norm": 1.2383958101272583,
      "learning_rate": 0.01,
      "loss": 4.416315078735352,
      "num_input_tokens_seen": 3760832,
      "step": 947,
      "train_runtime": 759.3402,
      "train_tokens_per_second": 4952.763
    },
    {
      "epoch": 72.84615384615384,
      "num_input_tokens_seen": 3760832,
      "step": 947,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 72.92307692307692,
      "grad_norm": 1.6494998931884766,
      "learning_rate": 0.01,
      "loss": 4.406277656555176,
      "num_input_tokens_seen": 3764928,
      "step": 948,
      "train_runtime": 760.1552,
      "train_tokens_per_second": 4952.841
    },
    {
      "epoch": 72.92307692307692,
      "num_input_tokens_seen": 3764928,
      "step": 948,
      "train_accuracy": 0.02280130237340927
    },
    {
      "epoch": 73.0,
      "grad_norm": 0.9242535829544067,
      "learning_rate": 0.01,
      "loss": 4.502875804901123,
      "num_input_tokens_seen": 3767384,
      "step": 949,
      "train_runtime": 760.6736,
      "train_tokens_per_second": 4952.695
    },
    {
      "epoch": 73.0,
      "num_input_tokens_seen": 3767384,
      "step": 949,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 73.07692307692308,
      "grad_norm": 0.8408144116401672,
      "learning_rate": 0.01,
      "loss": 4.431186676025391,
      "num_input_tokens_seen": 3771480,
      "step": 950,
      "train_runtime": 761.5042,
      "train_tokens_per_second": 4952.671
    },
    {
      "epoch": 73.07692307692308,
      "num_input_tokens_seen": 3771480,
      "step": 950,
      "train_accuracy": 0.041015625
    },
    {
      "epoch": 73.15384615384616,
      "grad_norm": 0.6841522455215454,
      "learning_rate": 0.01,
      "loss": 4.40715217590332,
      "num_input_tokens_seen": 3775576,
      "step": 951,
      "train_runtime": 762.3229,
      "train_tokens_per_second": 4952.725
    },
    {
      "epoch": 73.15384615384616,
      "num_input_tokens_seen": 3775576,
      "step": 951,
      "train_accuracy": 0.037109375
    },
    {
      "epoch": 73.23076923076923,
      "grad_norm": 0.39841896295547485,
      "learning_rate": 0.01,
      "loss": 4.319639682769775,
      "num_input_tokens_seen": 3779672,
      "step": 952,
      "train_runtime": 763.6838,
      "train_tokens_per_second": 4949.263
    },
    {
      "epoch": 73.23076923076923,
      "num_input_tokens_seen": 3779672,
      "step": 952,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 73.3076923076923,
      "grad_norm": 0.6825417876243591,
      "learning_rate": 0.01,
      "loss": 4.34348201751709,
      "num_input_tokens_seen": 3783768,
      "step": 953,
      "train_runtime": 764.5022,
      "train_tokens_per_second": 4949.323
    },
    {
      "epoch": 73.3076923076923,
      "num_input_tokens_seen": 3783768,
      "step": 953,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 73.38461538461539,
      "grad_norm": 0.6725438237190247,
      "learning_rate": 0.01,
      "loss": 4.354001998901367,
      "num_input_tokens_seen": 3787864,
      "step": 954,
      "train_runtime": 765.32,
      "train_tokens_per_second": 4949.386
    },
    {
      "epoch": 73.38461538461539,
      "num_input_tokens_seen": 3787864,
      "step": 954,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 73.46153846153847,
      "grad_norm": 1.6613304615020752,
      "learning_rate": 0.01,
      "loss": 4.371725559234619,
      "num_input_tokens_seen": 3791960,
      "step": 955,
      "train_runtime": 766.1378,
      "train_tokens_per_second": 4949.449
    },
    {
      "epoch": 73.46153846153847,
      "num_input_tokens_seen": 3791960,
      "step": 955,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 73.53846153846153,
      "grad_norm": 1.223771095275879,
      "learning_rate": 0.01,
      "loss": 4.372991561889648,
      "num_input_tokens_seen": 3796056,
      "step": 956,
      "train_runtime": 766.9556,
      "train_tokens_per_second": 4949.512
    },
    {
      "epoch": 73.53846153846153,
      "num_input_tokens_seen": 3796056,
      "step": 956,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 73.61538461538461,
      "grad_norm": 0.7996600866317749,
      "learning_rate": 0.01,
      "loss": 4.342113494873047,
      "num_input_tokens_seen": 3800152,
      "step": 957,
      "train_runtime": 767.7738,
      "train_tokens_per_second": 4949.572
    },
    {
      "epoch": 73.61538461538461,
      "num_input_tokens_seen": 3800152,
      "step": 957,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 73.6923076923077,
      "grad_norm": 1.256047010421753,
      "learning_rate": 0.01,
      "loss": 4.377371788024902,
      "num_input_tokens_seen": 3804248,
      "step": 958,
      "train_runtime": 768.5925,
      "train_tokens_per_second": 4949.629
    },
    {
      "epoch": 73.6923076923077,
      "num_input_tokens_seen": 3804248,
      "step": 958,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 73.76923076923077,
      "grad_norm": 0.7696335911750793,
      "learning_rate": 0.01,
      "loss": 4.350641250610352,
      "num_input_tokens_seen": 3808344,
      "step": 959,
      "train_runtime": 769.4112,
      "train_tokens_per_second": 4949.686
    },
    {
      "epoch": 73.76923076923077,
      "num_input_tokens_seen": 3808344,
      "step": 959,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 73.84615384615384,
      "grad_norm": 1.2821881771087646,
      "learning_rate": 0.01,
      "loss": 4.333581924438477,
      "num_input_tokens_seen": 3812440,
      "step": 960,
      "train_runtime": 770.2291,
      "train_tokens_per_second": 4949.748
    },
    {
      "epoch": 73.84615384615384,
      "num_input_tokens_seen": 3812440,
      "step": 960,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 73.92307692307692,
      "grad_norm": 0.6806766986846924,
      "learning_rate": 0.01,
      "loss": 4.393897533416748,
      "num_input_tokens_seen": 3816536,
      "step": 961,
      "train_runtime": 771.0434,
      "train_tokens_per_second": 4949.833
    },
    {
      "epoch": 73.92307692307692,
      "num_input_tokens_seen": 3816536,
      "step": 961,
      "train_accuracy": 0.019543973729014397
    },
    {
      "epoch": 74.0,
      "grad_norm": 0.998210608959198,
      "learning_rate": 0.01,
      "loss": 4.358442783355713,
      "num_input_tokens_seen": 3818992,
      "step": 962,
      "train_runtime": 771.5609,
      "train_tokens_per_second": 4949.696
    },
    {
      "epoch": 74.0,
      "num_input_tokens_seen": 3818992,
      "step": 962,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 74.07692307692308,
      "grad_norm": 0.5987136960029602,
      "learning_rate": 0.01,
      "loss": 4.374314308166504,
      "num_input_tokens_seen": 3823088,
      "step": 963,
      "train_runtime": 772.3907,
      "train_tokens_per_second": 4949.681
    },
    {
      "epoch": 74.07692307692308,
      "num_input_tokens_seen": 3823088,
      "step": 963,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 74.15384615384616,
      "grad_norm": 0.8599548935890198,
      "learning_rate": 0.01,
      "loss": 4.387784957885742,
      "num_input_tokens_seen": 3827184,
      "step": 964,
      "train_runtime": 773.2094,
      "train_tokens_per_second": 4949.738
    },
    {
      "epoch": 74.15384615384616,
      "num_input_tokens_seen": 3827184,
      "step": 964,
      "train_accuracy": 0.03125
    },
    {
      "epoch": 74.23076923076923,
      "grad_norm": 2.9651641845703125,
      "learning_rate": 0.01,
      "loss": 4.416442394256592,
      "num_input_tokens_seen": 3831280,
      "step": 965,
      "train_runtime": 774.0277,
      "train_tokens_per_second": 4949.797
    },
    {
      "epoch": 74.23076923076923,
      "num_input_tokens_seen": 3831280,
      "step": 965,
      "train_accuracy": 0.037109375
    },
    {
      "epoch": 74.3076923076923,
      "grad_norm": 0.8523027896881104,
      "learning_rate": 0.01,
      "loss": 4.401150703430176,
      "num_input_tokens_seen": 3835376,
      "step": 966,
      "train_runtime": 774.8466,
      "train_tokens_per_second": 4949.852
    },
    {
      "epoch": 74.3076923076923,
      "num_input_tokens_seen": 3835376,
      "step": 966,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 74.38461538461539,
      "grad_norm": 0.40367913246154785,
      "learning_rate": 0.01,
      "loss": 4.303525924682617,
      "num_input_tokens_seen": 3839472,
      "step": 967,
      "train_runtime": 775.6652,
      "train_tokens_per_second": 4949.909
    },
    {
      "epoch": 74.38461538461539,
      "num_input_tokens_seen": 3839472,
      "step": 967,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 74.46153846153847,
      "grad_norm": 1.1555758714675903,
      "learning_rate": 0.01,
      "loss": 4.327036380767822,
      "num_input_tokens_seen": 3843568,
      "step": 968,
      "train_runtime": 776.4845,
      "train_tokens_per_second": 4949.961
    },
    {
      "epoch": 74.46153846153847,
      "num_input_tokens_seen": 3843568,
      "step": 968,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 74.53846153846153,
      "grad_norm": 0.40165475010871887,
      "learning_rate": 0.01,
      "loss": 4.322185516357422,
      "num_input_tokens_seen": 3847664,
      "step": 969,
      "train_runtime": 777.3041,
      "train_tokens_per_second": 4950.011
    },
    {
      "epoch": 74.53846153846153,
      "num_input_tokens_seen": 3847664,
      "step": 969,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 74.61538461538461,
      "grad_norm": 14.481494903564453,
      "learning_rate": 0.01,
      "loss": 4.397060394287109,
      "num_input_tokens_seen": 3851760,
      "step": 970,
      "train_runtime": 778.1226,
      "train_tokens_per_second": 4950.068
    },
    {
      "epoch": 74.61538461538461,
      "num_input_tokens_seen": 3851760,
      "step": 970,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 74.6923076923077,
      "grad_norm": 1.5750373601913452,
      "learning_rate": 0.01,
      "loss": 4.374350070953369,
      "num_input_tokens_seen": 3855856,
      "step": 971,
      "train_runtime": 778.9408,
      "train_tokens_per_second": 4950.127
    },
    {
      "epoch": 74.6923076923077,
      "num_input_tokens_seen": 3855856,
      "step": 971,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 74.76923076923077,
      "grad_norm": 2.556546688079834,
      "learning_rate": 0.01,
      "loss": 4.3275556564331055,
      "num_input_tokens_seen": 3859952,
      "step": 972,
      "train_runtime": 779.7665,
      "train_tokens_per_second": 4950.138
    },
    {
      "epoch": 74.76923076923077,
      "num_input_tokens_seen": 3859952,
      "step": 972,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 74.84615384615384,
      "grad_norm": 0.40207862854003906,
      "learning_rate": 0.01,
      "loss": 4.291187286376953,
      "num_input_tokens_seen": 3864048,
      "step": 973,
      "train_runtime": 780.5871,
      "train_tokens_per_second": 4950.182
    },
    {
      "epoch": 74.84615384615384,
      "num_input_tokens_seen": 3864048,
      "step": 973,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 74.92307692307692,
      "grad_norm": 0.8552497029304504,
      "learning_rate": 0.01,
      "loss": 4.339695930480957,
      "num_input_tokens_seen": 3868144,
      "step": 974,
      "train_runtime": 781.4012,
      "train_tokens_per_second": 4950.266
    },
    {
      "epoch": 74.92307692307692,
      "num_input_tokens_seen": 3868144,
      "step": 974,
      "train_accuracy": 0.026058631017804146
    },
    {
      "epoch": 75.0,
      "grad_norm": 1.4550114870071411,
      "learning_rate": 0.01,
      "loss": 4.3495965003967285,
      "num_input_tokens_seen": 3870600,
      "step": 975,
      "train_runtime": 781.9189,
      "train_tokens_per_second": 4950.13
    },
    {
      "epoch": 75.0,
      "num_input_tokens_seen": 3870600,
      "step": 975,
      "train_accuracy": 0.04296875
    },
    {
      "epoch": 75.07692307692308,
      "grad_norm": 1.2355130910873413,
      "learning_rate": 0.01,
      "loss": 4.248445987701416,
      "num_input_tokens_seen": 3874696,
      "step": 976,
      "train_runtime": 782.7524,
      "train_tokens_per_second": 4950.092
    },
    {
      "epoch": 75.07692307692308,
      "num_input_tokens_seen": 3874696,
      "step": 976,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 75.15384615384616,
      "grad_norm": 1.122639536857605,
      "learning_rate": 0.01,
      "loss": 4.32548713684082,
      "num_input_tokens_seen": 3878792,
      "step": 977,
      "train_runtime": 783.5706,
      "train_tokens_per_second": 4950.15
    },
    {
      "epoch": 75.15384615384616,
      "num_input_tokens_seen": 3878792,
      "step": 977,
      "train_accuracy": 0.04296875
    },
    {
      "epoch": 75.23076923076923,
      "grad_norm": 0.5361711382865906,
      "learning_rate": 0.01,
      "loss": 4.248356342315674,
      "num_input_tokens_seen": 3882888,
      "step": 978,
      "train_runtime": 784.3897,
      "train_tokens_per_second": 4950.203
    },
    {
      "epoch": 75.23076923076923,
      "num_input_tokens_seen": 3882888,
      "step": 978,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 75.3076923076923,
      "grad_norm": 0.6754993200302124,
      "learning_rate": 0.01,
      "loss": 4.307168006896973,
      "num_input_tokens_seen": 3886984,
      "step": 979,
      "train_runtime": 785.209,
      "train_tokens_per_second": 4950.254
    },
    {
      "epoch": 75.3076923076923,
      "num_input_tokens_seen": 3886984,
      "step": 979,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 75.38461538461539,
      "grad_norm": 0.46616852283477783,
      "learning_rate": 0.01,
      "loss": 4.327777862548828,
      "num_input_tokens_seen": 3891080,
      "step": 980,
      "train_runtime": 786.0275,
      "train_tokens_per_second": 4950.31
    },
    {
      "epoch": 75.38461538461539,
      "num_input_tokens_seen": 3891080,
      "step": 980,
      "train_accuracy": 0.0390625
    },
    {
      "epoch": 75.46153846153847,
      "grad_norm": 1.2690057754516602,
      "learning_rate": 0.01,
      "loss": 4.262716293334961,
      "num_input_tokens_seen": 3895176,
      "step": 981,
      "train_runtime": 786.8466,
      "train_tokens_per_second": 4950.363
    },
    {
      "epoch": 75.46153846153847,
      "num_input_tokens_seen": 3895176,
      "step": 981,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 75.53846153846153,
      "grad_norm": 0.8771840333938599,
      "learning_rate": 0.01,
      "loss": 4.32326602935791,
      "num_input_tokens_seen": 3899272,
      "step": 982,
      "train_runtime": 787.6731,
      "train_tokens_per_second": 4950.368
    },
    {
      "epoch": 75.53846153846153,
      "num_input_tokens_seen": 3899272,
      "step": 982,
      "train_accuracy": 0.04296875
    },
    {
      "epoch": 75.61538461538461,
      "grad_norm": 7.643060207366943,
      "learning_rate": 0.01,
      "loss": 4.260503768920898,
      "num_input_tokens_seen": 3903368,
      "step": 983,
      "train_runtime": 788.4916,
      "train_tokens_per_second": 4950.425
    },
    {
      "epoch": 75.61538461538461,
      "num_input_tokens_seen": 3903368,
      "step": 983,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 75.6923076923077,
      "grad_norm": 0.5241610407829285,
      "learning_rate": 0.01,
      "loss": 4.345402240753174,
      "num_input_tokens_seen": 3907464,
      "step": 984,
      "train_runtime": 789.3117,
      "train_tokens_per_second": 4950.47
    },
    {
      "epoch": 75.6923076923077,
      "num_input_tokens_seen": 3907464,
      "step": 984,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 75.76923076923077,
      "grad_norm": 0.5066173672676086,
      "learning_rate": 0.01,
      "loss": 4.356078147888184,
      "num_input_tokens_seen": 3911560,
      "step": 985,
      "train_runtime": 790.1298,
      "train_tokens_per_second": 4950.528
    },
    {
      "epoch": 75.76923076923077,
      "num_input_tokens_seen": 3911560,
      "step": 985,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 75.84615384615384,
      "grad_norm": 0.6324288845062256,
      "learning_rate": 0.01,
      "loss": 4.339529037475586,
      "num_input_tokens_seen": 3915656,
      "step": 986,
      "train_runtime": 790.9508,
      "train_tokens_per_second": 4950.568
    },
    {
      "epoch": 75.84615384615384,
      "num_input_tokens_seen": 3915656,
      "step": 986,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 75.92307692307692,
      "grad_norm": 0.7764051556587219,
      "learning_rate": 0.01,
      "loss": 4.338321208953857,
      "num_input_tokens_seen": 3919752,
      "step": 987,
      "train_runtime": 791.7661,
      "train_tokens_per_second": 4950.644
    },
    {
      "epoch": 75.92307692307692,
      "num_input_tokens_seen": 3919752,
      "step": 987,
      "train_accuracy": 0.016286645084619522
    },
    {
      "epoch": 76.0,
      "grad_norm": 0.6350783109664917,
      "learning_rate": 0.01,
      "loss": 4.3816609382629395,
      "num_input_tokens_seen": 3922208,
      "step": 988,
      "train_runtime": 792.2841,
      "train_tokens_per_second": 4950.507
    },
    {
      "epoch": 76.0,
      "num_input_tokens_seen": 3922208,
      "step": 988,
      "train_accuracy": 0.03125
    },
    {
      "epoch": 76.07692307692308,
      "grad_norm": 2.2945821285247803,
      "learning_rate": 0.01,
      "loss": 4.2715864181518555,
      "num_input_tokens_seen": 3926304,
      "step": 989,
      "train_runtime": 793.116,
      "train_tokens_per_second": 4950.479
    },
    {
      "epoch": 76.07692307692308,
      "num_input_tokens_seen": 3926304,
      "step": 989,
      "train_accuracy": 0.0390625
    },
    {
      "epoch": 76.15384615384616,
      "grad_norm": 0.5859194397926331,
      "learning_rate": 0.01,
      "loss": 4.2348175048828125,
      "num_input_tokens_seen": 3930400,
      "step": 990,
      "train_runtime": 793.9358,
      "train_tokens_per_second": 4950.526
    },
    {
      "epoch": 76.15384615384616,
      "num_input_tokens_seen": 3930400,
      "step": 990,
      "train_accuracy": 0.04296875
    },
    {
      "epoch": 76.23076923076923,
      "grad_norm": 0.4957868754863739,
      "learning_rate": 0.01,
      "loss": 4.291881561279297,
      "num_input_tokens_seen": 3934496,
      "step": 991,
      "train_runtime": 794.7544,
      "train_tokens_per_second": 4950.581
    },
    {
      "epoch": 76.23076923076923,
      "num_input_tokens_seen": 3934496,
      "step": 991,
      "train_accuracy": 0.0390625
    },
    {
      "epoch": 76.3076923076923,
      "grad_norm": 0.5065727233886719,
      "learning_rate": 0.01,
      "loss": 4.309178829193115,
      "num_input_tokens_seen": 3938592,
      "step": 992,
      "train_runtime": 795.5729,
      "train_tokens_per_second": 4950.636
    },
    {
      "epoch": 76.3076923076923,
      "num_input_tokens_seen": 3938592,
      "step": 992,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 76.38461538461539,
      "grad_norm": 0.877784013748169,
      "learning_rate": 0.01,
      "loss": 4.307445049285889,
      "num_input_tokens_seen": 3942688,
      "step": 993,
      "train_runtime": 796.3908,
      "train_tokens_per_second": 4950.695
    },
    {
      "epoch": 76.38461538461539,
      "num_input_tokens_seen": 3942688,
      "step": 993,
      "train_accuracy": 0.03515625
    },
    {
      "epoch": 76.46153846153847,
      "grad_norm": 0.6491925120353699,
      "learning_rate": 0.01,
      "loss": 4.238561153411865,
      "num_input_tokens_seen": 3946784,
      "step": 994,
      "train_runtime": 797.2094,
      "train_tokens_per_second": 4950.749
    },
    {
      "epoch": 76.46153846153847,
      "num_input_tokens_seen": 3946784,
      "step": 994,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 76.53846153846153,
      "grad_norm": 0.8613327145576477,
      "learning_rate": 0.01,
      "loss": 4.259726047515869,
      "num_input_tokens_seen": 3950880,
      "step": 995,
      "train_runtime": 798.0284,
      "train_tokens_per_second": 4950.801
    },
    {
      "epoch": 76.53846153846153,
      "num_input_tokens_seen": 3950880,
      "step": 995,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 76.61538461538461,
      "grad_norm": 1.273020625114441,
      "learning_rate": 0.01,
      "loss": 4.271668910980225,
      "num_input_tokens_seen": 3954976,
      "step": 996,
      "train_runtime": 798.8477,
      "train_tokens_per_second": 4950.851
    },
    {
      "epoch": 76.61538461538461,
      "num_input_tokens_seen": 3954976,
      "step": 996,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 76.6923076923077,
      "grad_norm": 0.4400981664657593,
      "learning_rate": 0.01,
      "loss": 4.302046298980713,
      "num_input_tokens_seen": 3959072,
      "step": 997,
      "train_runtime": 799.6675,
      "train_tokens_per_second": 4950.898
    },
    {
      "epoch": 76.6923076923077,
      "num_input_tokens_seen": 3959072,
      "step": 997,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 76.76923076923077,
      "grad_norm": 0.5882481336593628,
      "learning_rate": 0.01,
      "loss": 4.22745418548584,
      "num_input_tokens_seen": 3963168,
      "step": 998,
      "train_runtime": 800.4862,
      "train_tokens_per_second": 4950.951
    },
    {
      "epoch": 76.76923076923077,
      "num_input_tokens_seen": 3963168,
      "step": 998,
      "train_accuracy": 0.04296875
    },
    {
      "epoch": 76.84615384615384,
      "grad_norm": 0.5601505637168884,
      "learning_rate": 0.01,
      "loss": 4.18474006652832,
      "num_input_tokens_seen": 3967264,
      "step": 999,
      "train_runtime": 801.305,
      "train_tokens_per_second": 4951.003
    },
    {
      "epoch": 76.84615384615384,
      "num_input_tokens_seen": 3967264,
      "step": 999,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 76.92307692307692,
      "grad_norm": 0.6990476250648499,
      "learning_rate": 0.01,
      "loss": 4.3541154861450195,
      "num_input_tokens_seen": 3971360,
      "step": 1000,
      "train_runtime": 802.1186,
      "train_tokens_per_second": 4951.088
    },
    {
      "epoch": 76.92307692307692,
      "eval_CMD_3_branch_eval_accuracy": 0.01735985533453888,
      "eval_CMD_3_branch_eval_loss": 4.395492076873779,
      "eval_CMD_3_branch_eval_runtime": 1.165,
      "eval_CMD_3_branch_eval_samples_per_second": 2373.337,
      "eval_CMD_3_branch_eval_steps_per_second": 5.15,
      "num_input_tokens_seen": 3971360,
      "step": 1000
    },
    {
      "epoch": 76.92307692307692,
      "num_input_tokens_seen": 3971360,
      "step": 1000,
      "train_accuracy": 0.06514658033847809
    },
    {
      "epoch": 77.0,
      "grad_norm": 0.8677210211753845,
      "learning_rate": 0.01,
      "loss": 4.1859660148620605,
      "num_input_tokens_seen": 3973816,
      "step": 1001,
      "train_runtime": 819.9952,
      "train_tokens_per_second": 4846.145
    },
    {
      "epoch": 77.0,
      "num_input_tokens_seen": 3973816,
      "step": 1001,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 77.07692307692308,
      "grad_norm": 0.6715117692947388,
      "learning_rate": 0.01,
      "loss": 4.3048624992370605,
      "num_input_tokens_seen": 3977912,
      "step": 1002,
      "train_runtime": 820.8133,
      "train_tokens_per_second": 4846.305
    },
    {
      "epoch": 77.07692307692308,
      "num_input_tokens_seen": 3977912,
      "step": 1002,
      "train_accuracy": 0.044921875
    },
    {
      "epoch": 77.15384615384616,
      "grad_norm": 1.224298119544983,
      "learning_rate": 0.01,
      "loss": 4.202672004699707,
      "num_input_tokens_seen": 3982008,
      "step": 1003,
      "train_runtime": 821.6194,
      "train_tokens_per_second": 4846.536
    },
    {
      "epoch": 77.15384615384616,
      "num_input_tokens_seen": 3982008,
      "step": 1003,
      "train_accuracy": 0.044921875
    },
    {
      "epoch": 77.23076923076923,
      "grad_norm": 0.4090309739112854,
      "learning_rate": 0.01,
      "loss": 4.234436988830566,
      "num_input_tokens_seen": 3986104,
      "step": 1004,
      "train_runtime": 822.4263,
      "train_tokens_per_second": 4846.761
    },
    {
      "epoch": 77.23076923076923,
      "num_input_tokens_seen": 3986104,
      "step": 1004,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 77.3076923076923,
      "grad_norm": 0.587716281414032,
      "learning_rate": 0.01,
      "loss": 4.283956527709961,
      "num_input_tokens_seen": 3990200,
      "step": 1005,
      "train_runtime": 823.2341,
      "train_tokens_per_second": 4846.981
    },
    {
      "epoch": 77.3076923076923,
      "num_input_tokens_seen": 3990200,
      "step": 1005,
      "train_accuracy": 0.03125
    },
    {
      "epoch": 77.38461538461539,
      "grad_norm": 0.46258968114852905,
      "learning_rate": 0.01,
      "loss": 4.266878128051758,
      "num_input_tokens_seen": 3994296,
      "step": 1006,
      "train_runtime": 824.0426,
      "train_tokens_per_second": 4847.196
    },
    {
      "epoch": 77.38461538461539,
      "num_input_tokens_seen": 3994296,
      "step": 1006,
      "train_accuracy": 0.0390625
    },
    {
      "epoch": 77.46153846153847,
      "grad_norm": 0.8795014023780823,
      "learning_rate": 0.01,
      "loss": 4.294693946838379,
      "num_input_tokens_seen": 3998392,
      "step": 1007,
      "train_runtime": 824.8502,
      "train_tokens_per_second": 4847.416
    },
    {
      "epoch": 77.46153846153847,
      "num_input_tokens_seen": 3998392,
      "step": 1007,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 77.53846153846153,
      "grad_norm": 0.5326021313667297,
      "learning_rate": 0.01,
      "loss": 4.288190841674805,
      "num_input_tokens_seen": 4002488,
      "step": 1008,
      "train_runtime": 825.6595,
      "train_tokens_per_second": 4847.626
    },
    {
      "epoch": 77.53846153846153,
      "num_input_tokens_seen": 4002488,
      "step": 1008,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 77.61538461538461,
      "grad_norm": 0.44118621945381165,
      "learning_rate": 0.01,
      "loss": 4.2714433670043945,
      "num_input_tokens_seen": 4006584,
      "step": 1009,
      "train_runtime": 826.4689,
      "train_tokens_per_second": 4847.834
    },
    {
      "epoch": 77.61538461538461,
      "num_input_tokens_seen": 4006584,
      "step": 1009,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 77.6923076923077,
      "grad_norm": 0.47735729813575745,
      "learning_rate": 0.01,
      "loss": 4.25751256942749,
      "num_input_tokens_seen": 4010680,
      "step": 1010,
      "train_runtime": 827.2768,
      "train_tokens_per_second": 4848.051
    },
    {
      "epoch": 77.6923076923077,
      "num_input_tokens_seen": 4010680,
      "step": 1010,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 77.76923076923077,
      "grad_norm": 0.5568310022354126,
      "learning_rate": 0.01,
      "loss": 4.27866792678833,
      "num_input_tokens_seen": 4014776,
      "step": 1011,
      "train_runtime": 828.0854,
      "train_tokens_per_second": 4848.263
    },
    {
      "epoch": 77.76923076923077,
      "num_input_tokens_seen": 4014776,
      "step": 1011,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 77.84615384615384,
      "grad_norm": 0.47787994146347046,
      "learning_rate": 0.01,
      "loss": 4.250924110412598,
      "num_input_tokens_seen": 4018872,
      "step": 1012,
      "train_runtime": 828.8949,
      "train_tokens_per_second": 4848.47
    },
    {
      "epoch": 77.84615384615384,
      "num_input_tokens_seen": 4018872,
      "step": 1012,
      "train_accuracy": 0.046875
    },
    {
      "epoch": 77.92307692307692,
      "grad_norm": 0.5050224661827087,
      "learning_rate": 0.01,
      "loss": 4.301644802093506,
      "num_input_tokens_seen": 4022968,
      "step": 1013,
      "train_runtime": 829.7,
      "train_tokens_per_second": 4848.702
    },
    {
      "epoch": 77.92307692307692,
      "num_input_tokens_seen": 4022968,
      "step": 1013,
      "train_accuracy": 0.016286645084619522
    },
    {
      "epoch": 78.0,
      "grad_norm": 1.0325103998184204,
      "learning_rate": 0.01,
      "loss": 4.303225040435791,
      "num_input_tokens_seen": 4025424,
      "step": 1014,
      "train_runtime": 830.2129,
      "train_tokens_per_second": 4848.665
    },
    {
      "epoch": 78.0,
      "num_input_tokens_seen": 4025424,
      "step": 1014,
      "train_accuracy": 0.041015625
    },
    {
      "epoch": 78.07692307692308,
      "grad_norm": 0.42929312586784363,
      "learning_rate": 0.01,
      "loss": 4.242375373840332,
      "num_input_tokens_seen": 4029520,
      "step": 1015,
      "train_runtime": 831.0341,
      "train_tokens_per_second": 4848.802
    },
    {
      "epoch": 78.07692307692308,
      "num_input_tokens_seen": 4029520,
      "step": 1015,
      "train_accuracy": 0.0390625
    },
    {
      "epoch": 78.15384615384616,
      "grad_norm": 1.1013460159301758,
      "learning_rate": 0.01,
      "loss": 4.244626045227051,
      "num_input_tokens_seen": 4033616,
      "step": 1016,
      "train_runtime": 831.8509,
      "train_tokens_per_second": 4848.965
    },
    {
      "epoch": 78.15384615384616,
      "num_input_tokens_seen": 4033616,
      "step": 1016,
      "train_accuracy": 0.04296875
    },
    {
      "epoch": 78.23076923076923,
      "grad_norm": 0.7049079537391663,
      "learning_rate": 0.01,
      "loss": 4.299908638000488,
      "num_input_tokens_seen": 4037712,
      "step": 1017,
      "train_runtime": 832.6647,
      "train_tokens_per_second": 4849.145
    },
    {
      "epoch": 78.23076923076923,
      "num_input_tokens_seen": 4037712,
      "step": 1017,
      "train_accuracy": 0.03125
    },
    {
      "epoch": 78.3076923076923,
      "grad_norm": 7.399558067321777,
      "learning_rate": 0.01,
      "loss": 4.2371745109558105,
      "num_input_tokens_seen": 4041808,
      "step": 1018,
      "train_runtime": 833.4753,
      "train_tokens_per_second": 4849.343
    },
    {
      "epoch": 78.3076923076923,
      "num_input_tokens_seen": 4041808,
      "step": 1018,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 78.38461538461539,
      "grad_norm": 0.5494755506515503,
      "learning_rate": 0.01,
      "loss": 4.262982368469238,
      "num_input_tokens_seen": 4045904,
      "step": 1019,
      "train_runtime": 834.2875,
      "train_tokens_per_second": 4849.532
    },
    {
      "epoch": 78.38461538461539,
      "num_input_tokens_seen": 4045904,
      "step": 1019,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 78.46153846153847,
      "grad_norm": 0.7562094330787659,
      "learning_rate": 0.01,
      "loss": 4.380556106567383,
      "num_input_tokens_seen": 4050000,
      "step": 1020,
      "train_runtime": 835.0982,
      "train_tokens_per_second": 4849.729
    },
    {
      "epoch": 78.46153846153847,
      "num_input_tokens_seen": 4050000,
      "step": 1020,
      "train_accuracy": 0.041015625
    },
    {
      "epoch": 78.53846153846153,
      "grad_norm": 0.7579846978187561,
      "learning_rate": 0.01,
      "loss": 4.199994087219238,
      "num_input_tokens_seen": 4054096,
      "step": 1021,
      "train_runtime": 835.9089,
      "train_tokens_per_second": 4849.926
    },
    {
      "epoch": 78.53846153846153,
      "num_input_tokens_seen": 4054096,
      "step": 1021,
      "train_accuracy": 0.037109375
    },
    {
      "epoch": 78.61538461538461,
      "grad_norm": 0.7965441942214966,
      "learning_rate": 0.01,
      "loss": 4.252337455749512,
      "num_input_tokens_seen": 4058192,
      "step": 1022,
      "train_runtime": 836.7224,
      "train_tokens_per_second": 4850.105
    },
    {
      "epoch": 78.61538461538461,
      "num_input_tokens_seen": 4058192,
      "step": 1022,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 78.6923076923077,
      "grad_norm": 0.5306705236434937,
      "learning_rate": 0.01,
      "loss": 4.2623090744018555,
      "num_input_tokens_seen": 4062288,
      "step": 1023,
      "train_runtime": 837.5341,
      "train_tokens_per_second": 4850.295
    },
    {
      "epoch": 78.6923076923077,
      "num_input_tokens_seen": 4062288,
      "step": 1023,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 78.76923076923077,
      "grad_norm": 0.5631991028785706,
      "learning_rate": 0.01,
      "loss": 4.307033538818359,
      "num_input_tokens_seen": 4066384,
      "step": 1024,
      "train_runtime": 838.3458,
      "train_tokens_per_second": 4850.485
    },
    {
      "epoch": 78.76923076923077,
      "num_input_tokens_seen": 4066384,
      "step": 1024,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 78.84615384615384,
      "grad_norm": 0.9251755475997925,
      "learning_rate": 0.01,
      "loss": 4.3464860916137695,
      "num_input_tokens_seen": 4070480,
      "step": 1025,
      "train_runtime": 839.1571,
      "train_tokens_per_second": 4850.677
    },
    {
      "epoch": 78.84615384615384,
      "num_input_tokens_seen": 4070480,
      "step": 1025,
      "train_accuracy": 0.03515625
    },
    {
      "epoch": 78.92307692307692,
      "grad_norm": 3.4484596252441406,
      "learning_rate": 0.01,
      "loss": 4.21057653427124,
      "num_input_tokens_seen": 4074576,
      "step": 1026,
      "train_runtime": 839.9638,
      "train_tokens_per_second": 4850.895
    },
    {
      "epoch": 78.92307692307692,
      "num_input_tokens_seen": 4074576,
      "step": 1026,
      "train_accuracy": 0.02931595966219902
    },
    {
      "epoch": 79.0,
      "grad_norm": 0.8657160401344299,
      "learning_rate": 0.01,
      "loss": 4.253246784210205,
      "num_input_tokens_seen": 4077032,
      "step": 1027,
      "train_runtime": 840.4776,
      "train_tokens_per_second": 4850.851
    },
    {
      "epoch": 79.0,
      "num_input_tokens_seen": 4077032,
      "step": 1027,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 79.07692307692308,
      "grad_norm": 1.3953666687011719,
      "learning_rate": 0.01,
      "loss": 4.2584757804870605,
      "num_input_tokens_seen": 4081128,
      "step": 1028,
      "train_runtime": 841.3018,
      "train_tokens_per_second": 4850.968
    },
    {
      "epoch": 79.07692307692308,
      "num_input_tokens_seen": 4081128,
      "step": 1028,
      "train_accuracy": 0.041015625
    },
    {
      "epoch": 79.15384615384616,
      "grad_norm": 0.5391897559165955,
      "learning_rate": 0.01,
      "loss": 4.180985450744629,
      "num_input_tokens_seen": 4085224,
      "step": 1029,
      "train_runtime": 842.113,
      "train_tokens_per_second": 4851.159
    },
    {
      "epoch": 79.15384615384616,
      "num_input_tokens_seen": 4085224,
      "step": 1029,
      "train_accuracy": 0.041015625
    },
    {
      "epoch": 79.23076923076923,
      "grad_norm": 0.47898000478744507,
      "learning_rate": 0.01,
      "loss": 4.175528526306152,
      "num_input_tokens_seen": 4089320,
      "step": 1030,
      "train_runtime": 842.9248,
      "train_tokens_per_second": 4851.346
    },
    {
      "epoch": 79.23076923076923,
      "num_input_tokens_seen": 4089320,
      "step": 1030,
      "train_accuracy": 0.03125
    },
    {
      "epoch": 79.3076923076923,
      "grad_norm": 0.6484806537628174,
      "learning_rate": 0.01,
      "loss": 4.273150444030762,
      "num_input_tokens_seen": 4093416,
      "step": 1031,
      "train_runtime": 843.7382,
      "train_tokens_per_second": 4851.524
    },
    {
      "epoch": 79.3076923076923,
      "num_input_tokens_seen": 4093416,
      "step": 1031,
      "train_accuracy": 0.0546875
    },
    {
      "epoch": 79.38461538461539,
      "grad_norm": 0.8459232449531555,
      "learning_rate": 0.01,
      "loss": 4.244646072387695,
      "num_input_tokens_seen": 4097512,
      "step": 1032,
      "train_runtime": 844.5509,
      "train_tokens_per_second": 4851.705
    },
    {
      "epoch": 79.38461538461539,
      "num_input_tokens_seen": 4097512,
      "step": 1032,
      "train_accuracy": 0.03125
    },
    {
      "epoch": 79.46153846153847,
      "grad_norm": 0.4148889482021332,
      "learning_rate": 0.01,
      "loss": 4.248509407043457,
      "num_input_tokens_seen": 4101608,
      "step": 1033,
      "train_runtime": 845.3638,
      "train_tokens_per_second": 4851.885
    },
    {
      "epoch": 79.46153846153847,
      "num_input_tokens_seen": 4101608,
      "step": 1033,
      "train_accuracy": 0.044921875
    },
    {
      "epoch": 79.53846153846153,
      "grad_norm": 0.5211220383644104,
      "learning_rate": 0.01,
      "loss": 4.183545112609863,
      "num_input_tokens_seen": 4105704,
      "step": 1034,
      "train_runtime": 846.1766,
      "train_tokens_per_second": 4852.065
    },
    {
      "epoch": 79.53846153846153,
      "num_input_tokens_seen": 4105704,
      "step": 1034,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 79.61538461538461,
      "grad_norm": 0.6639049649238586,
      "learning_rate": 0.01,
      "loss": 4.243070602416992,
      "num_input_tokens_seen": 4109800,
      "step": 1035,
      "train_runtime": 846.9891,
      "train_tokens_per_second": 4852.247
    },
    {
      "epoch": 79.61538461538461,
      "num_input_tokens_seen": 4109800,
      "step": 1035,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 79.6923076923077,
      "grad_norm": 1.5792025327682495,
      "learning_rate": 0.01,
      "loss": 4.212851047515869,
      "num_input_tokens_seen": 4113896,
      "step": 1036,
      "train_runtime": 847.801,
      "train_tokens_per_second": 4852.431
    },
    {
      "epoch": 79.6923076923077,
      "num_input_tokens_seen": 4113896,
      "step": 1036,
      "train_accuracy": 0.04296875
    },
    {
      "epoch": 79.76923076923077,
      "grad_norm": 0.7158578038215637,
      "learning_rate": 0.01,
      "loss": 4.255945682525635,
      "num_input_tokens_seen": 4117992,
      "step": 1037,
      "train_runtime": 848.6136,
      "train_tokens_per_second": 4852.611
    },
    {
      "epoch": 79.76923076923077,
      "num_input_tokens_seen": 4117992,
      "step": 1037,
      "train_accuracy": 0.046875
    },
    {
      "epoch": 79.84615384615384,
      "grad_norm": 0.35050055384635925,
      "learning_rate": 0.01,
      "loss": 4.240667343139648,
      "num_input_tokens_seen": 4122088,
      "step": 1038,
      "train_runtime": 849.4272,
      "train_tokens_per_second": 4852.786
    },
    {
      "epoch": 79.84615384615384,
      "num_input_tokens_seen": 4122088,
      "step": 1038,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 79.92307692307692,
      "grad_norm": 0.6327638030052185,
      "learning_rate": 0.01,
      "loss": 4.217525482177734,
      "num_input_tokens_seen": 4126184,
      "step": 1039,
      "train_runtime": 850.2363,
      "train_tokens_per_second": 4852.985
    },
    {
      "epoch": 79.92307692307692,
      "num_input_tokens_seen": 4126184,
      "step": 1039,
      "train_accuracy": 0.02280130237340927
    },
    {
      "epoch": 80.0,
      "grad_norm": 0.42952316999435425,
      "learning_rate": 0.01,
      "loss": 4.2495646476745605,
      "num_input_tokens_seen": 4128640,
      "step": 1040,
      "train_runtime": 850.7507,
      "train_tokens_per_second": 4852.937
    },
    {
      "epoch": 80.0,
      "num_input_tokens_seen": 4128640,
      "step": 1040,
      "train_accuracy": 0.044921875
    },
    {
      "epoch": 80.07692307692308,
      "grad_norm": 2.4387288093566895,
      "learning_rate": 0.01,
      "loss": 4.163417816162109,
      "num_input_tokens_seen": 4132736,
      "step": 1041,
      "train_runtime": 851.5759,
      "train_tokens_per_second": 4853.045
    },
    {
      "epoch": 80.07692307692308,
      "num_input_tokens_seen": 4132736,
      "step": 1041,
      "train_accuracy": 0.037109375
    },
    {
      "epoch": 80.15384615384616,
      "grad_norm": 0.8980920314788818,
      "learning_rate": 0.01,
      "loss": 4.2331037521362305,
      "num_input_tokens_seen": 4136832,
      "step": 1042,
      "train_runtime": 852.3892,
      "train_tokens_per_second": 4853.22
    },
    {
      "epoch": 80.15384615384616,
      "num_input_tokens_seen": 4136832,
      "step": 1042,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 80.23076923076923,
      "grad_norm": 1.5235867500305176,
      "learning_rate": 0.01,
      "loss": 4.24924373626709,
      "num_input_tokens_seen": 4140928,
      "step": 1043,
      "train_runtime": 853.2033,
      "train_tokens_per_second": 4853.389
    },
    {
      "epoch": 80.23076923076923,
      "num_input_tokens_seen": 4140928,
      "step": 1043,
      "train_accuracy": 0.044921875
    },
    {
      "epoch": 80.3076923076923,
      "grad_norm": 3.1564149856567383,
      "learning_rate": 0.01,
      "loss": 4.218427658081055,
      "num_input_tokens_seen": 4145024,
      "step": 1044,
      "train_runtime": 854.0173,
      "train_tokens_per_second": 4853.56
    },
    {
      "epoch": 80.3076923076923,
      "num_input_tokens_seen": 4145024,
      "step": 1044,
      "train_accuracy": 0.03515625
    },
    {
      "epoch": 80.38461538461539,
      "grad_norm": 0.7506723403930664,
      "learning_rate": 0.01,
      "loss": 4.248618125915527,
      "num_input_tokens_seen": 4149120,
      "step": 1045,
      "train_runtime": 854.8303,
      "train_tokens_per_second": 4853.735
    },
    {
      "epoch": 80.38461538461539,
      "num_input_tokens_seen": 4149120,
      "step": 1045,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 80.46153846153847,
      "grad_norm": 0.6711944937705994,
      "learning_rate": 0.01,
      "loss": 4.253600597381592,
      "num_input_tokens_seen": 4153216,
      "step": 1046,
      "train_runtime": 855.6443,
      "train_tokens_per_second": 4853.905
    },
    {
      "epoch": 80.46153846153847,
      "num_input_tokens_seen": 4153216,
      "step": 1046,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 80.53846153846153,
      "grad_norm": 0.5716758370399475,
      "learning_rate": 0.01,
      "loss": 4.272745132446289,
      "num_input_tokens_seen": 4157312,
      "step": 1047,
      "train_runtime": 856.4592,
      "train_tokens_per_second": 4854.069
    },
    {
      "epoch": 80.53846153846153,
      "num_input_tokens_seen": 4157312,
      "step": 1047,
      "train_accuracy": 0.041015625
    },
    {
      "epoch": 80.61538461538461,
      "grad_norm": 0.4625732898712158,
      "learning_rate": 0.01,
      "loss": 4.166475772857666,
      "num_input_tokens_seen": 4161408,
      "step": 1048,
      "train_runtime": 857.275,
      "train_tokens_per_second": 4854.228
    },
    {
      "epoch": 80.61538461538461,
      "num_input_tokens_seen": 4161408,
      "step": 1048,
      "train_accuracy": 0.041015625
    },
    {
      "epoch": 80.6923076923077,
      "grad_norm": 1.3532286882400513,
      "learning_rate": 0.01,
      "loss": 4.213746070861816,
      "num_input_tokens_seen": 4165504,
      "step": 1049,
      "train_runtime": 858.0885,
      "train_tokens_per_second": 4854.399
    },
    {
      "epoch": 80.6923076923077,
      "num_input_tokens_seen": 4165504,
      "step": 1049,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 80.76923076923077,
      "grad_norm": 0.7559223771095276,
      "learning_rate": 0.01,
      "loss": 4.3056559562683105,
      "num_input_tokens_seen": 4169600,
      "step": 1050,
      "train_runtime": 858.9036,
      "train_tokens_per_second": 4854.561
    },
    {
      "epoch": 80.76923076923077,
      "num_input_tokens_seen": 4169600,
      "step": 1050,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 80.84615384615384,
      "grad_norm": 2.033456325531006,
      "learning_rate": 0.01,
      "loss": 4.330817222595215,
      "num_input_tokens_seen": 4173696,
      "step": 1051,
      "train_runtime": 859.7179,
      "train_tokens_per_second": 4854.727
    },
    {
      "epoch": 80.84615384615384,
      "num_input_tokens_seen": 4173696,
      "step": 1051,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 80.92307692307692,
      "grad_norm": 2.785479784011841,
      "learning_rate": 0.01,
      "loss": 4.377662658691406,
      "num_input_tokens_seen": 4177792,
      "step": 1052,
      "train_runtime": 860.5279,
      "train_tokens_per_second": 4854.918
    },
    {
      "epoch": 80.92307692307692,
      "num_input_tokens_seen": 4177792,
      "step": 1052,
      "train_accuracy": 0.02931595966219902
    },
    {
      "epoch": 81.0,
      "grad_norm": 1.168498158454895,
      "learning_rate": 0.01,
      "loss": 4.308221340179443,
      "num_input_tokens_seen": 4180248,
      "step": 1053,
      "train_runtime": 861.0432,
      "train_tokens_per_second": 4854.864
    },
    {
      "epoch": 81.0,
      "num_input_tokens_seen": 4180248,
      "step": 1053,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 81.07692307692308,
      "grad_norm": 0.596804678440094,
      "learning_rate": 0.01,
      "loss": 4.257911205291748,
      "num_input_tokens_seen": 4184344,
      "step": 1054,
      "train_runtime": 861.8729,
      "train_tokens_per_second": 4854.943
    },
    {
      "epoch": 81.07692307692308,
      "num_input_tokens_seen": 4184344,
      "step": 1054,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 81.15384615384616,
      "grad_norm": 1.002935767173767,
      "learning_rate": 0.01,
      "loss": 4.293037414550781,
      "num_input_tokens_seen": 4188440,
      "step": 1055,
      "train_runtime": 862.687,
      "train_tokens_per_second": 4855.109
    },
    {
      "epoch": 81.15384615384616,
      "num_input_tokens_seen": 4188440,
      "step": 1055,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 81.23076923076923,
      "grad_norm": 0.7886541485786438,
      "learning_rate": 0.01,
      "loss": 4.286433219909668,
      "num_input_tokens_seen": 4192536,
      "step": 1056,
      "train_runtime": 863.5013,
      "train_tokens_per_second": 4855.274
    },
    {
      "epoch": 81.23076923076923,
      "num_input_tokens_seen": 4192536,
      "step": 1056,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 81.3076923076923,
      "grad_norm": 0.6340736150741577,
      "learning_rate": 0.01,
      "loss": 4.273266792297363,
      "num_input_tokens_seen": 4196632,
      "step": 1057,
      "train_runtime": 864.3156,
      "train_tokens_per_second": 4855.44
    },
    {
      "epoch": 81.3076923076923,
      "num_input_tokens_seen": 4196632,
      "step": 1057,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 81.38461538461539,
      "grad_norm": 1.1839436292648315,
      "learning_rate": 0.01,
      "loss": 4.253959655761719,
      "num_input_tokens_seen": 4200728,
      "step": 1058,
      "train_runtime": 865.1308,
      "train_tokens_per_second": 4855.599
    },
    {
      "epoch": 81.38461538461539,
      "num_input_tokens_seen": 4200728,
      "step": 1058,
      "train_accuracy": 0.037109375
    },
    {
      "epoch": 81.46153846153847,
      "grad_norm": 0.7244887948036194,
      "learning_rate": 0.01,
      "loss": 4.320132255554199,
      "num_input_tokens_seen": 4204824,
      "step": 1059,
      "train_runtime": 865.9461,
      "train_tokens_per_second": 4855.757
    },
    {
      "epoch": 81.46153846153847,
      "num_input_tokens_seen": 4204824,
      "step": 1059,
      "train_accuracy": 0.041015625
    },
    {
      "epoch": 81.53846153846153,
      "grad_norm": 4.3612847328186035,
      "learning_rate": 0.01,
      "loss": 4.247152805328369,
      "num_input_tokens_seen": 4208920,
      "step": 1060,
      "train_runtime": 866.7613,
      "train_tokens_per_second": 4855.916
    },
    {
      "epoch": 81.53846153846153,
      "num_input_tokens_seen": 4208920,
      "step": 1060,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 81.61538461538461,
      "grad_norm": 1.1570528745651245,
      "learning_rate": 0.01,
      "loss": 4.320521831512451,
      "num_input_tokens_seen": 4213016,
      "step": 1061,
      "train_runtime": 867.5766,
      "train_tokens_per_second": 4856.074
    },
    {
      "epoch": 81.61538461538461,
      "num_input_tokens_seen": 4213016,
      "step": 1061,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 81.6923076923077,
      "grad_norm": 0.8297848105430603,
      "learning_rate": 0.01,
      "loss": 4.2926530838012695,
      "num_input_tokens_seen": 4217112,
      "step": 1062,
      "train_runtime": 868.3917,
      "train_tokens_per_second": 4856.232
    },
    {
      "epoch": 81.6923076923077,
      "num_input_tokens_seen": 4217112,
      "step": 1062,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 81.76923076923077,
      "grad_norm": 1.384696364402771,
      "learning_rate": 0.01,
      "loss": 4.249601364135742,
      "num_input_tokens_seen": 4221208,
      "step": 1063,
      "train_runtime": 869.2062,
      "train_tokens_per_second": 4856.394
    },
    {
      "epoch": 81.76923076923077,
      "num_input_tokens_seen": 4221208,
      "step": 1063,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 81.84615384615384,
      "grad_norm": 0.4551432132720947,
      "learning_rate": 0.01,
      "loss": 4.318694114685059,
      "num_input_tokens_seen": 4225304,
      "step": 1064,
      "train_runtime": 870.0212,
      "train_tokens_per_second": 4856.553
    },
    {
      "epoch": 81.84615384615384,
      "num_input_tokens_seen": 4225304,
      "step": 1064,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 81.92307692307692,
      "grad_norm": 0.6914705038070679,
      "learning_rate": 0.01,
      "loss": 4.334219932556152,
      "num_input_tokens_seen": 4229400,
      "step": 1065,
      "train_runtime": 870.8319,
      "train_tokens_per_second": 4856.735
    },
    {
      "epoch": 81.92307692307692,
      "num_input_tokens_seen": 4229400,
      "step": 1065,
      "train_accuracy": 0.02280130237340927
    },
    {
      "epoch": 82.0,
      "grad_norm": 2.4079184532165527,
      "learning_rate": 0.01,
      "loss": 4.283096790313721,
      "num_input_tokens_seen": 4231856,
      "step": 1066,
      "train_runtime": 871.3474,
      "train_tokens_per_second": 4856.681
    },
    {
      "epoch": 82.0,
      "num_input_tokens_seen": 4231856,
      "step": 1066,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 82.07692307692308,
      "grad_norm": 1.5846831798553467,
      "learning_rate": 0.01,
      "loss": 4.151785850524902,
      "num_input_tokens_seen": 4235952,
      "step": 1067,
      "train_runtime": 872.1771,
      "train_tokens_per_second": 4856.757
    },
    {
      "epoch": 82.07692307692308,
      "num_input_tokens_seen": 4235952,
      "step": 1067,
      "train_accuracy": 0.037109375
    },
    {
      "epoch": 82.15384615384616,
      "grad_norm": 1.0147782564163208,
      "learning_rate": 0.01,
      "loss": 4.166957855224609,
      "num_input_tokens_seen": 4240048,
      "step": 1068,
      "train_runtime": 872.9922,
      "train_tokens_per_second": 4856.914
    },
    {
      "epoch": 82.15384615384616,
      "num_input_tokens_seen": 4240048,
      "step": 1068,
      "train_accuracy": 0.044921875
    },
    {
      "epoch": 82.23076923076923,
      "grad_norm": 1.1793670654296875,
      "learning_rate": 0.01,
      "loss": 4.211719989776611,
      "num_input_tokens_seen": 4244144,
      "step": 1069,
      "train_runtime": 873.8071,
      "train_tokens_per_second": 4857.072
    },
    {
      "epoch": 82.23076923076923,
      "num_input_tokens_seen": 4244144,
      "step": 1069,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 82.3076923076923,
      "grad_norm": 0.6302931904792786,
      "learning_rate": 0.01,
      "loss": 4.268909454345703,
      "num_input_tokens_seen": 4248240,
      "step": 1070,
      "train_runtime": 874.6225,
      "train_tokens_per_second": 4857.227
    },
    {
      "epoch": 82.3076923076923,
      "num_input_tokens_seen": 4248240,
      "step": 1070,
      "train_accuracy": 0.037109375
    },
    {
      "epoch": 82.38461538461539,
      "grad_norm": 0.5030584335327148,
      "learning_rate": 0.01,
      "loss": 4.18870210647583,
      "num_input_tokens_seen": 4252336,
      "step": 1071,
      "train_runtime": 875.4379,
      "train_tokens_per_second": 4857.382
    },
    {
      "epoch": 82.38461538461539,
      "num_input_tokens_seen": 4252336,
      "step": 1071,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 82.46153846153847,
      "grad_norm": 0.5036581754684448,
      "learning_rate": 0.01,
      "loss": 4.271892547607422,
      "num_input_tokens_seen": 4256432,
      "step": 1072,
      "train_runtime": 876.2539,
      "train_tokens_per_second": 4857.533
    },
    {
      "epoch": 82.46153846153847,
      "num_input_tokens_seen": 4256432,
      "step": 1072,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 82.53846153846153,
      "grad_norm": 0.5043556690216064,
      "learning_rate": 0.01,
      "loss": 4.25629997253418,
      "num_input_tokens_seen": 4260528,
      "step": 1073,
      "train_runtime": 877.0711,
      "train_tokens_per_second": 4857.677
    },
    {
      "epoch": 82.53846153846153,
      "num_input_tokens_seen": 4260528,
      "step": 1073,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 82.61538461538461,
      "grad_norm": 0.5362634658813477,
      "learning_rate": 0.01,
      "loss": 4.2652177810668945,
      "num_input_tokens_seen": 4264624,
      "step": 1074,
      "train_runtime": 877.8861,
      "train_tokens_per_second": 4857.833
    },
    {
      "epoch": 82.61538461538461,
      "num_input_tokens_seen": 4264624,
      "step": 1074,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 82.6923076923077,
      "grad_norm": 0.6997628211975098,
      "learning_rate": 0.01,
      "loss": 4.252917289733887,
      "num_input_tokens_seen": 4268720,
      "step": 1075,
      "train_runtime": 878.702,
      "train_tokens_per_second": 4857.984
    },
    {
      "epoch": 82.6923076923077,
      "num_input_tokens_seen": 4268720,
      "step": 1075,
      "train_accuracy": 0.037109375
    },
    {
      "epoch": 82.76923076923077,
      "grad_norm": 2.129366397857666,
      "learning_rate": 0.01,
      "loss": 4.2818684577941895,
      "num_input_tokens_seen": 4272816,
      "step": 1076,
      "train_runtime": 879.5181,
      "train_tokens_per_second": 4858.133
    },
    {
      "epoch": 82.76923076923077,
      "num_input_tokens_seen": 4272816,
      "step": 1076,
      "train_accuracy": 0.037109375
    },
    {
      "epoch": 82.84615384615384,
      "grad_norm": 0.822412371635437,
      "learning_rate": 0.01,
      "loss": 4.201027870178223,
      "num_input_tokens_seen": 4276912,
      "step": 1077,
      "train_runtime": 880.3335,
      "train_tokens_per_second": 4858.286
    },
    {
      "epoch": 82.84615384615384,
      "num_input_tokens_seen": 4276912,
      "step": 1077,
      "train_accuracy": 0.04296875
    },
    {
      "epoch": 82.92307692307692,
      "grad_norm": 0.6472997665405273,
      "learning_rate": 0.01,
      "loss": 4.227897644042969,
      "num_input_tokens_seen": 4281008,
      "step": 1078,
      "train_runtime": 881.1449,
      "train_tokens_per_second": 4858.461
    },
    {
      "epoch": 82.92307692307692,
      "num_input_tokens_seen": 4281008,
      "step": 1078,
      "train_accuracy": 0.032573290169239044
    },
    {
      "epoch": 83.0,
      "grad_norm": 5.376842021942139,
      "learning_rate": 0.01,
      "loss": 4.231244087219238,
      "num_input_tokens_seen": 4283464,
      "step": 1079,
      "train_runtime": 881.6618,
      "train_tokens_per_second": 4858.398
    },
    {
      "epoch": 83.0,
      "num_input_tokens_seen": 4283464,
      "step": 1079,
      "train_accuracy": 0.041015625
    },
    {
      "epoch": 83.07692307692308,
      "grad_norm": 1.4393701553344727,
      "learning_rate": 0.01,
      "loss": 4.188133716583252,
      "num_input_tokens_seen": 4287560,
      "step": 1080,
      "train_runtime": 882.4896,
      "train_tokens_per_second": 4858.482
    },
    {
      "epoch": 83.07692307692308,
      "num_input_tokens_seen": 4287560,
      "step": 1080,
      "train_accuracy": 0.052734375
    },
    {
      "epoch": 83.15384615384616,
      "grad_norm": 1.2249149084091187,
      "learning_rate": 0.01,
      "loss": 4.136630058288574,
      "num_input_tokens_seen": 4291656,
      "step": 1081,
      "train_runtime": 883.3058,
      "train_tokens_per_second": 4858.63
    },
    {
      "epoch": 83.15384615384616,
      "num_input_tokens_seen": 4291656,
      "step": 1081,
      "train_accuracy": 0.037109375
    },
    {
      "epoch": 83.23076923076923,
      "grad_norm": 0.4371238350868225,
      "learning_rate": 0.01,
      "loss": 4.125578880310059,
      "num_input_tokens_seen": 4295752,
      "step": 1082,
      "train_runtime": 884.1252,
      "train_tokens_per_second": 4858.76
    },
    {
      "epoch": 83.23076923076923,
      "num_input_tokens_seen": 4295752,
      "step": 1082,
      "train_accuracy": 0.037109375
    },
    {
      "epoch": 83.3076923076923,
      "grad_norm": 1.1341203451156616,
      "learning_rate": 0.01,
      "loss": 4.162911415100098,
      "num_input_tokens_seen": 4299848,
      "step": 1083,
      "train_runtime": 884.9411,
      "train_tokens_per_second": 4858.908
    },
    {
      "epoch": 83.3076923076923,
      "num_input_tokens_seen": 4299848,
      "step": 1083,
      "train_accuracy": 0.0390625
    },
    {
      "epoch": 83.38461538461539,
      "grad_norm": 1.4623558521270752,
      "learning_rate": 0.01,
      "loss": 4.171051979064941,
      "num_input_tokens_seen": 4303944,
      "step": 1084,
      "train_runtime": 885.7585,
      "train_tokens_per_second": 4859.049
    },
    {
      "epoch": 83.38461538461539,
      "num_input_tokens_seen": 4303944,
      "step": 1084,
      "train_accuracy": 0.03125
    },
    {
      "epoch": 83.46153846153847,
      "grad_norm": 1.3158528804779053,
      "learning_rate": 0.01,
      "loss": 4.242735862731934,
      "num_input_tokens_seen": 4308040,
      "step": 1085,
      "train_runtime": 886.5758,
      "train_tokens_per_second": 4859.19
    },
    {
      "epoch": 83.46153846153847,
      "num_input_tokens_seen": 4308040,
      "step": 1085,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 83.53846153846153,
      "grad_norm": 0.6823511123657227,
      "learning_rate": 0.01,
      "loss": 4.177392959594727,
      "num_input_tokens_seen": 4312136,
      "step": 1086,
      "train_runtime": 887.3927,
      "train_tokens_per_second": 4859.332
    },
    {
      "epoch": 83.53846153846153,
      "num_input_tokens_seen": 4312136,
      "step": 1086,
      "train_accuracy": 0.048828125
    },
    {
      "epoch": 83.61538461538461,
      "grad_norm": 0.42744797468185425,
      "learning_rate": 0.01,
      "loss": 4.140125751495361,
      "num_input_tokens_seen": 4316232,
      "step": 1087,
      "train_runtime": 888.2133,
      "train_tokens_per_second": 4859.454
    },
    {
      "epoch": 83.61538461538461,
      "num_input_tokens_seen": 4316232,
      "step": 1087,
      "train_accuracy": 0.03125
    },
    {
      "epoch": 83.6923076923077,
      "grad_norm": 1.4262641668319702,
      "learning_rate": 0.01,
      "loss": 4.256410121917725,
      "num_input_tokens_seen": 4320328,
      "step": 1088,
      "train_runtime": 889.0304,
      "train_tokens_per_second": 4859.595
    },
    {
      "epoch": 83.6923076923077,
      "num_input_tokens_seen": 4320328,
      "step": 1088,
      "train_accuracy": 0.0546875
    },
    {
      "epoch": 83.76923076923077,
      "grad_norm": 0.8702552914619446,
      "learning_rate": 0.01,
      "loss": 4.101435661315918,
      "num_input_tokens_seen": 4324424,
      "step": 1089,
      "train_runtime": 889.8488,
      "train_tokens_per_second": 4859.729
    },
    {
      "epoch": 83.76923076923077,
      "num_input_tokens_seen": 4324424,
      "step": 1089,
      "train_accuracy": 0.03125
    },
    {
      "epoch": 83.84615384615384,
      "grad_norm": 1.1348586082458496,
      "learning_rate": 0.01,
      "loss": 4.187271595001221,
      "num_input_tokens_seen": 4328520,
      "step": 1090,
      "train_runtime": 890.6691,
      "train_tokens_per_second": 4859.852
    },
    {
      "epoch": 83.84615384615384,
      "num_input_tokens_seen": 4328520,
      "step": 1090,
      "train_accuracy": 0.0390625
    },
    {
      "epoch": 83.92307692307692,
      "grad_norm": 1.5674711465835571,
      "learning_rate": 0.01,
      "loss": 4.217589378356934,
      "num_input_tokens_seen": 4332616,
      "step": 1091,
      "train_runtime": 891.4819,
      "train_tokens_per_second": 4860.016
    },
    {
      "epoch": 83.92307692307692,
      "num_input_tokens_seen": 4332616,
      "step": 1091,
      "train_accuracy": 0.03583061695098877
    },
    {
      "epoch": 84.0,
      "grad_norm": 0.7583575248718262,
      "learning_rate": 0.01,
      "loss": 4.197638511657715,
      "num_input_tokens_seen": 4335072,
      "step": 1092,
      "train_runtime": 891.9986,
      "train_tokens_per_second": 4859.954
    },
    {
      "epoch": 84.0,
      "num_input_tokens_seen": 4335072,
      "step": 1092,
      "train_accuracy": 0.044921875
    },
    {
      "epoch": 84.07692307692308,
      "grad_norm": 0.5050715804100037,
      "learning_rate": 0.01,
      "loss": 4.155641555786133,
      "num_input_tokens_seen": 4339168,
      "step": 1093,
      "train_runtime": 892.8272,
      "train_tokens_per_second": 4860.031
    },
    {
      "epoch": 84.07692307692308,
      "num_input_tokens_seen": 4339168,
      "step": 1093,
      "train_accuracy": 0.041015625
    },
    {
      "epoch": 84.15384615384616,
      "grad_norm": 0.6129341721534729,
      "learning_rate": 0.01,
      "loss": 4.215075492858887,
      "num_input_tokens_seen": 4343264,
      "step": 1094,
      "train_runtime": 893.6472,
      "train_tokens_per_second": 4860.155
    },
    {
      "epoch": 84.15384615384616,
      "num_input_tokens_seen": 4343264,
      "step": 1094,
      "train_accuracy": 0.03515625
    },
    {
      "epoch": 84.23076923076923,
      "grad_norm": 0.6293500065803528,
      "learning_rate": 0.01,
      "loss": 4.187466621398926,
      "num_input_tokens_seen": 4347360,
      "step": 1095,
      "train_runtime": 894.4682,
      "train_tokens_per_second": 4860.274
    },
    {
      "epoch": 84.23076923076923,
      "num_input_tokens_seen": 4347360,
      "step": 1095,
      "train_accuracy": 0.037109375
    },
    {
      "epoch": 84.3076923076923,
      "grad_norm": 1.3071644306182861,
      "learning_rate": 0.01,
      "loss": 4.187564373016357,
      "num_input_tokens_seen": 4351456,
      "step": 1096,
      "train_runtime": 895.2857,
      "train_tokens_per_second": 4860.411
    },
    {
      "epoch": 84.3076923076923,
      "num_input_tokens_seen": 4351456,
      "step": 1096,
      "train_accuracy": 0.03515625
    },
    {
      "epoch": 84.38461538461539,
      "grad_norm": 1.0470569133758545,
      "learning_rate": 0.01,
      "loss": 4.172420501708984,
      "num_input_tokens_seen": 4355552,
      "step": 1097,
      "train_runtime": 896.1034,
      "train_tokens_per_second": 4860.546
    },
    {
      "epoch": 84.38461538461539,
      "num_input_tokens_seen": 4355552,
      "step": 1097,
      "train_accuracy": 0.03515625
    },
    {
      "epoch": 84.46153846153847,
      "grad_norm": 0.9348244071006775,
      "learning_rate": 0.01,
      "loss": 4.170975685119629,
      "num_input_tokens_seen": 4359648,
      "step": 1098,
      "train_runtime": 896.9221,
      "train_tokens_per_second": 4860.676
    },
    {
      "epoch": 84.46153846153847,
      "num_input_tokens_seen": 4359648,
      "step": 1098,
      "train_accuracy": 0.0546875
    },
    {
      "epoch": 84.53846153846153,
      "grad_norm": 0.9920463562011719,
      "learning_rate": 0.01,
      "loss": 4.151240348815918,
      "num_input_tokens_seen": 4363744,
      "step": 1099,
      "train_runtime": 897.7409,
      "train_tokens_per_second": 4860.806
    },
    {
      "epoch": 84.53846153846153,
      "num_input_tokens_seen": 4363744,
      "step": 1099,
      "train_accuracy": 0.046875
    },
    {
      "epoch": 84.61538461538461,
      "grad_norm": 1.1034128665924072,
      "learning_rate": 0.01,
      "loss": 4.1717023849487305,
      "num_input_tokens_seen": 4367840,
      "step": 1100,
      "train_runtime": 898.5598,
      "train_tokens_per_second": 4860.934
    },
    {
      "epoch": 84.61538461538461,
      "eval_CMD_3_branch_eval_accuracy": 0.02061482820976492,
      "eval_CMD_3_branch_eval_loss": 4.35856294631958,
      "eval_CMD_3_branch_eval_runtime": 1.1583,
      "eval_CMD_3_branch_eval_samples_per_second": 2387.192,
      "eval_CMD_3_branch_eval_steps_per_second": 5.18,
      "num_input_tokens_seen": 4367840,
      "step": 1100
    },
    {
      "epoch": 84.61538461538461,
      "num_input_tokens_seen": 4367840,
      "step": 1100,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 84.6923076923077,
      "grad_norm": 0.9364305138587952,
      "learning_rate": 0.01,
      "loss": 4.090972900390625,
      "num_input_tokens_seen": 4371936,
      "step": 1101,
      "train_runtime": 900.5399,
      "train_tokens_per_second": 4854.794
    },
    {
      "epoch": 84.6923076923077,
      "num_input_tokens_seen": 4371936,
      "step": 1101,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 84.76923076923077,
      "grad_norm": 0.5586864948272705,
      "learning_rate": 0.01,
      "loss": 4.2509331703186035,
      "num_input_tokens_seen": 4376032,
      "step": 1102,
      "train_runtime": 901.3592,
      "train_tokens_per_second": 4854.926
    },
    {
      "epoch": 84.76923076923077,
      "num_input_tokens_seen": 4376032,
      "step": 1102,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 84.84615384615384,
      "grad_norm": 4.886346817016602,
      "learning_rate": 0.01,
      "loss": 4.229729652404785,
      "num_input_tokens_seen": 4380128,
      "step": 1103,
      "train_runtime": 902.1759,
      "train_tokens_per_second": 4855.071
    },
    {
      "epoch": 84.84615384615384,
      "num_input_tokens_seen": 4380128,
      "step": 1103,
      "train_accuracy": 0.03125
    },
    {
      "epoch": 84.92307692307692,
      "grad_norm": 0.8539766669273376,
      "learning_rate": 0.01,
      "loss": 4.093820571899414,
      "num_input_tokens_seen": 4384224,
      "step": 1104,
      "train_runtime": 902.9887,
      "train_tokens_per_second": 4855.237
    },
    {
      "epoch": 84.92307692307692,
      "num_input_tokens_seen": 4384224,
      "step": 1104,
      "train_accuracy": 0.03908794745802879
    },
    {
      "epoch": 85.0,
      "grad_norm": 0.5139070749282837,
      "learning_rate": 0.01,
      "loss": 4.261080265045166,
      "num_input_tokens_seen": 4386680,
      "step": 1105,
      "train_runtime": 903.506,
      "train_tokens_per_second": 4855.175
    },
    {
      "epoch": 85.0,
      "num_input_tokens_seen": 4386680,
      "step": 1105,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 85.07692307692308,
      "grad_norm": 1.144660234451294,
      "learning_rate": 0.01,
      "loss": 4.166398048400879,
      "num_input_tokens_seen": 4390776,
      "step": 1106,
      "train_runtime": 904.3353,
      "train_tokens_per_second": 4855.252
    },
    {
      "epoch": 85.07692307692308,
      "num_input_tokens_seen": 4390776,
      "step": 1106,
      "train_accuracy": 0.0390625
    },
    {
      "epoch": 85.15384615384616,
      "grad_norm": 0.8064528107643127,
      "learning_rate": 0.01,
      "loss": 4.142458438873291,
      "num_input_tokens_seen": 4394872,
      "step": 1107,
      "train_runtime": 905.1539,
      "train_tokens_per_second": 4855.387
    },
    {
      "epoch": 85.15384615384616,
      "num_input_tokens_seen": 4394872,
      "step": 1107,
      "train_accuracy": 0.041015625
    },
    {
      "epoch": 85.23076923076923,
      "grad_norm": 2.6371541023254395,
      "learning_rate": 0.01,
      "loss": 4.138216972351074,
      "num_input_tokens_seen": 4398968,
      "step": 1108,
      "train_runtime": 905.9709,
      "train_tokens_per_second": 4855.529
    },
    {
      "epoch": 85.23076923076923,
      "num_input_tokens_seen": 4398968,
      "step": 1108,
      "train_accuracy": 0.044921875
    },
    {
      "epoch": 85.3076923076923,
      "grad_norm": 0.7484170794487,
      "learning_rate": 0.01,
      "loss": 4.101762294769287,
      "num_input_tokens_seen": 4403064,
      "step": 1109,
      "train_runtime": 906.7888,
      "train_tokens_per_second": 4855.666
    },
    {
      "epoch": 85.3076923076923,
      "num_input_tokens_seen": 4403064,
      "step": 1109,
      "train_accuracy": 0.041015625
    },
    {
      "epoch": 85.38461538461539,
      "grad_norm": 1.0887620449066162,
      "learning_rate": 0.01,
      "loss": 4.148985385894775,
      "num_input_tokens_seen": 4407160,
      "step": 1110,
      "train_runtime": 907.6066,
      "train_tokens_per_second": 4855.804
    },
    {
      "epoch": 85.38461538461539,
      "num_input_tokens_seen": 4407160,
      "step": 1110,
      "train_accuracy": 0.03515625
    },
    {
      "epoch": 85.46153846153847,
      "grad_norm": 0.7943981885910034,
      "learning_rate": 0.01,
      "loss": 4.156397819519043,
      "num_input_tokens_seen": 4411256,
      "step": 1111,
      "train_runtime": 908.4247,
      "train_tokens_per_second": 4855.94
    },
    {
      "epoch": 85.46153846153847,
      "num_input_tokens_seen": 4411256,
      "step": 1111,
      "train_accuracy": 0.048828125
    },
    {
      "epoch": 85.53846153846153,
      "grad_norm": 0.9102615118026733,
      "learning_rate": 0.01,
      "loss": 4.130383491516113,
      "num_input_tokens_seen": 4415352,
      "step": 1112,
      "train_runtime": 909.2433,
      "train_tokens_per_second": 4856.073
    },
    {
      "epoch": 85.53846153846153,
      "num_input_tokens_seen": 4415352,
      "step": 1112,
      "train_accuracy": 0.048828125
    },
    {
      "epoch": 85.61538461538461,
      "grad_norm": 0.5536665916442871,
      "learning_rate": 0.01,
      "loss": 4.2096428871154785,
      "num_input_tokens_seen": 4419448,
      "step": 1113,
      "train_runtime": 910.0609,
      "train_tokens_per_second": 4856.212
    },
    {
      "epoch": 85.61538461538461,
      "num_input_tokens_seen": 4419448,
      "step": 1113,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 85.6923076923077,
      "grad_norm": 7.640713691711426,
      "learning_rate": 0.01,
      "loss": 4.195767402648926,
      "num_input_tokens_seen": 4423544,
      "step": 1114,
      "train_runtime": 910.8786,
      "train_tokens_per_second": 4856.348
    },
    {
      "epoch": 85.6923076923077,
      "num_input_tokens_seen": 4423544,
      "step": 1114,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 85.76923076923077,
      "grad_norm": 0.5668033361434937,
      "learning_rate": 0.01,
      "loss": 4.197161674499512,
      "num_input_tokens_seen": 4427640,
      "step": 1115,
      "train_runtime": 911.6979,
      "train_tokens_per_second": 4856.477
    },
    {
      "epoch": 85.76923076923077,
      "num_input_tokens_seen": 4427640,
      "step": 1115,
      "train_accuracy": 0.041015625
    },
    {
      "epoch": 85.84615384615384,
      "grad_norm": 1.2391166687011719,
      "learning_rate": 0.01,
      "loss": 4.19352912902832,
      "num_input_tokens_seen": 4431736,
      "step": 1116,
      "train_runtime": 912.5156,
      "train_tokens_per_second": 4856.614
    },
    {
      "epoch": 85.84615384615384,
      "num_input_tokens_seen": 4431736,
      "step": 1116,
      "train_accuracy": 0.0546875
    },
    {
      "epoch": 85.92307692307692,
      "grad_norm": 0.5270545482635498,
      "learning_rate": 0.01,
      "loss": 4.185867786407471,
      "num_input_tokens_seen": 4435832,
      "step": 1117,
      "train_runtime": 913.3308,
      "train_tokens_per_second": 4856.764
    },
    {
      "epoch": 85.92307692307692,
      "num_input_tokens_seen": 4435832,
      "step": 1117,
      "train_accuracy": 0.02931595966219902
    },
    {
      "epoch": 86.0,
      "grad_norm": 0.6924501657485962,
      "learning_rate": 0.01,
      "loss": 4.22587251663208,
      "num_input_tokens_seen": 4438288,
      "step": 1118,
      "train_runtime": 913.8483,
      "train_tokens_per_second": 4856.701
    },
    {
      "epoch": 86.0,
      "num_input_tokens_seen": 4438288,
      "step": 1118,
      "train_accuracy": 0.056640625
    },
    {
      "epoch": 86.07692307692308,
      "grad_norm": 1.070239543914795,
      "learning_rate": 0.01,
      "loss": 4.100295543670654,
      "num_input_tokens_seen": 4442384,
      "step": 1119,
      "train_runtime": 914.6796,
      "train_tokens_per_second": 4856.765
    },
    {
      "epoch": 86.07692307692308,
      "num_input_tokens_seen": 4442384,
      "step": 1119,
      "train_accuracy": 0.03515625
    },
    {
      "epoch": 86.15384615384616,
      "grad_norm": 1.8627127408981323,
      "learning_rate": 0.01,
      "loss": 4.123768329620361,
      "num_input_tokens_seen": 4446480,
      "step": 1120,
      "train_runtime": 915.4975,
      "train_tokens_per_second": 4856.9
    },
    {
      "epoch": 86.15384615384616,
      "num_input_tokens_seen": 4446480,
      "step": 1120,
      "train_accuracy": 0.05078125
    },
    {
      "epoch": 86.23076923076923,
      "grad_norm": 1.2123548984527588,
      "learning_rate": 0.01,
      "loss": 4.242486953735352,
      "num_input_tokens_seen": 4450576,
      "step": 1121,
      "train_runtime": 916.3143,
      "train_tokens_per_second": 4857.041
    },
    {
      "epoch": 86.23076923076923,
      "num_input_tokens_seen": 4450576,
      "step": 1121,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 86.3076923076923,
      "grad_norm": 0.9519500732421875,
      "learning_rate": 0.01,
      "loss": 4.170391082763672,
      "num_input_tokens_seen": 4454672,
      "step": 1122,
      "train_runtime": 917.1319,
      "train_tokens_per_second": 4857.177
    },
    {
      "epoch": 86.3076923076923,
      "num_input_tokens_seen": 4454672,
      "step": 1122,
      "train_accuracy": 0.03515625
    },
    {
      "epoch": 86.38461538461539,
      "grad_norm": 1.4011510610580444,
      "learning_rate": 0.01,
      "loss": 4.253750801086426,
      "num_input_tokens_seen": 4458768,
      "step": 1123,
      "train_runtime": 917.9483,
      "train_tokens_per_second": 4857.319
    },
    {
      "epoch": 86.38461538461539,
      "num_input_tokens_seen": 4458768,
      "step": 1123,
      "train_accuracy": 0.04296875
    },
    {
      "epoch": 86.46153846153847,
      "grad_norm": 1.1189138889312744,
      "learning_rate": 0.01,
      "loss": 4.405412673950195,
      "num_input_tokens_seen": 4462864,
      "step": 1124,
      "train_runtime": 918.765,
      "train_tokens_per_second": 4857.46
    },
    {
      "epoch": 86.46153846153847,
      "num_input_tokens_seen": 4462864,
      "step": 1124,
      "train_accuracy": 0.03515625
    },
    {
      "epoch": 86.53846153846153,
      "grad_norm": 11.288074493408203,
      "learning_rate": 0.01,
      "loss": 4.227231979370117,
      "num_input_tokens_seen": 4466960,
      "step": 1125,
      "train_runtime": 919.5836,
      "train_tokens_per_second": 4857.59
    },
    {
      "epoch": 86.53846153846153,
      "num_input_tokens_seen": 4466960,
      "step": 1125,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 86.61538461538461,
      "grad_norm": 0.8495609760284424,
      "learning_rate": 0.01,
      "loss": 4.3608198165893555,
      "num_input_tokens_seen": 4471056,
      "step": 1126,
      "train_runtime": 920.4005,
      "train_tokens_per_second": 4857.729
    },
    {
      "epoch": 86.61538461538461,
      "num_input_tokens_seen": 4471056,
      "step": 1126,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 86.6923076923077,
      "grad_norm": 0.7331825494766235,
      "learning_rate": 0.01,
      "loss": 4.32719087600708,
      "num_input_tokens_seen": 4475152,
      "step": 1127,
      "train_runtime": 921.2179,
      "train_tokens_per_second": 4857.865
    },
    {
      "epoch": 86.6923076923077,
      "num_input_tokens_seen": 4475152,
      "step": 1127,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 86.76923076923077,
      "grad_norm": 0.7106150388717651,
      "learning_rate": 0.01,
      "loss": 4.3882036209106445,
      "num_input_tokens_seen": 4479248,
      "step": 1128,
      "train_runtime": 922.0346,
      "train_tokens_per_second": 4858.004
    },
    {
      "epoch": 86.76923076923077,
      "num_input_tokens_seen": 4479248,
      "step": 1128,
      "train_accuracy": 0.03515625
    },
    {
      "epoch": 86.84615384615384,
      "grad_norm": 0.9769771099090576,
      "learning_rate": 0.01,
      "loss": 4.429689884185791,
      "num_input_tokens_seen": 4483344,
      "step": 1129,
      "train_runtime": 922.853,
      "train_tokens_per_second": 4858.134
    },
    {
      "epoch": 86.84615384615384,
      "num_input_tokens_seen": 4483344,
      "step": 1129,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 86.92307692307692,
      "grad_norm": 0.6609678268432617,
      "learning_rate": 0.01,
      "loss": 4.388575553894043,
      "num_input_tokens_seen": 4487440,
      "step": 1130,
      "train_runtime": 923.6671,
      "train_tokens_per_second": 4858.287
    },
    {
      "epoch": 86.92307692307692,
      "num_input_tokens_seen": 4487440,
      "step": 1130,
      "train_accuracy": 0.016286645084619522
    },
    {
      "epoch": 87.0,
      "grad_norm": 0.6923887133598328,
      "learning_rate": 0.01,
      "loss": 4.417357444763184,
      "num_input_tokens_seen": 4489896,
      "step": 1131,
      "train_runtime": 924.1832,
      "train_tokens_per_second": 4858.232
    },
    {
      "epoch": 87.0,
      "num_input_tokens_seen": 4489896,
      "step": 1131,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 87.07692307692308,
      "grad_norm": 2.304231643676758,
      "learning_rate": 0.01,
      "loss": 4.336197853088379,
      "num_input_tokens_seen": 4493992,
      "step": 1132,
      "train_runtime": 925.0116,
      "train_tokens_per_second": 4858.309
    },
    {
      "epoch": 87.07692307692308,
      "num_input_tokens_seen": 4493992,
      "step": 1132,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 87.15384615384616,
      "grad_norm": 1.368730068206787,
      "learning_rate": 0.01,
      "loss": 4.309045791625977,
      "num_input_tokens_seen": 4498088,
      "step": 1133,
      "train_runtime": 925.8289,
      "train_tokens_per_second": 4858.444
    },
    {
      "epoch": 87.15384615384616,
      "num_input_tokens_seen": 4498088,
      "step": 1133,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 87.23076923076923,
      "grad_norm": 1.7913775444030762,
      "learning_rate": 0.01,
      "loss": 4.4043869972229,
      "num_input_tokens_seen": 4502184,
      "step": 1134,
      "train_runtime": 926.647,
      "train_tokens_per_second": 4858.575
    },
    {
      "epoch": 87.23076923076923,
      "num_input_tokens_seen": 4502184,
      "step": 1134,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 87.3076923076923,
      "grad_norm": 2.1816532611846924,
      "learning_rate": 0.01,
      "loss": 4.301054954528809,
      "num_input_tokens_seen": 4506280,
      "step": 1135,
      "train_runtime": 927.4643,
      "train_tokens_per_second": 4858.71
    },
    {
      "epoch": 87.3076923076923,
      "num_input_tokens_seen": 4506280,
      "step": 1135,
      "train_accuracy": 0.0390625
    },
    {
      "epoch": 87.38461538461539,
      "grad_norm": 0.8795368075370789,
      "learning_rate": 0.01,
      "loss": 4.266789436340332,
      "num_input_tokens_seen": 4510376,
      "step": 1136,
      "train_runtime": 928.282,
      "train_tokens_per_second": 4858.843
    },
    {
      "epoch": 87.38461538461539,
      "num_input_tokens_seen": 4510376,
      "step": 1136,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 87.46153846153847,
      "grad_norm": 0.6091873645782471,
      "learning_rate": 0.01,
      "loss": 4.291828155517578,
      "num_input_tokens_seen": 4514472,
      "step": 1137,
      "train_runtime": 929.0994,
      "train_tokens_per_second": 4858.976
    },
    {
      "epoch": 87.46153846153847,
      "num_input_tokens_seen": 4514472,
      "step": 1137,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 87.53846153846153,
      "grad_norm": 0.9331724643707275,
      "learning_rate": 0.01,
      "loss": 4.287989139556885,
      "num_input_tokens_seen": 4518568,
      "step": 1138,
      "train_runtime": 929.9162,
      "train_tokens_per_second": 4859.113
    },
    {
      "epoch": 87.53846153846153,
      "num_input_tokens_seen": 4518568,
      "step": 1138,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 87.61538461538461,
      "grad_norm": 2.546476125717163,
      "learning_rate": 0.01,
      "loss": 4.350374221801758,
      "num_input_tokens_seen": 4522664,
      "step": 1139,
      "train_runtime": 930.7331,
      "train_tokens_per_second": 4859.249
    },
    {
      "epoch": 87.61538461538461,
      "num_input_tokens_seen": 4522664,
      "step": 1139,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 87.6923076923077,
      "grad_norm": 3.0142698287963867,
      "learning_rate": 0.01,
      "loss": 4.275198459625244,
      "num_input_tokens_seen": 4526760,
      "step": 1140,
      "train_runtime": 931.5505,
      "train_tokens_per_second": 4859.382
    },
    {
      "epoch": 87.6923076923077,
      "num_input_tokens_seen": 4526760,
      "step": 1140,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 87.76923076923077,
      "grad_norm": 0.6362144351005554,
      "learning_rate": 0.01,
      "loss": 4.3212080001831055,
      "num_input_tokens_seen": 4530856,
      "step": 1141,
      "train_runtime": 932.3678,
      "train_tokens_per_second": 4859.516
    },
    {
      "epoch": 87.76923076923077,
      "num_input_tokens_seen": 4530856,
      "step": 1141,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 87.84615384615384,
      "grad_norm": 0.9679552316665649,
      "learning_rate": 0.01,
      "loss": 4.337464332580566,
      "num_input_tokens_seen": 4534952,
      "step": 1142,
      "train_runtime": 933.1857,
      "train_tokens_per_second": 4859.646
    },
    {
      "epoch": 87.84615384615384,
      "num_input_tokens_seen": 4534952,
      "step": 1142,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 87.92307692307692,
      "grad_norm": 0.4788326621055603,
      "learning_rate": 0.01,
      "loss": 4.329340934753418,
      "num_input_tokens_seen": 4539048,
      "step": 1143,
      "train_runtime": 933.9975,
      "train_tokens_per_second": 4859.808
    },
    {
      "epoch": 87.92307692307692,
      "num_input_tokens_seen": 4539048,
      "step": 1143,
      "train_accuracy": 0.03583061695098877
    },
    {
      "epoch": 88.0,
      "grad_norm": 0.878277063369751,
      "learning_rate": 0.01,
      "loss": 4.35450553894043,
      "num_input_tokens_seen": 4541504,
      "step": 1144,
      "train_runtime": 934.5138,
      "train_tokens_per_second": 4859.751
    },
    {
      "epoch": 88.0,
      "num_input_tokens_seen": 4541504,
      "step": 1144,
      "train_accuracy": 0.037109375
    },
    {
      "epoch": 88.07692307692308,
      "grad_norm": 1.1310226917266846,
      "learning_rate": 0.01,
      "loss": 4.281495094299316,
      "num_input_tokens_seen": 4545600,
      "step": 1145,
      "train_runtime": 935.3421,
      "train_tokens_per_second": 4859.826
    },
    {
      "epoch": 88.07692307692308,
      "num_input_tokens_seen": 4545600,
      "step": 1145,
      "train_accuracy": 0.05078125
    },
    {
      "epoch": 88.15384615384616,
      "grad_norm": 0.765187680721283,
      "learning_rate": 0.01,
      "loss": 4.2237701416015625,
      "num_input_tokens_seen": 4549696,
      "step": 1146,
      "train_runtime": 936.16,
      "train_tokens_per_second": 4859.955
    },
    {
      "epoch": 88.15384615384616,
      "num_input_tokens_seen": 4549696,
      "step": 1146,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 88.23076923076923,
      "grad_norm": 0.8071278929710388,
      "learning_rate": 0.01,
      "loss": 4.272477149963379,
      "num_input_tokens_seen": 4553792,
      "step": 1147,
      "train_runtime": 936.9765,
      "train_tokens_per_second": 4860.092
    },
    {
      "epoch": 88.23076923076923,
      "num_input_tokens_seen": 4553792,
      "step": 1147,
      "train_accuracy": 0.03125
    },
    {
      "epoch": 88.3076923076923,
      "grad_norm": 0.5325653553009033,
      "learning_rate": 0.01,
      "loss": 4.258377552032471,
      "num_input_tokens_seen": 4557888,
      "step": 1148,
      "train_runtime": 937.7933,
      "train_tokens_per_second": 4860.227
    },
    {
      "epoch": 88.3076923076923,
      "num_input_tokens_seen": 4557888,
      "step": 1148,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 88.38461538461539,
      "grad_norm": 1.1729495525360107,
      "learning_rate": 0.01,
      "loss": 4.319500923156738,
      "num_input_tokens_seen": 4561984,
      "step": 1149,
      "train_runtime": 938.6102,
      "train_tokens_per_second": 4860.361
    },
    {
      "epoch": 88.38461538461539,
      "num_input_tokens_seen": 4561984,
      "step": 1149,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 88.46153846153847,
      "grad_norm": 0.8045989274978638,
      "learning_rate": 0.01,
      "loss": 4.301680564880371,
      "num_input_tokens_seen": 4566080,
      "step": 1150,
      "train_runtime": 939.4264,
      "train_tokens_per_second": 4860.498
    },
    {
      "epoch": 88.46153846153847,
      "num_input_tokens_seen": 4566080,
      "step": 1150,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 88.53846153846153,
      "grad_norm": 0.7991436719894409,
      "learning_rate": 0.01,
      "loss": 4.305707931518555,
      "num_input_tokens_seen": 4570176,
      "step": 1151,
      "train_runtime": 940.2443,
      "train_tokens_per_second": 4860.626
    },
    {
      "epoch": 88.53846153846153,
      "num_input_tokens_seen": 4570176,
      "step": 1151,
      "train_accuracy": 0.03125
    },
    {
      "epoch": 88.61538461538461,
      "grad_norm": 1.3253687620162964,
      "learning_rate": 0.01,
      "loss": 4.271323204040527,
      "num_input_tokens_seen": 4574272,
      "step": 1152,
      "train_runtime": 941.0624,
      "train_tokens_per_second": 4860.753
    },
    {
      "epoch": 88.61538461538461,
      "num_input_tokens_seen": 4574272,
      "step": 1152,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 88.6923076923077,
      "grad_norm": 1.1081411838531494,
      "learning_rate": 0.01,
      "loss": 4.294156074523926,
      "num_input_tokens_seen": 4578368,
      "step": 1153,
      "train_runtime": 941.8836,
      "train_tokens_per_second": 4860.864
    },
    {
      "epoch": 88.6923076923077,
      "num_input_tokens_seen": 4578368,
      "step": 1153,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 88.76923076923077,
      "grad_norm": 0.7775978446006775,
      "learning_rate": 0.01,
      "loss": 4.2814836502075195,
      "num_input_tokens_seen": 4582464,
      "step": 1154,
      "train_runtime": 942.7009,
      "train_tokens_per_second": 4860.995
    },
    {
      "epoch": 88.76923076923077,
      "num_input_tokens_seen": 4582464,
      "step": 1154,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 88.84615384615384,
      "grad_norm": 1.447250247001648,
      "learning_rate": 0.01,
      "loss": 4.294820785522461,
      "num_input_tokens_seen": 4586560,
      "step": 1155,
      "train_runtime": 943.5175,
      "train_tokens_per_second": 4861.128
    },
    {
      "epoch": 88.84615384615384,
      "num_input_tokens_seen": 4586560,
      "step": 1155,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 88.92307692307692,
      "grad_norm": 0.6836444139480591,
      "learning_rate": 0.01,
      "loss": 4.362523078918457,
      "num_input_tokens_seen": 4590656,
      "step": 1156,
      "train_runtime": 944.3298,
      "train_tokens_per_second": 4861.285
    },
    {
      "epoch": 88.92307692307692,
      "num_input_tokens_seen": 4590656,
      "step": 1156,
      "train_accuracy": 0.02931595966219902
    },
    {
      "epoch": 89.0,
      "grad_norm": 1.3853188753128052,
      "learning_rate": 0.01,
      "loss": 4.322265148162842,
      "num_input_tokens_seen": 4593112,
      "step": 1157,
      "train_runtime": 944.8459,
      "train_tokens_per_second": 4861.229
    },
    {
      "epoch": 89.0,
      "num_input_tokens_seen": 4593112,
      "step": 1157,
      "train_accuracy": 0.037109375
    },
    {
      "epoch": 89.07692307692308,
      "grad_norm": 1.3838152885437012,
      "learning_rate": 0.01,
      "loss": 4.244259834289551,
      "num_input_tokens_seen": 4597208,
      "step": 1158,
      "train_runtime": 945.6807,
      "train_tokens_per_second": 4861.269
    },
    {
      "epoch": 89.07692307692308,
      "num_input_tokens_seen": 4597208,
      "step": 1158,
      "train_accuracy": 0.0390625
    },
    {
      "epoch": 89.15384615384616,
      "grad_norm": 3.4085545539855957,
      "learning_rate": 0.01,
      "loss": 4.246095180511475,
      "num_input_tokens_seen": 4601304,
      "step": 1159,
      "train_runtime": 946.4989,
      "train_tokens_per_second": 4861.394
    },
    {
      "epoch": 89.15384615384616,
      "num_input_tokens_seen": 4601304,
      "step": 1159,
      "train_accuracy": 0.03515625
    },
    {
      "epoch": 89.23076923076923,
      "grad_norm": 0.8216474652290344,
      "learning_rate": 0.01,
      "loss": 4.261091232299805,
      "num_input_tokens_seen": 4605400,
      "step": 1160,
      "train_runtime": 947.318,
      "train_tokens_per_second": 4861.514
    },
    {
      "epoch": 89.23076923076923,
      "num_input_tokens_seen": 4605400,
      "step": 1160,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 89.3076923076923,
      "grad_norm": 2.4385972023010254,
      "learning_rate": 0.01,
      "loss": 4.18310022354126,
      "num_input_tokens_seen": 4609496,
      "step": 1161,
      "train_runtime": 948.1347,
      "train_tokens_per_second": 4861.647
    },
    {
      "epoch": 89.3076923076923,
      "num_input_tokens_seen": 4609496,
      "step": 1161,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 89.38461538461539,
      "grad_norm": 0.588120698928833,
      "learning_rate": 0.01,
      "loss": 4.27022123336792,
      "num_input_tokens_seen": 4613592,
      "step": 1162,
      "train_runtime": 948.9517,
      "train_tokens_per_second": 4861.778
    },
    {
      "epoch": 89.38461538461539,
      "num_input_tokens_seen": 4613592,
      "step": 1162,
      "train_accuracy": 0.03125
    },
    {
      "epoch": 89.46153846153847,
      "grad_norm": 0.4883492588996887,
      "learning_rate": 0.01,
      "loss": 4.272211074829102,
      "num_input_tokens_seen": 4617688,
      "step": 1163,
      "train_runtime": 949.7689,
      "train_tokens_per_second": 4861.907
    },
    {
      "epoch": 89.46153846153847,
      "num_input_tokens_seen": 4617688,
      "step": 1163,
      "train_accuracy": 0.0390625
    },
    {
      "epoch": 89.53846153846153,
      "grad_norm": 2.5225794315338135,
      "learning_rate": 0.01,
      "loss": 4.269200801849365,
      "num_input_tokens_seen": 4621784,
      "step": 1164,
      "train_runtime": 950.5863,
      "train_tokens_per_second": 4862.035
    },
    {
      "epoch": 89.53846153846153,
      "num_input_tokens_seen": 4621784,
      "step": 1164,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 89.61538461538461,
      "grad_norm": 8.58218002319336,
      "learning_rate": 0.01,
      "loss": 4.250087738037109,
      "num_input_tokens_seen": 4625880,
      "step": 1165,
      "train_runtime": 951.4028,
      "train_tokens_per_second": 4862.168
    },
    {
      "epoch": 89.61538461538461,
      "num_input_tokens_seen": 4625880,
      "step": 1165,
      "train_accuracy": 0.04296875
    },
    {
      "epoch": 89.6923076923077,
      "grad_norm": 1.266677737236023,
      "learning_rate": 0.01,
      "loss": 4.277984619140625,
      "num_input_tokens_seen": 4629976,
      "step": 1166,
      "train_runtime": 952.2202,
      "train_tokens_per_second": 4862.296
    },
    {
      "epoch": 89.6923076923077,
      "num_input_tokens_seen": 4629976,
      "step": 1166,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 89.76923076923077,
      "grad_norm": 0.9045878052711487,
      "learning_rate": 0.01,
      "loss": 4.283880233764648,
      "num_input_tokens_seen": 4634072,
      "step": 1167,
      "train_runtime": 953.0376,
      "train_tokens_per_second": 4862.423
    },
    {
      "epoch": 89.76923076923077,
      "num_input_tokens_seen": 4634072,
      "step": 1167,
      "train_accuracy": 0.0390625
    },
    {
      "epoch": 89.84615384615384,
      "grad_norm": 4.231438159942627,
      "learning_rate": 0.01,
      "loss": 4.293394088745117,
      "num_input_tokens_seen": 4638168,
      "step": 1168,
      "train_runtime": 953.8557,
      "train_tokens_per_second": 4862.547
    },
    {
      "epoch": 89.84615384615384,
      "num_input_tokens_seen": 4638168,
      "step": 1168,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 89.92307692307692,
      "grad_norm": 1.1969122886657715,
      "learning_rate": 0.01,
      "loss": 4.2695770263671875,
      "num_input_tokens_seen": 4642264,
      "step": 1169,
      "train_runtime": 954.6697,
      "train_tokens_per_second": 4862.691
    },
    {
      "epoch": 89.92307692307692,
      "num_input_tokens_seen": 4642264,
      "step": 1169,
      "train_accuracy": 0.02931595966219902
    },
    {
      "epoch": 90.0,
      "grad_norm": 0.8131340742111206,
      "learning_rate": 0.01,
      "loss": 4.261406898498535,
      "num_input_tokens_seen": 4644720,
      "step": 1170,
      "train_runtime": 955.1865,
      "train_tokens_per_second": 4862.632
    },
    {
      "epoch": 90.0,
      "num_input_tokens_seen": 4644720,
      "step": 1170,
      "train_accuracy": 0.04296875
    },
    {
      "epoch": 90.07692307692308,
      "grad_norm": 0.33925577998161316,
      "learning_rate": 0.01,
      "loss": 4.161164283752441,
      "num_input_tokens_seen": 4648816,
      "step": 1171,
      "train_runtime": 956.018,
      "train_tokens_per_second": 4862.687
    },
    {
      "epoch": 90.07692307692308,
      "num_input_tokens_seen": 4648816,
      "step": 1171,
      "train_accuracy": 0.044921875
    },
    {
      "epoch": 90.15384615384616,
      "grad_norm": 0.6678487062454224,
      "learning_rate": 0.01,
      "loss": 4.15652322769165,
      "num_input_tokens_seen": 4652912,
      "step": 1172,
      "train_runtime": 956.8349,
      "train_tokens_per_second": 4862.816
    },
    {
      "epoch": 90.15384615384616,
      "num_input_tokens_seen": 4652912,
      "step": 1172,
      "train_accuracy": 0.0390625
    },
    {
      "epoch": 90.23076923076923,
      "grad_norm": 1.120121955871582,
      "learning_rate": 0.01,
      "loss": 4.300291061401367,
      "num_input_tokens_seen": 4657008,
      "step": 1173,
      "train_runtime": 957.6523,
      "train_tokens_per_second": 4862.942
    },
    {
      "epoch": 90.23076923076923,
      "num_input_tokens_seen": 4657008,
      "step": 1173,
      "train_accuracy": 0.037109375
    },
    {
      "epoch": 90.3076923076923,
      "grad_norm": 6.495781421661377,
      "learning_rate": 0.01,
      "loss": 4.2706098556518555,
      "num_input_tokens_seen": 4661104,
      "step": 1174,
      "train_runtime": 958.4728,
      "train_tokens_per_second": 4863.053
    },
    {
      "epoch": 90.3076923076923,
      "num_input_tokens_seen": 4661104,
      "step": 1174,
      "train_accuracy": 0.03515625
    },
    {
      "epoch": 90.38461538461539,
      "grad_norm": 0.7740323543548584,
      "learning_rate": 0.01,
      "loss": 4.182338714599609,
      "num_input_tokens_seen": 4665200,
      "step": 1175,
      "train_runtime": 959.2896,
      "train_tokens_per_second": 4863.182
    },
    {
      "epoch": 90.38461538461539,
      "num_input_tokens_seen": 4665200,
      "step": 1175,
      "train_accuracy": 0.04296875
    },
    {
      "epoch": 90.46153846153847,
      "grad_norm": 0.553860604763031,
      "learning_rate": 0.01,
      "loss": 4.159100532531738,
      "num_input_tokens_seen": 4669296,
      "step": 1176,
      "train_runtime": 960.1061,
      "train_tokens_per_second": 4863.313
    },
    {
      "epoch": 90.46153846153847,
      "num_input_tokens_seen": 4669296,
      "step": 1176,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 90.53846153846153,
      "grad_norm": 0.43262046575546265,
      "learning_rate": 0.01,
      "loss": 4.254860877990723,
      "num_input_tokens_seen": 4673392,
      "step": 1177,
      "train_runtime": 960.9233,
      "train_tokens_per_second": 4863.439
    },
    {
      "epoch": 90.53846153846153,
      "num_input_tokens_seen": 4673392,
      "step": 1177,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 90.61538461538461,
      "grad_norm": 3.9053070545196533,
      "learning_rate": 0.01,
      "loss": 4.224895000457764,
      "num_input_tokens_seen": 4677488,
      "step": 1178,
      "train_runtime": 961.7407,
      "train_tokens_per_second": 4863.565
    },
    {
      "epoch": 90.61538461538461,
      "num_input_tokens_seen": 4677488,
      "step": 1178,
      "train_accuracy": 0.0390625
    },
    {
      "epoch": 90.6923076923077,
      "grad_norm": 1.961185336112976,
      "learning_rate": 0.01,
      "loss": 4.247408866882324,
      "num_input_tokens_seen": 4681584,
      "step": 1179,
      "train_runtime": 962.5576,
      "train_tokens_per_second": 4863.692
    },
    {
      "epoch": 90.6923076923077,
      "num_input_tokens_seen": 4681584,
      "step": 1179,
      "train_accuracy": 0.041015625
    },
    {
      "epoch": 90.76923076923077,
      "grad_norm": 0.4231705665588379,
      "learning_rate": 0.01,
      "loss": 4.19923734664917,
      "num_input_tokens_seen": 4685680,
      "step": 1180,
      "train_runtime": 963.3754,
      "train_tokens_per_second": 4863.815
    },
    {
      "epoch": 90.76923076923077,
      "num_input_tokens_seen": 4685680,
      "step": 1180,
      "train_accuracy": 0.044921875
    },
    {
      "epoch": 90.84615384615384,
      "grad_norm": 1.0635826587677002,
      "learning_rate": 0.01,
      "loss": 4.2024383544921875,
      "num_input_tokens_seen": 4689776,
      "step": 1181,
      "train_runtime": 964.192,
      "train_tokens_per_second": 4863.944
    },
    {
      "epoch": 90.84615384615384,
      "num_input_tokens_seen": 4689776,
      "step": 1181,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 90.92307692307692,
      "grad_norm": 0.7891145944595337,
      "learning_rate": 0.01,
      "loss": 4.261574745178223,
      "num_input_tokens_seen": 4693872,
      "step": 1182,
      "train_runtime": 965.0041,
      "train_tokens_per_second": 4864.095
    },
    {
      "epoch": 90.92307692307692,
      "num_input_tokens_seen": 4693872,
      "step": 1182,
      "train_accuracy": 0.026058631017804146
    },
    {
      "epoch": 91.0,
      "grad_norm": 1.3556636571884155,
      "learning_rate": 0.01,
      "loss": 4.212470531463623,
      "num_input_tokens_seen": 4696328,
      "step": 1183,
      "train_runtime": 965.5211,
      "train_tokens_per_second": 4864.035
    },
    {
      "epoch": 91.0,
      "num_input_tokens_seen": 4696328,
      "step": 1183,
      "train_accuracy": 0.041015625
    },
    {
      "epoch": 91.07692307692308,
      "grad_norm": 0.7309383749961853,
      "learning_rate": 0.01,
      "loss": 4.167816162109375,
      "num_input_tokens_seen": 4700424,
      "step": 1184,
      "train_runtime": 966.3509,
      "train_tokens_per_second": 4864.097
    },
    {
      "epoch": 91.07692307692308,
      "num_input_tokens_seen": 4700424,
      "step": 1184,
      "train_accuracy": 0.046875
    },
    {
      "epoch": 91.15384615384616,
      "grad_norm": 0.39610621333122253,
      "learning_rate": 0.01,
      "loss": 4.186983108520508,
      "num_input_tokens_seen": 4704520,
      "step": 1185,
      "train_runtime": 967.1719,
      "train_tokens_per_second": 4864.203
    },
    {
      "epoch": 91.15384615384616,
      "num_input_tokens_seen": 4704520,
      "step": 1185,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 91.23076923076923,
      "grad_norm": 0.5263922810554504,
      "learning_rate": 0.01,
      "loss": 4.1591973304748535,
      "num_input_tokens_seen": 4708616,
      "step": 1186,
      "train_runtime": 967.9892,
      "train_tokens_per_second": 4864.327
    },
    {
      "epoch": 91.23076923076923,
      "num_input_tokens_seen": 4708616,
      "step": 1186,
      "train_accuracy": 0.037109375
    },
    {
      "epoch": 91.3076923076923,
      "grad_norm": 0.4648636281490326,
      "learning_rate": 0.01,
      "loss": 4.180056571960449,
      "num_input_tokens_seen": 4712712,
      "step": 1187,
      "train_runtime": 968.8056,
      "train_tokens_per_second": 4864.456
    },
    {
      "epoch": 91.3076923076923,
      "num_input_tokens_seen": 4712712,
      "step": 1187,
      "train_accuracy": 0.04296875
    },
    {
      "epoch": 91.38461538461539,
      "grad_norm": 1.0477451086044312,
      "learning_rate": 0.01,
      "loss": 4.164782524108887,
      "num_input_tokens_seen": 4716808,
      "step": 1188,
      "train_runtime": 969.623,
      "train_tokens_per_second": 4864.579
    },
    {
      "epoch": 91.38461538461539,
      "num_input_tokens_seen": 4716808,
      "step": 1188,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 91.46153846153847,
      "grad_norm": 0.5782698392868042,
      "learning_rate": 0.01,
      "loss": 4.202156066894531,
      "num_input_tokens_seen": 4720904,
      "step": 1189,
      "train_runtime": 970.4405,
      "train_tokens_per_second": 4864.702
    },
    {
      "epoch": 91.46153846153847,
      "num_input_tokens_seen": 4720904,
      "step": 1189,
      "train_accuracy": 0.052734375
    },
    {
      "epoch": 91.53846153846153,
      "grad_norm": 1.6004842519760132,
      "learning_rate": 0.01,
      "loss": 4.210166931152344,
      "num_input_tokens_seen": 4725000,
      "step": 1190,
      "train_runtime": 971.2583,
      "train_tokens_per_second": 4864.823
    },
    {
      "epoch": 91.53846153846153,
      "num_input_tokens_seen": 4725000,
      "step": 1190,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 91.61538461538461,
      "grad_norm": 1.2771614789962769,
      "learning_rate": 0.01,
      "loss": 4.170170307159424,
      "num_input_tokens_seen": 4729096,
      "step": 1191,
      "train_runtime": 972.0773,
      "train_tokens_per_second": 4864.938
    },
    {
      "epoch": 91.61538461538461,
      "num_input_tokens_seen": 4729096,
      "step": 1191,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 91.6923076923077,
      "grad_norm": 0.6172093152999878,
      "learning_rate": 0.01,
      "loss": 4.237349987030029,
      "num_input_tokens_seen": 4733192,
      "step": 1192,
      "train_runtime": 972.8944,
      "train_tokens_per_second": 4865.062
    },
    {
      "epoch": 91.6923076923077,
      "num_input_tokens_seen": 4733192,
      "step": 1192,
      "train_accuracy": 0.03125
    },
    {
      "epoch": 91.76923076923077,
      "grad_norm": 1.2454875707626343,
      "learning_rate": 0.01,
      "loss": 4.218174457550049,
      "num_input_tokens_seen": 4737288,
      "step": 1193,
      "train_runtime": 973.7112,
      "train_tokens_per_second": 4865.188
    },
    {
      "epoch": 91.76923076923077,
      "num_input_tokens_seen": 4737288,
      "step": 1193,
      "train_accuracy": 0.05078125
    },
    {
      "epoch": 91.84615384615384,
      "grad_norm": 0.7239828109741211,
      "learning_rate": 0.01,
      "loss": 4.134560585021973,
      "num_input_tokens_seen": 4741384,
      "step": 1194,
      "train_runtime": 974.5288,
      "train_tokens_per_second": 4865.309
    },
    {
      "epoch": 91.84615384615384,
      "num_input_tokens_seen": 4741384,
      "step": 1194,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 91.92307692307692,
      "grad_norm": 1.593626856803894,
      "learning_rate": 0.01,
      "loss": 4.236126899719238,
      "num_input_tokens_seen": 4745480,
      "step": 1195,
      "train_runtime": 975.3411,
      "train_tokens_per_second": 4865.457
    },
    {
      "epoch": 91.92307692307692,
      "num_input_tokens_seen": 4745480,
      "step": 1195,
      "train_accuracy": 0.032573290169239044
    },
    {
      "epoch": 92.0,
      "grad_norm": 1.4044294357299805,
      "learning_rate": 0.01,
      "loss": 4.203438758850098,
      "num_input_tokens_seen": 4747936,
      "step": 1196,
      "train_runtime": 975.8595,
      "train_tokens_per_second": 4865.389
    },
    {
      "epoch": 92.0,
      "num_input_tokens_seen": 4747936,
      "step": 1196,
      "train_accuracy": 0.0390625
    },
    {
      "epoch": 92.07692307692308,
      "grad_norm": 1.451888084411621,
      "learning_rate": 0.01,
      "loss": 4.142299652099609,
      "num_input_tokens_seen": 4752032,
      "step": 1197,
      "train_runtime": 976.6879,
      "train_tokens_per_second": 4865.456
    },
    {
      "epoch": 92.07692307692308,
      "num_input_tokens_seen": 4752032,
      "step": 1197,
      "train_accuracy": 0.046875
    },
    {
      "epoch": 92.15384615384616,
      "grad_norm": 0.4934719502925873,
      "learning_rate": 0.01,
      "loss": 4.190487384796143,
      "num_input_tokens_seen": 4756128,
      "step": 1198,
      "train_runtime": 977.5047,
      "train_tokens_per_second": 4865.581
    },
    {
      "epoch": 92.15384615384616,
      "num_input_tokens_seen": 4756128,
      "step": 1198,
      "train_accuracy": 0.0390625
    },
    {
      "epoch": 92.23076923076923,
      "grad_norm": 0.6478560566902161,
      "learning_rate": 0.01,
      "loss": 4.193492889404297,
      "num_input_tokens_seen": 4760224,
      "step": 1199,
      "train_runtime": 978.3219,
      "train_tokens_per_second": 4865.703
    },
    {
      "epoch": 92.23076923076923,
      "num_input_tokens_seen": 4760224,
      "step": 1199,
      "train_accuracy": 0.04296875
    },
    {
      "epoch": 92.3076923076923,
      "grad_norm": 1.190477728843689,
      "learning_rate": 0.01,
      "loss": 4.161839485168457,
      "num_input_tokens_seen": 4764320,
      "step": 1200,
      "train_runtime": 979.1392,
      "train_tokens_per_second": 4865.825
    },
    {
      "epoch": 92.3076923076923,
      "eval_CMD_3_branch_eval_accuracy": 0.019891500904159132,
      "eval_CMD_3_branch_eval_loss": 4.357153415679932,
      "eval_CMD_3_branch_eval_runtime": 1.1592,
      "eval_CMD_3_branch_eval_samples_per_second": 2385.288,
      "eval_CMD_3_branch_eval_steps_per_second": 5.176,
      "num_input_tokens_seen": 4764320,
      "step": 1200
    },
    {
      "epoch": 92.3076923076923,
      "num_input_tokens_seen": 4764320,
      "step": 1200,
      "train_accuracy": 0.04296875
    },
    {
      "epoch": 92.38461538461539,
      "grad_norm": 0.6409584879875183,
      "learning_rate": 0.01,
      "loss": 4.188936233520508,
      "num_input_tokens_seen": 4768416,
      "step": 1201,
      "train_runtime": 981.1193,
      "train_tokens_per_second": 4860.18
    },
    {
      "epoch": 92.38461538461539,
      "num_input_tokens_seen": 4768416,
      "step": 1201,
      "train_accuracy": 0.041015625
    },
    {
      "epoch": 92.46153846153847,
      "grad_norm": 0.6553785800933838,
      "learning_rate": 0.01,
      "loss": 4.207437038421631,
      "num_input_tokens_seen": 4772512,
      "step": 1202,
      "train_runtime": 981.9361,
      "train_tokens_per_second": 4860.308
    },
    {
      "epoch": 92.46153846153847,
      "num_input_tokens_seen": 4772512,
      "step": 1202,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 92.53846153846153,
      "grad_norm": 0.4285067915916443,
      "learning_rate": 0.01,
      "loss": 4.174130916595459,
      "num_input_tokens_seen": 4776608,
      "step": 1203,
      "train_runtime": 982.7549,
      "train_tokens_per_second": 4860.427
    },
    {
      "epoch": 92.53846153846153,
      "num_input_tokens_seen": 4776608,
      "step": 1203,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 92.61538461538461,
      "grad_norm": 0.7696940898895264,
      "learning_rate": 0.01,
      "loss": 4.180700302124023,
      "num_input_tokens_seen": 4780704,
      "step": 1204,
      "train_runtime": 983.5726,
      "train_tokens_per_second": 4860.55
    },
    {
      "epoch": 92.61538461538461,
      "num_input_tokens_seen": 4780704,
      "step": 1204,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 92.6923076923077,
      "grad_norm": 0.4185137152671814,
      "learning_rate": 0.01,
      "loss": 4.337588310241699,
      "num_input_tokens_seen": 4784800,
      "step": 1205,
      "train_runtime": 984.3898,
      "train_tokens_per_second": 4860.676
    },
    {
      "epoch": 92.6923076923077,
      "num_input_tokens_seen": 4784800,
      "step": 1205,
      "train_accuracy": 0.03515625
    },
    {
      "epoch": 92.76923076923077,
      "grad_norm": 0.5836478471755981,
      "learning_rate": 0.01,
      "loss": 4.2482075691223145,
      "num_input_tokens_seen": 4788896,
      "step": 1206,
      "train_runtime": 985.2067,
      "train_tokens_per_second": 4860.803
    },
    {
      "epoch": 92.76923076923077,
      "num_input_tokens_seen": 4788896,
      "step": 1206,
      "train_accuracy": 0.0390625
    },
    {
      "epoch": 92.84615384615384,
      "grad_norm": 0.9326725006103516,
      "learning_rate": 0.01,
      "loss": 4.2536773681640625,
      "num_input_tokens_seen": 4792992,
      "step": 1207,
      "train_runtime": 986.0242,
      "train_tokens_per_second": 4860.927
    },
    {
      "epoch": 92.84615384615384,
      "num_input_tokens_seen": 4792992,
      "step": 1207,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 92.92307692307692,
      "grad_norm": 2.190541982650757,
      "learning_rate": 0.01,
      "loss": 4.291766166687012,
      "num_input_tokens_seen": 4797088,
      "step": 1208,
      "train_runtime": 986.8369,
      "train_tokens_per_second": 4861.075
    },
    {
      "epoch": 92.92307692307692,
      "num_input_tokens_seen": 4797088,
      "step": 1208,
      "train_accuracy": 0.032573290169239044
    },
    {
      "epoch": 93.0,
      "grad_norm": 0.9914067387580872,
      "learning_rate": 0.01,
      "loss": 4.260455131530762,
      "num_input_tokens_seen": 4799544,
      "step": 1209,
      "train_runtime": 987.354,
      "train_tokens_per_second": 4861.016
    },
    {
      "epoch": 93.0,
      "num_input_tokens_seen": 4799544,
      "step": 1209,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 93.07692307692308,
      "grad_norm": 1.1241073608398438,
      "learning_rate": 0.01,
      "loss": 4.217317581176758,
      "num_input_tokens_seen": 4803640,
      "step": 1210,
      "train_runtime": 988.1836,
      "train_tokens_per_second": 4861.08
    },
    {
      "epoch": 93.07692307692308,
      "num_input_tokens_seen": 4803640,
      "step": 1210,
      "train_accuracy": 0.046875
    },
    {
      "epoch": 93.15384615384616,
      "grad_norm": 0.898827850818634,
      "learning_rate": 0.01,
      "loss": 4.246952056884766,
      "num_input_tokens_seen": 4807736,
      "step": 1211,
      "train_runtime": 988.9996,
      "train_tokens_per_second": 4861.211
    },
    {
      "epoch": 93.15384615384616,
      "num_input_tokens_seen": 4807736,
      "step": 1211,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 93.23076923076923,
      "grad_norm": 0.7480000257492065,
      "learning_rate": 0.01,
      "loss": 4.2528276443481445,
      "num_input_tokens_seen": 4811832,
      "step": 1212,
      "train_runtime": 989.8158,
      "train_tokens_per_second": 4861.341
    },
    {
      "epoch": 93.23076923076923,
      "num_input_tokens_seen": 4811832,
      "step": 1212,
      "train_accuracy": 0.03515625
    },
    {
      "epoch": 93.3076923076923,
      "grad_norm": 2.9047458171844482,
      "learning_rate": 0.01,
      "loss": 4.243381023406982,
      "num_input_tokens_seen": 4815928,
      "step": 1213,
      "train_runtime": 990.632,
      "train_tokens_per_second": 4861.47
    },
    {
      "epoch": 93.3076923076923,
      "num_input_tokens_seen": 4815928,
      "step": 1213,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 93.38461538461539,
      "grad_norm": 0.5581874847412109,
      "learning_rate": 0.01,
      "loss": 4.258604049682617,
      "num_input_tokens_seen": 4820024,
      "step": 1214,
      "train_runtime": 991.4509,
      "train_tokens_per_second": 4861.586
    },
    {
      "epoch": 93.38461538461539,
      "num_input_tokens_seen": 4820024,
      "step": 1214,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 93.46153846153847,
      "grad_norm": 1.116030216217041,
      "learning_rate": 0.01,
      "loss": 4.252742767333984,
      "num_input_tokens_seen": 4824120,
      "step": 1215,
      "train_runtime": 992.2696,
      "train_tokens_per_second": 4861.703
    },
    {
      "epoch": 93.46153846153847,
      "num_input_tokens_seen": 4824120,
      "step": 1215,
      "train_accuracy": 0.041015625
    },
    {
      "epoch": 93.53846153846153,
      "grad_norm": 0.8851171731948853,
      "learning_rate": 0.01,
      "loss": 4.307258605957031,
      "num_input_tokens_seen": 4828216,
      "step": 1216,
      "train_runtime": 993.0861,
      "train_tokens_per_second": 4861.83
    },
    {
      "epoch": 93.53846153846153,
      "num_input_tokens_seen": 4828216,
      "step": 1216,
      "train_accuracy": 0.03515625
    },
    {
      "epoch": 93.61538461538461,
      "grad_norm": 0.4048510193824768,
      "learning_rate": 0.01,
      "loss": 4.233102798461914,
      "num_input_tokens_seen": 4832312,
      "step": 1217,
      "train_runtime": 993.9029,
      "train_tokens_per_second": 4861.956
    },
    {
      "epoch": 93.61538461538461,
      "num_input_tokens_seen": 4832312,
      "step": 1217,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 93.6923076923077,
      "grad_norm": 0.5768749117851257,
      "learning_rate": 0.01,
      "loss": 4.241782188415527,
      "num_input_tokens_seen": 4836408,
      "step": 1218,
      "train_runtime": 994.7203,
      "train_tokens_per_second": 4862.078
    },
    {
      "epoch": 93.6923076923077,
      "num_input_tokens_seen": 4836408,
      "step": 1218,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 93.76923076923077,
      "grad_norm": 0.7128937840461731,
      "learning_rate": 0.01,
      "loss": 4.259064674377441,
      "num_input_tokens_seen": 4840504,
      "step": 1219,
      "train_runtime": 995.5372,
      "train_tokens_per_second": 4862.203
    },
    {
      "epoch": 93.76923076923077,
      "num_input_tokens_seen": 4840504,
      "step": 1219,
      "train_accuracy": 0.03515625
    },
    {
      "epoch": 93.84615384615384,
      "grad_norm": 8.663732528686523,
      "learning_rate": 0.01,
      "loss": 4.251834392547607,
      "num_input_tokens_seen": 4844600,
      "step": 1220,
      "train_runtime": 996.3553,
      "train_tokens_per_second": 4862.322
    },
    {
      "epoch": 93.84615384615384,
      "num_input_tokens_seen": 4844600,
      "step": 1220,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 93.92307692307692,
      "grad_norm": 7.0520243644714355,
      "learning_rate": 0.01,
      "loss": 4.247310161590576,
      "num_input_tokens_seen": 4848696,
      "step": 1221,
      "train_runtime": 997.167,
      "train_tokens_per_second": 4862.471
    },
    {
      "epoch": 93.92307692307692,
      "num_input_tokens_seen": 4848696,
      "step": 1221,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 94.0,
      "grad_norm": 1.0234301090240479,
      "learning_rate": 0.01,
      "loss": 4.274682998657227,
      "num_input_tokens_seen": 4851152,
      "step": 1222,
      "train_runtime": 997.6844,
      "train_tokens_per_second": 4862.411
    },
    {
      "epoch": 94.0,
      "num_input_tokens_seen": 4851152,
      "step": 1222,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 94.07692307692308,
      "grad_norm": 0.6315427422523499,
      "learning_rate": 0.01,
      "loss": 4.229619979858398,
      "num_input_tokens_seen": 4855248,
      "step": 1223,
      "train_runtime": 998.513,
      "train_tokens_per_second": 4862.478
    },
    {
      "epoch": 94.07692307692308,
      "num_input_tokens_seen": 4855248,
      "step": 1223,
      "train_accuracy": 0.041015625
    },
    {
      "epoch": 94.15384615384616,
      "grad_norm": 1.369253396987915,
      "learning_rate": 0.01,
      "loss": 4.144454479217529,
      "num_input_tokens_seen": 4859344,
      "step": 1224,
      "train_runtime": 999.33,
      "train_tokens_per_second": 4862.602
    },
    {
      "epoch": 94.15384615384616,
      "num_input_tokens_seen": 4859344,
      "step": 1224,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 94.23076923076923,
      "grad_norm": 1.8054120540618896,
      "learning_rate": 0.01,
      "loss": 4.198246002197266,
      "num_input_tokens_seen": 4863440,
      "step": 1225,
      "train_runtime": 1000.1484,
      "train_tokens_per_second": 4862.718
    },
    {
      "epoch": 94.23076923076923,
      "num_input_tokens_seen": 4863440,
      "step": 1225,
      "train_accuracy": 0.0546875
    },
    {
      "epoch": 94.3076923076923,
      "grad_norm": 1.6095513105392456,
      "learning_rate": 0.01,
      "loss": 4.210019588470459,
      "num_input_tokens_seen": 4867536,
      "step": 1226,
      "train_runtime": 1000.977,
      "train_tokens_per_second": 4862.785
    },
    {
      "epoch": 94.3076923076923,
      "num_input_tokens_seen": 4867536,
      "step": 1226,
      "train_accuracy": 0.0390625
    },
    {
      "epoch": 94.38461538461539,
      "grad_norm": 1.6851614713668823,
      "learning_rate": 0.01,
      "loss": 4.179011344909668,
      "num_input_tokens_seen": 4871632,
      "step": 1227,
      "train_runtime": 1001.7935,
      "train_tokens_per_second": 4862.91
    },
    {
      "epoch": 94.38461538461539,
      "num_input_tokens_seen": 4871632,
      "step": 1227,
      "train_accuracy": 0.044921875
    },
    {
      "epoch": 94.46153846153847,
      "grad_norm": 0.6386091113090515,
      "learning_rate": 0.01,
      "loss": 4.238643646240234,
      "num_input_tokens_seen": 4875728,
      "step": 1228,
      "train_runtime": 1002.6101,
      "train_tokens_per_second": 4863.035
    },
    {
      "epoch": 94.46153846153847,
      "num_input_tokens_seen": 4875728,
      "step": 1228,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 94.53846153846153,
      "grad_norm": 3.2731964588165283,
      "learning_rate": 0.01,
      "loss": 4.1840972900390625,
      "num_input_tokens_seen": 4879824,
      "step": 1229,
      "train_runtime": 1003.4266,
      "train_tokens_per_second": 4863.16
    },
    {
      "epoch": 94.53846153846153,
      "num_input_tokens_seen": 4879824,
      "step": 1229,
      "train_accuracy": 0.05859375
    },
    {
      "epoch": 94.61538461538461,
      "grad_norm": 0.926724374294281,
      "learning_rate": 0.01,
      "loss": 4.11090087890625,
      "num_input_tokens_seen": 4883920,
      "step": 1230,
      "train_runtime": 1004.2444,
      "train_tokens_per_second": 4863.278
    },
    {
      "epoch": 94.61538461538461,
      "num_input_tokens_seen": 4883920,
      "step": 1230,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 94.6923076923077,
      "grad_norm": 0.9720467329025269,
      "learning_rate": 0.01,
      "loss": 4.310583114624023,
      "num_input_tokens_seen": 4888016,
      "step": 1231,
      "train_runtime": 1005.0629,
      "train_tokens_per_second": 4863.393
    },
    {
      "epoch": 94.6923076923077,
      "num_input_tokens_seen": 4888016,
      "step": 1231,
      "train_accuracy": 0.037109375
    },
    {
      "epoch": 94.76923076923077,
      "grad_norm": 0.7912406921386719,
      "learning_rate": 0.01,
      "loss": 4.284384727478027,
      "num_input_tokens_seen": 4892112,
      "step": 1232,
      "train_runtime": 1005.8813,
      "train_tokens_per_second": 4863.508
    },
    {
      "epoch": 94.76923076923077,
      "num_input_tokens_seen": 4892112,
      "step": 1232,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 94.84615384615384,
      "grad_norm": 1.2726390361785889,
      "learning_rate": 0.01,
      "loss": 4.27509880065918,
      "num_input_tokens_seen": 4896208,
      "step": 1233,
      "train_runtime": 1006.6981,
      "train_tokens_per_second": 4863.631
    },
    {
      "epoch": 94.84615384615384,
      "num_input_tokens_seen": 4896208,
      "step": 1233,
      "train_accuracy": 0.041015625
    },
    {
      "epoch": 94.92307692307692,
      "grad_norm": 0.6728102564811707,
      "learning_rate": 0.01,
      "loss": 4.257051467895508,
      "num_input_tokens_seen": 4900304,
      "step": 1234,
      "train_runtime": 1007.5108,
      "train_tokens_per_second": 4863.773
    },
    {
      "epoch": 94.92307692307692,
      "num_input_tokens_seen": 4900304,
      "step": 1234,
      "train_accuracy": 0.04885993152856827
    },
    {
      "epoch": 95.0,
      "grad_norm": 1.2158743143081665,
      "learning_rate": 0.01,
      "loss": 4.241811752319336,
      "num_input_tokens_seen": 4902760,
      "step": 1235,
      "train_runtime": 1008.0279,
      "train_tokens_per_second": 4863.715
    },
    {
      "epoch": 95.0,
      "num_input_tokens_seen": 4902760,
      "step": 1235,
      "train_accuracy": 0.044921875
    },
    {
      "epoch": 95.07692307692308,
      "grad_norm": 0.5960951447486877,
      "learning_rate": 0.01,
      "loss": 4.13967227935791,
      "num_input_tokens_seen": 4906856,
      "step": 1236,
      "train_runtime": 1008.8579,
      "train_tokens_per_second": 4863.773
    },
    {
      "epoch": 95.07692307692308,
      "num_input_tokens_seen": 4906856,
      "step": 1236,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 95.15384615384616,
      "grad_norm": 0.6546602249145508,
      "learning_rate": 0.01,
      "loss": 4.273693084716797,
      "num_input_tokens_seen": 4910952,
      "step": 1237,
      "train_runtime": 1009.6769,
      "train_tokens_per_second": 4863.885
    },
    {
      "epoch": 95.15384615384616,
      "num_input_tokens_seen": 4910952,
      "step": 1237,
      "train_accuracy": 0.041015625
    },
    {
      "epoch": 95.23076923076923,
      "grad_norm": 0.6968697905540466,
      "learning_rate": 0.01,
      "loss": 4.230842590332031,
      "num_input_tokens_seen": 4915048,
      "step": 1238,
      "train_runtime": 1010.4945,
      "train_tokens_per_second": 4864.003
    },
    {
      "epoch": 95.23076923076923,
      "num_input_tokens_seen": 4915048,
      "step": 1238,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 95.3076923076923,
      "grad_norm": 1.1342315673828125,
      "learning_rate": 0.01,
      "loss": 4.240490436553955,
      "num_input_tokens_seen": 4919144,
      "step": 1239,
      "train_runtime": 1011.3106,
      "train_tokens_per_second": 4864.128
    },
    {
      "epoch": 95.3076923076923,
      "num_input_tokens_seen": 4919144,
      "step": 1239,
      "train_accuracy": 0.03125
    },
    {
      "epoch": 95.38461538461539,
      "grad_norm": 1.3191282749176025,
      "learning_rate": 0.01,
      "loss": 4.234597206115723,
      "num_input_tokens_seen": 4923240,
      "step": 1240,
      "train_runtime": 1012.1279,
      "train_tokens_per_second": 4864.247
    },
    {
      "epoch": 95.38461538461539,
      "num_input_tokens_seen": 4923240,
      "step": 1240,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 95.46153846153847,
      "grad_norm": 0.9121236801147461,
      "learning_rate": 0.01,
      "loss": 4.307871341705322,
      "num_input_tokens_seen": 4927336,
      "step": 1241,
      "train_runtime": 1012.9444,
      "train_tokens_per_second": 4864.369
    },
    {
      "epoch": 95.46153846153847,
      "num_input_tokens_seen": 4927336,
      "step": 1241,
      "train_accuracy": 0.03125
    },
    {
      "epoch": 95.53846153846153,
      "grad_norm": 0.9026360511779785,
      "learning_rate": 0.01,
      "loss": 4.26276969909668,
      "num_input_tokens_seen": 4931432,
      "step": 1242,
      "train_runtime": 1013.7636,
      "train_tokens_per_second": 4864.479
    },
    {
      "epoch": 95.53846153846153,
      "num_input_tokens_seen": 4931432,
      "step": 1242,
      "train_accuracy": 0.033203125
    },
    {
      "epoch": 95.61538461538461,
      "grad_norm": 1.45496666431427,
      "learning_rate": 0.01,
      "loss": 4.283936500549316,
      "num_input_tokens_seen": 4935528,
      "step": 1243,
      "train_runtime": 1014.5802,
      "train_tokens_per_second": 4864.601
    },
    {
      "epoch": 95.61538461538461,
      "num_input_tokens_seen": 4935528,
      "step": 1243,
      "train_accuracy": 0.03515625
    },
    {
      "epoch": 95.6923076923077,
      "grad_norm": 2.2112810611724854,
      "learning_rate": 0.01,
      "loss": 4.249178886413574,
      "num_input_tokens_seen": 4939624,
      "step": 1244,
      "train_runtime": 1015.3964,
      "train_tokens_per_second": 4864.725
    },
    {
      "epoch": 95.6923076923077,
      "num_input_tokens_seen": 4939624,
      "step": 1244,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 95.76923076923077,
      "grad_norm": 1.4071255922317505,
      "learning_rate": 0.01,
      "loss": 4.245018005371094,
      "num_input_tokens_seen": 4943720,
      "step": 1245,
      "train_runtime": 1016.2131,
      "train_tokens_per_second": 4864.846
    },
    {
      "epoch": 95.76923076923077,
      "num_input_tokens_seen": 4943720,
      "step": 1245,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 95.84615384615384,
      "grad_norm": 0.9206699132919312,
      "learning_rate": 0.01,
      "loss": 4.237589359283447,
      "num_input_tokens_seen": 4947816,
      "step": 1246,
      "train_runtime": 1017.0292,
      "train_tokens_per_second": 4864.969
    },
    {
      "epoch": 95.84615384615384,
      "num_input_tokens_seen": 4947816,
      "step": 1246,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 95.92307692307692,
      "grad_norm": 1.309809923171997,
      "learning_rate": 0.01,
      "loss": 4.217013359069824,
      "num_input_tokens_seen": 4951912,
      "step": 1247,
      "train_runtime": 1017.8425,
      "train_tokens_per_second": 4865.106
    },
    {
      "epoch": 95.92307692307692,
      "num_input_tokens_seen": 4951912,
      "step": 1247,
      "train_accuracy": 0.026058631017804146
    },
    {
      "epoch": 96.0,
      "grad_norm": 4.887404918670654,
      "learning_rate": 0.01,
      "loss": 4.300361156463623,
      "num_input_tokens_seen": 4954368,
      "step": 1248,
      "train_runtime": 1018.3623,
      "train_tokens_per_second": 4865.035
    },
    {
      "epoch": 96.0,
      "num_input_tokens_seen": 4954368,
      "step": 1248,
      "train_accuracy": 0.03515625
    },
    {
      "epoch": 96.07692307692308,
      "grad_norm": 1.9327681064605713,
      "learning_rate": 0.01,
      "loss": 4.225034713745117,
      "num_input_tokens_seen": 4958464,
      "step": 1249,
      "train_runtime": 1019.1909,
      "train_tokens_per_second": 4865.098
    },
    {
      "epoch": 96.07692307692308,
      "num_input_tokens_seen": 4958464,
      "step": 1249,
      "train_accuracy": 0.041015625
    },
    {
      "epoch": 96.15384615384616,
      "grad_norm": 0.612694501876831,
      "learning_rate": 0.01,
      "loss": 4.222924709320068,
      "num_input_tokens_seen": 4962560,
      "step": 1250,
      "train_runtime": 1020.0066,
      "train_tokens_per_second": 4865.224
    },
    {
      "epoch": 96.15384615384616,
      "num_input_tokens_seen": 4962560,
      "step": 1250,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 96.23076923076923,
      "grad_norm": 4.84871244430542,
      "learning_rate": 0.01,
      "loss": 4.2540507316589355,
      "num_input_tokens_seen": 4966656,
      "step": 1251,
      "train_runtime": 1020.8235,
      "train_tokens_per_second": 4865.342
    },
    {
      "epoch": 96.23076923076923,
      "num_input_tokens_seen": 4966656,
      "step": 1251,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 96.3076923076923,
      "grad_norm": 5.11841344833374,
      "learning_rate": 0.01,
      "loss": 4.731534957885742,
      "num_input_tokens_seen": 4970752,
      "step": 1252,
      "train_runtime": 1021.6401,
      "train_tokens_per_second": 4865.463
    },
    {
      "epoch": 96.3076923076923,
      "num_input_tokens_seen": 4970752,
      "step": 1252,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 96.38461538461539,
      "grad_norm": 1.5406798124313354,
      "learning_rate": 0.01,
      "loss": 4.5885820388793945,
      "num_input_tokens_seen": 4974848,
      "step": 1253,
      "train_runtime": 1022.4602,
      "train_tokens_per_second": 4865.566
    },
    {
      "epoch": 96.38461538461539,
      "num_input_tokens_seen": 4974848,
      "step": 1253,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 96.46153846153847,
      "grad_norm": 0.45561483502388,
      "learning_rate": 0.01,
      "loss": 4.525201797485352,
      "num_input_tokens_seen": 4978944,
      "step": 1254,
      "train_runtime": 1023.2769,
      "train_tokens_per_second": 4865.686
    },
    {
      "epoch": 96.46153846153847,
      "num_input_tokens_seen": 4978944,
      "step": 1254,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 96.53846153846153,
      "grad_norm": 4.504939079284668,
      "learning_rate": 0.01,
      "loss": 4.626739501953125,
      "num_input_tokens_seen": 4983040,
      "step": 1255,
      "train_runtime": 1024.0923,
      "train_tokens_per_second": 4865.811
    },
    {
      "epoch": 96.53846153846153,
      "num_input_tokens_seen": 4983040,
      "step": 1255,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 96.61538461538461,
      "grad_norm": 0.6294952034950256,
      "learning_rate": 0.01,
      "loss": 4.591184616088867,
      "num_input_tokens_seen": 4987136,
      "step": 1256,
      "train_runtime": 1024.9081,
      "train_tokens_per_second": 4865.935
    },
    {
      "epoch": 96.61538461538461,
      "num_input_tokens_seen": 4987136,
      "step": 1256,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 96.6923076923077,
      "grad_norm": 0.6702298521995544,
      "learning_rate": 0.01,
      "loss": 4.538229942321777,
      "num_input_tokens_seen": 4991232,
      "step": 1257,
      "train_runtime": 1025.725,
      "train_tokens_per_second": 4866.053
    },
    {
      "epoch": 96.6923076923077,
      "num_input_tokens_seen": 4991232,
      "step": 1257,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 96.76923076923077,
      "grad_norm": 1.0073717832565308,
      "learning_rate": 0.01,
      "loss": 4.5416107177734375,
      "num_input_tokens_seen": 4995328,
      "step": 1258,
      "train_runtime": 1026.5412,
      "train_tokens_per_second": 4866.174
    },
    {
      "epoch": 96.76923076923077,
      "num_input_tokens_seen": 4995328,
      "step": 1258,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 96.84615384615384,
      "grad_norm": 0.7780309915542603,
      "learning_rate": 0.01,
      "loss": 4.570990562438965,
      "num_input_tokens_seen": 4999424,
      "step": 1259,
      "train_runtime": 1027.3592,
      "train_tokens_per_second": 4866.286
    },
    {
      "epoch": 96.84615384615384,
      "num_input_tokens_seen": 4999424,
      "step": 1259,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 96.92307692307692,
      "grad_norm": 7.325876712799072,
      "learning_rate": 0.01,
      "loss": 4.555654525756836,
      "num_input_tokens_seen": 5003520,
      "step": 1260,
      "train_runtime": 1028.1706,
      "train_tokens_per_second": 4866.43
    },
    {
      "epoch": 96.92307692307692,
      "num_input_tokens_seen": 5003520,
      "step": 1260,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 97.0,
      "grad_norm": 1.0936717987060547,
      "learning_rate": 0.01,
      "loss": 4.580423831939697,
      "num_input_tokens_seen": 5005976,
      "step": 1261,
      "train_runtime": 1028.6866,
      "train_tokens_per_second": 4866.376
    },
    {
      "epoch": 97.0,
      "num_input_tokens_seen": 5005976,
      "step": 1261,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 97.07692307692308,
      "grad_norm": 9.796006202697754,
      "learning_rate": 0.01,
      "loss": 4.472999095916748,
      "num_input_tokens_seen": 5010072,
      "step": 1262,
      "train_runtime": 1029.514,
      "train_tokens_per_second": 4866.444
    },
    {
      "epoch": 97.07692307692308,
      "num_input_tokens_seen": 5010072,
      "step": 1262,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 97.15384615384616,
      "grad_norm": 1.7899651527404785,
      "learning_rate": 0.01,
      "loss": 4.557491779327393,
      "num_input_tokens_seen": 5014168,
      "step": 1263,
      "train_runtime": 1030.3311,
      "train_tokens_per_second": 4866.56
    },
    {
      "epoch": 97.15384615384616,
      "num_input_tokens_seen": 5014168,
      "step": 1263,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 97.23076923076923,
      "grad_norm": 0.8044832348823547,
      "learning_rate": 0.01,
      "loss": 4.589766979217529,
      "num_input_tokens_seen": 5018264,
      "step": 1264,
      "train_runtime": 1031.1464,
      "train_tokens_per_second": 4866.684
    },
    {
      "epoch": 97.23076923076923,
      "num_input_tokens_seen": 5018264,
      "step": 1264,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 97.3076923076923,
      "grad_norm": 0.7345467805862427,
      "learning_rate": 0.01,
      "loss": 4.575111389160156,
      "num_input_tokens_seen": 5022360,
      "step": 1265,
      "train_runtime": 1032.3531,
      "train_tokens_per_second": 4864.964
    },
    {
      "epoch": 97.3076923076923,
      "num_input_tokens_seen": 5022360,
      "step": 1265,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 97.38461538461539,
      "grad_norm": 1.5114500522613525,
      "learning_rate": 0.01,
      "loss": 4.474294662475586,
      "num_input_tokens_seen": 5026456,
      "step": 1266,
      "train_runtime": 1033.1672,
      "train_tokens_per_second": 4865.094
    },
    {
      "epoch": 97.38461538461539,
      "num_input_tokens_seen": 5026456,
      "step": 1266,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 97.46153846153847,
      "grad_norm": 0.6166411638259888,
      "learning_rate": 0.01,
      "loss": 4.488404273986816,
      "num_input_tokens_seen": 5030552,
      "step": 1267,
      "train_runtime": 1033.9825,
      "train_tokens_per_second": 4865.219
    },
    {
      "epoch": 97.46153846153847,
      "num_input_tokens_seen": 5030552,
      "step": 1267,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 97.53846153846153,
      "grad_norm": 1.085659146308899,
      "learning_rate": 0.01,
      "loss": 4.550913333892822,
      "num_input_tokens_seen": 5034648,
      "step": 1268,
      "train_runtime": 1034.7978,
      "train_tokens_per_second": 4865.344
    },
    {
      "epoch": 97.53846153846153,
      "num_input_tokens_seen": 5034648,
      "step": 1268,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 97.61538461538461,
      "grad_norm": 0.6593523621559143,
      "learning_rate": 0.01,
      "loss": 4.613582611083984,
      "num_input_tokens_seen": 5038744,
      "step": 1269,
      "train_runtime": 1035.6128,
      "train_tokens_per_second": 4865.471
    },
    {
      "epoch": 97.61538461538461,
      "num_input_tokens_seen": 5038744,
      "step": 1269,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 97.6923076923077,
      "grad_norm": 0.41679027676582336,
      "learning_rate": 0.01,
      "loss": 4.539933681488037,
      "num_input_tokens_seen": 5042840,
      "step": 1270,
      "train_runtime": 1036.4274,
      "train_tokens_per_second": 4865.599
    },
    {
      "epoch": 97.6923076923077,
      "num_input_tokens_seen": 5042840,
      "step": 1270,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 97.76923076923077,
      "grad_norm": 0.3402726650238037,
      "learning_rate": 0.01,
      "loss": 4.557590484619141,
      "num_input_tokens_seen": 5046936,
      "step": 1271,
      "train_runtime": 1037.2435,
      "train_tokens_per_second": 4865.72
    },
    {
      "epoch": 97.76923076923077,
      "num_input_tokens_seen": 5046936,
      "step": 1271,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 97.84615384615384,
      "grad_norm": 0.5146462917327881,
      "learning_rate": 0.01,
      "loss": 4.5818963050842285,
      "num_input_tokens_seen": 5051032,
      "step": 1272,
      "train_runtime": 1038.0618,
      "train_tokens_per_second": 4865.83
    },
    {
      "epoch": 97.84615384615384,
      "num_input_tokens_seen": 5051032,
      "step": 1272,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 97.92307692307692,
      "grad_norm": 0.4620774984359741,
      "learning_rate": 0.01,
      "loss": 4.5395660400390625,
      "num_input_tokens_seen": 5055128,
      "step": 1273,
      "train_runtime": 1038.8735,
      "train_tokens_per_second": 4865.971
    },
    {
      "epoch": 97.92307692307692,
      "num_input_tokens_seen": 5055128,
      "step": 1273,
      "train_accuracy": 0.013029315508902073
    },
    {
      "epoch": 98.0,
      "grad_norm": 0.6541016697883606,
      "learning_rate": 0.01,
      "loss": 4.595381259918213,
      "num_input_tokens_seen": 5057584,
      "step": 1274,
      "train_runtime": 1039.3895,
      "train_tokens_per_second": 4865.918
    },
    {
      "epoch": 98.0,
      "num_input_tokens_seen": 5057584,
      "step": 1274,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 98.07692307692308,
      "grad_norm": 0.6528225541114807,
      "learning_rate": 0.01,
      "loss": 4.497180461883545,
      "num_input_tokens_seen": 5061680,
      "step": 1275,
      "train_runtime": 1040.2178,
      "train_tokens_per_second": 4865.981
    },
    {
      "epoch": 98.07692307692308,
      "num_input_tokens_seen": 5061680,
      "step": 1275,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 98.15384615384616,
      "grad_norm": 0.3280772566795349,
      "learning_rate": 0.01,
      "loss": 4.5014801025390625,
      "num_input_tokens_seen": 5065776,
      "step": 1276,
      "train_runtime": 1041.0325,
      "train_tokens_per_second": 4866.107
    },
    {
      "epoch": 98.15384615384616,
      "num_input_tokens_seen": 5065776,
      "step": 1276,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 98.23076923076923,
      "grad_norm": 0.6081762909889221,
      "learning_rate": 0.01,
      "loss": 4.539563179016113,
      "num_input_tokens_seen": 5069872,
      "step": 1277,
      "train_runtime": 1041.8497,
      "train_tokens_per_second": 4866.222
    },
    {
      "epoch": 98.23076923076923,
      "num_input_tokens_seen": 5069872,
      "step": 1277,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 98.3076923076923,
      "grad_norm": 0.46605730056762695,
      "learning_rate": 0.01,
      "loss": 4.5333404541015625,
      "num_input_tokens_seen": 5073968,
      "step": 1278,
      "train_runtime": 1042.6659,
      "train_tokens_per_second": 4866.341
    },
    {
      "epoch": 98.3076923076923,
      "num_input_tokens_seen": 5073968,
      "step": 1278,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 98.38461538461539,
      "grad_norm": 0.4803134500980377,
      "learning_rate": 0.01,
      "loss": 4.540291786193848,
      "num_input_tokens_seen": 5078064,
      "step": 1279,
      "train_runtime": 1043.4819,
      "train_tokens_per_second": 4866.461
    },
    {
      "epoch": 98.38461538461539,
      "num_input_tokens_seen": 5078064,
      "step": 1279,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 98.46153846153847,
      "grad_norm": 0.37552720308303833,
      "learning_rate": 0.01,
      "loss": 4.5177812576293945,
      "num_input_tokens_seen": 5082160,
      "step": 1280,
      "train_runtime": 1044.298,
      "train_tokens_per_second": 4866.58
    },
    {
      "epoch": 98.46153846153847,
      "num_input_tokens_seen": 5082160,
      "step": 1280,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 98.53846153846153,
      "grad_norm": 0.6628419160842896,
      "learning_rate": 0.01,
      "loss": 4.5339179039001465,
      "num_input_tokens_seen": 5086256,
      "step": 1281,
      "train_runtime": 1045.1136,
      "train_tokens_per_second": 4866.702
    },
    {
      "epoch": 98.53846153846153,
      "num_input_tokens_seen": 5086256,
      "step": 1281,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 98.61538461538461,
      "grad_norm": 0.3709065914154053,
      "learning_rate": 0.01,
      "loss": 4.5074076652526855,
      "num_input_tokens_seen": 5090352,
      "step": 1282,
      "train_runtime": 1045.93,
      "train_tokens_per_second": 4866.819
    },
    {
      "epoch": 98.61538461538461,
      "num_input_tokens_seen": 5090352,
      "step": 1282,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 98.6923076923077,
      "grad_norm": 1.0720301866531372,
      "learning_rate": 0.01,
      "loss": 4.541929244995117,
      "num_input_tokens_seen": 5094448,
      "step": 1283,
      "train_runtime": 1046.7458,
      "train_tokens_per_second": 4866.939
    },
    {
      "epoch": 98.6923076923077,
      "num_input_tokens_seen": 5094448,
      "step": 1283,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 98.76923076923077,
      "grad_norm": 0.4565328359603882,
      "learning_rate": 0.01,
      "loss": 4.486786842346191,
      "num_input_tokens_seen": 5098544,
      "step": 1284,
      "train_runtime": 1047.5621,
      "train_tokens_per_second": 4867.057
    },
    {
      "epoch": 98.76923076923077,
      "num_input_tokens_seen": 5098544,
      "step": 1284,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 98.84615384615384,
      "grad_norm": 0.5800324082374573,
      "learning_rate": 0.01,
      "loss": 4.52537727355957,
      "num_input_tokens_seen": 5102640,
      "step": 1285,
      "train_runtime": 1048.3791,
      "train_tokens_per_second": 4867.171
    },
    {
      "epoch": 98.84615384615384,
      "num_input_tokens_seen": 5102640,
      "step": 1285,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 98.92307692307692,
      "grad_norm": 3.7269937992095947,
      "learning_rate": 0.01,
      "loss": 4.4630842208862305,
      "num_input_tokens_seen": 5106736,
      "step": 1286,
      "train_runtime": 1049.1917,
      "train_tokens_per_second": 4867.305
    },
    {
      "epoch": 98.92307692307692,
      "num_input_tokens_seen": 5106736,
      "step": 1286,
      "train_accuracy": 0.019543973729014397
    },
    {
      "epoch": 99.0,
      "grad_norm": 12.668912887573242,
      "learning_rate": 0.01,
      "loss": 4.502503395080566,
      "num_input_tokens_seen": 5109192,
      "step": 1287,
      "train_runtime": 1049.7089,
      "train_tokens_per_second": 4867.247
    },
    {
      "epoch": 99.0,
      "num_input_tokens_seen": 5109192,
      "step": 1287,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 99.07692307692308,
      "grad_norm": 0.6205081343650818,
      "learning_rate": 0.01,
      "loss": 4.450595855712891,
      "num_input_tokens_seen": 5113288,
      "step": 1288,
      "train_runtime": 1050.5416,
      "train_tokens_per_second": 4867.288
    },
    {
      "epoch": 99.07692307692308,
      "num_input_tokens_seen": 5113288,
      "step": 1288,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 99.15384615384616,
      "grad_norm": 0.45831891894340515,
      "learning_rate": 0.01,
      "loss": 4.471729755401611,
      "num_input_tokens_seen": 5117384,
      "step": 1289,
      "train_runtime": 1051.3585,
      "train_tokens_per_second": 4867.402
    },
    {
      "epoch": 99.15384615384616,
      "num_input_tokens_seen": 5117384,
      "step": 1289,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 99.23076923076923,
      "grad_norm": 0.5405982136726379,
      "learning_rate": 0.01,
      "loss": 4.478852272033691,
      "num_input_tokens_seen": 5121480,
      "step": 1290,
      "train_runtime": 1052.1749,
      "train_tokens_per_second": 4867.518
    },
    {
      "epoch": 99.23076923076923,
      "num_input_tokens_seen": 5121480,
      "step": 1290,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 99.3076923076923,
      "grad_norm": 0.46504953503608704,
      "learning_rate": 0.01,
      "loss": 4.49061393737793,
      "num_input_tokens_seen": 5125576,
      "step": 1291,
      "train_runtime": 1052.9918,
      "train_tokens_per_second": 4867.631
    },
    {
      "epoch": 99.3076923076923,
      "num_input_tokens_seen": 5125576,
      "step": 1291,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 99.38461538461539,
      "grad_norm": 1.3747881650924683,
      "learning_rate": 0.01,
      "loss": 4.48637580871582,
      "num_input_tokens_seen": 5129672,
      "step": 1292,
      "train_runtime": 1053.8107,
      "train_tokens_per_second": 4867.736
    },
    {
      "epoch": 99.38461538461539,
      "num_input_tokens_seen": 5129672,
      "step": 1292,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 99.46153846153847,
      "grad_norm": 1.103324294090271,
      "learning_rate": 0.01,
      "loss": 4.496672630310059,
      "num_input_tokens_seen": 5133768,
      "step": 1293,
      "train_runtime": 1054.6266,
      "train_tokens_per_second": 4867.854
    },
    {
      "epoch": 99.46153846153847,
      "num_input_tokens_seen": 5133768,
      "step": 1293,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 99.53846153846153,
      "grad_norm": 1.8743489980697632,
      "learning_rate": 0.01,
      "loss": 4.473363876342773,
      "num_input_tokens_seen": 5137864,
      "step": 1294,
      "train_runtime": 1055.4449,
      "train_tokens_per_second": 4867.96
    },
    {
      "epoch": 99.53846153846153,
      "num_input_tokens_seen": 5137864,
      "step": 1294,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 99.61538461538461,
      "grad_norm": 0.6432539224624634,
      "learning_rate": 0.01,
      "loss": 4.562296390533447,
      "num_input_tokens_seen": 5141960,
      "step": 1295,
      "train_runtime": 1056.265,
      "train_tokens_per_second": 4868.059
    },
    {
      "epoch": 99.61538461538461,
      "num_input_tokens_seen": 5141960,
      "step": 1295,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 99.6923076923077,
      "grad_norm": 2.6819510459899902,
      "learning_rate": 0.01,
      "loss": 4.490937232971191,
      "num_input_tokens_seen": 5146056,
      "step": 1296,
      "train_runtime": 1057.0814,
      "train_tokens_per_second": 4868.174
    },
    {
      "epoch": 99.6923076923077,
      "num_input_tokens_seen": 5146056,
      "step": 1296,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 99.76923076923077,
      "grad_norm": 0.6963559985160828,
      "learning_rate": 0.01,
      "loss": 4.513890266418457,
      "num_input_tokens_seen": 5150152,
      "step": 1297,
      "train_runtime": 1057.8976,
      "train_tokens_per_second": 4868.29
    },
    {
      "epoch": 99.76923076923077,
      "num_input_tokens_seen": 5150152,
      "step": 1297,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 99.84615384615384,
      "grad_norm": 0.3832263946533203,
      "learning_rate": 0.01,
      "loss": 4.526019096374512,
      "num_input_tokens_seen": 5154248,
      "step": 1298,
      "train_runtime": 1058.7143,
      "train_tokens_per_second": 4868.403
    },
    {
      "epoch": 99.84615384615384,
      "num_input_tokens_seen": 5154248,
      "step": 1298,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 99.92307692307692,
      "grad_norm": 2.4339518547058105,
      "learning_rate": 0.01,
      "loss": 4.52780818939209,
      "num_input_tokens_seen": 5158344,
      "step": 1299,
      "train_runtime": 1059.5272,
      "train_tokens_per_second": 4868.534
    },
    {
      "epoch": 99.92307692307692,
      "num_input_tokens_seen": 5158344,
      "step": 1299,
      "train_accuracy": 0.016286645084619522
    },
    {
      "epoch": 100.0,
      "grad_norm": 1.640897512435913,
      "learning_rate": 0.01,
      "loss": 4.473927021026611,
      "num_input_tokens_seen": 5160800,
      "step": 1300,
      "train_runtime": 1060.0452,
      "train_tokens_per_second": 4868.472
    },
    {
      "epoch": 100.0,
      "eval_CMD_3_branch_eval_accuracy": 0.012296564195298372,
      "eval_CMD_3_branch_eval_loss": 4.5511088371276855,
      "eval_CMD_3_branch_eval_runtime": 1.1574,
      "eval_CMD_3_branch_eval_samples_per_second": 2389.023,
      "eval_CMD_3_branch_eval_steps_per_second": 5.184,
      "num_input_tokens_seen": 5160800,
      "step": 1300
    },
    {
      "epoch": 100.0,
      "num_input_tokens_seen": 5160800,
      "step": 1300,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 100.07692307692308,
      "grad_norm": 0.892929196357727,
      "learning_rate": 0.01,
      "loss": 4.47678279876709,
      "num_input_tokens_seen": 5164896,
      "step": 1301,
      "train_runtime": 1062.035,
      "train_tokens_per_second": 4863.207
    },
    {
      "epoch": 100.07692307692308,
      "num_input_tokens_seen": 5164896,
      "step": 1301,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 100.15384615384616,
      "grad_norm": 0.2669127583503723,
      "learning_rate": 0.01,
      "loss": 4.469684600830078,
      "num_input_tokens_seen": 5168992,
      "step": 1302,
      "train_runtime": 1062.8533,
      "train_tokens_per_second": 4863.316
    },
    {
      "epoch": 100.15384615384616,
      "num_input_tokens_seen": 5168992,
      "step": 1302,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 100.23076923076923,
      "grad_norm": 0.6381169557571411,
      "learning_rate": 0.01,
      "loss": 4.53653621673584,
      "num_input_tokens_seen": 5173088,
      "step": 1303,
      "train_runtime": 1063.67,
      "train_tokens_per_second": 4863.433
    },
    {
      "epoch": 100.23076923076923,
      "num_input_tokens_seen": 5173088,
      "step": 1303,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 100.3076923076923,
      "grad_norm": 0.6963146328926086,
      "learning_rate": 0.01,
      "loss": 4.516428470611572,
      "num_input_tokens_seen": 5177184,
      "step": 1304,
      "train_runtime": 1064.4852,
      "train_tokens_per_second": 4863.557
    },
    {
      "epoch": 100.3076923076923,
      "num_input_tokens_seen": 5177184,
      "step": 1304,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 100.38461538461539,
      "grad_norm": 0.42733877897262573,
      "learning_rate": 0.01,
      "loss": 4.487287521362305,
      "num_input_tokens_seen": 5181280,
      "step": 1305,
      "train_runtime": 1065.3018,
      "train_tokens_per_second": 4863.673
    },
    {
      "epoch": 100.38461538461539,
      "num_input_tokens_seen": 5181280,
      "step": 1305,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 100.46153846153847,
      "grad_norm": 0.500443160533905,
      "learning_rate": 0.01,
      "loss": 4.478186130523682,
      "num_input_tokens_seen": 5185376,
      "step": 1306,
      "train_runtime": 1066.1183,
      "train_tokens_per_second": 4863.79
    },
    {
      "epoch": 100.46153846153847,
      "num_input_tokens_seen": 5185376,
      "step": 1306,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 100.53846153846153,
      "grad_norm": 1.2631031274795532,
      "learning_rate": 0.01,
      "loss": 4.496086597442627,
      "num_input_tokens_seen": 5189472,
      "step": 1307,
      "train_runtime": 1066.935,
      "train_tokens_per_second": 4863.906
    },
    {
      "epoch": 100.53846153846153,
      "num_input_tokens_seen": 5189472,
      "step": 1307,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 100.61538461538461,
      "grad_norm": 0.4480014443397522,
      "learning_rate": 0.01,
      "loss": 4.502488136291504,
      "num_input_tokens_seen": 5193568,
      "step": 1308,
      "train_runtime": 1067.751,
      "train_tokens_per_second": 4864.026
    },
    {
      "epoch": 100.61538461538461,
      "num_input_tokens_seen": 5193568,
      "step": 1308,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 100.6923076923077,
      "grad_norm": 0.9957578778266907,
      "learning_rate": 0.01,
      "loss": 4.496293544769287,
      "num_input_tokens_seen": 5197664,
      "step": 1309,
      "train_runtime": 1068.568,
      "train_tokens_per_second": 4864.14
    },
    {
      "epoch": 100.6923076923077,
      "num_input_tokens_seen": 5197664,
      "step": 1309,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 100.76923076923077,
      "grad_norm": 0.6487727761268616,
      "learning_rate": 0.01,
      "loss": 4.521688461303711,
      "num_input_tokens_seen": 5201760,
      "step": 1310,
      "train_runtime": 1069.3839,
      "train_tokens_per_second": 4864.259
    },
    {
      "epoch": 100.76923076923077,
      "num_input_tokens_seen": 5201760,
      "step": 1310,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 100.84615384615384,
      "grad_norm": 0.38405242562294006,
      "learning_rate": 0.01,
      "loss": 4.537778854370117,
      "num_input_tokens_seen": 5205856,
      "step": 1311,
      "train_runtime": 1070.2009,
      "train_tokens_per_second": 4864.373
    },
    {
      "epoch": 100.84615384615384,
      "num_input_tokens_seen": 5205856,
      "step": 1311,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 100.92307692307692,
      "grad_norm": 0.465504914522171,
      "learning_rate": 0.01,
      "loss": 4.50172233581543,
      "num_input_tokens_seen": 5209952,
      "step": 1312,
      "train_runtime": 1071.0125,
      "train_tokens_per_second": 4864.511
    },
    {
      "epoch": 100.92307692307692,
      "num_input_tokens_seen": 5209952,
      "step": 1312,
      "train_accuracy": 0.013029315508902073
    },
    {
      "epoch": 101.0,
      "grad_norm": 0.4227706491947174,
      "learning_rate": 0.01,
      "loss": 4.505374431610107,
      "num_input_tokens_seen": 5212408,
      "step": 1313,
      "train_runtime": 1071.5284,
      "train_tokens_per_second": 4864.461
    },
    {
      "epoch": 101.0,
      "num_input_tokens_seen": 5212408,
      "step": 1313,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 101.07692307692308,
      "grad_norm": 0.5651357173919678,
      "learning_rate": 0.01,
      "loss": 4.4439520835876465,
      "num_input_tokens_seen": 5216504,
      "step": 1314,
      "train_runtime": 1072.3567,
      "train_tokens_per_second": 4864.523
    },
    {
      "epoch": 101.07692307692308,
      "num_input_tokens_seen": 5216504,
      "step": 1314,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 101.15384615384616,
      "grad_norm": 0.5506910085678101,
      "learning_rate": 0.01,
      "loss": 4.473283767700195,
      "num_input_tokens_seen": 5220600,
      "step": 1315,
      "train_runtime": 1073.1745,
      "train_tokens_per_second": 4864.633
    },
    {
      "epoch": 101.15384615384616,
      "num_input_tokens_seen": 5220600,
      "step": 1315,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 101.23076923076923,
      "grad_norm": 2.351944923400879,
      "learning_rate": 0.01,
      "loss": 4.432819366455078,
      "num_input_tokens_seen": 5224696,
      "step": 1316,
      "train_runtime": 1073.9905,
      "train_tokens_per_second": 4864.751
    },
    {
      "epoch": 101.23076923076923,
      "num_input_tokens_seen": 5224696,
      "step": 1316,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 101.3076923076923,
      "grad_norm": 1.3694813251495361,
      "learning_rate": 0.01,
      "loss": 4.481587886810303,
      "num_input_tokens_seen": 5228792,
      "step": 1317,
      "train_runtime": 1074.8067,
      "train_tokens_per_second": 4864.867
    },
    {
      "epoch": 101.3076923076923,
      "num_input_tokens_seen": 5228792,
      "step": 1317,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 101.38461538461539,
      "grad_norm": 0.8915128707885742,
      "learning_rate": 0.01,
      "loss": 4.507000923156738,
      "num_input_tokens_seen": 5232888,
      "step": 1318,
      "train_runtime": 1075.623,
      "train_tokens_per_second": 4864.983
    },
    {
      "epoch": 101.38461538461539,
      "num_input_tokens_seen": 5232888,
      "step": 1318,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 101.46153846153847,
      "grad_norm": 0.43306732177734375,
      "learning_rate": 0.01,
      "loss": 4.518030643463135,
      "num_input_tokens_seen": 5236984,
      "step": 1319,
      "train_runtime": 1076.4391,
      "train_tokens_per_second": 4865.1
    },
    {
      "epoch": 101.46153846153847,
      "num_input_tokens_seen": 5236984,
      "step": 1319,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 101.53846153846153,
      "grad_norm": 0.8570546507835388,
      "learning_rate": 0.01,
      "loss": 4.507914066314697,
      "num_input_tokens_seen": 5241080,
      "step": 1320,
      "train_runtime": 1077.257,
      "train_tokens_per_second": 4865.209
    },
    {
      "epoch": 101.53846153846153,
      "num_input_tokens_seen": 5241080,
      "step": 1320,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 101.61538461538461,
      "grad_norm": 4.387678623199463,
      "learning_rate": 0.01,
      "loss": 4.4336137771606445,
      "num_input_tokens_seen": 5245176,
      "step": 1321,
      "train_runtime": 1078.0734,
      "train_tokens_per_second": 4865.323
    },
    {
      "epoch": 101.61538461538461,
      "num_input_tokens_seen": 5245176,
      "step": 1321,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 101.6923076923077,
      "grad_norm": 0.5083701014518738,
      "learning_rate": 0.01,
      "loss": 4.486014366149902,
      "num_input_tokens_seen": 5249272,
      "step": 1322,
      "train_runtime": 1078.8906,
      "train_tokens_per_second": 4865.435
    },
    {
      "epoch": 101.6923076923077,
      "num_input_tokens_seen": 5249272,
      "step": 1322,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 101.76923076923077,
      "grad_norm": 0.6799424886703491,
      "learning_rate": 0.01,
      "loss": 4.509057998657227,
      "num_input_tokens_seen": 5253368,
      "step": 1323,
      "train_runtime": 1079.7068,
      "train_tokens_per_second": 4865.551
    },
    {
      "epoch": 101.76923076923077,
      "num_input_tokens_seen": 5253368,
      "step": 1323,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 101.84615384615384,
      "grad_norm": 0.2695162892341614,
      "learning_rate": 0.01,
      "loss": 4.512124061584473,
      "num_input_tokens_seen": 5257464,
      "step": 1324,
      "train_runtime": 1080.5231,
      "train_tokens_per_second": 4865.666
    },
    {
      "epoch": 101.84615384615384,
      "num_input_tokens_seen": 5257464,
      "step": 1324,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 101.92307692307692,
      "grad_norm": 0.6636524796485901,
      "learning_rate": 0.01,
      "loss": 4.5006232261657715,
      "num_input_tokens_seen": 5261560,
      "step": 1325,
      "train_runtime": 1081.3348,
      "train_tokens_per_second": 4865.801
    },
    {
      "epoch": 101.92307692307692,
      "num_input_tokens_seen": 5261560,
      "step": 1325,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 102.0,
      "grad_norm": 1.4672373533248901,
      "learning_rate": 0.01,
      "loss": 4.47036600112915,
      "num_input_tokens_seen": 5264016,
      "step": 1326,
      "train_runtime": 1081.8505,
      "train_tokens_per_second": 4865.752
    },
    {
      "epoch": 102.0,
      "num_input_tokens_seen": 5264016,
      "step": 1326,
      "train_accuracy": 0.04296875
    },
    {
      "epoch": 102.07692307692308,
      "grad_norm": 0.7811271548271179,
      "learning_rate": 0.01,
      "loss": 4.485345840454102,
      "num_input_tokens_seen": 5268112,
      "step": 1327,
      "train_runtime": 1082.6787,
      "train_tokens_per_second": 4865.813
    },
    {
      "epoch": 102.07692307692308,
      "num_input_tokens_seen": 5268112,
      "step": 1327,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 102.15384615384616,
      "grad_norm": 1.2522523403167725,
      "learning_rate": 0.01,
      "loss": 4.4972944259643555,
      "num_input_tokens_seen": 5272208,
      "step": 1328,
      "train_runtime": 1083.4943,
      "train_tokens_per_second": 4865.93
    },
    {
      "epoch": 102.15384615384616,
      "num_input_tokens_seen": 5272208,
      "step": 1328,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 102.23076923076923,
      "grad_norm": 0.5090000629425049,
      "learning_rate": 0.01,
      "loss": 4.478704452514648,
      "num_input_tokens_seen": 5276304,
      "step": 1329,
      "train_runtime": 1084.3101,
      "train_tokens_per_second": 4866.047
    },
    {
      "epoch": 102.23076923076923,
      "num_input_tokens_seen": 5276304,
      "step": 1329,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 102.3076923076923,
      "grad_norm": 1.3681564331054688,
      "learning_rate": 0.01,
      "loss": 4.505438804626465,
      "num_input_tokens_seen": 5280400,
      "step": 1330,
      "train_runtime": 1085.126,
      "train_tokens_per_second": 4866.163
    },
    {
      "epoch": 102.3076923076923,
      "num_input_tokens_seen": 5280400,
      "step": 1330,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 102.38461538461539,
      "grad_norm": 0.4047395586967468,
      "learning_rate": 0.01,
      "loss": 4.525785446166992,
      "num_input_tokens_seen": 5284496,
      "step": 1331,
      "train_runtime": 1085.9414,
      "train_tokens_per_second": 4866.281
    },
    {
      "epoch": 102.38461538461539,
      "num_input_tokens_seen": 5284496,
      "step": 1331,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 102.46153846153847,
      "grad_norm": 0.3903910517692566,
      "learning_rate": 0.01,
      "loss": 4.525243759155273,
      "num_input_tokens_seen": 5288592,
      "step": 1332,
      "train_runtime": 1086.7583,
      "train_tokens_per_second": 4866.392
    },
    {
      "epoch": 102.46153846153847,
      "num_input_tokens_seen": 5288592,
      "step": 1332,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 102.53846153846153,
      "grad_norm": 0.5233359336853027,
      "learning_rate": 0.01,
      "loss": 4.5092363357543945,
      "num_input_tokens_seen": 5292688,
      "step": 1333,
      "train_runtime": 1087.5731,
      "train_tokens_per_second": 4866.512
    },
    {
      "epoch": 102.53846153846153,
      "num_input_tokens_seen": 5292688,
      "step": 1333,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 102.61538461538461,
      "grad_norm": 0.8171457648277283,
      "learning_rate": 0.01,
      "loss": 4.50815486907959,
      "num_input_tokens_seen": 5296784,
      "step": 1334,
      "train_runtime": 1088.3892,
      "train_tokens_per_second": 4866.627
    },
    {
      "epoch": 102.61538461538461,
      "num_input_tokens_seen": 5296784,
      "step": 1334,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 102.6923076923077,
      "grad_norm": 0.47790205478668213,
      "learning_rate": 0.01,
      "loss": 4.499876976013184,
      "num_input_tokens_seen": 5300880,
      "step": 1335,
      "train_runtime": 1089.2054,
      "train_tokens_per_second": 4866.74
    },
    {
      "epoch": 102.6923076923077,
      "num_input_tokens_seen": 5300880,
      "step": 1335,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 102.76923076923077,
      "grad_norm": 0.4069991111755371,
      "learning_rate": 0.01,
      "loss": 4.518918991088867,
      "num_input_tokens_seen": 5304976,
      "step": 1336,
      "train_runtime": 1090.0216,
      "train_tokens_per_second": 4866.854
    },
    {
      "epoch": 102.76923076923077,
      "num_input_tokens_seen": 5304976,
      "step": 1336,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 102.84615384615384,
      "grad_norm": 0.507952094078064,
      "learning_rate": 0.01,
      "loss": 4.630908012390137,
      "num_input_tokens_seen": 5309072,
      "step": 1337,
      "train_runtime": 1090.8369,
      "train_tokens_per_second": 4866.971
    },
    {
      "epoch": 102.84615384615384,
      "num_input_tokens_seen": 5309072,
      "step": 1337,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 102.92307692307692,
      "grad_norm": 0.5505064129829407,
      "learning_rate": 0.01,
      "loss": 4.5251851081848145,
      "num_input_tokens_seen": 5313168,
      "step": 1338,
      "train_runtime": 1091.6491,
      "train_tokens_per_second": 4867.102
    },
    {
      "epoch": 102.92307692307692,
      "num_input_tokens_seen": 5313168,
      "step": 1338,
      "train_accuracy": 0.019543973729014397
    },
    {
      "epoch": 103.0,
      "grad_norm": 1.5184812545776367,
      "learning_rate": 0.01,
      "loss": 4.507806777954102,
      "num_input_tokens_seen": 5315624,
      "step": 1339,
      "train_runtime": 1092.1656,
      "train_tokens_per_second": 4867.049
    },
    {
      "epoch": 103.0,
      "num_input_tokens_seen": 5315624,
      "step": 1339,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 103.07692307692308,
      "grad_norm": 0.5381852984428406,
      "learning_rate": 0.01,
      "loss": 4.502768516540527,
      "num_input_tokens_seen": 5319720,
      "step": 1340,
      "train_runtime": 1092.9927,
      "train_tokens_per_second": 4867.114
    },
    {
      "epoch": 103.07692307692308,
      "num_input_tokens_seen": 5319720,
      "step": 1340,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 103.15384615384616,
      "grad_norm": 1.031249761581421,
      "learning_rate": 0.01,
      "loss": 4.466507911682129,
      "num_input_tokens_seen": 5323816,
      "step": 1341,
      "train_runtime": 1093.8079,
      "train_tokens_per_second": 4867.231
    },
    {
      "epoch": 103.15384615384616,
      "num_input_tokens_seen": 5323816,
      "step": 1341,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 103.23076923076923,
      "grad_norm": 0.4926789402961731,
      "learning_rate": 0.01,
      "loss": 4.506186485290527,
      "num_input_tokens_seen": 5327912,
      "step": 1342,
      "train_runtime": 1094.6238,
      "train_tokens_per_second": 4867.345
    },
    {
      "epoch": 103.23076923076923,
      "num_input_tokens_seen": 5327912,
      "step": 1342,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 103.3076923076923,
      "grad_norm": 1.5908927917480469,
      "learning_rate": 0.01,
      "loss": 4.501346588134766,
      "num_input_tokens_seen": 5332008,
      "step": 1343,
      "train_runtime": 1095.4394,
      "train_tokens_per_second": 4867.461
    },
    {
      "epoch": 103.3076923076923,
      "num_input_tokens_seen": 5332008,
      "step": 1343,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 103.38461538461539,
      "grad_norm": 1.313116431236267,
      "learning_rate": 0.01,
      "loss": 4.512948036193848,
      "num_input_tokens_seen": 5336104,
      "step": 1344,
      "train_runtime": 1096.2558,
      "train_tokens_per_second": 4867.572
    },
    {
      "epoch": 103.38461538461539,
      "num_input_tokens_seen": 5336104,
      "step": 1344,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 103.46153846153847,
      "grad_norm": 0.5012454986572266,
      "learning_rate": 0.01,
      "loss": 4.510312080383301,
      "num_input_tokens_seen": 5340200,
      "step": 1345,
      "train_runtime": 1097.0737,
      "train_tokens_per_second": 4867.677
    },
    {
      "epoch": 103.46153846153847,
      "num_input_tokens_seen": 5340200,
      "step": 1345,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 103.53846153846153,
      "grad_norm": 0.806168258190155,
      "learning_rate": 0.01,
      "loss": 4.508563995361328,
      "num_input_tokens_seen": 5344296,
      "step": 1346,
      "train_runtime": 1097.8883,
      "train_tokens_per_second": 4867.796
    },
    {
      "epoch": 103.53846153846153,
      "num_input_tokens_seen": 5344296,
      "step": 1346,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 103.61538461538461,
      "grad_norm": 0.4300622045993805,
      "learning_rate": 0.01,
      "loss": 4.521967887878418,
      "num_input_tokens_seen": 5348392,
      "step": 1347,
      "train_runtime": 1098.7041,
      "train_tokens_per_second": 4867.909
    },
    {
      "epoch": 103.61538461538461,
      "num_input_tokens_seen": 5348392,
      "step": 1347,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 103.6923076923077,
      "grad_norm": 0.5178453326225281,
      "learning_rate": 0.01,
      "loss": 4.544081687927246,
      "num_input_tokens_seen": 5352488,
      "step": 1348,
      "train_runtime": 1099.5192,
      "train_tokens_per_second": 4868.026
    },
    {
      "epoch": 103.6923076923077,
      "num_input_tokens_seen": 5352488,
      "step": 1348,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 103.76923076923077,
      "grad_norm": 0.5010653734207153,
      "learning_rate": 0.01,
      "loss": 4.522707939147949,
      "num_input_tokens_seen": 5356584,
      "step": 1349,
      "train_runtime": 1100.3345,
      "train_tokens_per_second": 4868.141
    },
    {
      "epoch": 103.76923076923077,
      "num_input_tokens_seen": 5356584,
      "step": 1349,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 103.84615384615384,
      "grad_norm": 0.40772804617881775,
      "learning_rate": 0.01,
      "loss": 4.531052589416504,
      "num_input_tokens_seen": 5360680,
      "step": 1350,
      "train_runtime": 1101.1514,
      "train_tokens_per_second": 4868.25
    },
    {
      "epoch": 103.84615384615384,
      "num_input_tokens_seen": 5360680,
      "step": 1350,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 103.92307692307692,
      "grad_norm": 1.338511347770691,
      "learning_rate": 0.01,
      "loss": 4.557485580444336,
      "num_input_tokens_seen": 5364776,
      "step": 1351,
      "train_runtime": 1101.9631,
      "train_tokens_per_second": 4868.381
    },
    {
      "epoch": 103.92307692307692,
      "num_input_tokens_seen": 5364776,
      "step": 1351,
      "train_accuracy": 0.013029315508902073
    },
    {
      "epoch": 104.0,
      "grad_norm": 1.1316109895706177,
      "learning_rate": 0.01,
      "loss": 4.522695541381836,
      "num_input_tokens_seen": 5367232,
      "step": 1352,
      "train_runtime": 1102.4797,
      "train_tokens_per_second": 4868.327
    },
    {
      "epoch": 104.0,
      "num_input_tokens_seen": 5367232,
      "step": 1352,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 104.07692307692308,
      "grad_norm": 0.2996485233306885,
      "learning_rate": 0.01,
      "loss": 4.48882532119751,
      "num_input_tokens_seen": 5371328,
      "step": 1353,
      "train_runtime": 1103.3075,
      "train_tokens_per_second": 4868.387
    },
    {
      "epoch": 104.07692307692308,
      "num_input_tokens_seen": 5371328,
      "step": 1353,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 104.15384615384616,
      "grad_norm": 1.3333604335784912,
      "learning_rate": 0.01,
      "loss": 4.502570629119873,
      "num_input_tokens_seen": 5375424,
      "step": 1354,
      "train_runtime": 1104.1237,
      "train_tokens_per_second": 4868.498
    },
    {
      "epoch": 104.15384615384616,
      "num_input_tokens_seen": 5375424,
      "step": 1354,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 104.23076923076923,
      "grad_norm": 0.4529602527618408,
      "learning_rate": 0.01,
      "loss": 4.515133857727051,
      "num_input_tokens_seen": 5379520,
      "step": 1355,
      "train_runtime": 1104.9438,
      "train_tokens_per_second": 4868.592
    },
    {
      "epoch": 104.23076923076923,
      "num_input_tokens_seen": 5379520,
      "step": 1355,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 104.3076923076923,
      "grad_norm": 0.45579686760902405,
      "learning_rate": 0.01,
      "loss": 4.46575927734375,
      "num_input_tokens_seen": 5383616,
      "step": 1356,
      "train_runtime": 1105.7616,
      "train_tokens_per_second": 4868.695
    },
    {
      "epoch": 104.3076923076923,
      "num_input_tokens_seen": 5383616,
      "step": 1356,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 104.38461538461539,
      "grad_norm": 0.5557273030281067,
      "learning_rate": 0.01,
      "loss": 4.466312408447266,
      "num_input_tokens_seen": 5387712,
      "step": 1357,
      "train_runtime": 1106.5764,
      "train_tokens_per_second": 4868.812
    },
    {
      "epoch": 104.38461538461539,
      "num_input_tokens_seen": 5387712,
      "step": 1357,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 104.46153846153847,
      "grad_norm": 0.46682295203208923,
      "learning_rate": 0.01,
      "loss": 4.481467247009277,
      "num_input_tokens_seen": 5391808,
      "step": 1358,
      "train_runtime": 1107.3915,
      "train_tokens_per_second": 4868.927
    },
    {
      "epoch": 104.46153846153847,
      "num_input_tokens_seen": 5391808,
      "step": 1358,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 104.53846153846153,
      "grad_norm": 0.6216288208961487,
      "learning_rate": 0.01,
      "loss": 4.555716037750244,
      "num_input_tokens_seen": 5395904,
      "step": 1359,
      "train_runtime": 1108.2062,
      "train_tokens_per_second": 4869.043
    },
    {
      "epoch": 104.53846153846153,
      "num_input_tokens_seen": 5395904,
      "step": 1359,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 104.61538461538461,
      "grad_norm": 0.5530970692634583,
      "learning_rate": 0.01,
      "loss": 4.535384178161621,
      "num_input_tokens_seen": 5400000,
      "step": 1360,
      "train_runtime": 1109.0219,
      "train_tokens_per_second": 4869.156
    },
    {
      "epoch": 104.61538461538461,
      "num_input_tokens_seen": 5400000,
      "step": 1360,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 104.6923076923077,
      "grad_norm": 0.2619735300540924,
      "learning_rate": 0.01,
      "loss": 4.543535232543945,
      "num_input_tokens_seen": 5404096,
      "step": 1361,
      "train_runtime": 1109.8376,
      "train_tokens_per_second": 4869.267
    },
    {
      "epoch": 104.6923076923077,
      "num_input_tokens_seen": 5404096,
      "step": 1361,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 104.76923076923077,
      "grad_norm": 1.2103418111801147,
      "learning_rate": 0.01,
      "loss": 4.506714820861816,
      "num_input_tokens_seen": 5408192,
      "step": 1362,
      "train_runtime": 1110.6543,
      "train_tokens_per_second": 4869.375
    },
    {
      "epoch": 104.76923076923077,
      "num_input_tokens_seen": 5408192,
      "step": 1362,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 104.84615384615384,
      "grad_norm": 0.7848754525184631,
      "learning_rate": 0.01,
      "loss": 4.537437438964844,
      "num_input_tokens_seen": 5412288,
      "step": 1363,
      "train_runtime": 1111.4723,
      "train_tokens_per_second": 4869.476
    },
    {
      "epoch": 104.84615384615384,
      "num_input_tokens_seen": 5412288,
      "step": 1363,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 104.92307692307692,
      "grad_norm": 1.4974429607391357,
      "learning_rate": 0.01,
      "loss": 4.492596626281738,
      "num_input_tokens_seen": 5416384,
      "step": 1364,
      "train_runtime": 1112.2834,
      "train_tokens_per_second": 4869.608
    },
    {
      "epoch": 104.92307692307692,
      "num_input_tokens_seen": 5416384,
      "step": 1364,
      "train_accuracy": 0.0032573288772255182
    },
    {
      "epoch": 105.0,
      "grad_norm": 0.6197435259819031,
      "learning_rate": 0.01,
      "loss": 4.542846202850342,
      "num_input_tokens_seen": 5418840,
      "step": 1365,
      "train_runtime": 1112.7997,
      "train_tokens_per_second": 4869.556
    },
    {
      "epoch": 105.0,
      "num_input_tokens_seen": 5418840,
      "step": 1365,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 105.07692307692308,
      "grad_norm": 5.34373140335083,
      "learning_rate": 0.01,
      "loss": 4.490141868591309,
      "num_input_tokens_seen": 5422936,
      "step": 1366,
      "train_runtime": 1113.6259,
      "train_tokens_per_second": 4869.621
    },
    {
      "epoch": 105.07692307692308,
      "num_input_tokens_seen": 5422936,
      "step": 1366,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 105.15384615384616,
      "grad_norm": 1.6519452333450317,
      "learning_rate": 0.01,
      "loss": 4.50926399230957,
      "num_input_tokens_seen": 5427032,
      "step": 1367,
      "train_runtime": 1114.4419,
      "train_tokens_per_second": 4869.731
    },
    {
      "epoch": 105.15384615384616,
      "num_input_tokens_seen": 5427032,
      "step": 1367,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 105.23076923076923,
      "grad_norm": 1.499245524406433,
      "learning_rate": 0.01,
      "loss": 4.476587295532227,
      "num_input_tokens_seen": 5431128,
      "step": 1368,
      "train_runtime": 1115.2575,
      "train_tokens_per_second": 4869.842
    },
    {
      "epoch": 105.23076923076923,
      "num_input_tokens_seen": 5431128,
      "step": 1368,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 105.3076923076923,
      "grad_norm": 0.3605392575263977,
      "learning_rate": 0.01,
      "loss": 4.499443054199219,
      "num_input_tokens_seen": 5435224,
      "step": 1369,
      "train_runtime": 1116.0729,
      "train_tokens_per_second": 4869.954
    },
    {
      "epoch": 105.3076923076923,
      "num_input_tokens_seen": 5435224,
      "step": 1369,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 105.38461538461539,
      "grad_norm": 0.5438734292984009,
      "learning_rate": 0.01,
      "loss": 4.489776611328125,
      "num_input_tokens_seen": 5439320,
      "step": 1370,
      "train_runtime": 1116.8867,
      "train_tokens_per_second": 4870.073
    },
    {
      "epoch": 105.38461538461539,
      "num_input_tokens_seen": 5439320,
      "step": 1370,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 105.46153846153847,
      "grad_norm": 0.3976382613182068,
      "learning_rate": 0.01,
      "loss": 4.520956993103027,
      "num_input_tokens_seen": 5443416,
      "step": 1371,
      "train_runtime": 1117.7014,
      "train_tokens_per_second": 4870.188
    },
    {
      "epoch": 105.46153846153847,
      "num_input_tokens_seen": 5443416,
      "step": 1371,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 105.53846153846153,
      "grad_norm": 0.591706395149231,
      "learning_rate": 0.01,
      "loss": 4.4757585525512695,
      "num_input_tokens_seen": 5447512,
      "step": 1372,
      "train_runtime": 1118.5156,
      "train_tokens_per_second": 4870.305
    },
    {
      "epoch": 105.53846153846153,
      "num_input_tokens_seen": 5447512,
      "step": 1372,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 105.61538461538461,
      "grad_norm": 0.7065641283988953,
      "learning_rate": 0.01,
      "loss": 4.515710830688477,
      "num_input_tokens_seen": 5451608,
      "step": 1373,
      "train_runtime": 1119.3306,
      "train_tokens_per_second": 4870.418
    },
    {
      "epoch": 105.61538461538461,
      "num_input_tokens_seen": 5451608,
      "step": 1373,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 105.6923076923077,
      "grad_norm": 0.2678771913051605,
      "learning_rate": 0.01,
      "loss": 4.480876922607422,
      "num_input_tokens_seen": 5455704,
      "step": 1374,
      "train_runtime": 1120.145,
      "train_tokens_per_second": 4870.534
    },
    {
      "epoch": 105.6923076923077,
      "num_input_tokens_seen": 5455704,
      "step": 1374,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 105.76923076923077,
      "grad_norm": 0.9524516463279724,
      "learning_rate": 0.01,
      "loss": 4.506048202514648,
      "num_input_tokens_seen": 5459800,
      "step": 1375,
      "train_runtime": 1120.9612,
      "train_tokens_per_second": 4870.641
    },
    {
      "epoch": 105.76923076923077,
      "num_input_tokens_seen": 5459800,
      "step": 1375,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 105.84615384615384,
      "grad_norm": 1.7505069971084595,
      "learning_rate": 0.01,
      "loss": 4.470991611480713,
      "num_input_tokens_seen": 5463896,
      "step": 1376,
      "train_runtime": 1121.7753,
      "train_tokens_per_second": 4870.758
    },
    {
      "epoch": 105.84615384615384,
      "num_input_tokens_seen": 5463896,
      "step": 1376,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 105.92307692307692,
      "grad_norm": 3.8635685443878174,
      "learning_rate": 0.01,
      "loss": 4.509904861450195,
      "num_input_tokens_seen": 5467992,
      "step": 1377,
      "train_runtime": 1122.5855,
      "train_tokens_per_second": 4870.892
    },
    {
      "epoch": 105.92307692307692,
      "num_input_tokens_seen": 5467992,
      "step": 1377,
      "train_accuracy": 0.02280130237340927
    },
    {
      "epoch": 106.0,
      "grad_norm": 0.8702496290206909,
      "learning_rate": 0.01,
      "loss": 4.480625629425049,
      "num_input_tokens_seen": 5470448,
      "step": 1378,
      "train_runtime": 1123.1018,
      "train_tokens_per_second": 4870.839
    },
    {
      "epoch": 106.0,
      "num_input_tokens_seen": 5470448,
      "step": 1378,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 106.07692307692308,
      "grad_norm": 0.7744153141975403,
      "learning_rate": 0.01,
      "loss": 4.470782279968262,
      "num_input_tokens_seen": 5474544,
      "step": 1379,
      "train_runtime": 1123.9284,
      "train_tokens_per_second": 4870.901
    },
    {
      "epoch": 106.07692307692308,
      "num_input_tokens_seen": 5474544,
      "step": 1379,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 106.15384615384616,
      "grad_norm": 1.867272973060608,
      "learning_rate": 0.01,
      "loss": 4.535487174987793,
      "num_input_tokens_seen": 5478640,
      "step": 1380,
      "train_runtime": 1124.7423,
      "train_tokens_per_second": 4871.018
    },
    {
      "epoch": 106.15384615384616,
      "num_input_tokens_seen": 5478640,
      "step": 1380,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 106.23076923076923,
      "grad_norm": 0.63252854347229,
      "learning_rate": 0.01,
      "loss": 4.462093353271484,
      "num_input_tokens_seen": 5482736,
      "step": 1381,
      "train_runtime": 1125.5573,
      "train_tokens_per_second": 4871.13
    },
    {
      "epoch": 106.23076923076923,
      "num_input_tokens_seen": 5482736,
      "step": 1381,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 106.3076923076923,
      "grad_norm": 0.937783420085907,
      "learning_rate": 0.01,
      "loss": 4.492024898529053,
      "num_input_tokens_seen": 5486832,
      "step": 1382,
      "train_runtime": 1126.3719,
      "train_tokens_per_second": 4871.243
    },
    {
      "epoch": 106.3076923076923,
      "num_input_tokens_seen": 5486832,
      "step": 1382,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 106.38461538461539,
      "grad_norm": 0.43855607509613037,
      "learning_rate": 0.01,
      "loss": 4.493302822113037,
      "num_input_tokens_seen": 5490928,
      "step": 1383,
      "train_runtime": 1127.1865,
      "train_tokens_per_second": 4871.357
    },
    {
      "epoch": 106.38461538461539,
      "num_input_tokens_seen": 5490928,
      "step": 1383,
      "train_accuracy": 0.03125
    },
    {
      "epoch": 106.46153846153847,
      "grad_norm": 0.45569297671318054,
      "learning_rate": 0.01,
      "loss": 4.524416923522949,
      "num_input_tokens_seen": 5495024,
      "step": 1384,
      "train_runtime": 1128.0007,
      "train_tokens_per_second": 4871.472
    },
    {
      "epoch": 106.46153846153847,
      "num_input_tokens_seen": 5495024,
      "step": 1384,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 106.53846153846153,
      "grad_norm": 1.6600298881530762,
      "learning_rate": 0.01,
      "loss": 4.5299072265625,
      "num_input_tokens_seen": 5499120,
      "step": 1385,
      "train_runtime": 1128.8157,
      "train_tokens_per_second": 4871.583
    },
    {
      "epoch": 106.53846153846153,
      "num_input_tokens_seen": 5499120,
      "step": 1385,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 106.61538461538461,
      "grad_norm": 1.8410983085632324,
      "learning_rate": 0.01,
      "loss": 4.5285539627075195,
      "num_input_tokens_seen": 5503216,
      "step": 1386,
      "train_runtime": 1129.6315,
      "train_tokens_per_second": 4871.691
    },
    {
      "epoch": 106.61538461538461,
      "num_input_tokens_seen": 5503216,
      "step": 1386,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 106.6923076923077,
      "grad_norm": 2.97233247756958,
      "learning_rate": 0.01,
      "loss": 4.5334272384643555,
      "num_input_tokens_seen": 5507312,
      "step": 1387,
      "train_runtime": 1130.4471,
      "train_tokens_per_second": 4871.8
    },
    {
      "epoch": 106.6923076923077,
      "num_input_tokens_seen": 5507312,
      "step": 1387,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 106.76923076923077,
      "grad_norm": 0.9362927675247192,
      "learning_rate": 0.01,
      "loss": 4.569294452667236,
      "num_input_tokens_seen": 5511408,
      "step": 1388,
      "train_runtime": 1131.2628,
      "train_tokens_per_second": 4871.908
    },
    {
      "epoch": 106.76923076923077,
      "num_input_tokens_seen": 5511408,
      "step": 1388,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 106.84615384615384,
      "grad_norm": 0.379261314868927,
      "learning_rate": 0.01,
      "loss": 4.547344207763672,
      "num_input_tokens_seen": 5515504,
      "step": 1389,
      "train_runtime": 1132.0776,
      "train_tokens_per_second": 4872.019
    },
    {
      "epoch": 106.84615384615384,
      "num_input_tokens_seen": 5515504,
      "step": 1389,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 106.92307692307692,
      "grad_norm": 0.4333992600440979,
      "learning_rate": 0.01,
      "loss": 4.518441200256348,
      "num_input_tokens_seen": 5519600,
      "step": 1390,
      "train_runtime": 1132.8871,
      "train_tokens_per_second": 4872.154
    },
    {
      "epoch": 106.92307692307692,
      "num_input_tokens_seen": 5519600,
      "step": 1390,
      "train_accuracy": 0.016286645084619522
    },
    {
      "epoch": 107.0,
      "grad_norm": 1.052661418914795,
      "learning_rate": 0.01,
      "loss": 4.530174255371094,
      "num_input_tokens_seen": 5522056,
      "step": 1391,
      "train_runtime": 1133.4022,
      "train_tokens_per_second": 4872.106
    },
    {
      "epoch": 107.0,
      "num_input_tokens_seen": 5522056,
      "step": 1391,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 107.07692307692308,
      "grad_norm": 0.3952656686306,
      "learning_rate": 0.01,
      "loss": 4.527536392211914,
      "num_input_tokens_seen": 5526152,
      "step": 1392,
      "train_runtime": 1134.2269,
      "train_tokens_per_second": 4872.175
    },
    {
      "epoch": 107.07692307692308,
      "num_input_tokens_seen": 5526152,
      "step": 1392,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 107.15384615384616,
      "grad_norm": 0.43859216570854187,
      "learning_rate": 0.01,
      "loss": 4.522705554962158,
      "num_input_tokens_seen": 5530248,
      "step": 1393,
      "train_runtime": 1135.0413,
      "train_tokens_per_second": 4872.288
    },
    {
      "epoch": 107.15384615384616,
      "num_input_tokens_seen": 5530248,
      "step": 1393,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 107.23076923076923,
      "grad_norm": 0.5607250332832336,
      "learning_rate": 0.01,
      "loss": 4.516456604003906,
      "num_input_tokens_seen": 5534344,
      "step": 1394,
      "train_runtime": 1135.8572,
      "train_tokens_per_second": 4872.394
    },
    {
      "epoch": 107.23076923076923,
      "num_input_tokens_seen": 5534344,
      "step": 1394,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 107.3076923076923,
      "grad_norm": 0.3729170858860016,
      "learning_rate": 0.01,
      "loss": 4.525454998016357,
      "num_input_tokens_seen": 5538440,
      "step": 1395,
      "train_runtime": 1136.6725,
      "train_tokens_per_second": 4872.503
    },
    {
      "epoch": 107.3076923076923,
      "num_input_tokens_seen": 5538440,
      "step": 1395,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 107.38461538461539,
      "grad_norm": 0.334216445684433,
      "learning_rate": 0.01,
      "loss": 4.485908508300781,
      "num_input_tokens_seen": 5542536,
      "step": 1396,
      "train_runtime": 1137.4868,
      "train_tokens_per_second": 4872.616
    },
    {
      "epoch": 107.38461538461539,
      "num_input_tokens_seen": 5542536,
      "step": 1396,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 107.46153846153847,
      "grad_norm": 0.5450636744499207,
      "learning_rate": 0.01,
      "loss": 4.540503025054932,
      "num_input_tokens_seen": 5546632,
      "step": 1397,
      "train_runtime": 1138.3031,
      "train_tokens_per_second": 4872.72
    },
    {
      "epoch": 107.46153846153847,
      "num_input_tokens_seen": 5546632,
      "step": 1397,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 107.53846153846153,
      "grad_norm": 0.5269734859466553,
      "learning_rate": 0.01,
      "loss": 4.557070255279541,
      "num_input_tokens_seen": 5550728,
      "step": 1398,
      "train_runtime": 1139.118,
      "train_tokens_per_second": 4872.83
    },
    {
      "epoch": 107.53846153846153,
      "num_input_tokens_seen": 5550728,
      "step": 1398,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 107.61538461538461,
      "grad_norm": 0.5704514384269714,
      "learning_rate": 0.01,
      "loss": 4.560052871704102,
      "num_input_tokens_seen": 5554824,
      "step": 1399,
      "train_runtime": 1139.933,
      "train_tokens_per_second": 4872.939
    },
    {
      "epoch": 107.61538461538461,
      "num_input_tokens_seen": 5554824,
      "step": 1399,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 107.6923076923077,
      "grad_norm": 0.6849747896194458,
      "learning_rate": 0.01,
      "loss": 4.529512405395508,
      "num_input_tokens_seen": 5558920,
      "step": 1400,
      "train_runtime": 1140.7474,
      "train_tokens_per_second": 4873.051
    },
    {
      "epoch": 107.6923076923077,
      "eval_CMD_3_branch_eval_accuracy": 0.00976491862567812,
      "eval_CMD_3_branch_eval_loss": 4.564810276031494,
      "eval_CMD_3_branch_eval_runtime": 1.1518,
      "eval_CMD_3_branch_eval_samples_per_second": 2400.514,
      "eval_CMD_3_branch_eval_steps_per_second": 5.209,
      "num_input_tokens_seen": 5558920,
      "step": 1400
    },
    {
      "epoch": 107.6923076923077,
      "num_input_tokens_seen": 5558920,
      "step": 1400,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 107.76923076923077,
      "grad_norm": 0.39933985471725464,
      "learning_rate": 0.01,
      "loss": 4.526744842529297,
      "num_input_tokens_seen": 5563016,
      "step": 1401,
      "train_runtime": 1142.7177,
      "train_tokens_per_second": 4868.233
    },
    {
      "epoch": 107.76923076923077,
      "num_input_tokens_seen": 5563016,
      "step": 1401,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 107.84615384615384,
      "grad_norm": 0.5478185415267944,
      "learning_rate": 0.01,
      "loss": 4.520236968994141,
      "num_input_tokens_seen": 5567112,
      "step": 1402,
      "train_runtime": 1143.5327,
      "train_tokens_per_second": 4868.345
    },
    {
      "epoch": 107.84615384615384,
      "num_input_tokens_seen": 5567112,
      "step": 1402,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 107.92307692307692,
      "grad_norm": 0.36211809515953064,
      "learning_rate": 0.01,
      "loss": 4.5181403160095215,
      "num_input_tokens_seen": 5571208,
      "step": 1403,
      "train_runtime": 1144.3426,
      "train_tokens_per_second": 4868.479
    },
    {
      "epoch": 107.92307692307692,
      "num_input_tokens_seen": 5571208,
      "step": 1403,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 108.0,
      "grad_norm": 0.6052718758583069,
      "learning_rate": 0.01,
      "loss": 4.526625633239746,
      "num_input_tokens_seen": 5573664,
      "step": 1404,
      "train_runtime": 1144.8593,
      "train_tokens_per_second": 4868.427
    },
    {
      "epoch": 108.0,
      "num_input_tokens_seen": 5573664,
      "step": 1404,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 108.07692307692308,
      "grad_norm": 0.41202157735824585,
      "learning_rate": 0.01,
      "loss": 4.466943740844727,
      "num_input_tokens_seen": 5577760,
      "step": 1405,
      "train_runtime": 1145.687,
      "train_tokens_per_second": 4868.485
    },
    {
      "epoch": 108.07692307692308,
      "num_input_tokens_seen": 5577760,
      "step": 1405,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 108.15384615384616,
      "grad_norm": 0.4319964051246643,
      "learning_rate": 0.01,
      "loss": 4.516010284423828,
      "num_input_tokens_seen": 5581856,
      "step": 1406,
      "train_runtime": 1146.5027,
      "train_tokens_per_second": 4868.594
    },
    {
      "epoch": 108.15384615384616,
      "num_input_tokens_seen": 5581856,
      "step": 1406,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 108.23076923076923,
      "grad_norm": 0.34288203716278076,
      "learning_rate": 0.01,
      "loss": 4.507363796234131,
      "num_input_tokens_seen": 5585952,
      "step": 1407,
      "train_runtime": 1147.3191,
      "train_tokens_per_second": 4868.7
    },
    {
      "epoch": 108.23076923076923,
      "num_input_tokens_seen": 5585952,
      "step": 1407,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 108.3076923076923,
      "grad_norm": 0.7752683758735657,
      "learning_rate": 0.01,
      "loss": 4.497525215148926,
      "num_input_tokens_seen": 5590048,
      "step": 1408,
      "train_runtime": 1148.1337,
      "train_tokens_per_second": 4868.813
    },
    {
      "epoch": 108.3076923076923,
      "num_input_tokens_seen": 5590048,
      "step": 1408,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 108.38461538461539,
      "grad_norm": 1.1515727043151855,
      "learning_rate": 0.01,
      "loss": 4.53062105178833,
      "num_input_tokens_seen": 5594144,
      "step": 1409,
      "train_runtime": 1148.9479,
      "train_tokens_per_second": 4868.928
    },
    {
      "epoch": 108.38461538461539,
      "num_input_tokens_seen": 5594144,
      "step": 1409,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 108.46153846153847,
      "grad_norm": 0.6937203407287598,
      "learning_rate": 0.01,
      "loss": 4.519171237945557,
      "num_input_tokens_seen": 5598240,
      "step": 1410,
      "train_runtime": 1149.7623,
      "train_tokens_per_second": 4869.041
    },
    {
      "epoch": 108.46153846153847,
      "num_input_tokens_seen": 5598240,
      "step": 1410,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 108.53846153846153,
      "grad_norm": 0.5314910411834717,
      "learning_rate": 0.01,
      "loss": 4.530661106109619,
      "num_input_tokens_seen": 5602336,
      "step": 1411,
      "train_runtime": 1150.5771,
      "train_tokens_per_second": 4869.153
    },
    {
      "epoch": 108.53846153846153,
      "num_input_tokens_seen": 5602336,
      "step": 1411,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 108.61538461538461,
      "grad_norm": 1.1272478103637695,
      "learning_rate": 0.01,
      "loss": 4.53105354309082,
      "num_input_tokens_seen": 5606432,
      "step": 1412,
      "train_runtime": 1151.3918,
      "train_tokens_per_second": 4869.265
    },
    {
      "epoch": 108.61538461538461,
      "num_input_tokens_seen": 5606432,
      "step": 1412,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 108.6923076923077,
      "grad_norm": 0.9376735687255859,
      "learning_rate": 0.01,
      "loss": 4.627614498138428,
      "num_input_tokens_seen": 5610528,
      "step": 1413,
      "train_runtime": 1152.2063,
      "train_tokens_per_second": 4869.378
    },
    {
      "epoch": 108.6923076923077,
      "num_input_tokens_seen": 5610528,
      "step": 1413,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 108.76923076923077,
      "grad_norm": 0.747143566608429,
      "learning_rate": 0.01,
      "loss": 4.748932838439941,
      "num_input_tokens_seen": 5614624,
      "step": 1414,
      "train_runtime": 1153.0215,
      "train_tokens_per_second": 4869.488
    },
    {
      "epoch": 108.76923076923077,
      "num_input_tokens_seen": 5614624,
      "step": 1414,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 108.84615384615384,
      "grad_norm": 0.4503246247768402,
      "learning_rate": 0.01,
      "loss": 4.613273620605469,
      "num_input_tokens_seen": 5618720,
      "step": 1415,
      "train_runtime": 1153.8353,
      "train_tokens_per_second": 4869.603
    },
    {
      "epoch": 108.84615384615384,
      "num_input_tokens_seen": 5618720,
      "step": 1415,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 108.92307692307692,
      "grad_norm": 1.1499062776565552,
      "learning_rate": 0.01,
      "loss": 4.5845417976379395,
      "num_input_tokens_seen": 5622816,
      "step": 1416,
      "train_runtime": 1154.646,
      "train_tokens_per_second": 4869.732
    },
    {
      "epoch": 108.92307692307692,
      "num_input_tokens_seen": 5622816,
      "step": 1416,
      "train_accuracy": 0.013029315508902073
    },
    {
      "epoch": 109.0,
      "grad_norm": 0.33534467220306396,
      "learning_rate": 0.01,
      "loss": 4.5637311935424805,
      "num_input_tokens_seen": 5625272,
      "step": 1417,
      "train_runtime": 1155.1613,
      "train_tokens_per_second": 4869.685
    },
    {
      "epoch": 109.0,
      "num_input_tokens_seen": 5625272,
      "step": 1417,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 109.07692307692308,
      "grad_norm": 0.4022475481033325,
      "learning_rate": 0.01,
      "loss": 4.535723686218262,
      "num_input_tokens_seen": 5629368,
      "step": 1418,
      "train_runtime": 1155.9864,
      "train_tokens_per_second": 4869.753
    },
    {
      "epoch": 109.07692307692308,
      "num_input_tokens_seen": 5629368,
      "step": 1418,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 109.15384615384616,
      "grad_norm": 0.2704116106033325,
      "learning_rate": 0.01,
      "loss": 4.557677268981934,
      "num_input_tokens_seen": 5633464,
      "step": 1419,
      "train_runtime": 1156.7996,
      "train_tokens_per_second": 4869.87
    },
    {
      "epoch": 109.15384615384616,
      "num_input_tokens_seen": 5633464,
      "step": 1419,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 109.23076923076923,
      "grad_norm": 0.604241132736206,
      "learning_rate": 0.01,
      "loss": 4.562737464904785,
      "num_input_tokens_seen": 5637560,
      "step": 1420,
      "train_runtime": 1157.6133,
      "train_tokens_per_second": 4869.985
    },
    {
      "epoch": 109.23076923076923,
      "num_input_tokens_seen": 5637560,
      "step": 1420,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 109.3076923076923,
      "grad_norm": 0.5170044302940369,
      "learning_rate": 0.01,
      "loss": 4.5539093017578125,
      "num_input_tokens_seen": 5641656,
      "step": 1421,
      "train_runtime": 1158.4268,
      "train_tokens_per_second": 4870.101
    },
    {
      "epoch": 109.3076923076923,
      "num_input_tokens_seen": 5641656,
      "step": 1421,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 109.38461538461539,
      "grad_norm": 0.4709688723087311,
      "learning_rate": 0.01,
      "loss": 4.597808837890625,
      "num_input_tokens_seen": 5645752,
      "step": 1422,
      "train_runtime": 1159.2413,
      "train_tokens_per_second": 4870.213
    },
    {
      "epoch": 109.38461538461539,
      "num_input_tokens_seen": 5645752,
      "step": 1422,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 109.46153846153847,
      "grad_norm": 0.43736159801483154,
      "learning_rate": 0.01,
      "loss": 4.583188056945801,
      "num_input_tokens_seen": 5649848,
      "step": 1423,
      "train_runtime": 1160.0589,
      "train_tokens_per_second": 4870.312
    },
    {
      "epoch": 109.46153846153847,
      "num_input_tokens_seen": 5649848,
      "step": 1423,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 109.53846153846153,
      "grad_norm": 0.465040385723114,
      "learning_rate": 0.01,
      "loss": 4.599481582641602,
      "num_input_tokens_seen": 5653944,
      "step": 1424,
      "train_runtime": 1160.8744,
      "train_tokens_per_second": 4870.418
    },
    {
      "epoch": 109.53846153846153,
      "num_input_tokens_seen": 5653944,
      "step": 1424,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 109.61538461538461,
      "grad_norm": 0.5211769342422485,
      "learning_rate": 0.01,
      "loss": 4.568815231323242,
      "num_input_tokens_seen": 5658040,
      "step": 1425,
      "train_runtime": 1161.6874,
      "train_tokens_per_second": 4870.536
    },
    {
      "epoch": 109.61538461538461,
      "num_input_tokens_seen": 5658040,
      "step": 1425,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 109.6923076923077,
      "grad_norm": 0.398411363363266,
      "learning_rate": 0.01,
      "loss": 4.551018714904785,
      "num_input_tokens_seen": 5662136,
      "step": 1426,
      "train_runtime": 1162.505,
      "train_tokens_per_second": 4870.634
    },
    {
      "epoch": 109.6923076923077,
      "num_input_tokens_seen": 5662136,
      "step": 1426,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 109.76923076923077,
      "grad_norm": 4.086050510406494,
      "learning_rate": 0.01,
      "loss": 4.555219650268555,
      "num_input_tokens_seen": 5666232,
      "step": 1427,
      "train_runtime": 1163.3179,
      "train_tokens_per_second": 4870.751
    },
    {
      "epoch": 109.76923076923077,
      "num_input_tokens_seen": 5666232,
      "step": 1427,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 109.84615384615384,
      "grad_norm": 0.28065016865730286,
      "learning_rate": 0.01,
      "loss": 4.5995683670043945,
      "num_input_tokens_seen": 5670328,
      "step": 1428,
      "train_runtime": 1164.1324,
      "train_tokens_per_second": 4870.862
    },
    {
      "epoch": 109.84615384615384,
      "num_input_tokens_seen": 5670328,
      "step": 1428,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 109.92307692307692,
      "grad_norm": 0.1984911561012268,
      "learning_rate": 0.01,
      "loss": 4.5547380447387695,
      "num_input_tokens_seen": 5674424,
      "step": 1429,
      "train_runtime": 1164.9414,
      "train_tokens_per_second": 4870.995
    },
    {
      "epoch": 109.92307692307692,
      "num_input_tokens_seen": 5674424,
      "step": 1429,
      "train_accuracy": 0.019543973729014397
    },
    {
      "epoch": 110.0,
      "grad_norm": 0.3055298328399658,
      "learning_rate": 0.01,
      "loss": 4.543587684631348,
      "num_input_tokens_seen": 5676880,
      "step": 1430,
      "train_runtime": 1165.4563,
      "train_tokens_per_second": 4870.951
    },
    {
      "epoch": 110.0,
      "num_input_tokens_seen": 5676880,
      "step": 1430,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 110.07692307692308,
      "grad_norm": 0.1978810727596283,
      "learning_rate": 0.01,
      "loss": 4.543689250946045,
      "num_input_tokens_seen": 5680976,
      "step": 1431,
      "train_runtime": 1166.2828,
      "train_tokens_per_second": 4871.011
    },
    {
      "epoch": 110.07692307692308,
      "num_input_tokens_seen": 5680976,
      "step": 1431,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 110.15384615384616,
      "grad_norm": 0.37786683440208435,
      "learning_rate": 0.01,
      "loss": 4.540652275085449,
      "num_input_tokens_seen": 5685072,
      "step": 1432,
      "train_runtime": 1167.0982,
      "train_tokens_per_second": 4871.117
    },
    {
      "epoch": 110.15384615384616,
      "num_input_tokens_seen": 5685072,
      "step": 1432,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 110.23076923076923,
      "grad_norm": 0.2681422531604767,
      "learning_rate": 0.01,
      "loss": 4.540897846221924,
      "num_input_tokens_seen": 5689168,
      "step": 1433,
      "train_runtime": 1167.9118,
      "train_tokens_per_second": 4871.231
    },
    {
      "epoch": 110.23076923076923,
      "num_input_tokens_seen": 5689168,
      "step": 1433,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 110.3076923076923,
      "grad_norm": 0.34770667552948,
      "learning_rate": 0.01,
      "loss": 4.5776166915893555,
      "num_input_tokens_seen": 5693264,
      "step": 1434,
      "train_runtime": 1168.7269,
      "train_tokens_per_second": 4871.338
    },
    {
      "epoch": 110.3076923076923,
      "num_input_tokens_seen": 5693264,
      "step": 1434,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 110.38461538461539,
      "grad_norm": 0.35044267773628235,
      "learning_rate": 0.01,
      "loss": 4.558687210083008,
      "num_input_tokens_seen": 5697360,
      "step": 1435,
      "train_runtime": 1169.5418,
      "train_tokens_per_second": 4871.446
    },
    {
      "epoch": 110.38461538461539,
      "num_input_tokens_seen": 5697360,
      "step": 1435,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 110.46153846153847,
      "grad_norm": 0.27794668078422546,
      "learning_rate": 0.01,
      "loss": 4.535930156707764,
      "num_input_tokens_seen": 5701456,
      "step": 1436,
      "train_runtime": 1170.3581,
      "train_tokens_per_second": 4871.548
    },
    {
      "epoch": 110.46153846153847,
      "num_input_tokens_seen": 5701456,
      "step": 1436,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 110.53846153846153,
      "grad_norm": 0.39730581641197205,
      "learning_rate": 0.01,
      "loss": 4.593379974365234,
      "num_input_tokens_seen": 5705552,
      "step": 1437,
      "train_runtime": 1171.1722,
      "train_tokens_per_second": 4871.659
    },
    {
      "epoch": 110.53846153846153,
      "num_input_tokens_seen": 5705552,
      "step": 1437,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 110.61538461538461,
      "grad_norm": 0.33294665813446045,
      "learning_rate": 0.01,
      "loss": 4.538558006286621,
      "num_input_tokens_seen": 5709648,
      "step": 1438,
      "train_runtime": 1171.986,
      "train_tokens_per_second": 4871.772
    },
    {
      "epoch": 110.61538461538461,
      "num_input_tokens_seen": 5709648,
      "step": 1438,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 110.6923076923077,
      "grad_norm": 0.33621513843536377,
      "learning_rate": 0.01,
      "loss": 4.540549278259277,
      "num_input_tokens_seen": 5713744,
      "step": 1439,
      "train_runtime": 1172.8,
      "train_tokens_per_second": 4871.883
    },
    {
      "epoch": 110.6923076923077,
      "num_input_tokens_seen": 5713744,
      "step": 1439,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 110.76923076923077,
      "grad_norm": 0.3213450610637665,
      "learning_rate": 0.01,
      "loss": 4.538887977600098,
      "num_input_tokens_seen": 5717840,
      "step": 1440,
      "train_runtime": 1173.6146,
      "train_tokens_per_second": 4871.991
    },
    {
      "epoch": 110.76923076923077,
      "num_input_tokens_seen": 5717840,
      "step": 1440,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 110.84615384615384,
      "grad_norm": 0.2600676417350769,
      "learning_rate": 0.01,
      "loss": 4.570079803466797,
      "num_input_tokens_seen": 5721936,
      "step": 1441,
      "train_runtime": 1174.4289,
      "train_tokens_per_second": 4872.101
    },
    {
      "epoch": 110.84615384615384,
      "num_input_tokens_seen": 5721936,
      "step": 1441,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 110.92307692307692,
      "grad_norm": 12.039423942565918,
      "learning_rate": 0.01,
      "loss": 4.559736728668213,
      "num_input_tokens_seen": 5726032,
      "step": 1442,
      "train_runtime": 1175.2394,
      "train_tokens_per_second": 4872.226
    },
    {
      "epoch": 110.92307692307692,
      "num_input_tokens_seen": 5726032,
      "step": 1442,
      "train_accuracy": 0.016286645084619522
    },
    {
      "epoch": 111.0,
      "grad_norm": 0.5153545141220093,
      "learning_rate": 0.01,
      "loss": 4.533757209777832,
      "num_input_tokens_seen": 5728488,
      "step": 1443,
      "train_runtime": 1175.7551,
      "train_tokens_per_second": 4872.178
    },
    {
      "epoch": 111.0,
      "num_input_tokens_seen": 5728488,
      "step": 1443,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 111.07692307692308,
      "grad_norm": 0.49407270550727844,
      "learning_rate": 0.01,
      "loss": 4.524785995483398,
      "num_input_tokens_seen": 5732584,
      "step": 1444,
      "train_runtime": 1176.5822,
      "train_tokens_per_second": 4872.234
    },
    {
      "epoch": 111.07692307692308,
      "num_input_tokens_seen": 5732584,
      "step": 1444,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 111.15384615384616,
      "grad_norm": 4.117186546325684,
      "learning_rate": 0.01,
      "loss": 4.543494701385498,
      "num_input_tokens_seen": 5736680,
      "step": 1445,
      "train_runtime": 1177.3963,
      "train_tokens_per_second": 4872.344
    },
    {
      "epoch": 111.15384615384616,
      "num_input_tokens_seen": 5736680,
      "step": 1445,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 111.23076923076923,
      "grad_norm": 0.6634436249732971,
      "learning_rate": 0.01,
      "loss": 4.539254188537598,
      "num_input_tokens_seen": 5740776,
      "step": 1446,
      "train_runtime": 1178.2103,
      "train_tokens_per_second": 4872.455
    },
    {
      "epoch": 111.23076923076923,
      "num_input_tokens_seen": 5740776,
      "step": 1446,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 111.3076923076923,
      "grad_norm": 0.4485829174518585,
      "learning_rate": 0.01,
      "loss": 4.550300598144531,
      "num_input_tokens_seen": 5744872,
      "step": 1447,
      "train_runtime": 1179.0239,
      "train_tokens_per_second": 4872.566
    },
    {
      "epoch": 111.3076923076923,
      "num_input_tokens_seen": 5744872,
      "step": 1447,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 111.38461538461539,
      "grad_norm": 0.3060574233531952,
      "learning_rate": 0.01,
      "loss": 4.542709827423096,
      "num_input_tokens_seen": 5748968,
      "step": 1448,
      "train_runtime": 1179.8382,
      "train_tokens_per_second": 4872.675
    },
    {
      "epoch": 111.38461538461539,
      "num_input_tokens_seen": 5748968,
      "step": 1448,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 111.46153846153847,
      "grad_norm": 1.7807625532150269,
      "learning_rate": 0.01,
      "loss": 4.533871650695801,
      "num_input_tokens_seen": 5753064,
      "step": 1449,
      "train_runtime": 1180.6555,
      "train_tokens_per_second": 4872.771
    },
    {
      "epoch": 111.46153846153847,
      "num_input_tokens_seen": 5753064,
      "step": 1449,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 111.53846153846153,
      "grad_norm": 0.7963235974311829,
      "learning_rate": 0.01,
      "loss": 4.5543951988220215,
      "num_input_tokens_seen": 5757160,
      "step": 1450,
      "train_runtime": 1181.4708,
      "train_tokens_per_second": 4872.875
    },
    {
      "epoch": 111.53846153846153,
      "num_input_tokens_seen": 5757160,
      "step": 1450,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 111.61538461538461,
      "grad_norm": 0.29571372270584106,
      "learning_rate": 0.01,
      "loss": 4.5262908935546875,
      "num_input_tokens_seen": 5761256,
      "step": 1451,
      "train_runtime": 1182.2854,
      "train_tokens_per_second": 4872.982
    },
    {
      "epoch": 111.61538461538461,
      "num_input_tokens_seen": 5761256,
      "step": 1451,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 111.6923076923077,
      "grad_norm": 0.38416430354118347,
      "learning_rate": 0.01,
      "loss": 4.546797752380371,
      "num_input_tokens_seen": 5765352,
      "step": 1452,
      "train_runtime": 1183.1004,
      "train_tokens_per_second": 4873.088
    },
    {
      "epoch": 111.6923076923077,
      "num_input_tokens_seen": 5765352,
      "step": 1452,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 111.76923076923077,
      "grad_norm": 0.24193359911441803,
      "learning_rate": 0.01,
      "loss": 4.554234981536865,
      "num_input_tokens_seen": 5769448,
      "step": 1453,
      "train_runtime": 1183.9152,
      "train_tokens_per_second": 4873.194
    },
    {
      "epoch": 111.76923076923077,
      "num_input_tokens_seen": 5769448,
      "step": 1453,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 111.84615384615384,
      "grad_norm": 0.5733964443206787,
      "learning_rate": 0.01,
      "loss": 4.5706682205200195,
      "num_input_tokens_seen": 5773544,
      "step": 1454,
      "train_runtime": 1184.7307,
      "train_tokens_per_second": 4873.296
    },
    {
      "epoch": 111.84615384615384,
      "num_input_tokens_seen": 5773544,
      "step": 1454,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 111.92307692307692,
      "grad_norm": 0.2824409604072571,
      "learning_rate": 0.01,
      "loss": 4.55647087097168,
      "num_input_tokens_seen": 5777640,
      "step": 1455,
      "train_runtime": 1185.5406,
      "train_tokens_per_second": 4873.422
    },
    {
      "epoch": 111.92307692307692,
      "num_input_tokens_seen": 5777640,
      "step": 1455,
      "train_accuracy": 0.0032573288772255182
    },
    {
      "epoch": 112.0,
      "grad_norm": 0.31641995906829834,
      "learning_rate": 0.01,
      "loss": 4.55803108215332,
      "num_input_tokens_seen": 5780096,
      "step": 1456,
      "train_runtime": 1186.0571,
      "train_tokens_per_second": 4873.371
    },
    {
      "epoch": 112.0,
      "num_input_tokens_seen": 5780096,
      "step": 1456,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 112.07692307692308,
      "grad_norm": 0.21771086752414703,
      "learning_rate": 0.01,
      "loss": 4.534934043884277,
      "num_input_tokens_seen": 5784192,
      "step": 1457,
      "train_runtime": 1186.8848,
      "train_tokens_per_second": 4873.423
    },
    {
      "epoch": 112.07692307692308,
      "num_input_tokens_seen": 5784192,
      "step": 1457,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 112.15384615384616,
      "grad_norm": 0.29175636172294617,
      "learning_rate": 0.01,
      "loss": 4.509088516235352,
      "num_input_tokens_seen": 5788288,
      "step": 1458,
      "train_runtime": 1187.699,
      "train_tokens_per_second": 4873.531
    },
    {
      "epoch": 112.15384615384616,
      "num_input_tokens_seen": 5788288,
      "step": 1458,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 112.23076923076923,
      "grad_norm": 0.35472649335861206,
      "learning_rate": 0.01,
      "loss": 4.531649589538574,
      "num_input_tokens_seen": 5792384,
      "step": 1459,
      "train_runtime": 1188.5124,
      "train_tokens_per_second": 4873.642
    },
    {
      "epoch": 112.23076923076923,
      "num_input_tokens_seen": 5792384,
      "step": 1459,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 112.3076923076923,
      "grad_norm": 0.5154550075531006,
      "learning_rate": 0.01,
      "loss": 4.5805745124816895,
      "num_input_tokens_seen": 5796480,
      "step": 1460,
      "train_runtime": 1189.3254,
      "train_tokens_per_second": 4873.754
    },
    {
      "epoch": 112.3076923076923,
      "num_input_tokens_seen": 5796480,
      "step": 1460,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 112.38461538461539,
      "grad_norm": 0.5546479225158691,
      "learning_rate": 0.01,
      "loss": 4.542065143585205,
      "num_input_tokens_seen": 5800576,
      "step": 1461,
      "train_runtime": 1190.1392,
      "train_tokens_per_second": 4873.864
    },
    {
      "epoch": 112.38461538461539,
      "num_input_tokens_seen": 5800576,
      "step": 1461,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 112.46153846153847,
      "grad_norm": 0.4114753305912018,
      "learning_rate": 0.01,
      "loss": 4.521361351013184,
      "num_input_tokens_seen": 5804672,
      "step": 1462,
      "train_runtime": 1190.9546,
      "train_tokens_per_second": 4873.966
    },
    {
      "epoch": 112.46153846153847,
      "num_input_tokens_seen": 5804672,
      "step": 1462,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 112.53846153846153,
      "grad_norm": 0.3744247257709503,
      "learning_rate": 0.01,
      "loss": 4.540783882141113,
      "num_input_tokens_seen": 5808768,
      "step": 1463,
      "train_runtime": 1191.7683,
      "train_tokens_per_second": 4874.075
    },
    {
      "epoch": 112.53846153846153,
      "num_input_tokens_seen": 5808768,
      "step": 1463,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 112.61538461538461,
      "grad_norm": 0.17463082075119019,
      "learning_rate": 0.01,
      "loss": 4.549612045288086,
      "num_input_tokens_seen": 5812864,
      "step": 1464,
      "train_runtime": 1192.5821,
      "train_tokens_per_second": 4874.184
    },
    {
      "epoch": 112.61538461538461,
      "num_input_tokens_seen": 5812864,
      "step": 1464,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 112.6923076923077,
      "grad_norm": 0.4204837381839752,
      "learning_rate": 0.01,
      "loss": 4.537134170532227,
      "num_input_tokens_seen": 5816960,
      "step": 1465,
      "train_runtime": 1193.396,
      "train_tokens_per_second": 4874.291
    },
    {
      "epoch": 112.6923076923077,
      "num_input_tokens_seen": 5816960,
      "step": 1465,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 112.76923076923077,
      "grad_norm": 0.6896317005157471,
      "learning_rate": 0.01,
      "loss": 4.549983024597168,
      "num_input_tokens_seen": 5821056,
      "step": 1466,
      "train_runtime": 1194.2101,
      "train_tokens_per_second": 4874.398
    },
    {
      "epoch": 112.76923076923077,
      "num_input_tokens_seen": 5821056,
      "step": 1466,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 112.84615384615384,
      "grad_norm": 0.37896084785461426,
      "learning_rate": 0.01,
      "loss": 4.571794509887695,
      "num_input_tokens_seen": 5825152,
      "step": 1467,
      "train_runtime": 1195.0239,
      "train_tokens_per_second": 4874.507
    },
    {
      "epoch": 112.84615384615384,
      "num_input_tokens_seen": 5825152,
      "step": 1467,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 112.92307692307692,
      "grad_norm": 0.21988117694854736,
      "learning_rate": 0.01,
      "loss": 4.540635108947754,
      "num_input_tokens_seen": 5829248,
      "step": 1468,
      "train_runtime": 1195.8336,
      "train_tokens_per_second": 4874.631
    },
    {
      "epoch": 112.92307692307692,
      "num_input_tokens_seen": 5829248,
      "step": 1468,
      "train_accuracy": 0.0
    },
    {
      "epoch": 113.0,
      "grad_norm": 0.5901161432266235,
      "learning_rate": 0.01,
      "loss": 4.515387058258057,
      "num_input_tokens_seen": 5831704,
      "step": 1469,
      "train_runtime": 1196.348,
      "train_tokens_per_second": 4874.588
    },
    {
      "epoch": 113.0,
      "num_input_tokens_seen": 5831704,
      "step": 1469,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 113.07692307692308,
      "grad_norm": 0.3445182144641876,
      "learning_rate": 0.01,
      "loss": 4.550724983215332,
      "num_input_tokens_seen": 5835800,
      "step": 1470,
      "train_runtime": 1197.1769,
      "train_tokens_per_second": 4874.635
    },
    {
      "epoch": 113.07692307692308,
      "num_input_tokens_seen": 5835800,
      "step": 1470,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 113.15384615384616,
      "grad_norm": 1.3338204622268677,
      "learning_rate": 0.01,
      "loss": 4.518813610076904,
      "num_input_tokens_seen": 5839896,
      "step": 1471,
      "train_runtime": 1197.9906,
      "train_tokens_per_second": 4874.743
    },
    {
      "epoch": 113.15384615384616,
      "num_input_tokens_seen": 5839896,
      "step": 1471,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 113.23076923076923,
      "grad_norm": 0.37299713492393494,
      "learning_rate": 0.01,
      "loss": 4.526444911956787,
      "num_input_tokens_seen": 5843992,
      "step": 1472,
      "train_runtime": 1198.8038,
      "train_tokens_per_second": 4874.853
    },
    {
      "epoch": 113.23076923076923,
      "num_input_tokens_seen": 5843992,
      "step": 1472,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 113.3076923076923,
      "grad_norm": 0.4445403218269348,
      "learning_rate": 0.01,
      "loss": 4.503165245056152,
      "num_input_tokens_seen": 5848088,
      "step": 1473,
      "train_runtime": 1199.618,
      "train_tokens_per_second": 4874.959
    },
    {
      "epoch": 113.3076923076923,
      "num_input_tokens_seen": 5848088,
      "step": 1473,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 113.38461538461539,
      "grad_norm": 0.579134464263916,
      "learning_rate": 0.01,
      "loss": 4.5347747802734375,
      "num_input_tokens_seen": 5852184,
      "step": 1474,
      "train_runtime": 1200.4319,
      "train_tokens_per_second": 4875.065
    },
    {
      "epoch": 113.38461538461539,
      "num_input_tokens_seen": 5852184,
      "step": 1474,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 113.46153846153847,
      "grad_norm": 0.3748880922794342,
      "learning_rate": 0.01,
      "loss": 4.542248249053955,
      "num_input_tokens_seen": 5856280,
      "step": 1475,
      "train_runtime": 1201.2472,
      "train_tokens_per_second": 4875.166
    },
    {
      "epoch": 113.46153846153847,
      "num_input_tokens_seen": 5856280,
      "step": 1475,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 113.53846153846153,
      "grad_norm": 0.3015724718570709,
      "learning_rate": 0.01,
      "loss": 4.5192670822143555,
      "num_input_tokens_seen": 5860376,
      "step": 1476,
      "train_runtime": 1202.07,
      "train_tokens_per_second": 4875.237
    },
    {
      "epoch": 113.53846153846153,
      "num_input_tokens_seen": 5860376,
      "step": 1476,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 113.61538461538461,
      "grad_norm": 0.3007107079029083,
      "learning_rate": 0.01,
      "loss": 4.521328926086426,
      "num_input_tokens_seen": 5864472,
      "step": 1477,
      "train_runtime": 1202.8841,
      "train_tokens_per_second": 4875.343
    },
    {
      "epoch": 113.61538461538461,
      "num_input_tokens_seen": 5864472,
      "step": 1477,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 113.6923076923077,
      "grad_norm": 0.42604634165763855,
      "learning_rate": 0.01,
      "loss": 4.530096054077148,
      "num_input_tokens_seen": 5868568,
      "step": 1478,
      "train_runtime": 1203.6976,
      "train_tokens_per_second": 4875.451
    },
    {
      "epoch": 113.6923076923077,
      "num_input_tokens_seen": 5868568,
      "step": 1478,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 113.76923076923077,
      "grad_norm": 0.37296631932258606,
      "learning_rate": 0.01,
      "loss": 4.533902645111084,
      "num_input_tokens_seen": 5872664,
      "step": 1479,
      "train_runtime": 1204.5124,
      "train_tokens_per_second": 4875.553
    },
    {
      "epoch": 113.76923076923077,
      "num_input_tokens_seen": 5872664,
      "step": 1479,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 113.84615384615384,
      "grad_norm": 0.2979356348514557,
      "learning_rate": 0.01,
      "loss": 4.533270835876465,
      "num_input_tokens_seen": 5876760,
      "step": 1480,
      "train_runtime": 1205.3272,
      "train_tokens_per_second": 4875.655
    },
    {
      "epoch": 113.84615384615384,
      "num_input_tokens_seen": 5876760,
      "step": 1480,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 113.92307692307692,
      "grad_norm": 0.414663165807724,
      "learning_rate": 0.01,
      "loss": 4.535632133483887,
      "num_input_tokens_seen": 5880856,
      "step": 1481,
      "train_runtime": 1206.1372,
      "train_tokens_per_second": 4875.777
    },
    {
      "epoch": 113.92307692307692,
      "num_input_tokens_seen": 5880856,
      "step": 1481,
      "train_accuracy": 0.0065146577544510365
    },
    {
      "epoch": 114.0,
      "grad_norm": 0.33322423696517944,
      "learning_rate": 0.01,
      "loss": 4.562551021575928,
      "num_input_tokens_seen": 5883312,
      "step": 1482,
      "train_runtime": 1206.6525,
      "train_tokens_per_second": 4875.73
    },
    {
      "epoch": 114.0,
      "num_input_tokens_seen": 5883312,
      "step": 1482,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 114.07692307692308,
      "grad_norm": 4.517330646514893,
      "learning_rate": 0.01,
      "loss": 4.516086578369141,
      "num_input_tokens_seen": 5887408,
      "step": 1483,
      "train_runtime": 1207.4785,
      "train_tokens_per_second": 4875.787
    },
    {
      "epoch": 114.07692307692308,
      "num_input_tokens_seen": 5887408,
      "step": 1483,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 114.15384615384616,
      "grad_norm": 0.2041095793247223,
      "learning_rate": 0.01,
      "loss": 4.498183250427246,
      "num_input_tokens_seen": 5891504,
      "step": 1484,
      "train_runtime": 1208.2922,
      "train_tokens_per_second": 4875.893
    },
    {
      "epoch": 114.15384615384616,
      "num_input_tokens_seen": 5891504,
      "step": 1484,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 114.23076923076923,
      "grad_norm": 0.6149014830589294,
      "learning_rate": 0.01,
      "loss": 4.543646812438965,
      "num_input_tokens_seen": 5895600,
      "step": 1485,
      "train_runtime": 1209.1061,
      "train_tokens_per_second": 4875.999
    },
    {
      "epoch": 114.23076923076923,
      "num_input_tokens_seen": 5895600,
      "step": 1485,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 114.3076923076923,
      "grad_norm": 0.29106220602989197,
      "learning_rate": 0.01,
      "loss": 4.506340980529785,
      "num_input_tokens_seen": 5899696,
      "step": 1486,
      "train_runtime": 1209.9198,
      "train_tokens_per_second": 4876.105
    },
    {
      "epoch": 114.3076923076923,
      "num_input_tokens_seen": 5899696,
      "step": 1486,
      "train_accuracy": 0.001953125
    },
    {
      "epoch": 114.38461538461539,
      "grad_norm": 0.44996026158332825,
      "learning_rate": 0.01,
      "loss": 4.5403008460998535,
      "num_input_tokens_seen": 5903792,
      "step": 1487,
      "train_runtime": 1210.7341,
      "train_tokens_per_second": 4876.209
    },
    {
      "epoch": 114.38461538461539,
      "num_input_tokens_seen": 5903792,
      "step": 1487,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 114.46153846153847,
      "grad_norm": 0.2865310609340668,
      "learning_rate": 0.01,
      "loss": 4.520112991333008,
      "num_input_tokens_seen": 5907888,
      "step": 1488,
      "train_runtime": 1211.5477,
      "train_tokens_per_second": 4876.315
    },
    {
      "epoch": 114.46153846153847,
      "num_input_tokens_seen": 5907888,
      "step": 1488,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 114.53846153846153,
      "grad_norm": 0.42982539534568787,
      "learning_rate": 0.01,
      "loss": 4.546148777008057,
      "num_input_tokens_seen": 5911984,
      "step": 1489,
      "train_runtime": 1212.3627,
      "train_tokens_per_second": 4876.415
    },
    {
      "epoch": 114.53846153846153,
      "num_input_tokens_seen": 5911984,
      "step": 1489,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 114.61538461538461,
      "grad_norm": 0.4213695824146271,
      "learning_rate": 0.01,
      "loss": 4.540425777435303,
      "num_input_tokens_seen": 5916080,
      "step": 1490,
      "train_runtime": 1213.1755,
      "train_tokens_per_second": 4876.524
    },
    {
      "epoch": 114.61538461538461,
      "num_input_tokens_seen": 5916080,
      "step": 1490,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 114.6923076923077,
      "grad_norm": 0.24819839000701904,
      "learning_rate": 0.01,
      "loss": 4.515840530395508,
      "num_input_tokens_seen": 5920176,
      "step": 1491,
      "train_runtime": 1213.9879,
      "train_tokens_per_second": 4876.635
    },
    {
      "epoch": 114.6923076923077,
      "num_input_tokens_seen": 5920176,
      "step": 1491,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 114.76923076923077,
      "grad_norm": 1.2839123010635376,
      "learning_rate": 0.01,
      "loss": 4.551513671875,
      "num_input_tokens_seen": 5924272,
      "step": 1492,
      "train_runtime": 1214.8004,
      "train_tokens_per_second": 4876.745
    },
    {
      "epoch": 114.76923076923077,
      "num_input_tokens_seen": 5924272,
      "step": 1492,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 114.84615384615384,
      "grad_norm": 0.31538495421409607,
      "learning_rate": 0.01,
      "loss": 4.560274600982666,
      "num_input_tokens_seen": 5928368,
      "step": 1493,
      "train_runtime": 1215.6135,
      "train_tokens_per_second": 4876.853
    },
    {
      "epoch": 114.84615384615384,
      "num_input_tokens_seen": 5928368,
      "step": 1493,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 114.92307692307692,
      "grad_norm": 0.21700944006443024,
      "learning_rate": 0.01,
      "loss": 4.535152435302734,
      "num_input_tokens_seen": 5932464,
      "step": 1494,
      "train_runtime": 1216.4219,
      "train_tokens_per_second": 4876.979
    },
    {
      "epoch": 114.92307692307692,
      "num_input_tokens_seen": 5932464,
      "step": 1494,
      "train_accuracy": 0.013029315508902073
    },
    {
      "epoch": 115.0,
      "grad_norm": 4.251666069030762,
      "learning_rate": 0.01,
      "loss": 4.5423455238342285,
      "num_input_tokens_seen": 5934920,
      "step": 1495,
      "train_runtime": 1216.936,
      "train_tokens_per_second": 4876.937
    },
    {
      "epoch": 115.0,
      "num_input_tokens_seen": 5934920,
      "step": 1495,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 115.07692307692308,
      "grad_norm": 0.25329291820526123,
      "learning_rate": 0.01,
      "loss": 4.501167297363281,
      "num_input_tokens_seen": 5939016,
      "step": 1496,
      "train_runtime": 1217.7617,
      "train_tokens_per_second": 4876.994
    },
    {
      "epoch": 115.07692307692308,
      "num_input_tokens_seen": 5939016,
      "step": 1496,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 115.15384615384616,
      "grad_norm": 0.3199782371520996,
      "learning_rate": 0.01,
      "loss": 4.542051315307617,
      "num_input_tokens_seen": 5943112,
      "step": 1497,
      "train_runtime": 1218.5764,
      "train_tokens_per_second": 4877.094
    },
    {
      "epoch": 115.15384615384616,
      "num_input_tokens_seen": 5943112,
      "step": 1497,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 115.23076923076923,
      "grad_norm": 1.3273310661315918,
      "learning_rate": 0.01,
      "loss": 4.527716636657715,
      "num_input_tokens_seen": 5947208,
      "step": 1498,
      "train_runtime": 1219.3912,
      "train_tokens_per_second": 4877.194
    },
    {
      "epoch": 115.23076923076923,
      "num_input_tokens_seen": 5947208,
      "step": 1498,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 115.3076923076923,
      "grad_norm": 0.22500677406787872,
      "learning_rate": 0.01,
      "loss": 4.54163932800293,
      "num_input_tokens_seen": 5951304,
      "step": 1499,
      "train_runtime": 1220.2031,
      "train_tokens_per_second": 4877.306
    },
    {
      "epoch": 115.3076923076923,
      "num_input_tokens_seen": 5951304,
      "step": 1499,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 115.38461538461539,
      "grad_norm": 0.42557427287101746,
      "learning_rate": 0.01,
      "loss": 4.538112640380859,
      "num_input_tokens_seen": 5955400,
      "step": 1500,
      "train_runtime": 1221.0157,
      "train_tokens_per_second": 4877.415
    },
    {
      "epoch": 115.38461538461539,
      "eval_CMD_3_branch_eval_accuracy": 0.006871609403254973,
      "eval_CMD_3_branch_eval_loss": 4.5631232261657715,
      "eval_CMD_3_branch_eval_runtime": 1.1547,
      "eval_CMD_3_branch_eval_samples_per_second": 2394.649,
      "eval_CMD_3_branch_eval_steps_per_second": 5.196,
      "num_input_tokens_seen": 5955400,
      "step": 1500
    },
    {
      "epoch": 115.38461538461539,
      "num_input_tokens_seen": 5955400,
      "step": 1500,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 115.46153846153847,
      "grad_norm": 0.6366361379623413,
      "learning_rate": 0.01,
      "loss": 4.517929553985596,
      "num_input_tokens_seen": 5959496,
      "step": 1501,
      "train_runtime": 1222.9858,
      "train_tokens_per_second": 4872.907
    },
    {
      "epoch": 115.46153846153847,
      "num_input_tokens_seen": 5959496,
      "step": 1501,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 115.53846153846153,
      "grad_norm": 0.578191876411438,
      "learning_rate": 0.01,
      "loss": 4.528219223022461,
      "num_input_tokens_seen": 5963592,
      "step": 1502,
      "train_runtime": 1223.7985,
      "train_tokens_per_second": 4873.018
    },
    {
      "epoch": 115.53846153846153,
      "num_input_tokens_seen": 5963592,
      "step": 1502,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 115.61538461538461,
      "grad_norm": 0.2566545903682709,
      "learning_rate": 0.01,
      "loss": 4.548543453216553,
      "num_input_tokens_seen": 5967688,
      "step": 1503,
      "train_runtime": 1224.6105,
      "train_tokens_per_second": 4873.132
    },
    {
      "epoch": 115.61538461538461,
      "num_input_tokens_seen": 5967688,
      "step": 1503,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 115.6923076923077,
      "grad_norm": 0.41449329257011414,
      "learning_rate": 0.01,
      "loss": 4.538558006286621,
      "num_input_tokens_seen": 5971784,
      "step": 1504,
      "train_runtime": 1225.4234,
      "train_tokens_per_second": 4873.241
    },
    {
      "epoch": 115.6923076923077,
      "num_input_tokens_seen": 5971784,
      "step": 1504,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 115.76923076923077,
      "grad_norm": 2.3327584266662598,
      "learning_rate": 0.01,
      "loss": 4.521422386169434,
      "num_input_tokens_seen": 5975880,
      "step": 1505,
      "train_runtime": 1226.2365,
      "train_tokens_per_second": 4873.35
    },
    {
      "epoch": 115.76923076923077,
      "num_input_tokens_seen": 5975880,
      "step": 1505,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 115.84615384615384,
      "grad_norm": 0.3173248767852783,
      "learning_rate": 0.01,
      "loss": 4.513408184051514,
      "num_input_tokens_seen": 5979976,
      "step": 1506,
      "train_runtime": 1227.0522,
      "train_tokens_per_second": 4873.449
    },
    {
      "epoch": 115.84615384615384,
      "num_input_tokens_seen": 5979976,
      "step": 1506,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 115.92307692307692,
      "grad_norm": 0.22287431359291077,
      "learning_rate": 0.01,
      "loss": 4.497661590576172,
      "num_input_tokens_seen": 5984072,
      "step": 1507,
      "train_runtime": 1227.8623,
      "train_tokens_per_second": 4873.569
    },
    {
      "epoch": 115.92307692307692,
      "num_input_tokens_seen": 5984072,
      "step": 1507,
      "train_accuracy": 0.019543973729014397
    },
    {
      "epoch": 116.0,
      "grad_norm": 0.6383256912231445,
      "learning_rate": 0.01,
      "loss": 4.531820774078369,
      "num_input_tokens_seen": 5986528,
      "step": 1508,
      "train_runtime": 1228.3769,
      "train_tokens_per_second": 4873.527
    },
    {
      "epoch": 116.0,
      "num_input_tokens_seen": 5986528,
      "step": 1508,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 116.07692307692308,
      "grad_norm": 0.3919311761856079,
      "learning_rate": 0.01,
      "loss": 4.509546279907227,
      "num_input_tokens_seen": 5990624,
      "step": 1509,
      "train_runtime": 1229.2043,
      "train_tokens_per_second": 4873.579
    },
    {
      "epoch": 116.07692307692308,
      "num_input_tokens_seen": 5990624,
      "step": 1509,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 116.15384615384616,
      "grad_norm": 25.548311233520508,
      "learning_rate": 0.01,
      "loss": 4.636435508728027,
      "num_input_tokens_seen": 5994720,
      "step": 1510,
      "train_runtime": 1230.0168,
      "train_tokens_per_second": 4873.69
    },
    {
      "epoch": 116.15384615384616,
      "num_input_tokens_seen": 5994720,
      "step": 1510,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 116.23076923076923,
      "grad_norm": 1.890404462814331,
      "learning_rate": 0.01,
      "loss": 4.6568450927734375,
      "num_input_tokens_seen": 5998816,
      "step": 1511,
      "train_runtime": 1230.8311,
      "train_tokens_per_second": 4873.793
    },
    {
      "epoch": 116.23076923076923,
      "num_input_tokens_seen": 5998816,
      "step": 1511,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 116.3076923076923,
      "grad_norm": 0.3822583556175232,
      "learning_rate": 0.01,
      "loss": 4.55836820602417,
      "num_input_tokens_seen": 6002912,
      "step": 1512,
      "train_runtime": 1231.6454,
      "train_tokens_per_second": 4873.896
    },
    {
      "epoch": 116.3076923076923,
      "num_input_tokens_seen": 6002912,
      "step": 1512,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 116.38461538461539,
      "grad_norm": 0.28379833698272705,
      "learning_rate": 0.01,
      "loss": 4.556880950927734,
      "num_input_tokens_seen": 6007008,
      "step": 1513,
      "train_runtime": 1232.4596,
      "train_tokens_per_second": 4874.0
    },
    {
      "epoch": 116.38461538461539,
      "num_input_tokens_seen": 6007008,
      "step": 1513,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 116.46153846153847,
      "grad_norm": 0.6593945026397705,
      "learning_rate": 0.01,
      "loss": 4.567363739013672,
      "num_input_tokens_seen": 6011104,
      "step": 1514,
      "train_runtime": 1233.2732,
      "train_tokens_per_second": 4874.106
    },
    {
      "epoch": 116.46153846153847,
      "num_input_tokens_seen": 6011104,
      "step": 1514,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 116.53846153846153,
      "grad_norm": 0.5276172757148743,
      "learning_rate": 0.01,
      "loss": 4.560641765594482,
      "num_input_tokens_seen": 6015200,
      "step": 1515,
      "train_runtime": 1234.0845,
      "train_tokens_per_second": 4874.221
    },
    {
      "epoch": 116.53846153846153,
      "num_input_tokens_seen": 6015200,
      "step": 1515,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 116.61538461538461,
      "grad_norm": 1.2776269912719727,
      "learning_rate": 0.01,
      "loss": 4.544119834899902,
      "num_input_tokens_seen": 6019296,
      "step": 1516,
      "train_runtime": 1234.897,
      "train_tokens_per_second": 4874.33
    },
    {
      "epoch": 116.61538461538461,
      "num_input_tokens_seen": 6019296,
      "step": 1516,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 116.6923076923077,
      "grad_norm": 0.8084747195243835,
      "learning_rate": 0.01,
      "loss": 4.5611724853515625,
      "num_input_tokens_seen": 6023392,
      "step": 1517,
      "train_runtime": 1235.7099,
      "train_tokens_per_second": 4874.439
    },
    {
      "epoch": 116.6923076923077,
      "num_input_tokens_seen": 6023392,
      "step": 1517,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 116.76923076923077,
      "grad_norm": 0.2594858705997467,
      "learning_rate": 0.01,
      "loss": 4.562878608703613,
      "num_input_tokens_seen": 6027488,
      "step": 1518,
      "train_runtime": 1236.5224,
      "train_tokens_per_second": 4874.548
    },
    {
      "epoch": 116.76923076923077,
      "num_input_tokens_seen": 6027488,
      "step": 1518,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 116.84615384615384,
      "grad_norm": 0.6123083829879761,
      "learning_rate": 0.01,
      "loss": 4.556002616882324,
      "num_input_tokens_seen": 6031584,
      "step": 1519,
      "train_runtime": 1237.335,
      "train_tokens_per_second": 4874.657
    },
    {
      "epoch": 116.84615384615384,
      "num_input_tokens_seen": 6031584,
      "step": 1519,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 116.92307692307692,
      "grad_norm": 0.2527475655078888,
      "learning_rate": 0.01,
      "loss": 4.575403213500977,
      "num_input_tokens_seen": 6035680,
      "step": 1520,
      "train_runtime": 1238.1428,
      "train_tokens_per_second": 4874.785
    },
    {
      "epoch": 116.92307692307692,
      "num_input_tokens_seen": 6035680,
      "step": 1520,
      "train_accuracy": 0.026058631017804146
    },
    {
      "epoch": 117.0,
      "grad_norm": 0.5278998017311096,
      "learning_rate": 0.01,
      "loss": 4.529703617095947,
      "num_input_tokens_seen": 6038136,
      "step": 1521,
      "train_runtime": 1238.6569,
      "train_tokens_per_second": 4874.745
    },
    {
      "epoch": 117.0,
      "num_input_tokens_seen": 6038136,
      "step": 1521,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 117.07692307692308,
      "grad_norm": 0.2524627149105072,
      "learning_rate": 0.01,
      "loss": 4.556354522705078,
      "num_input_tokens_seen": 6042232,
      "step": 1522,
      "train_runtime": 1239.4804,
      "train_tokens_per_second": 4874.81
    },
    {
      "epoch": 117.07692307692308,
      "num_input_tokens_seen": 6042232,
      "step": 1522,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 117.15384615384616,
      "grad_norm": 0.5767359137535095,
      "learning_rate": 0.01,
      "loss": 4.554250240325928,
      "num_input_tokens_seen": 6046328,
      "step": 1523,
      "train_runtime": 1240.2925,
      "train_tokens_per_second": 4874.921
    },
    {
      "epoch": 117.15384615384616,
      "num_input_tokens_seen": 6046328,
      "step": 1523,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 117.23076923076923,
      "grad_norm": 0.40972158312797546,
      "learning_rate": 0.01,
      "loss": 4.57272481918335,
      "num_input_tokens_seen": 6050424,
      "step": 1524,
      "train_runtime": 1241.1045,
      "train_tokens_per_second": 4875.032
    },
    {
      "epoch": 117.23076923076923,
      "num_input_tokens_seen": 6050424,
      "step": 1524,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 117.3076923076923,
      "grad_norm": 0.4189790189266205,
      "learning_rate": 0.01,
      "loss": 4.536509037017822,
      "num_input_tokens_seen": 6054520,
      "step": 1525,
      "train_runtime": 1241.9162,
      "train_tokens_per_second": 4875.144
    },
    {
      "epoch": 117.3076923076923,
      "num_input_tokens_seen": 6054520,
      "step": 1525,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 117.38461538461539,
      "grad_norm": 0.3326832950115204,
      "learning_rate": 0.01,
      "loss": 4.555102825164795,
      "num_input_tokens_seen": 6058616,
      "step": 1526,
      "train_runtime": 1242.7287,
      "train_tokens_per_second": 4875.252
    },
    {
      "epoch": 117.38461538461539,
      "num_input_tokens_seen": 6058616,
      "step": 1526,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 117.46153846153847,
      "grad_norm": 0.21425527334213257,
      "learning_rate": 0.01,
      "loss": 4.541358947753906,
      "num_input_tokens_seen": 6062712,
      "step": 1527,
      "train_runtime": 1243.5408,
      "train_tokens_per_second": 4875.362
    },
    {
      "epoch": 117.46153846153847,
      "num_input_tokens_seen": 6062712,
      "step": 1527,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 117.53846153846153,
      "grad_norm": 0.28938883543014526,
      "learning_rate": 0.01,
      "loss": 4.534416198730469,
      "num_input_tokens_seen": 6066808,
      "step": 1528,
      "train_runtime": 1244.3541,
      "train_tokens_per_second": 4875.468
    },
    {
      "epoch": 117.53846153846153,
      "num_input_tokens_seen": 6066808,
      "step": 1528,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 117.61538461538461,
      "grad_norm": 0.8966801166534424,
      "learning_rate": 0.01,
      "loss": 4.54386043548584,
      "num_input_tokens_seen": 6070904,
      "step": 1529,
      "train_runtime": 1245.167,
      "train_tokens_per_second": 4875.574
    },
    {
      "epoch": 117.61538461538461,
      "num_input_tokens_seen": 6070904,
      "step": 1529,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 117.6923076923077,
      "grad_norm": 0.39462676644325256,
      "learning_rate": 0.01,
      "loss": 4.542434215545654,
      "num_input_tokens_seen": 6075000,
      "step": 1530,
      "train_runtime": 1245.98,
      "train_tokens_per_second": 4875.68
    },
    {
      "epoch": 117.6923076923077,
      "num_input_tokens_seen": 6075000,
      "step": 1530,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 117.76923076923077,
      "grad_norm": 0.5106446146965027,
      "learning_rate": 0.01,
      "loss": 4.5446319580078125,
      "num_input_tokens_seen": 6079096,
      "step": 1531,
      "train_runtime": 1246.7934,
      "train_tokens_per_second": 4875.784
    },
    {
      "epoch": 117.76923076923077,
      "num_input_tokens_seen": 6079096,
      "step": 1531,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 117.84615384615384,
      "grad_norm": 0.5330077409744263,
      "learning_rate": 0.01,
      "loss": 4.5503411293029785,
      "num_input_tokens_seen": 6083192,
      "step": 1532,
      "train_runtime": 1247.6063,
      "train_tokens_per_second": 4875.891
    },
    {
      "epoch": 117.84615384615384,
      "num_input_tokens_seen": 6083192,
      "step": 1532,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 117.92307692307692,
      "grad_norm": 0.9235372543334961,
      "learning_rate": 0.01,
      "loss": 4.517930030822754,
      "num_input_tokens_seen": 6087288,
      "step": 1533,
      "train_runtime": 1248.4155,
      "train_tokens_per_second": 4876.011
    },
    {
      "epoch": 117.92307692307692,
      "num_input_tokens_seen": 6087288,
      "step": 1533,
      "train_accuracy": 0.0065146577544510365
    },
    {
      "epoch": 118.0,
      "grad_norm": 0.6051939725875854,
      "learning_rate": 0.01,
      "loss": 4.540765285491943,
      "num_input_tokens_seen": 6089744,
      "step": 1534,
      "train_runtime": 1248.9301,
      "train_tokens_per_second": 4875.968
    },
    {
      "epoch": 118.0,
      "num_input_tokens_seen": 6089744,
      "step": 1534,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 118.07692307692308,
      "grad_norm": 23.663299560546875,
      "learning_rate": 0.01,
      "loss": 4.546059608459473,
      "num_input_tokens_seen": 6093840,
      "step": 1535,
      "train_runtime": 1249.7557,
      "train_tokens_per_second": 4876.025
    },
    {
      "epoch": 118.07692307692308,
      "num_input_tokens_seen": 6093840,
      "step": 1535,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 118.15384615384616,
      "grad_norm": 0.5285918712615967,
      "learning_rate": 0.01,
      "loss": 4.551592826843262,
      "num_input_tokens_seen": 6097936,
      "step": 1536,
      "train_runtime": 1250.569,
      "train_tokens_per_second": 4876.129
    },
    {
      "epoch": 118.15384615384616,
      "num_input_tokens_seen": 6097936,
      "step": 1536,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 118.23076923076923,
      "grad_norm": 0.24745753407478333,
      "learning_rate": 0.01,
      "loss": 4.568268299102783,
      "num_input_tokens_seen": 6102032,
      "step": 1537,
      "train_runtime": 1251.3827,
      "train_tokens_per_second": 4876.232
    },
    {
      "epoch": 118.23076923076923,
      "num_input_tokens_seen": 6102032,
      "step": 1537,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 118.3076923076923,
      "grad_norm": 0.2812255620956421,
      "learning_rate": 0.01,
      "loss": 4.563979148864746,
      "num_input_tokens_seen": 6106128,
      "step": 1538,
      "train_runtime": 1252.196,
      "train_tokens_per_second": 4876.336
    },
    {
      "epoch": 118.3076923076923,
      "num_input_tokens_seen": 6106128,
      "step": 1538,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 118.38461538461539,
      "grad_norm": 0.2536022961139679,
      "learning_rate": 0.01,
      "loss": 4.585928916931152,
      "num_input_tokens_seen": 6110224,
      "step": 1539,
      "train_runtime": 1253.0086,
      "train_tokens_per_second": 4876.442
    },
    {
      "epoch": 118.38461538461539,
      "num_input_tokens_seen": 6110224,
      "step": 1539,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 118.46153846153847,
      "grad_norm": 0.2379070371389389,
      "learning_rate": 0.01,
      "loss": 4.58247709274292,
      "num_input_tokens_seen": 6114320,
      "step": 1540,
      "train_runtime": 1253.8218,
      "train_tokens_per_second": 4876.546
    },
    {
      "epoch": 118.46153846153847,
      "num_input_tokens_seen": 6114320,
      "step": 1540,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 118.53846153846153,
      "grad_norm": 0.2722030282020569,
      "learning_rate": 0.01,
      "loss": 4.601729869842529,
      "num_input_tokens_seen": 6118416,
      "step": 1541,
      "train_runtime": 1254.6355,
      "train_tokens_per_second": 4876.648
    },
    {
      "epoch": 118.53846153846153,
      "num_input_tokens_seen": 6118416,
      "step": 1541,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 118.61538461538461,
      "grad_norm": 0.3251706659793854,
      "learning_rate": 0.01,
      "loss": 4.564596652984619,
      "num_input_tokens_seen": 6122512,
      "step": 1542,
      "train_runtime": 1255.4519,
      "train_tokens_per_second": 4876.74
    },
    {
      "epoch": 118.61538461538461,
      "num_input_tokens_seen": 6122512,
      "step": 1542,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 118.6923076923077,
      "grad_norm": 0.36576464772224426,
      "learning_rate": 0.01,
      "loss": 4.585412502288818,
      "num_input_tokens_seen": 6126608,
      "step": 1543,
      "train_runtime": 1256.2656,
      "train_tokens_per_second": 4876.841
    },
    {
      "epoch": 118.6923076923077,
      "num_input_tokens_seen": 6126608,
      "step": 1543,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 118.76923076923077,
      "grad_norm": 0.2578296661376953,
      "learning_rate": 0.01,
      "loss": 4.5692644119262695,
      "num_input_tokens_seen": 6130704,
      "step": 1544,
      "train_runtime": 1257.0786,
      "train_tokens_per_second": 4876.946
    },
    {
      "epoch": 118.76923076923077,
      "num_input_tokens_seen": 6130704,
      "step": 1544,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 118.84615384615384,
      "grad_norm": 0.3710087239742279,
      "learning_rate": 0.01,
      "loss": 4.584305286407471,
      "num_input_tokens_seen": 6134800,
      "step": 1545,
      "train_runtime": 1257.8904,
      "train_tokens_per_second": 4877.055
    },
    {
      "epoch": 118.84615384615384,
      "num_input_tokens_seen": 6134800,
      "step": 1545,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 118.92307692307692,
      "grad_norm": 0.2943045496940613,
      "learning_rate": 0.01,
      "loss": 4.585121154785156,
      "num_input_tokens_seen": 6138896,
      "step": 1546,
      "train_runtime": 1258.6997,
      "train_tokens_per_second": 4877.173
    },
    {
      "epoch": 118.92307692307692,
      "num_input_tokens_seen": 6138896,
      "step": 1546,
      "train_accuracy": 0.02280130237340927
    },
    {
      "epoch": 119.0,
      "grad_norm": 1.1209063529968262,
      "learning_rate": 0.01,
      "loss": 4.567545413970947,
      "num_input_tokens_seen": 6141352,
      "step": 1547,
      "train_runtime": 1259.2139,
      "train_tokens_per_second": 4877.132
    },
    {
      "epoch": 119.0,
      "num_input_tokens_seen": 6141352,
      "step": 1547,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 119.07692307692308,
      "grad_norm": 0.3971582055091858,
      "learning_rate": 0.01,
      "loss": 4.567508697509766,
      "num_input_tokens_seen": 6145448,
      "step": 1548,
      "train_runtime": 1260.0386,
      "train_tokens_per_second": 4877.19
    },
    {
      "epoch": 119.07692307692308,
      "num_input_tokens_seen": 6145448,
      "step": 1548,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 119.15384615384616,
      "grad_norm": 0.2880050837993622,
      "learning_rate": 0.01,
      "loss": 4.574665546417236,
      "num_input_tokens_seen": 6149544,
      "step": 1549,
      "train_runtime": 1260.8528,
      "train_tokens_per_second": 4877.29
    },
    {
      "epoch": 119.15384615384616,
      "num_input_tokens_seen": 6149544,
      "step": 1549,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 119.23076923076923,
      "grad_norm": 0.2553396224975586,
      "learning_rate": 0.01,
      "loss": 4.554409980773926,
      "num_input_tokens_seen": 6153640,
      "step": 1550,
      "train_runtime": 1261.6679,
      "train_tokens_per_second": 4877.385
    },
    {
      "epoch": 119.23076923076923,
      "num_input_tokens_seen": 6153640,
      "step": 1550,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 119.3076923076923,
      "grad_norm": 0.6814112663269043,
      "learning_rate": 0.01,
      "loss": 4.573980331420898,
      "num_input_tokens_seen": 6157736,
      "step": 1551,
      "train_runtime": 1262.4817,
      "train_tokens_per_second": 4877.485
    },
    {
      "epoch": 119.3076923076923,
      "num_input_tokens_seen": 6157736,
      "step": 1551,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 119.38461538461539,
      "grad_norm": 0.2922833561897278,
      "learning_rate": 0.01,
      "loss": 4.581429958343506,
      "num_input_tokens_seen": 6161832,
      "step": 1552,
      "train_runtime": 1263.2957,
      "train_tokens_per_second": 4877.585
    },
    {
      "epoch": 119.38461538461539,
      "num_input_tokens_seen": 6161832,
      "step": 1552,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 119.46153846153847,
      "grad_norm": 0.27610599994659424,
      "learning_rate": 0.01,
      "loss": 4.571043968200684,
      "num_input_tokens_seen": 6165928,
      "step": 1553,
      "train_runtime": 1264.109,
      "train_tokens_per_second": 4877.687
    },
    {
      "epoch": 119.46153846153847,
      "num_input_tokens_seen": 6165928,
      "step": 1553,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 119.53846153846153,
      "grad_norm": 0.2627987563610077,
      "learning_rate": 0.01,
      "loss": 4.556126594543457,
      "num_input_tokens_seen": 6170024,
      "step": 1554,
      "train_runtime": 1264.922,
      "train_tokens_per_second": 4877.79
    },
    {
      "epoch": 119.53846153846153,
      "num_input_tokens_seen": 6170024,
      "step": 1554,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 119.61538461538461,
      "grad_norm": 0.36993029713630676,
      "learning_rate": 0.01,
      "loss": 4.565702438354492,
      "num_input_tokens_seen": 6174120,
      "step": 1555,
      "train_runtime": 1265.7352,
      "train_tokens_per_second": 4877.892
    },
    {
      "epoch": 119.61538461538461,
      "num_input_tokens_seen": 6174120,
      "step": 1555,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 119.6923076923077,
      "grad_norm": 0.2819044291973114,
      "learning_rate": 0.01,
      "loss": 4.570314884185791,
      "num_input_tokens_seen": 6178216,
      "step": 1556,
      "train_runtime": 1266.8144,
      "train_tokens_per_second": 4876.97
    },
    {
      "epoch": 119.6923076923077,
      "num_input_tokens_seen": 6178216,
      "step": 1556,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 119.76923076923077,
      "grad_norm": 0.3040362298488617,
      "learning_rate": 0.01,
      "loss": 4.568306922912598,
      "num_input_tokens_seen": 6182312,
      "step": 1557,
      "train_runtime": 1267.6271,
      "train_tokens_per_second": 4877.075
    },
    {
      "epoch": 119.76923076923077,
      "num_input_tokens_seen": 6182312,
      "step": 1557,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 119.84615384615384,
      "grad_norm": 0.393155574798584,
      "learning_rate": 0.01,
      "loss": 4.59574031829834,
      "num_input_tokens_seen": 6186408,
      "step": 1558,
      "train_runtime": 1268.4395,
      "train_tokens_per_second": 4877.18
    },
    {
      "epoch": 119.84615384615384,
      "num_input_tokens_seen": 6186408,
      "step": 1558,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 119.92307692307692,
      "grad_norm": 0.46554693579673767,
      "learning_rate": 0.01,
      "loss": 4.5911054611206055,
      "num_input_tokens_seen": 6190504,
      "step": 1559,
      "train_runtime": 1269.2489,
      "train_tokens_per_second": 4877.297
    },
    {
      "epoch": 119.92307692307692,
      "num_input_tokens_seen": 6190504,
      "step": 1559,
      "train_accuracy": 0.019543973729014397
    },
    {
      "epoch": 120.0,
      "grad_norm": 0.3963291049003601,
      "learning_rate": 0.01,
      "loss": 4.585737705230713,
      "num_input_tokens_seen": 6192960,
      "step": 1560,
      "train_runtime": 1269.7632,
      "train_tokens_per_second": 4877.256
    },
    {
      "epoch": 120.0,
      "num_input_tokens_seen": 6192960,
      "step": 1560,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 120.07692307692308,
      "grad_norm": 0.18459410965442657,
      "learning_rate": 0.01,
      "loss": 4.553705215454102,
      "num_input_tokens_seen": 6197056,
      "step": 1561,
      "train_runtime": 1270.5909,
      "train_tokens_per_second": 4877.302
    },
    {
      "epoch": 120.07692307692308,
      "num_input_tokens_seen": 6197056,
      "step": 1561,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 120.15384615384616,
      "grad_norm": 0.27707549929618835,
      "learning_rate": 0.01,
      "loss": 4.553025245666504,
      "num_input_tokens_seen": 6201152,
      "step": 1562,
      "train_runtime": 1271.4035,
      "train_tokens_per_second": 4877.407
    },
    {
      "epoch": 120.15384615384616,
      "num_input_tokens_seen": 6201152,
      "step": 1562,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 120.23076923076923,
      "grad_norm": 0.3159920573234558,
      "learning_rate": 0.01,
      "loss": 4.586309432983398,
      "num_input_tokens_seen": 6205248,
      "step": 1563,
      "train_runtime": 1272.217,
      "train_tokens_per_second": 4877.507
    },
    {
      "epoch": 120.23076923076923,
      "num_input_tokens_seen": 6205248,
      "step": 1563,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 120.3076923076923,
      "grad_norm": 0.40838930010795593,
      "learning_rate": 0.01,
      "loss": 4.560218811035156,
      "num_input_tokens_seen": 6209344,
      "step": 1564,
      "train_runtime": 1273.03,
      "train_tokens_per_second": 4877.61
    },
    {
      "epoch": 120.3076923076923,
      "num_input_tokens_seen": 6209344,
      "step": 1564,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 120.38461538461539,
      "grad_norm": 0.36436623334884644,
      "learning_rate": 0.01,
      "loss": 4.558023929595947,
      "num_input_tokens_seen": 6213440,
      "step": 1565,
      "train_runtime": 1273.8434,
      "train_tokens_per_second": 4877.711
    },
    {
      "epoch": 120.38461538461539,
      "num_input_tokens_seen": 6213440,
      "step": 1565,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 120.46153846153847,
      "grad_norm": 0.30659976601600647,
      "learning_rate": 0.01,
      "loss": 4.53167724609375,
      "num_input_tokens_seen": 6217536,
      "step": 1566,
      "train_runtime": 1274.6589,
      "train_tokens_per_second": 4877.804
    },
    {
      "epoch": 120.46153846153847,
      "num_input_tokens_seen": 6217536,
      "step": 1566,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 120.53846153846153,
      "grad_norm": 0.36716049909591675,
      "learning_rate": 0.01,
      "loss": 4.546405792236328,
      "num_input_tokens_seen": 6221632,
      "step": 1567,
      "train_runtime": 1275.474,
      "train_tokens_per_second": 4877.898
    },
    {
      "epoch": 120.53846153846153,
      "num_input_tokens_seen": 6221632,
      "step": 1567,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 120.61538461538461,
      "grad_norm": 0.19651320576667786,
      "learning_rate": 0.01,
      "loss": 4.566239356994629,
      "num_input_tokens_seen": 6225728,
      "step": 1568,
      "train_runtime": 1276.2858,
      "train_tokens_per_second": 4878.004
    },
    {
      "epoch": 120.61538461538461,
      "num_input_tokens_seen": 6225728,
      "step": 1568,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 120.6923076923077,
      "grad_norm": 0.18629644811153412,
      "learning_rate": 0.01,
      "loss": 4.558133125305176,
      "num_input_tokens_seen": 6229824,
      "step": 1569,
      "train_runtime": 1277.0979,
      "train_tokens_per_second": 4878.11
    },
    {
      "epoch": 120.6923076923077,
      "num_input_tokens_seen": 6229824,
      "step": 1569,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 120.76923076923077,
      "grad_norm": 0.26930102705955505,
      "learning_rate": 0.01,
      "loss": 4.562258720397949,
      "num_input_tokens_seen": 6233920,
      "step": 1570,
      "train_runtime": 1277.9112,
      "train_tokens_per_second": 4878.211
    },
    {
      "epoch": 120.76923076923077,
      "num_input_tokens_seen": 6233920,
      "step": 1570,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 120.84615384615384,
      "grad_norm": 0.3980523347854614,
      "learning_rate": 0.01,
      "loss": 4.559825420379639,
      "num_input_tokens_seen": 6238016,
      "step": 1571,
      "train_runtime": 1278.7232,
      "train_tokens_per_second": 4878.316
    },
    {
      "epoch": 120.84615384615384,
      "num_input_tokens_seen": 6238016,
      "step": 1571,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 120.92307692307692,
      "grad_norm": 0.3914523422718048,
      "learning_rate": 0.01,
      "loss": 4.574028968811035,
      "num_input_tokens_seen": 6242112,
      "step": 1572,
      "train_runtime": 1279.5318,
      "train_tokens_per_second": 4878.434
    },
    {
      "epoch": 120.92307692307692,
      "num_input_tokens_seen": 6242112,
      "step": 1572,
      "train_accuracy": 0.0065146577544510365
    },
    {
      "epoch": 121.0,
      "grad_norm": 0.30136576294898987,
      "learning_rate": 0.01,
      "loss": 4.564297199249268,
      "num_input_tokens_seen": 6244568,
      "step": 1573,
      "train_runtime": 1280.0471,
      "train_tokens_per_second": 4878.389
    },
    {
      "epoch": 121.0,
      "num_input_tokens_seen": 6244568,
      "step": 1573,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 121.07692307692308,
      "grad_norm": 0.21680454909801483,
      "learning_rate": 0.01,
      "loss": 4.55159854888916,
      "num_input_tokens_seen": 6248664,
      "step": 1574,
      "train_runtime": 1280.8751,
      "train_tokens_per_second": 4878.433
    },
    {
      "epoch": 121.07692307692308,
      "num_input_tokens_seen": 6248664,
      "step": 1574,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 121.15384615384616,
      "grad_norm": 0.230843186378479,
      "learning_rate": 0.01,
      "loss": 4.535984039306641,
      "num_input_tokens_seen": 6252760,
      "step": 1575,
      "train_runtime": 1281.6876,
      "train_tokens_per_second": 4878.537
    },
    {
      "epoch": 121.15384615384616,
      "num_input_tokens_seen": 6252760,
      "step": 1575,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 121.23076923076923,
      "grad_norm": 0.19835351407527924,
      "learning_rate": 0.01,
      "loss": 4.533585548400879,
      "num_input_tokens_seen": 6256856,
      "step": 1576,
      "train_runtime": 1282.5011,
      "train_tokens_per_second": 4878.636
    },
    {
      "epoch": 121.23076923076923,
      "num_input_tokens_seen": 6256856,
      "step": 1576,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 121.3076923076923,
      "grad_norm": 0.5837871432304382,
      "learning_rate": 0.01,
      "loss": 4.528852939605713,
      "num_input_tokens_seen": 6260952,
      "step": 1577,
      "train_runtime": 1283.3139,
      "train_tokens_per_second": 4878.738
    },
    {
      "epoch": 121.3076923076923,
      "num_input_tokens_seen": 6260952,
      "step": 1577,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 121.38461538461539,
      "grad_norm": 0.8927022814750671,
      "learning_rate": 0.01,
      "loss": 4.5188751220703125,
      "num_input_tokens_seen": 6265048,
      "step": 1578,
      "train_runtime": 1284.1272,
      "train_tokens_per_second": 4878.837
    },
    {
      "epoch": 121.38461538461539,
      "num_input_tokens_seen": 6265048,
      "step": 1578,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 121.46153846153847,
      "grad_norm": 2.586885452270508,
      "learning_rate": 0.01,
      "loss": 4.525791168212891,
      "num_input_tokens_seen": 6269144,
      "step": 1579,
      "train_runtime": 1284.94,
      "train_tokens_per_second": 4878.939
    },
    {
      "epoch": 121.46153846153847,
      "num_input_tokens_seen": 6269144,
      "step": 1579,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 121.53846153846153,
      "grad_norm": 1.0155558586120605,
      "learning_rate": 0.01,
      "loss": 4.515064239501953,
      "num_input_tokens_seen": 6273240,
      "step": 1580,
      "train_runtime": 1285.7563,
      "train_tokens_per_second": 4879.027
    },
    {
      "epoch": 121.53846153846153,
      "num_input_tokens_seen": 6273240,
      "step": 1580,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 121.61538461538461,
      "grad_norm": 0.36298471689224243,
      "learning_rate": 0.01,
      "loss": 4.528317451477051,
      "num_input_tokens_seen": 6277336,
      "step": 1581,
      "train_runtime": 1286.5716,
      "train_tokens_per_second": 4879.119
    },
    {
      "epoch": 121.61538461538461,
      "num_input_tokens_seen": 6277336,
      "step": 1581,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 121.6923076923077,
      "grad_norm": 1.5292681455612183,
      "learning_rate": 0.01,
      "loss": 4.569167137145996,
      "num_input_tokens_seen": 6281432,
      "step": 1582,
      "train_runtime": 1287.3859,
      "train_tokens_per_second": 4879.214
    },
    {
      "epoch": 121.6923076923077,
      "num_input_tokens_seen": 6281432,
      "step": 1582,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 121.76923076923077,
      "grad_norm": 0.34332823753356934,
      "learning_rate": 0.01,
      "loss": 4.535788536071777,
      "num_input_tokens_seen": 6285528,
      "step": 1583,
      "train_runtime": 1288.1992,
      "train_tokens_per_second": 4879.314
    },
    {
      "epoch": 121.76923076923077,
      "num_input_tokens_seen": 6285528,
      "step": 1583,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 121.84615384615384,
      "grad_norm": 0.4531383216381073,
      "learning_rate": 0.01,
      "loss": 4.5581207275390625,
      "num_input_tokens_seen": 6289624,
      "step": 1584,
      "train_runtime": 1289.0122,
      "train_tokens_per_second": 4879.414
    },
    {
      "epoch": 121.84615384615384,
      "num_input_tokens_seen": 6289624,
      "step": 1584,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 121.92307692307692,
      "grad_norm": 0.27923622727394104,
      "learning_rate": 0.01,
      "loss": 4.559006690979004,
      "num_input_tokens_seen": 6293720,
      "step": 1585,
      "train_runtime": 1289.8213,
      "train_tokens_per_second": 4879.529
    },
    {
      "epoch": 121.92307692307692,
      "num_input_tokens_seen": 6293720,
      "step": 1585,
      "train_accuracy": 0.013029315508902073
    },
    {
      "epoch": 122.0,
      "grad_norm": 0.5589063167572021,
      "learning_rate": 0.01,
      "loss": 4.544114589691162,
      "num_input_tokens_seen": 6296176,
      "step": 1586,
      "train_runtime": 1290.3365,
      "train_tokens_per_second": 4879.484
    },
    {
      "epoch": 122.0,
      "num_input_tokens_seen": 6296176,
      "step": 1586,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 122.07692307692308,
      "grad_norm": 6.564464092254639,
      "learning_rate": 0.01,
      "loss": 4.536205768585205,
      "num_input_tokens_seen": 6300272,
      "step": 1587,
      "train_runtime": 1291.1625,
      "train_tokens_per_second": 4879.534
    },
    {
      "epoch": 122.07692307692308,
      "num_input_tokens_seen": 6300272,
      "step": 1587,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 122.15384615384616,
      "grad_norm": 0.27498263120651245,
      "learning_rate": 0.01,
      "loss": 4.545239448547363,
      "num_input_tokens_seen": 6304368,
      "step": 1588,
      "train_runtime": 1291.9743,
      "train_tokens_per_second": 4879.639
    },
    {
      "epoch": 122.15384615384616,
      "num_input_tokens_seen": 6304368,
      "step": 1588,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 122.23076923076923,
      "grad_norm": 0.24746458232402802,
      "learning_rate": 0.01,
      "loss": 4.548259735107422,
      "num_input_tokens_seen": 6308464,
      "step": 1589,
      "train_runtime": 1292.786,
      "train_tokens_per_second": 4879.743
    },
    {
      "epoch": 122.23076923076923,
      "num_input_tokens_seen": 6308464,
      "step": 1589,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 122.3076923076923,
      "grad_norm": 0.2766924798488617,
      "learning_rate": 0.01,
      "loss": 4.5435686111450195,
      "num_input_tokens_seen": 6312560,
      "step": 1590,
      "train_runtime": 1293.5982,
      "train_tokens_per_second": 4879.846
    },
    {
      "epoch": 122.3076923076923,
      "num_input_tokens_seen": 6312560,
      "step": 1590,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 122.38461538461539,
      "grad_norm": 0.45969757437705994,
      "learning_rate": 0.01,
      "loss": 4.556778907775879,
      "num_input_tokens_seen": 6316656,
      "step": 1591,
      "train_runtime": 1294.4097,
      "train_tokens_per_second": 4879.951
    },
    {
      "epoch": 122.38461538461539,
      "num_input_tokens_seen": 6316656,
      "step": 1591,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 122.46153846153847,
      "grad_norm": 0.3670724034309387,
      "learning_rate": 0.01,
      "loss": 4.519824981689453,
      "num_input_tokens_seen": 6320752,
      "step": 1592,
      "train_runtime": 1295.2225,
      "train_tokens_per_second": 4880.051
    },
    {
      "epoch": 122.46153846153847,
      "num_input_tokens_seen": 6320752,
      "step": 1592,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 122.53846153846153,
      "grad_norm": 0.39103612303733826,
      "learning_rate": 0.01,
      "loss": 4.520946025848389,
      "num_input_tokens_seen": 6324848,
      "step": 1593,
      "train_runtime": 1296.0353,
      "train_tokens_per_second": 4880.151
    },
    {
      "epoch": 122.53846153846153,
      "num_input_tokens_seen": 6324848,
      "step": 1593,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 122.61538461538461,
      "grad_norm": 0.3757348954677582,
      "learning_rate": 0.01,
      "loss": 4.541189670562744,
      "num_input_tokens_seen": 6328944,
      "step": 1594,
      "train_runtime": 1296.8499,
      "train_tokens_per_second": 4880.244
    },
    {
      "epoch": 122.61538461538461,
      "num_input_tokens_seen": 6328944,
      "step": 1594,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 122.6923076923077,
      "grad_norm": 0.4918854534626007,
      "learning_rate": 0.01,
      "loss": 4.520682334899902,
      "num_input_tokens_seen": 6333040,
      "step": 1595,
      "train_runtime": 1297.6675,
      "train_tokens_per_second": 4880.326
    },
    {
      "epoch": 122.6923076923077,
      "num_input_tokens_seen": 6333040,
      "step": 1595,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 122.76923076923077,
      "grad_norm": 0.24628299474716187,
      "learning_rate": 0.01,
      "loss": 4.551218032836914,
      "num_input_tokens_seen": 6337136,
      "step": 1596,
      "train_runtime": 1298.4828,
      "train_tokens_per_second": 4880.416
    },
    {
      "epoch": 122.76923076923077,
      "num_input_tokens_seen": 6337136,
      "step": 1596,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 122.84615384615384,
      "grad_norm": 0.2693702280521393,
      "learning_rate": 0.01,
      "loss": 4.543005466461182,
      "num_input_tokens_seen": 6341232,
      "step": 1597,
      "train_runtime": 1299.2991,
      "train_tokens_per_second": 4880.502
    },
    {
      "epoch": 122.84615384615384,
      "num_input_tokens_seen": 6341232,
      "step": 1597,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 122.92307692307692,
      "grad_norm": 0.22442473471164703,
      "learning_rate": 0.01,
      "loss": 4.543325901031494,
      "num_input_tokens_seen": 6345328,
      "step": 1598,
      "train_runtime": 1300.1073,
      "train_tokens_per_second": 4880.619
    },
    {
      "epoch": 122.92307692307692,
      "num_input_tokens_seen": 6345328,
      "step": 1598,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 123.0,
      "grad_norm": 0.2987178564071655,
      "learning_rate": 0.01,
      "loss": 4.5363359451293945,
      "num_input_tokens_seen": 6347784,
      "step": 1599,
      "train_runtime": 1300.6223,
      "train_tokens_per_second": 4880.575
    },
    {
      "epoch": 123.0,
      "num_input_tokens_seen": 6347784,
      "step": 1599,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 123.07692307692308,
      "grad_norm": 0.3728465139865875,
      "learning_rate": 0.01,
      "loss": 4.529542446136475,
      "num_input_tokens_seen": 6351880,
      "step": 1600,
      "train_runtime": 1301.447,
      "train_tokens_per_second": 4880.629
    },
    {
      "epoch": 123.07692307692308,
      "eval_CMD_3_branch_eval_accuracy": 0.00976491862567812,
      "eval_CMD_3_branch_eval_loss": 4.556027412414551,
      "eval_CMD_3_branch_eval_runtime": 1.1547,
      "eval_CMD_3_branch_eval_samples_per_second": 2394.554,
      "eval_CMD_3_branch_eval_steps_per_second": 5.196,
      "num_input_tokens_seen": 6351880,
      "step": 1600
    },
    {
      "epoch": 123.07692307692308,
      "num_input_tokens_seen": 6351880,
      "step": 1600,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 123.15384615384616,
      "grad_norm": 0.7718985676765442,
      "learning_rate": 0.01,
      "loss": 4.517962455749512,
      "num_input_tokens_seen": 6355976,
      "step": 1601,
      "train_runtime": 1303.4195,
      "train_tokens_per_second": 4876.385
    },
    {
      "epoch": 123.15384615384616,
      "num_input_tokens_seen": 6355976,
      "step": 1601,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 123.23076923076923,
      "grad_norm": 0.4242638945579529,
      "learning_rate": 0.01,
      "loss": 4.509787559509277,
      "num_input_tokens_seen": 6360072,
      "step": 1602,
      "train_runtime": 1304.2336,
      "train_tokens_per_second": 4876.482
    },
    {
      "epoch": 123.23076923076923,
      "num_input_tokens_seen": 6360072,
      "step": 1602,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 123.3076923076923,
      "grad_norm": 0.2297019511461258,
      "learning_rate": 0.01,
      "loss": 4.508452415466309,
      "num_input_tokens_seen": 6364168,
      "step": 1603,
      "train_runtime": 1305.0504,
      "train_tokens_per_second": 4876.569
    },
    {
      "epoch": 123.3076923076923,
      "num_input_tokens_seen": 6364168,
      "step": 1603,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 123.38461538461539,
      "grad_norm": 3.918861150741577,
      "learning_rate": 0.01,
      "loss": 4.52266788482666,
      "num_input_tokens_seen": 6368264,
      "step": 1604,
      "train_runtime": 1305.8653,
      "train_tokens_per_second": 4876.662
    },
    {
      "epoch": 123.38461538461539,
      "num_input_tokens_seen": 6368264,
      "step": 1604,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 123.46153846153847,
      "grad_norm": 0.28163081407546997,
      "learning_rate": 0.01,
      "loss": 4.511225700378418,
      "num_input_tokens_seen": 6372360,
      "step": 1605,
      "train_runtime": 1306.6785,
      "train_tokens_per_second": 4876.762
    },
    {
      "epoch": 123.46153846153847,
      "num_input_tokens_seen": 6372360,
      "step": 1605,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 123.53846153846153,
      "grad_norm": 0.8108887672424316,
      "learning_rate": 0.01,
      "loss": 4.5190958976745605,
      "num_input_tokens_seen": 6376456,
      "step": 1606,
      "train_runtime": 1307.4919,
      "train_tokens_per_second": 4876.861
    },
    {
      "epoch": 123.53846153846153,
      "num_input_tokens_seen": 6376456,
      "step": 1606,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 123.61538461538461,
      "grad_norm": 0.43255141377449036,
      "learning_rate": 0.01,
      "loss": 4.543148517608643,
      "num_input_tokens_seen": 6380552,
      "step": 1607,
      "train_runtime": 1308.3067,
      "train_tokens_per_second": 4876.954
    },
    {
      "epoch": 123.61538461538461,
      "num_input_tokens_seen": 6380552,
      "step": 1607,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 123.6923076923077,
      "grad_norm": 0.43423449993133545,
      "learning_rate": 0.01,
      "loss": 4.50079345703125,
      "num_input_tokens_seen": 6384648,
      "step": 1608,
      "train_runtime": 1309.1205,
      "train_tokens_per_second": 4877.051
    },
    {
      "epoch": 123.6923076923077,
      "num_input_tokens_seen": 6384648,
      "step": 1608,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 123.76923076923077,
      "grad_norm": 0.3461111783981323,
      "learning_rate": 0.01,
      "loss": 4.5239973068237305,
      "num_input_tokens_seen": 6388744,
      "step": 1609,
      "train_runtime": 1309.9338,
      "train_tokens_per_second": 4877.15
    },
    {
      "epoch": 123.76923076923077,
      "num_input_tokens_seen": 6388744,
      "step": 1609,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 123.84615384615384,
      "grad_norm": 0.2603447139263153,
      "learning_rate": 0.01,
      "loss": 4.500923156738281,
      "num_input_tokens_seen": 6392840,
      "step": 1610,
      "train_runtime": 1310.7472,
      "train_tokens_per_second": 4877.249
    },
    {
      "epoch": 123.84615384615384,
      "num_input_tokens_seen": 6392840,
      "step": 1610,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 123.92307692307692,
      "grad_norm": 0.3868979513645172,
      "learning_rate": 0.01,
      "loss": 4.53810977935791,
      "num_input_tokens_seen": 6396936,
      "step": 1611,
      "train_runtime": 1311.5584,
      "train_tokens_per_second": 4877.355
    },
    {
      "epoch": 123.92307692307692,
      "num_input_tokens_seen": 6396936,
      "step": 1611,
      "train_accuracy": 0.0032573288772255182
    },
    {
      "epoch": 124.0,
      "grad_norm": 0.3463993966579437,
      "learning_rate": 0.01,
      "loss": 4.504649639129639,
      "num_input_tokens_seen": 6399392,
      "step": 1612,
      "train_runtime": 1312.0732,
      "train_tokens_per_second": 4877.313
    },
    {
      "epoch": 124.0,
      "num_input_tokens_seen": 6399392,
      "step": 1612,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 124.07692307692308,
      "grad_norm": 0.2502296566963196,
      "learning_rate": 0.01,
      "loss": 4.480804920196533,
      "num_input_tokens_seen": 6403488,
      "step": 1613,
      "train_runtime": 1312.8993,
      "train_tokens_per_second": 4877.364
    },
    {
      "epoch": 124.07692307692308,
      "num_input_tokens_seen": 6403488,
      "step": 1613,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 124.15384615384616,
      "grad_norm": 0.3520311117172241,
      "learning_rate": 0.01,
      "loss": 4.5011162757873535,
      "num_input_tokens_seen": 6407584,
      "step": 1614,
      "train_runtime": 1313.7128,
      "train_tokens_per_second": 4877.462
    },
    {
      "epoch": 124.15384615384616,
      "num_input_tokens_seen": 6407584,
      "step": 1614,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 124.23076923076923,
      "grad_norm": 0.31740519404411316,
      "learning_rate": 0.01,
      "loss": 4.49839973449707,
      "num_input_tokens_seen": 6411680,
      "step": 1615,
      "train_runtime": 1314.5269,
      "train_tokens_per_second": 4877.557
    },
    {
      "epoch": 124.23076923076923,
      "num_input_tokens_seen": 6411680,
      "step": 1615,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 124.3076923076923,
      "grad_norm": 1.2244733572006226,
      "learning_rate": 0.01,
      "loss": 4.507872104644775,
      "num_input_tokens_seen": 6415776,
      "step": 1616,
      "train_runtime": 1315.3402,
      "train_tokens_per_second": 4877.655
    },
    {
      "epoch": 124.3076923076923,
      "num_input_tokens_seen": 6415776,
      "step": 1616,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 124.38461538461539,
      "grad_norm": 0.23134058713912964,
      "learning_rate": 0.01,
      "loss": 4.503739356994629,
      "num_input_tokens_seen": 6419872,
      "step": 1617,
      "train_runtime": 1316.1544,
      "train_tokens_per_second": 4877.75
    },
    {
      "epoch": 124.38461538461539,
      "num_input_tokens_seen": 6419872,
      "step": 1617,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 124.46153846153847,
      "grad_norm": 0.23125620186328888,
      "learning_rate": 0.01,
      "loss": 4.487939834594727,
      "num_input_tokens_seen": 6423968,
      "step": 1618,
      "train_runtime": 1316.9681,
      "train_tokens_per_second": 4877.846
    },
    {
      "epoch": 124.46153846153847,
      "num_input_tokens_seen": 6423968,
      "step": 1618,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 124.53846153846153,
      "grad_norm": 0.38358980417251587,
      "learning_rate": 0.01,
      "loss": 4.493551254272461,
      "num_input_tokens_seen": 6428064,
      "step": 1619,
      "train_runtime": 1317.7821,
      "train_tokens_per_second": 4877.941
    },
    {
      "epoch": 124.53846153846153,
      "num_input_tokens_seen": 6428064,
      "step": 1619,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 124.61538461538461,
      "grad_norm": 0.2873987555503845,
      "learning_rate": 0.01,
      "loss": 4.499876976013184,
      "num_input_tokens_seen": 6432160,
      "step": 1620,
      "train_runtime": 1318.5957,
      "train_tokens_per_second": 4878.038
    },
    {
      "epoch": 124.61538461538461,
      "num_input_tokens_seen": 6432160,
      "step": 1620,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 124.6923076923077,
      "grad_norm": 0.34333235025405884,
      "learning_rate": 0.01,
      "loss": 4.523139953613281,
      "num_input_tokens_seen": 6436256,
      "step": 1621,
      "train_runtime": 1319.4087,
      "train_tokens_per_second": 4878.137
    },
    {
      "epoch": 124.6923076923077,
      "num_input_tokens_seen": 6436256,
      "step": 1621,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 124.76923076923077,
      "grad_norm": 0.2574313282966614,
      "learning_rate": 0.01,
      "loss": 4.504465103149414,
      "num_input_tokens_seen": 6440352,
      "step": 1622,
      "train_runtime": 1320.2226,
      "train_tokens_per_second": 4878.232
    },
    {
      "epoch": 124.76923076923077,
      "num_input_tokens_seen": 6440352,
      "step": 1622,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 124.84615384615384,
      "grad_norm": 0.42629674077033997,
      "learning_rate": 0.01,
      "loss": 4.479188919067383,
      "num_input_tokens_seen": 6444448,
      "step": 1623,
      "train_runtime": 1321.0355,
      "train_tokens_per_second": 4878.331
    },
    {
      "epoch": 124.84615384615384,
      "num_input_tokens_seen": 6444448,
      "step": 1623,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 124.92307692307692,
      "grad_norm": 0.4172653257846832,
      "learning_rate": 0.01,
      "loss": 4.5147705078125,
      "num_input_tokens_seen": 6448544,
      "step": 1624,
      "train_runtime": 1321.8485,
      "train_tokens_per_second": 4878.429
    },
    {
      "epoch": 124.92307692307692,
      "num_input_tokens_seen": 6448544,
      "step": 1624,
      "train_accuracy": 0.013029315508902073
    },
    {
      "epoch": 125.0,
      "grad_norm": 0.489302396774292,
      "learning_rate": 0.01,
      "loss": 4.570430278778076,
      "num_input_tokens_seen": 6451000,
      "step": 1625,
      "train_runtime": 1322.3633,
      "train_tokens_per_second": 4878.387
    },
    {
      "epoch": 125.0,
      "num_input_tokens_seen": 6451000,
      "step": 1625,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 125.07692307692308,
      "grad_norm": 0.25735050439834595,
      "learning_rate": 0.01,
      "loss": 4.496250629425049,
      "num_input_tokens_seen": 6455096,
      "step": 1626,
      "train_runtime": 1323.1893,
      "train_tokens_per_second": 4878.437
    },
    {
      "epoch": 125.07692307692308,
      "num_input_tokens_seen": 6455096,
      "step": 1626,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 125.15384615384616,
      "grad_norm": 0.3413771390914917,
      "learning_rate": 0.01,
      "loss": 4.487084865570068,
      "num_input_tokens_seen": 6459192,
      "step": 1627,
      "train_runtime": 1324.0028,
      "train_tokens_per_second": 4878.534
    },
    {
      "epoch": 125.15384615384616,
      "num_input_tokens_seen": 6459192,
      "step": 1627,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 125.23076923076923,
      "grad_norm": 0.3020739257335663,
      "learning_rate": 0.01,
      "loss": 4.504093170166016,
      "num_input_tokens_seen": 6463288,
      "step": 1628,
      "train_runtime": 1324.817,
      "train_tokens_per_second": 4878.627
    },
    {
      "epoch": 125.23076923076923,
      "num_input_tokens_seen": 6463288,
      "step": 1628,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 125.3076923076923,
      "grad_norm": 0.5350129008293152,
      "learning_rate": 0.01,
      "loss": 4.505185127258301,
      "num_input_tokens_seen": 6467384,
      "step": 1629,
      "train_runtime": 1325.6306,
      "train_tokens_per_second": 4878.723
    },
    {
      "epoch": 125.3076923076923,
      "num_input_tokens_seen": 6467384,
      "step": 1629,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 125.38461538461539,
      "grad_norm": 0.29904764890670776,
      "learning_rate": 0.01,
      "loss": 4.511266231536865,
      "num_input_tokens_seen": 6471480,
      "step": 1630,
      "train_runtime": 1326.4447,
      "train_tokens_per_second": 4878.816
    },
    {
      "epoch": 125.38461538461539,
      "num_input_tokens_seen": 6471480,
      "step": 1630,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 125.46153846153847,
      "grad_norm": 0.3054788410663605,
      "learning_rate": 0.01,
      "loss": 4.485260009765625,
      "num_input_tokens_seen": 6475576,
      "step": 1631,
      "train_runtime": 1327.2603,
      "train_tokens_per_second": 4878.904
    },
    {
      "epoch": 125.46153846153847,
      "num_input_tokens_seen": 6475576,
      "step": 1631,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 125.53846153846153,
      "grad_norm": 0.34672224521636963,
      "learning_rate": 0.01,
      "loss": 4.496123790740967,
      "num_input_tokens_seen": 6479672,
      "step": 1632,
      "train_runtime": 1328.0741,
      "train_tokens_per_second": 4878.999
    },
    {
      "epoch": 125.53846153846153,
      "num_input_tokens_seen": 6479672,
      "step": 1632,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 125.61538461538461,
      "grad_norm": 0.33455249667167664,
      "learning_rate": 0.01,
      "loss": 4.514712333679199,
      "num_input_tokens_seen": 6483768,
      "step": 1633,
      "train_runtime": 1328.8871,
      "train_tokens_per_second": 4879.096
    },
    {
      "epoch": 125.61538461538461,
      "num_input_tokens_seen": 6483768,
      "step": 1633,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 125.6923076923077,
      "grad_norm": 0.3615611493587494,
      "learning_rate": 0.01,
      "loss": 4.473272800445557,
      "num_input_tokens_seen": 6487864,
      "step": 1634,
      "train_runtime": 1329.7004,
      "train_tokens_per_second": 4879.192
    },
    {
      "epoch": 125.6923076923077,
      "num_input_tokens_seen": 6487864,
      "step": 1634,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 125.76923076923077,
      "grad_norm": 0.3291076421737671,
      "learning_rate": 0.01,
      "loss": 4.481527805328369,
      "num_input_tokens_seen": 6491960,
      "step": 1635,
      "train_runtime": 1330.5145,
      "train_tokens_per_second": 4879.285
    },
    {
      "epoch": 125.76923076923077,
      "num_input_tokens_seen": 6491960,
      "step": 1635,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 125.84615384615384,
      "grad_norm": 0.27003857493400574,
      "learning_rate": 0.01,
      "loss": 4.5121259689331055,
      "num_input_tokens_seen": 6496056,
      "step": 1636,
      "train_runtime": 1331.3279,
      "train_tokens_per_second": 4879.381
    },
    {
      "epoch": 125.84615384615384,
      "num_input_tokens_seen": 6496056,
      "step": 1636,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 125.92307692307692,
      "grad_norm": 1.236617922782898,
      "learning_rate": 0.01,
      "loss": 4.4990410804748535,
      "num_input_tokens_seen": 6500152,
      "step": 1637,
      "train_runtime": 1332.1368,
      "train_tokens_per_second": 4879.493
    },
    {
      "epoch": 125.92307692307692,
      "num_input_tokens_seen": 6500152,
      "step": 1637,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 126.0,
      "grad_norm": 0.4936729669570923,
      "learning_rate": 0.01,
      "loss": 4.497000217437744,
      "num_input_tokens_seen": 6502608,
      "step": 1638,
      "train_runtime": 1332.6521,
      "train_tokens_per_second": 4879.449
    },
    {
      "epoch": 126.0,
      "num_input_tokens_seen": 6502608,
      "step": 1638,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 126.07692307692308,
      "grad_norm": 1.1737216711044312,
      "learning_rate": 0.01,
      "loss": 4.479070663452148,
      "num_input_tokens_seen": 6506704,
      "step": 1639,
      "train_runtime": 1333.4797,
      "train_tokens_per_second": 4879.492
    },
    {
      "epoch": 126.07692307692308,
      "num_input_tokens_seen": 6506704,
      "step": 1639,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 126.15384615384616,
      "grad_norm": 0.24445922672748566,
      "learning_rate": 0.01,
      "loss": 4.45222282409668,
      "num_input_tokens_seen": 6510800,
      "step": 1640,
      "train_runtime": 1334.2924,
      "train_tokens_per_second": 4879.59
    },
    {
      "epoch": 126.15384615384616,
      "num_input_tokens_seen": 6510800,
      "step": 1640,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 126.23076923076923,
      "grad_norm": 0.37953752279281616,
      "learning_rate": 0.01,
      "loss": 4.512228488922119,
      "num_input_tokens_seen": 6514896,
      "step": 1641,
      "train_runtime": 1335.1062,
      "train_tokens_per_second": 4879.684
    },
    {
      "epoch": 126.23076923076923,
      "num_input_tokens_seen": 6514896,
      "step": 1641,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 126.3076923076923,
      "grad_norm": 0.27803337574005127,
      "learning_rate": 0.01,
      "loss": 4.438938140869141,
      "num_input_tokens_seen": 6518992,
      "step": 1642,
      "train_runtime": 1335.9201,
      "train_tokens_per_second": 4879.777
    },
    {
      "epoch": 126.3076923076923,
      "num_input_tokens_seen": 6518992,
      "step": 1642,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 126.38461538461539,
      "grad_norm": 0.5527097582817078,
      "learning_rate": 0.01,
      "loss": 4.486285209655762,
      "num_input_tokens_seen": 6523088,
      "step": 1643,
      "train_runtime": 1336.7334,
      "train_tokens_per_second": 4879.872
    },
    {
      "epoch": 126.38461538461539,
      "num_input_tokens_seen": 6523088,
      "step": 1643,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 126.46153846153847,
      "grad_norm": 0.2564176917076111,
      "learning_rate": 0.01,
      "loss": 4.4830732345581055,
      "num_input_tokens_seen": 6527184,
      "step": 1644,
      "train_runtime": 1337.5463,
      "train_tokens_per_second": 4879.969
    },
    {
      "epoch": 126.46153846153847,
      "num_input_tokens_seen": 6527184,
      "step": 1644,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 126.53846153846153,
      "grad_norm": 0.5148802995681763,
      "learning_rate": 0.01,
      "loss": 4.4817304611206055,
      "num_input_tokens_seen": 6531280,
      "step": 1645,
      "train_runtime": 1338.3606,
      "train_tokens_per_second": 4880.06
    },
    {
      "epoch": 126.53846153846153,
      "num_input_tokens_seen": 6531280,
      "step": 1645,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 126.61538461538461,
      "grad_norm": 0.3720771074295044,
      "learning_rate": 0.01,
      "loss": 4.476485252380371,
      "num_input_tokens_seen": 6535376,
      "step": 1646,
      "train_runtime": 1339.1734,
      "train_tokens_per_second": 4880.157
    },
    {
      "epoch": 126.61538461538461,
      "num_input_tokens_seen": 6535376,
      "step": 1646,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 126.6923076923077,
      "grad_norm": 0.225808784365654,
      "learning_rate": 0.01,
      "loss": 4.528336524963379,
      "num_input_tokens_seen": 6539472,
      "step": 1647,
      "train_runtime": 1339.9879,
      "train_tokens_per_second": 4880.247
    },
    {
      "epoch": 126.6923076923077,
      "num_input_tokens_seen": 6539472,
      "step": 1647,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 126.76923076923077,
      "grad_norm": 0.620141863822937,
      "learning_rate": 0.01,
      "loss": 4.514115333557129,
      "num_input_tokens_seen": 6543568,
      "step": 1648,
      "train_runtime": 1340.8025,
      "train_tokens_per_second": 4880.337
    },
    {
      "epoch": 126.76923076923077,
      "num_input_tokens_seen": 6543568,
      "step": 1648,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 126.84615384615384,
      "grad_norm": 0.39430055022239685,
      "learning_rate": 0.01,
      "loss": 4.453020095825195,
      "num_input_tokens_seen": 6547664,
      "step": 1649,
      "train_runtime": 1341.6178,
      "train_tokens_per_second": 4880.424
    },
    {
      "epoch": 126.84615384615384,
      "num_input_tokens_seen": 6547664,
      "step": 1649,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 126.92307692307692,
      "grad_norm": 0.3218206465244293,
      "learning_rate": 0.01,
      "loss": 4.488650321960449,
      "num_input_tokens_seen": 6551760,
      "step": 1650,
      "train_runtime": 1342.4271,
      "train_tokens_per_second": 4880.533
    },
    {
      "epoch": 126.92307692307692,
      "num_input_tokens_seen": 6551760,
      "step": 1650,
      "train_accuracy": 0.016286645084619522
    },
    {
      "epoch": 127.0,
      "grad_norm": 0.4372219443321228,
      "learning_rate": 0.01,
      "loss": 4.472595691680908,
      "num_input_tokens_seen": 6554216,
      "step": 1651,
      "train_runtime": 1342.9429,
      "train_tokens_per_second": 4880.488
    },
    {
      "epoch": 127.0,
      "num_input_tokens_seen": 6554216,
      "step": 1651,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 127.07692307692308,
      "grad_norm": 0.565567135810852,
      "learning_rate": 0.01,
      "loss": 4.468566417694092,
      "num_input_tokens_seen": 6558312,
      "step": 1652,
      "train_runtime": 1343.7698,
      "train_tokens_per_second": 4880.532
    },
    {
      "epoch": 127.07692307692308,
      "num_input_tokens_seen": 6558312,
      "step": 1652,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 127.15384615384616,
      "grad_norm": 0.4018939435482025,
      "learning_rate": 0.01,
      "loss": 4.468479633331299,
      "num_input_tokens_seen": 6562408,
      "step": 1653,
      "train_runtime": 1344.5841,
      "train_tokens_per_second": 4880.623
    },
    {
      "epoch": 127.15384615384616,
      "num_input_tokens_seen": 6562408,
      "step": 1653,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 127.23076923076923,
      "grad_norm": 0.44568854570388794,
      "learning_rate": 0.01,
      "loss": 4.407977104187012,
      "num_input_tokens_seen": 6566504,
      "step": 1654,
      "train_runtime": 1345.3978,
      "train_tokens_per_second": 4880.716
    },
    {
      "epoch": 127.23076923076923,
      "num_input_tokens_seen": 6566504,
      "step": 1654,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 127.3076923076923,
      "grad_norm": 0.45499709248542786,
      "learning_rate": 0.01,
      "loss": 4.49755334854126,
      "num_input_tokens_seen": 6570600,
      "step": 1655,
      "train_runtime": 1346.2113,
      "train_tokens_per_second": 4880.809
    },
    {
      "epoch": 127.3076923076923,
      "num_input_tokens_seen": 6570600,
      "step": 1655,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 127.38461538461539,
      "grad_norm": 0.35654014348983765,
      "learning_rate": 0.01,
      "loss": 4.417347431182861,
      "num_input_tokens_seen": 6574696,
      "step": 1656,
      "train_runtime": 1347.0263,
      "train_tokens_per_second": 4880.897
    },
    {
      "epoch": 127.38461538461539,
      "num_input_tokens_seen": 6574696,
      "step": 1656,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 127.46153846153847,
      "grad_norm": 1.5767470598220825,
      "learning_rate": 0.01,
      "loss": 4.4421610832214355,
      "num_input_tokens_seen": 6578792,
      "step": 1657,
      "train_runtime": 1347.8417,
      "train_tokens_per_second": 4880.983
    },
    {
      "epoch": 127.46153846153847,
      "num_input_tokens_seen": 6578792,
      "step": 1657,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 127.53846153846153,
      "grad_norm": 2.769388198852539,
      "learning_rate": 0.01,
      "loss": 4.470029830932617,
      "num_input_tokens_seen": 6582888,
      "step": 1658,
      "train_runtime": 1348.6593,
      "train_tokens_per_second": 4881.061
    },
    {
      "epoch": 127.53846153846153,
      "num_input_tokens_seen": 6582888,
      "step": 1658,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 127.61538461538461,
      "grad_norm": 0.3703688979148865,
      "learning_rate": 0.01,
      "loss": 4.554350852966309,
      "num_input_tokens_seen": 6586984,
      "step": 1659,
      "train_runtime": 1349.4813,
      "train_tokens_per_second": 4881.123
    },
    {
      "epoch": 127.61538461538461,
      "num_input_tokens_seen": 6586984,
      "step": 1659,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 127.6923076923077,
      "grad_norm": 0.28809821605682373,
      "learning_rate": 0.01,
      "loss": 4.53988790512085,
      "num_input_tokens_seen": 6591080,
      "step": 1660,
      "train_runtime": 1350.2957,
      "train_tokens_per_second": 4881.212
    },
    {
      "epoch": 127.6923076923077,
      "num_input_tokens_seen": 6591080,
      "step": 1660,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 127.76923076923077,
      "grad_norm": 0.20119811594486237,
      "learning_rate": 0.01,
      "loss": 4.546500205993652,
      "num_input_tokens_seen": 6595176,
      "step": 1661,
      "train_runtime": 1351.1102,
      "train_tokens_per_second": 4881.301
    },
    {
      "epoch": 127.76923076923077,
      "num_input_tokens_seen": 6595176,
      "step": 1661,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 127.84615384615384,
      "grad_norm": 0.4292847514152527,
      "learning_rate": 0.01,
      "loss": 4.514669895172119,
      "num_input_tokens_seen": 6599272,
      "step": 1662,
      "train_runtime": 1351.9254,
      "train_tokens_per_second": 4881.388
    },
    {
      "epoch": 127.84615384615384,
      "num_input_tokens_seen": 6599272,
      "step": 1662,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 127.92307692307692,
      "grad_norm": 0.2840721309185028,
      "learning_rate": 0.01,
      "loss": 4.520134449005127,
      "num_input_tokens_seen": 6603368,
      "step": 1663,
      "train_runtime": 1352.7368,
      "train_tokens_per_second": 4881.488
    },
    {
      "epoch": 127.92307692307692,
      "num_input_tokens_seen": 6603368,
      "step": 1663,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 128.0,
      "grad_norm": 0.2847785949707031,
      "learning_rate": 0.01,
      "loss": 4.505764961242676,
      "num_input_tokens_seen": 6605824,
      "step": 1664,
      "train_runtime": 1353.2521,
      "train_tokens_per_second": 4881.444
    },
    {
      "epoch": 128.0,
      "num_input_tokens_seen": 6605824,
      "step": 1664,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 128.07692307692307,
      "grad_norm": 0.2890031933784485,
      "learning_rate": 0.01,
      "loss": 4.50120735168457,
      "num_input_tokens_seen": 6609920,
      "step": 1665,
      "train_runtime": 1354.0795,
      "train_tokens_per_second": 4881.486
    },
    {
      "epoch": 128.07692307692307,
      "num_input_tokens_seen": 6609920,
      "step": 1665,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 128.15384615384616,
      "grad_norm": 0.46287044882774353,
      "learning_rate": 0.01,
      "loss": 4.448619842529297,
      "num_input_tokens_seen": 6614016,
      "step": 1666,
      "train_runtime": 1354.8941,
      "train_tokens_per_second": 4881.574
    },
    {
      "epoch": 128.15384615384616,
      "num_input_tokens_seen": 6614016,
      "step": 1666,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 128.23076923076923,
      "grad_norm": 0.3689666986465454,
      "learning_rate": 0.01,
      "loss": 4.424829006195068,
      "num_input_tokens_seen": 6618112,
      "step": 1667,
      "train_runtime": 1355.7089,
      "train_tokens_per_second": 4881.661
    },
    {
      "epoch": 128.23076923076923,
      "num_input_tokens_seen": 6618112,
      "step": 1667,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 128.30769230769232,
      "grad_norm": 0.36388009786605835,
      "learning_rate": 0.01,
      "loss": 4.468934059143066,
      "num_input_tokens_seen": 6622208,
      "step": 1668,
      "train_runtime": 1356.5239,
      "train_tokens_per_second": 4881.748
    },
    {
      "epoch": 128.30769230769232,
      "num_input_tokens_seen": 6622208,
      "step": 1668,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 128.3846153846154,
      "grad_norm": 0.8547375202178955,
      "learning_rate": 0.01,
      "loss": 4.457394599914551,
      "num_input_tokens_seen": 6626304,
      "step": 1669,
      "train_runtime": 1357.3392,
      "train_tokens_per_second": 4881.833
    },
    {
      "epoch": 128.3846153846154,
      "num_input_tokens_seen": 6626304,
      "step": 1669,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 128.46153846153845,
      "grad_norm": 0.9688699245452881,
      "learning_rate": 0.01,
      "loss": 4.502564907073975,
      "num_input_tokens_seen": 6630400,
      "step": 1670,
      "train_runtime": 1358.1563,
      "train_tokens_per_second": 4881.912
    },
    {
      "epoch": 128.46153846153845,
      "num_input_tokens_seen": 6630400,
      "step": 1670,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 128.53846153846155,
      "grad_norm": 0.8407202363014221,
      "learning_rate": 0.01,
      "loss": 4.473555088043213,
      "num_input_tokens_seen": 6634496,
      "step": 1671,
      "train_runtime": 1358.9712,
      "train_tokens_per_second": 4881.999
    },
    {
      "epoch": 128.53846153846155,
      "num_input_tokens_seen": 6634496,
      "step": 1671,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 128.6153846153846,
      "grad_norm": 3.261981964111328,
      "learning_rate": 0.01,
      "loss": 4.496661186218262,
      "num_input_tokens_seen": 6638592,
      "step": 1672,
      "train_runtime": 1359.7869,
      "train_tokens_per_second": 4882.083
    },
    {
      "epoch": 128.6153846153846,
      "num_input_tokens_seen": 6638592,
      "step": 1672,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 128.69230769230768,
      "grad_norm": 0.4284020662307739,
      "learning_rate": 0.01,
      "loss": 4.4803972244262695,
      "num_input_tokens_seen": 6642688,
      "step": 1673,
      "train_runtime": 1360.6013,
      "train_tokens_per_second": 4882.171
    },
    {
      "epoch": 128.69230769230768,
      "num_input_tokens_seen": 6642688,
      "step": 1673,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 128.76923076923077,
      "grad_norm": 0.5555926561355591,
      "learning_rate": 0.01,
      "loss": 4.493105888366699,
      "num_input_tokens_seen": 6646784,
      "step": 1674,
      "train_runtime": 1361.4161,
      "train_tokens_per_second": 4882.258
    },
    {
      "epoch": 128.76923076923077,
      "num_input_tokens_seen": 6646784,
      "step": 1674,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 128.84615384615384,
      "grad_norm": 0.38842296600341797,
      "learning_rate": 0.01,
      "loss": 4.494626998901367,
      "num_input_tokens_seen": 6650880,
      "step": 1675,
      "train_runtime": 1362.23,
      "train_tokens_per_second": 4882.347
    },
    {
      "epoch": 128.84615384615384,
      "num_input_tokens_seen": 6650880,
      "step": 1675,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 128.92307692307693,
      "grad_norm": 0.278170108795166,
      "learning_rate": 0.01,
      "loss": 4.515069484710693,
      "num_input_tokens_seen": 6654976,
      "step": 1676,
      "train_runtime": 1363.0406,
      "train_tokens_per_second": 4882.449
    },
    {
      "epoch": 128.92307692307693,
      "num_input_tokens_seen": 6654976,
      "step": 1676,
      "train_accuracy": 0.0065146577544510365
    },
    {
      "epoch": 129.0,
      "grad_norm": 0.36831656098365784,
      "learning_rate": 0.01,
      "loss": 4.511685848236084,
      "num_input_tokens_seen": 6657432,
      "step": 1677,
      "train_runtime": 1363.5563,
      "train_tokens_per_second": 4882.403
    },
    {
      "epoch": 129.0,
      "num_input_tokens_seen": 6657432,
      "step": 1677,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 129.07692307692307,
      "grad_norm": 0.5876153111457825,
      "learning_rate": 0.01,
      "loss": 4.464977264404297,
      "num_input_tokens_seen": 6661528,
      "step": 1678,
      "train_runtime": 1364.3856,
      "train_tokens_per_second": 4882.438
    },
    {
      "epoch": 129.07692307692307,
      "num_input_tokens_seen": 6661528,
      "step": 1678,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 129.15384615384616,
      "grad_norm": 0.46501055359840393,
      "learning_rate": 0.01,
      "loss": 4.496500015258789,
      "num_input_tokens_seen": 6665624,
      "step": 1679,
      "train_runtime": 1365.1985,
      "train_tokens_per_second": 4882.531
    },
    {
      "epoch": 129.15384615384616,
      "num_input_tokens_seen": 6665624,
      "step": 1679,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 129.23076923076923,
      "grad_norm": 0.42770761251449585,
      "learning_rate": 0.01,
      "loss": 4.488858222961426,
      "num_input_tokens_seen": 6669720,
      "step": 1680,
      "train_runtime": 1366.0122,
      "train_tokens_per_second": 4882.621
    },
    {
      "epoch": 129.23076923076923,
      "num_input_tokens_seen": 6669720,
      "step": 1680,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 129.30769230769232,
      "grad_norm": 0.3145144283771515,
      "learning_rate": 0.01,
      "loss": 4.476947784423828,
      "num_input_tokens_seen": 6673816,
      "step": 1681,
      "train_runtime": 1366.8265,
      "train_tokens_per_second": 4882.709
    },
    {
      "epoch": 129.30769230769232,
      "num_input_tokens_seen": 6673816,
      "step": 1681,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 129.3846153846154,
      "grad_norm": 0.47921332716941833,
      "learning_rate": 0.01,
      "loss": 4.460630416870117,
      "num_input_tokens_seen": 6677912,
      "step": 1682,
      "train_runtime": 1367.6421,
      "train_tokens_per_second": 4882.792
    },
    {
      "epoch": 129.3846153846154,
      "num_input_tokens_seen": 6677912,
      "step": 1682,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 129.46153846153845,
      "grad_norm": 1.500560998916626,
      "learning_rate": 0.01,
      "loss": 4.608785629272461,
      "num_input_tokens_seen": 6682008,
      "step": 1683,
      "train_runtime": 1368.4605,
      "train_tokens_per_second": 4882.865
    },
    {
      "epoch": 129.46153846153845,
      "num_input_tokens_seen": 6682008,
      "step": 1683,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 129.53846153846155,
      "grad_norm": 0.43431711196899414,
      "learning_rate": 0.01,
      "loss": 4.492959022521973,
      "num_input_tokens_seen": 6686104,
      "step": 1684,
      "train_runtime": 1369.2764,
      "train_tokens_per_second": 4882.947
    },
    {
      "epoch": 129.53846153846155,
      "num_input_tokens_seen": 6686104,
      "step": 1684,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 129.6153846153846,
      "grad_norm": 0.39037230610847473,
      "learning_rate": 0.01,
      "loss": 4.5267791748046875,
      "num_input_tokens_seen": 6690200,
      "step": 1685,
      "train_runtime": 1370.0904,
      "train_tokens_per_second": 4883.035
    },
    {
      "epoch": 129.6153846153846,
      "num_input_tokens_seen": 6690200,
      "step": 1685,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 129.69230769230768,
      "grad_norm": 0.47450190782546997,
      "learning_rate": 0.01,
      "loss": 4.51047420501709,
      "num_input_tokens_seen": 6694296,
      "step": 1686,
      "train_runtime": 1370.9053,
      "train_tokens_per_second": 4883.121
    },
    {
      "epoch": 129.69230769230768,
      "num_input_tokens_seen": 6694296,
      "step": 1686,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 129.76923076923077,
      "grad_norm": 0.4482632279396057,
      "learning_rate": 0.01,
      "loss": 4.5030364990234375,
      "num_input_tokens_seen": 6698392,
      "step": 1687,
      "train_runtime": 1371.7198,
      "train_tokens_per_second": 4883.207
    },
    {
      "epoch": 129.76923076923077,
      "num_input_tokens_seen": 6698392,
      "step": 1687,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 129.84615384615384,
      "grad_norm": 3.1814825534820557,
      "learning_rate": 0.01,
      "loss": 4.50639009475708,
      "num_input_tokens_seen": 6702488,
      "step": 1688,
      "train_runtime": 1372.5346,
      "train_tokens_per_second": 4883.293
    },
    {
      "epoch": 129.84615384615384,
      "num_input_tokens_seen": 6702488,
      "step": 1688,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 129.92307692307693,
      "grad_norm": 0.31933286786079407,
      "learning_rate": 0.01,
      "loss": 4.529560089111328,
      "num_input_tokens_seen": 6706584,
      "step": 1689,
      "train_runtime": 1373.3442,
      "train_tokens_per_second": 4883.396
    },
    {
      "epoch": 129.92307692307693,
      "num_input_tokens_seen": 6706584,
      "step": 1689,
      "train_accuracy": 0.019543973729014397
    },
    {
      "epoch": 130.0,
      "grad_norm": 0.2602151930332184,
      "learning_rate": 0.01,
      "loss": 4.536754131317139,
      "num_input_tokens_seen": 6709040,
      "step": 1690,
      "train_runtime": 1373.8607,
      "train_tokens_per_second": 4883.348
    },
    {
      "epoch": 130.0,
      "num_input_tokens_seen": 6709040,
      "step": 1690,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 130.07692307692307,
      "grad_norm": 0.2738721966743469,
      "learning_rate": 0.01,
      "loss": 4.500980854034424,
      "num_input_tokens_seen": 6713136,
      "step": 1691,
      "train_runtime": 1374.6886,
      "train_tokens_per_second": 4883.387
    },
    {
      "epoch": 130.07692307692307,
      "num_input_tokens_seen": 6713136,
      "step": 1691,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 130.15384615384616,
      "grad_norm": 0.39055389165878296,
      "learning_rate": 0.01,
      "loss": 4.494011878967285,
      "num_input_tokens_seen": 6717232,
      "step": 1692,
      "train_runtime": 1375.505,
      "train_tokens_per_second": 4883.466
    },
    {
      "epoch": 130.15384615384616,
      "num_input_tokens_seen": 6717232,
      "step": 1692,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 130.23076923076923,
      "grad_norm": 0.28793901205062866,
      "learning_rate": 0.01,
      "loss": 4.490740776062012,
      "num_input_tokens_seen": 6721328,
      "step": 1693,
      "train_runtime": 1376.3196,
      "train_tokens_per_second": 4883.552
    },
    {
      "epoch": 130.23076923076923,
      "num_input_tokens_seen": 6721328,
      "step": 1693,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 130.30769230769232,
      "grad_norm": 0.8072336912155151,
      "learning_rate": 0.01,
      "loss": 4.515347480773926,
      "num_input_tokens_seen": 6725424,
      "step": 1694,
      "train_runtime": 1377.1347,
      "train_tokens_per_second": 4883.636
    },
    {
      "epoch": 130.30769230769232,
      "num_input_tokens_seen": 6725424,
      "step": 1694,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 130.3846153846154,
      "grad_norm": 0.5644285082817078,
      "learning_rate": 0.01,
      "loss": 4.502502918243408,
      "num_input_tokens_seen": 6729520,
      "step": 1695,
      "train_runtime": 1377.9489,
      "train_tokens_per_second": 4883.723
    },
    {
      "epoch": 130.3846153846154,
      "num_input_tokens_seen": 6729520,
      "step": 1695,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 130.46153846153845,
      "grad_norm": 2.5769872665405273,
      "learning_rate": 0.01,
      "loss": 4.527018070220947,
      "num_input_tokens_seen": 6733616,
      "step": 1696,
      "train_runtime": 1378.7649,
      "train_tokens_per_second": 4883.803
    },
    {
      "epoch": 130.46153846153845,
      "num_input_tokens_seen": 6733616,
      "step": 1696,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 130.53846153846155,
      "grad_norm": 0.38207435607910156,
      "learning_rate": 0.01,
      "loss": 4.5115509033203125,
      "num_input_tokens_seen": 6737712,
      "step": 1697,
      "train_runtime": 1379.5795,
      "train_tokens_per_second": 4883.888
    },
    {
      "epoch": 130.53846153846155,
      "num_input_tokens_seen": 6737712,
      "step": 1697,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 130.6153846153846,
      "grad_norm": 0.3670922815799713,
      "learning_rate": 0.01,
      "loss": 4.5531768798828125,
      "num_input_tokens_seen": 6741808,
      "step": 1698,
      "train_runtime": 1380.3946,
      "train_tokens_per_second": 4883.972
    },
    {
      "epoch": 130.6153846153846,
      "num_input_tokens_seen": 6741808,
      "step": 1698,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 130.69230769230768,
      "grad_norm": 0.7520430684089661,
      "learning_rate": 0.01,
      "loss": 4.529556751251221,
      "num_input_tokens_seen": 6745904,
      "step": 1699,
      "train_runtime": 1381.2083,
      "train_tokens_per_second": 4884.06
    },
    {
      "epoch": 130.69230769230768,
      "num_input_tokens_seen": 6745904,
      "step": 1699,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 130.76923076923077,
      "grad_norm": 0.993793785572052,
      "learning_rate": 0.01,
      "loss": 4.509443283081055,
      "num_input_tokens_seen": 6750000,
      "step": 1700,
      "train_runtime": 1382.0215,
      "train_tokens_per_second": 4884.15
    },
    {
      "epoch": 130.76923076923077,
      "eval_CMD_3_branch_eval_accuracy": 0.00976491862567812,
      "eval_CMD_3_branch_eval_loss": 4.546371936798096,
      "eval_CMD_3_branch_eval_runtime": 1.1534,
      "eval_CMD_3_branch_eval_samples_per_second": 2397.362,
      "eval_CMD_3_branch_eval_steps_per_second": 5.202,
      "num_input_tokens_seen": 6750000,
      "step": 1700
    },
    {
      "epoch": 130.76923076923077,
      "num_input_tokens_seen": 6750000,
      "step": 1700,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 130.84615384615384,
      "grad_norm": 0.4083004891872406,
      "learning_rate": 0.01,
      "loss": 4.506755828857422,
      "num_input_tokens_seen": 6754096,
      "step": 1701,
      "train_runtime": 1383.9929,
      "train_tokens_per_second": 4880.152
    },
    {
      "epoch": 130.84615384615384,
      "num_input_tokens_seen": 6754096,
      "step": 1701,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 130.92307692307693,
      "grad_norm": 0.3738880455493927,
      "learning_rate": 0.01,
      "loss": 4.514200210571289,
      "num_input_tokens_seen": 6758192,
      "step": 1702,
      "train_runtime": 1384.8029,
      "train_tokens_per_second": 4880.256
    },
    {
      "epoch": 130.92307692307693,
      "num_input_tokens_seen": 6758192,
      "step": 1702,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 131.0,
      "grad_norm": 1.534274935722351,
      "learning_rate": 0.01,
      "loss": 4.540097713470459,
      "num_input_tokens_seen": 6760648,
      "step": 1703,
      "train_runtime": 1385.3176,
      "train_tokens_per_second": 4880.215
    },
    {
      "epoch": 131.0,
      "num_input_tokens_seen": 6760648,
      "step": 1703,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 131.07692307692307,
      "grad_norm": 0.7651125192642212,
      "learning_rate": 0.01,
      "loss": 4.548482894897461,
      "num_input_tokens_seen": 6764744,
      "step": 1704,
      "train_runtime": 1386.1445,
      "train_tokens_per_second": 4880.259
    },
    {
      "epoch": 131.07692307692307,
      "num_input_tokens_seen": 6764744,
      "step": 1704,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 131.15384615384616,
      "grad_norm": 0.36869847774505615,
      "learning_rate": 0.01,
      "loss": 4.5239410400390625,
      "num_input_tokens_seen": 6768840,
      "step": 1705,
      "train_runtime": 1386.9647,
      "train_tokens_per_second": 4880.326
    },
    {
      "epoch": 131.15384615384616,
      "num_input_tokens_seen": 6768840,
      "step": 1705,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 131.23076923076923,
      "grad_norm": 0.2955624759197235,
      "learning_rate": 0.01,
      "loss": 4.503084182739258,
      "num_input_tokens_seen": 6772936,
      "step": 1706,
      "train_runtime": 1387.7791,
      "train_tokens_per_second": 4880.414
    },
    {
      "epoch": 131.23076923076923,
      "num_input_tokens_seen": 6772936,
      "step": 1706,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 131.30769230769232,
      "grad_norm": 0.39183342456817627,
      "learning_rate": 0.01,
      "loss": 4.5019731521606445,
      "num_input_tokens_seen": 6777032,
      "step": 1707,
      "train_runtime": 1388.5941,
      "train_tokens_per_second": 4880.499
    },
    {
      "epoch": 131.30769230769232,
      "num_input_tokens_seen": 6777032,
      "step": 1707,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 131.3846153846154,
      "grad_norm": 0.3188590407371521,
      "learning_rate": 0.01,
      "loss": 4.503392219543457,
      "num_input_tokens_seen": 6781128,
      "step": 1708,
      "train_runtime": 1389.4072,
      "train_tokens_per_second": 4880.591
    },
    {
      "epoch": 131.3846153846154,
      "num_input_tokens_seen": 6781128,
      "step": 1708,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 131.46153846153845,
      "grad_norm": 0.32286202907562256,
      "learning_rate": 0.01,
      "loss": 4.52543830871582,
      "num_input_tokens_seen": 6785224,
      "step": 1709,
      "train_runtime": 1390.2221,
      "train_tokens_per_second": 4880.676
    },
    {
      "epoch": 131.46153846153845,
      "num_input_tokens_seen": 6785224,
      "step": 1709,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 131.53846153846155,
      "grad_norm": 0.3742615580558777,
      "learning_rate": 0.01,
      "loss": 4.537827491760254,
      "num_input_tokens_seen": 6789320,
      "step": 1710,
      "train_runtime": 1391.0365,
      "train_tokens_per_second": 4880.763
    },
    {
      "epoch": 131.53846153846155,
      "num_input_tokens_seen": 6789320,
      "step": 1710,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 131.6153846153846,
      "grad_norm": 0.2734524607658386,
      "learning_rate": 0.01,
      "loss": 4.555187225341797,
      "num_input_tokens_seen": 6793416,
      "step": 1711,
      "train_runtime": 1391.8522,
      "train_tokens_per_second": 4880.846
    },
    {
      "epoch": 131.6153846153846,
      "num_input_tokens_seen": 6793416,
      "step": 1711,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 131.69230769230768,
      "grad_norm": 0.4041428864002228,
      "learning_rate": 0.01,
      "loss": 4.540947437286377,
      "num_input_tokens_seen": 6797512,
      "step": 1712,
      "train_runtime": 1392.6674,
      "train_tokens_per_second": 4880.93
    },
    {
      "epoch": 131.69230769230768,
      "num_input_tokens_seen": 6797512,
      "step": 1712,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 131.76923076923077,
      "grad_norm": 1.1585386991500854,
      "learning_rate": 0.01,
      "loss": 4.52167272567749,
      "num_input_tokens_seen": 6801608,
      "step": 1713,
      "train_runtime": 1393.4821,
      "train_tokens_per_second": 4881.016
    },
    {
      "epoch": 131.76923076923077,
      "num_input_tokens_seen": 6801608,
      "step": 1713,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 131.84615384615384,
      "grad_norm": 1.229591727256775,
      "learning_rate": 0.01,
      "loss": 4.545475006103516,
      "num_input_tokens_seen": 6805704,
      "step": 1714,
      "train_runtime": 1394.2963,
      "train_tokens_per_second": 4881.103
    },
    {
      "epoch": 131.84615384615384,
      "num_input_tokens_seen": 6805704,
      "step": 1714,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 131.92307692307693,
      "grad_norm": 0.4838937819004059,
      "learning_rate": 0.01,
      "loss": 4.523252487182617,
      "num_input_tokens_seen": 6809800,
      "step": 1715,
      "train_runtime": 1395.107,
      "train_tokens_per_second": 4881.203
    },
    {
      "epoch": 131.92307692307693,
      "num_input_tokens_seen": 6809800,
      "step": 1715,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 132.0,
      "grad_norm": 0.3217167258262634,
      "learning_rate": 0.01,
      "loss": 4.476624488830566,
      "num_input_tokens_seen": 6812256,
      "step": 1716,
      "train_runtime": 1395.6229,
      "train_tokens_per_second": 4881.158
    },
    {
      "epoch": 132.0,
      "num_input_tokens_seen": 6812256,
      "step": 1716,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 132.07692307692307,
      "grad_norm": 0.3933837115764618,
      "learning_rate": 0.01,
      "loss": 4.501726150512695,
      "num_input_tokens_seen": 6816352,
      "step": 1717,
      "train_runtime": 1396.4515,
      "train_tokens_per_second": 4881.195
    },
    {
      "epoch": 132.07692307692307,
      "num_input_tokens_seen": 6816352,
      "step": 1717,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 132.15384615384616,
      "grad_norm": 0.74774569272995,
      "learning_rate": 0.01,
      "loss": 4.478212356567383,
      "num_input_tokens_seen": 6820448,
      "step": 1718,
      "train_runtime": 1397.2654,
      "train_tokens_per_second": 4881.283
    },
    {
      "epoch": 132.15384615384616,
      "num_input_tokens_seen": 6820448,
      "step": 1718,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 132.23076923076923,
      "grad_norm": 0.578484058380127,
      "learning_rate": 0.01,
      "loss": 4.514402389526367,
      "num_input_tokens_seen": 6824544,
      "step": 1719,
      "train_runtime": 1398.0798,
      "train_tokens_per_second": 4881.37
    },
    {
      "epoch": 132.23076923076923,
      "num_input_tokens_seen": 6824544,
      "step": 1719,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 132.30769230769232,
      "grad_norm": 0.7361067533493042,
      "learning_rate": 0.01,
      "loss": 4.499946594238281,
      "num_input_tokens_seen": 6828640,
      "step": 1720,
      "train_runtime": 1398.8941,
      "train_tokens_per_second": 4881.456
    },
    {
      "epoch": 132.30769230769232,
      "num_input_tokens_seen": 6828640,
      "step": 1720,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 132.3846153846154,
      "grad_norm": 1.0104666948318481,
      "learning_rate": 0.01,
      "loss": 4.525514125823975,
      "num_input_tokens_seen": 6832736,
      "step": 1721,
      "train_runtime": 1399.7085,
      "train_tokens_per_second": 4881.542
    },
    {
      "epoch": 132.3846153846154,
      "num_input_tokens_seen": 6832736,
      "step": 1721,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 132.46153846153845,
      "grad_norm": 0.2834595739841461,
      "learning_rate": 0.01,
      "loss": 4.531540870666504,
      "num_input_tokens_seen": 6836832,
      "step": 1722,
      "train_runtime": 1400.5231,
      "train_tokens_per_second": 4881.627
    },
    {
      "epoch": 132.46153846153845,
      "num_input_tokens_seen": 6836832,
      "step": 1722,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 132.53846153846155,
      "grad_norm": 0.5259590744972229,
      "learning_rate": 0.01,
      "loss": 4.53350830078125,
      "num_input_tokens_seen": 6840928,
      "step": 1723,
      "train_runtime": 1401.3372,
      "train_tokens_per_second": 4881.714
    },
    {
      "epoch": 132.53846153846155,
      "num_input_tokens_seen": 6840928,
      "step": 1723,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 132.6153846153846,
      "grad_norm": 0.33686336874961853,
      "learning_rate": 0.01,
      "loss": 4.540597438812256,
      "num_input_tokens_seen": 6845024,
      "step": 1724,
      "train_runtime": 1402.1517,
      "train_tokens_per_second": 4881.8
    },
    {
      "epoch": 132.6153846153846,
      "num_input_tokens_seen": 6845024,
      "step": 1724,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 132.69230769230768,
      "grad_norm": 1.2727960348129272,
      "learning_rate": 0.01,
      "loss": 4.526486396789551,
      "num_input_tokens_seen": 6849120,
      "step": 1725,
      "train_runtime": 1402.9655,
      "train_tokens_per_second": 4881.888
    },
    {
      "epoch": 132.69230769230768,
      "num_input_tokens_seen": 6849120,
      "step": 1725,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 132.76923076923077,
      "grad_norm": 0.23274320363998413,
      "learning_rate": 0.01,
      "loss": 4.526761531829834,
      "num_input_tokens_seen": 6853216,
      "step": 1726,
      "train_runtime": 1403.7797,
      "train_tokens_per_second": 4881.974
    },
    {
      "epoch": 132.76923076923077,
      "num_input_tokens_seen": 6853216,
      "step": 1726,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 132.84615384615384,
      "grad_norm": 0.22988763451576233,
      "learning_rate": 0.01,
      "loss": 4.542788505554199,
      "num_input_tokens_seen": 6857312,
      "step": 1727,
      "train_runtime": 1404.5943,
      "train_tokens_per_second": 4882.059
    },
    {
      "epoch": 132.84615384615384,
      "num_input_tokens_seen": 6857312,
      "step": 1727,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 132.92307692307693,
      "grad_norm": 0.47999700903892517,
      "learning_rate": 0.01,
      "loss": 4.517338752746582,
      "num_input_tokens_seen": 6861408,
      "step": 1728,
      "train_runtime": 1405.404,
      "train_tokens_per_second": 4882.16
    },
    {
      "epoch": 132.92307692307693,
      "num_input_tokens_seen": 6861408,
      "step": 1728,
      "train_accuracy": 0.019543973729014397
    },
    {
      "epoch": 133.0,
      "grad_norm": 0.39151906967163086,
      "learning_rate": 0.01,
      "loss": 4.529682636260986,
      "num_input_tokens_seen": 6863864,
      "step": 1729,
      "train_runtime": 1405.9191,
      "train_tokens_per_second": 4882.119
    },
    {
      "epoch": 133.0,
      "num_input_tokens_seen": 6863864,
      "step": 1729,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 133.07692307692307,
      "grad_norm": 0.6525077223777771,
      "learning_rate": 0.01,
      "loss": 4.528639316558838,
      "num_input_tokens_seen": 6867960,
      "step": 1730,
      "train_runtime": 1406.7438,
      "train_tokens_per_second": 4882.168
    },
    {
      "epoch": 133.07692307692307,
      "num_input_tokens_seen": 6867960,
      "step": 1730,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 133.15384615384616,
      "grad_norm": 0.3770185708999634,
      "learning_rate": 0.01,
      "loss": 4.490137100219727,
      "num_input_tokens_seen": 6872056,
      "step": 1731,
      "train_runtime": 1407.5578,
      "train_tokens_per_second": 4882.255
    },
    {
      "epoch": 133.15384615384616,
      "num_input_tokens_seen": 6872056,
      "step": 1731,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 133.23076923076923,
      "grad_norm": 0.49798017740249634,
      "learning_rate": 0.01,
      "loss": 4.516364574432373,
      "num_input_tokens_seen": 6876152,
      "step": 1732,
      "train_runtime": 1408.373,
      "train_tokens_per_second": 4882.337
    },
    {
      "epoch": 133.23076923076923,
      "num_input_tokens_seen": 6876152,
      "step": 1732,
      "train_accuracy": 0.03125
    },
    {
      "epoch": 133.30769230769232,
      "grad_norm": 0.2344183474779129,
      "learning_rate": 0.01,
      "loss": 4.481263637542725,
      "num_input_tokens_seen": 6880248,
      "step": 1733,
      "train_runtime": 1409.1878,
      "train_tokens_per_second": 4882.421
    },
    {
      "epoch": 133.30769230769232,
      "num_input_tokens_seen": 6880248,
      "step": 1733,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 133.3846153846154,
      "grad_norm": 0.3261810839176178,
      "learning_rate": 0.01,
      "loss": 4.528553009033203,
      "num_input_tokens_seen": 6884344,
      "step": 1734,
      "train_runtime": 1410.002,
      "train_tokens_per_second": 4882.507
    },
    {
      "epoch": 133.3846153846154,
      "num_input_tokens_seen": 6884344,
      "step": 1734,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 133.46153846153845,
      "grad_norm": 0.25679534673690796,
      "learning_rate": 0.01,
      "loss": 4.499446868896484,
      "num_input_tokens_seen": 6888440,
      "step": 1735,
      "train_runtime": 1410.8196,
      "train_tokens_per_second": 4882.58
    },
    {
      "epoch": 133.46153846153845,
      "num_input_tokens_seen": 6888440,
      "step": 1735,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 133.53846153846155,
      "grad_norm": 0.2007618248462677,
      "learning_rate": 0.01,
      "loss": 4.516724586486816,
      "num_input_tokens_seen": 6892536,
      "step": 1736,
      "train_runtime": 1411.6327,
      "train_tokens_per_second": 4882.67
    },
    {
      "epoch": 133.53846153846155,
      "num_input_tokens_seen": 6892536,
      "step": 1736,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 133.6153846153846,
      "grad_norm": 0.6148881316184998,
      "learning_rate": 0.01,
      "loss": 4.537455081939697,
      "num_input_tokens_seen": 6896632,
      "step": 1737,
      "train_runtime": 1412.4481,
      "train_tokens_per_second": 4882.751
    },
    {
      "epoch": 133.6153846153846,
      "num_input_tokens_seen": 6896632,
      "step": 1737,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 133.69230769230768,
      "grad_norm": 0.28915512561798096,
      "learning_rate": 0.01,
      "loss": 4.51920223236084,
      "num_input_tokens_seen": 6900728,
      "step": 1738,
      "train_runtime": 1413.2709,
      "train_tokens_per_second": 4882.806
    },
    {
      "epoch": 133.69230769230768,
      "num_input_tokens_seen": 6900728,
      "step": 1738,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 133.76923076923077,
      "grad_norm": 0.2960025370121002,
      "learning_rate": 0.01,
      "loss": 4.52116584777832,
      "num_input_tokens_seen": 6904824,
      "step": 1739,
      "train_runtime": 1414.0848,
      "train_tokens_per_second": 4882.892
    },
    {
      "epoch": 133.76923076923077,
      "num_input_tokens_seen": 6904824,
      "step": 1739,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 133.84615384615384,
      "grad_norm": 0.27149006724357605,
      "learning_rate": 0.01,
      "loss": 4.521803855895996,
      "num_input_tokens_seen": 6908920,
      "step": 1740,
      "train_runtime": 1414.899,
      "train_tokens_per_second": 4882.977
    },
    {
      "epoch": 133.84615384615384,
      "num_input_tokens_seen": 6908920,
      "step": 1740,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 133.92307692307693,
      "grad_norm": 0.26008468866348267,
      "learning_rate": 0.01,
      "loss": 4.5001749992370605,
      "num_input_tokens_seen": 6913016,
      "step": 1741,
      "train_runtime": 1415.7098,
      "train_tokens_per_second": 4883.074
    },
    {
      "epoch": 133.92307692307693,
      "num_input_tokens_seen": 6913016,
      "step": 1741,
      "train_accuracy": 0.0065146577544510365
    },
    {
      "epoch": 134.0,
      "grad_norm": 0.32249686121940613,
      "learning_rate": 0.01,
      "loss": 4.509130954742432,
      "num_input_tokens_seen": 6915472,
      "step": 1742,
      "train_runtime": 1416.2248,
      "train_tokens_per_second": 4883.033
    },
    {
      "epoch": 134.0,
      "num_input_tokens_seen": 6915472,
      "step": 1742,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 134.07692307692307,
      "grad_norm": 2.5880963802337646,
      "learning_rate": 0.01,
      "loss": 4.508394718170166,
      "num_input_tokens_seen": 6919568,
      "step": 1743,
      "train_runtime": 1417.0524,
      "train_tokens_per_second": 4883.071
    },
    {
      "epoch": 134.07692307692307,
      "num_input_tokens_seen": 6919568,
      "step": 1743,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 134.15384615384616,
      "grad_norm": 5.311941146850586,
      "learning_rate": 0.01,
      "loss": 4.494456768035889,
      "num_input_tokens_seen": 6923664,
      "step": 1744,
      "train_runtime": 1417.8671,
      "train_tokens_per_second": 4883.154
    },
    {
      "epoch": 134.15384615384616,
      "num_input_tokens_seen": 6923664,
      "step": 1744,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 134.23076923076923,
      "grad_norm": 0.3410141170024872,
      "learning_rate": 0.01,
      "loss": 4.491735458374023,
      "num_input_tokens_seen": 6927760,
      "step": 1745,
      "train_runtime": 1418.6872,
      "train_tokens_per_second": 4883.219
    },
    {
      "epoch": 134.23076923076923,
      "num_input_tokens_seen": 6927760,
      "step": 1745,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 134.30769230769232,
      "grad_norm": 0.31180036067962646,
      "learning_rate": 0.01,
      "loss": 4.476294040679932,
      "num_input_tokens_seen": 6931856,
      "step": 1746,
      "train_runtime": 1419.5048,
      "train_tokens_per_second": 4883.292
    },
    {
      "epoch": 134.30769230769232,
      "num_input_tokens_seen": 6931856,
      "step": 1746,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 134.3846153846154,
      "grad_norm": 0.7802852988243103,
      "learning_rate": 0.01,
      "loss": 4.52705192565918,
      "num_input_tokens_seen": 6935952,
      "step": 1747,
      "train_runtime": 1420.3185,
      "train_tokens_per_second": 4883.378
    },
    {
      "epoch": 134.3846153846154,
      "num_input_tokens_seen": 6935952,
      "step": 1747,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 134.46153846153845,
      "grad_norm": 0.41046053171157837,
      "learning_rate": 0.01,
      "loss": 4.528984546661377,
      "num_input_tokens_seen": 6940048,
      "step": 1748,
      "train_runtime": 1421.1322,
      "train_tokens_per_second": 4883.464
    },
    {
      "epoch": 134.46153846153845,
      "num_input_tokens_seen": 6940048,
      "step": 1748,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 134.53846153846155,
      "grad_norm": 1.0420804023742676,
      "learning_rate": 0.01,
      "loss": 4.518743515014648,
      "num_input_tokens_seen": 6944144,
      "step": 1749,
      "train_runtime": 1421.9475,
      "train_tokens_per_second": 4883.545
    },
    {
      "epoch": 134.53846153846155,
      "num_input_tokens_seen": 6944144,
      "step": 1749,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 134.6153846153846,
      "grad_norm": 0.2821536064147949,
      "learning_rate": 0.01,
      "loss": 4.530234336853027,
      "num_input_tokens_seen": 6948240,
      "step": 1750,
      "train_runtime": 1422.7625,
      "train_tokens_per_second": 4883.626
    },
    {
      "epoch": 134.6153846153846,
      "num_input_tokens_seen": 6948240,
      "step": 1750,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 134.69230769230768,
      "grad_norm": 0.5711707472801208,
      "learning_rate": 0.01,
      "loss": 4.520828723907471,
      "num_input_tokens_seen": 6952336,
      "step": 1751,
      "train_runtime": 1423.5769,
      "train_tokens_per_second": 4883.71
    },
    {
      "epoch": 134.69230769230768,
      "num_input_tokens_seen": 6952336,
      "step": 1751,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 134.76923076923077,
      "grad_norm": 0.22541077435016632,
      "learning_rate": 0.01,
      "loss": 4.508856296539307,
      "num_input_tokens_seen": 6956432,
      "step": 1752,
      "train_runtime": 1424.3914,
      "train_tokens_per_second": 4883.792
    },
    {
      "epoch": 134.76923076923077,
      "num_input_tokens_seen": 6956432,
      "step": 1752,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 134.84615384615384,
      "grad_norm": 0.2409822940826416,
      "learning_rate": 0.01,
      "loss": 4.5196943283081055,
      "num_input_tokens_seen": 6960528,
      "step": 1753,
      "train_runtime": 1425.2052,
      "train_tokens_per_second": 4883.878
    },
    {
      "epoch": 134.84615384615384,
      "num_input_tokens_seen": 6960528,
      "step": 1753,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 134.92307692307693,
      "grad_norm": 0.30738699436187744,
      "learning_rate": 0.01,
      "loss": 4.521076202392578,
      "num_input_tokens_seen": 6964624,
      "step": 1754,
      "train_runtime": 1426.0142,
      "train_tokens_per_second": 4883.979
    },
    {
      "epoch": 134.92307692307693,
      "num_input_tokens_seen": 6964624,
      "step": 1754,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 135.0,
      "grad_norm": 0.3633999824523926,
      "learning_rate": 0.01,
      "loss": 4.502272605895996,
      "num_input_tokens_seen": 6967080,
      "step": 1755,
      "train_runtime": 1426.5293,
      "train_tokens_per_second": 4883.938
    },
    {
      "epoch": 135.0,
      "num_input_tokens_seen": 6967080,
      "step": 1755,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 135.07692307692307,
      "grad_norm": 0.45537644624710083,
      "learning_rate": 0.01,
      "loss": 4.493215560913086,
      "num_input_tokens_seen": 6971176,
      "step": 1756,
      "train_runtime": 1427.364,
      "train_tokens_per_second": 4883.951
    },
    {
      "epoch": 135.07692307692307,
      "num_input_tokens_seen": 6971176,
      "step": 1756,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 135.15384615384616,
      "grad_norm": 0.909066379070282,
      "learning_rate": 0.01,
      "loss": 4.514348030090332,
      "num_input_tokens_seen": 6975272,
      "step": 1757,
      "train_runtime": 1428.1779,
      "train_tokens_per_second": 4884.036
    },
    {
      "epoch": 135.15384615384616,
      "num_input_tokens_seen": 6975272,
      "step": 1757,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 135.23076923076923,
      "grad_norm": 0.3726130723953247,
      "learning_rate": 0.01,
      "loss": 4.503530502319336,
      "num_input_tokens_seen": 6979368,
      "step": 1758,
      "train_runtime": 1428.9931,
      "train_tokens_per_second": 4884.116
    },
    {
      "epoch": 135.23076923076923,
      "num_input_tokens_seen": 6979368,
      "step": 1758,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 135.30769230769232,
      "grad_norm": 0.30177444219589233,
      "learning_rate": 0.01,
      "loss": 4.507753372192383,
      "num_input_tokens_seen": 6983464,
      "step": 1759,
      "train_runtime": 1429.8073,
      "train_tokens_per_second": 4884.199
    },
    {
      "epoch": 135.30769230769232,
      "num_input_tokens_seen": 6983464,
      "step": 1759,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 135.3846153846154,
      "grad_norm": 0.39806854724884033,
      "learning_rate": 0.01,
      "loss": 4.487634658813477,
      "num_input_tokens_seen": 6987560,
      "step": 1760,
      "train_runtime": 1430.621,
      "train_tokens_per_second": 4884.285
    },
    {
      "epoch": 135.3846153846154,
      "num_input_tokens_seen": 6987560,
      "step": 1760,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 135.46153846153845,
      "grad_norm": 0.353312611579895,
      "learning_rate": 0.01,
      "loss": 4.508077621459961,
      "num_input_tokens_seen": 6991656,
      "step": 1761,
      "train_runtime": 1431.4341,
      "train_tokens_per_second": 4884.371
    },
    {
      "epoch": 135.46153846153845,
      "num_input_tokens_seen": 6991656,
      "step": 1761,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 135.53846153846155,
      "grad_norm": 0.37810611724853516,
      "learning_rate": 0.01,
      "loss": 4.5164103507995605,
      "num_input_tokens_seen": 6995752,
      "step": 1762,
      "train_runtime": 1432.2494,
      "train_tokens_per_second": 4884.451
    },
    {
      "epoch": 135.53846153846155,
      "num_input_tokens_seen": 6995752,
      "step": 1762,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 135.6153846153846,
      "grad_norm": 0.2685927152633667,
      "learning_rate": 0.01,
      "loss": 4.478910446166992,
      "num_input_tokens_seen": 6999848,
      "step": 1763,
      "train_runtime": 1433.0636,
      "train_tokens_per_second": 4884.534
    },
    {
      "epoch": 135.6153846153846,
      "num_input_tokens_seen": 6999848,
      "step": 1763,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 135.69230769230768,
      "grad_norm": 0.48998892307281494,
      "learning_rate": 0.01,
      "loss": 4.526169300079346,
      "num_input_tokens_seen": 7003944,
      "step": 1764,
      "train_runtime": 1433.8791,
      "train_tokens_per_second": 4884.613
    },
    {
      "epoch": 135.69230769230768,
      "num_input_tokens_seen": 7003944,
      "step": 1764,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 135.76923076923077,
      "grad_norm": 0.24435654282569885,
      "learning_rate": 0.01,
      "loss": 4.500778675079346,
      "num_input_tokens_seen": 7008040,
      "step": 1765,
      "train_runtime": 1434.6929,
      "train_tokens_per_second": 4884.697
    },
    {
      "epoch": 135.76923076923077,
      "num_input_tokens_seen": 7008040,
      "step": 1765,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 135.84615384615384,
      "grad_norm": 0.5288189053535461,
      "learning_rate": 0.01,
      "loss": 4.510227203369141,
      "num_input_tokens_seen": 7012136,
      "step": 1766,
      "train_runtime": 1435.507,
      "train_tokens_per_second": 4884.78
    },
    {
      "epoch": 135.84615384615384,
      "num_input_tokens_seen": 7012136,
      "step": 1766,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 135.92307692307693,
      "grad_norm": 0.6351887583732605,
      "learning_rate": 0.01,
      "loss": 4.50391149520874,
      "num_input_tokens_seen": 7016232,
      "step": 1767,
      "train_runtime": 1436.3172,
      "train_tokens_per_second": 4884.877
    },
    {
      "epoch": 135.92307692307693,
      "num_input_tokens_seen": 7016232,
      "step": 1767,
      "train_accuracy": 0.04560260474681854
    },
    {
      "epoch": 136.0,
      "grad_norm": 0.6438477635383606,
      "learning_rate": 0.01,
      "loss": 4.502982139587402,
      "num_input_tokens_seen": 7018688,
      "step": 1768,
      "train_runtime": 1436.8311,
      "train_tokens_per_second": 4884.839
    },
    {
      "epoch": 136.0,
      "num_input_tokens_seen": 7018688,
      "step": 1768,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 136.07692307692307,
      "grad_norm": 0.27569079399108887,
      "learning_rate": 0.01,
      "loss": 4.499819755554199,
      "num_input_tokens_seen": 7022784,
      "step": 1769,
      "train_runtime": 1437.6571,
      "train_tokens_per_second": 4884.881
    },
    {
      "epoch": 136.07692307692307,
      "num_input_tokens_seen": 7022784,
      "step": 1769,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 136.15384615384616,
      "grad_norm": 0.5427327156066895,
      "learning_rate": 0.01,
      "loss": 4.490644454956055,
      "num_input_tokens_seen": 7026880,
      "step": 1770,
      "train_runtime": 1438.4713,
      "train_tokens_per_second": 4884.964
    },
    {
      "epoch": 136.15384615384616,
      "num_input_tokens_seen": 7026880,
      "step": 1770,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 136.23076923076923,
      "grad_norm": 0.2863885462284088,
      "learning_rate": 0.01,
      "loss": 4.516340732574463,
      "num_input_tokens_seen": 7030976,
      "step": 1771,
      "train_runtime": 1439.2844,
      "train_tokens_per_second": 4885.05
    },
    {
      "epoch": 136.23076923076923,
      "num_input_tokens_seen": 7030976,
      "step": 1771,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 136.30769230769232,
      "grad_norm": 0.1990378051996231,
      "learning_rate": 0.01,
      "loss": 4.494849681854248,
      "num_input_tokens_seen": 7035072,
      "step": 1772,
      "train_runtime": 1440.0979,
      "train_tokens_per_second": 4885.135
    },
    {
      "epoch": 136.30769230769232,
      "num_input_tokens_seen": 7035072,
      "step": 1772,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 136.3846153846154,
      "grad_norm": 0.2878986597061157,
      "learning_rate": 0.01,
      "loss": 4.487600803375244,
      "num_input_tokens_seen": 7039168,
      "step": 1773,
      "train_runtime": 1440.9111,
      "train_tokens_per_second": 4885.22
    },
    {
      "epoch": 136.3846153846154,
      "num_input_tokens_seen": 7039168,
      "step": 1773,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 136.46153846153845,
      "grad_norm": 0.3415122330188751,
      "learning_rate": 0.01,
      "loss": 4.510580539703369,
      "num_input_tokens_seen": 7043264,
      "step": 1774,
      "train_runtime": 1441.7254,
      "train_tokens_per_second": 4885.302
    },
    {
      "epoch": 136.46153846153845,
      "num_input_tokens_seen": 7043264,
      "step": 1774,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 136.53846153846155,
      "grad_norm": 0.2500530183315277,
      "learning_rate": 0.01,
      "loss": 4.51226282119751,
      "num_input_tokens_seen": 7047360,
      "step": 1775,
      "train_runtime": 1442.5399,
      "train_tokens_per_second": 4885.383
    },
    {
      "epoch": 136.53846153846155,
      "num_input_tokens_seen": 7047360,
      "step": 1775,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 136.6153846153846,
      "grad_norm": 0.31919777393341064,
      "learning_rate": 0.01,
      "loss": 4.492679595947266,
      "num_input_tokens_seen": 7051456,
      "step": 1776,
      "train_runtime": 1443.354,
      "train_tokens_per_second": 4885.465
    },
    {
      "epoch": 136.6153846153846,
      "num_input_tokens_seen": 7051456,
      "step": 1776,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 136.69230769230768,
      "grad_norm": 1.1813175678253174,
      "learning_rate": 0.01,
      "loss": 4.50545597076416,
      "num_input_tokens_seen": 7055552,
      "step": 1777,
      "train_runtime": 1444.1677,
      "train_tokens_per_second": 4885.549
    },
    {
      "epoch": 136.69230769230768,
      "num_input_tokens_seen": 7055552,
      "step": 1777,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 136.76923076923077,
      "grad_norm": 0.2754706144332886,
      "learning_rate": 0.01,
      "loss": 4.506965160369873,
      "num_input_tokens_seen": 7059648,
      "step": 1778,
      "train_runtime": 1444.9817,
      "train_tokens_per_second": 4885.631
    },
    {
      "epoch": 136.76923076923077,
      "num_input_tokens_seen": 7059648,
      "step": 1778,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 136.84615384615384,
      "grad_norm": 0.22163443267345428,
      "learning_rate": 0.01,
      "loss": 4.506715297698975,
      "num_input_tokens_seen": 7063744,
      "step": 1779,
      "train_runtime": 1445.7946,
      "train_tokens_per_second": 4885.718
    },
    {
      "epoch": 136.84615384615384,
      "num_input_tokens_seen": 7063744,
      "step": 1779,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 136.92307692307693,
      "grad_norm": 0.2083691954612732,
      "learning_rate": 0.01,
      "loss": 4.498074531555176,
      "num_input_tokens_seen": 7067840,
      "step": 1780,
      "train_runtime": 1446.6031,
      "train_tokens_per_second": 4885.819
    },
    {
      "epoch": 136.92307692307693,
      "num_input_tokens_seen": 7067840,
      "step": 1780,
      "train_accuracy": 0.0065146577544510365
    },
    {
      "epoch": 137.0,
      "grad_norm": 0.557692289352417,
      "learning_rate": 0.01,
      "loss": 4.558379173278809,
      "num_input_tokens_seen": 7070296,
      "step": 1781,
      "train_runtime": 1447.1173,
      "train_tokens_per_second": 4885.779
    },
    {
      "epoch": 137.0,
      "num_input_tokens_seen": 7070296,
      "step": 1781,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 137.07692307692307,
      "grad_norm": 0.31935107707977295,
      "learning_rate": 0.01,
      "loss": 4.525882244110107,
      "num_input_tokens_seen": 7074392,
      "step": 1782,
      "train_runtime": 1447.9424,
      "train_tokens_per_second": 4885.824
    },
    {
      "epoch": 137.07692307692307,
      "num_input_tokens_seen": 7074392,
      "step": 1782,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 137.15384615384616,
      "grad_norm": 0.675654947757721,
      "learning_rate": 0.01,
      "loss": 4.525753974914551,
      "num_input_tokens_seen": 7078488,
      "step": 1783,
      "train_runtime": 1448.7568,
      "train_tokens_per_second": 4885.905
    },
    {
      "epoch": 137.15384615384616,
      "num_input_tokens_seen": 7078488,
      "step": 1783,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 137.23076923076923,
      "grad_norm": 1.5541125535964966,
      "learning_rate": 0.01,
      "loss": 4.4922380447387695,
      "num_input_tokens_seen": 7082584,
      "step": 1784,
      "train_runtime": 1449.5691,
      "train_tokens_per_second": 4885.993
    },
    {
      "epoch": 137.23076923076923,
      "num_input_tokens_seen": 7082584,
      "step": 1784,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 137.30769230769232,
      "grad_norm": 0.3508233428001404,
      "learning_rate": 0.01,
      "loss": 4.518599987030029,
      "num_input_tokens_seen": 7086680,
      "step": 1785,
      "train_runtime": 1450.3817,
      "train_tokens_per_second": 4886.079
    },
    {
      "epoch": 137.30769230769232,
      "num_input_tokens_seen": 7086680,
      "step": 1785,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 137.3846153846154,
      "grad_norm": 0.20834478735923767,
      "learning_rate": 0.01,
      "loss": 4.44829797744751,
      "num_input_tokens_seen": 7090776,
      "step": 1786,
      "train_runtime": 1451.1961,
      "train_tokens_per_second": 4886.16
    },
    {
      "epoch": 137.3846153846154,
      "num_input_tokens_seen": 7090776,
      "step": 1786,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 137.46153846153845,
      "grad_norm": 1.0349732637405396,
      "learning_rate": 0.01,
      "loss": 4.531538009643555,
      "num_input_tokens_seen": 7094872,
      "step": 1787,
      "train_runtime": 1452.0082,
      "train_tokens_per_second": 4886.248
    },
    {
      "epoch": 137.46153846153845,
      "num_input_tokens_seen": 7094872,
      "step": 1787,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 137.53846153846155,
      "grad_norm": 0.3144688308238983,
      "learning_rate": 0.01,
      "loss": 4.505922317504883,
      "num_input_tokens_seen": 7098968,
      "step": 1788,
      "train_runtime": 1452.8218,
      "train_tokens_per_second": 4886.331
    },
    {
      "epoch": 137.53846153846155,
      "num_input_tokens_seen": 7098968,
      "step": 1788,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 137.6153846153846,
      "grad_norm": 0.3917436897754669,
      "learning_rate": 0.01,
      "loss": 4.52239990234375,
      "num_input_tokens_seen": 7103064,
      "step": 1789,
      "train_runtime": 1453.6347,
      "train_tokens_per_second": 4886.416
    },
    {
      "epoch": 137.6153846153846,
      "num_input_tokens_seen": 7103064,
      "step": 1789,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 137.69230769230768,
      "grad_norm": 0.5549172163009644,
      "learning_rate": 0.01,
      "loss": 4.514933109283447,
      "num_input_tokens_seen": 7107160,
      "step": 1790,
      "train_runtime": 1454.4497,
      "train_tokens_per_second": 4886.494
    },
    {
      "epoch": 137.69230769230768,
      "num_input_tokens_seen": 7107160,
      "step": 1790,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 137.76923076923077,
      "grad_norm": 5.873458385467529,
      "learning_rate": 0.01,
      "loss": 4.490239143371582,
      "num_input_tokens_seen": 7111256,
      "step": 1791,
      "train_runtime": 1455.2627,
      "train_tokens_per_second": 4886.579
    },
    {
      "epoch": 137.76923076923077,
      "num_input_tokens_seen": 7111256,
      "step": 1791,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 137.84615384615384,
      "grad_norm": 0.34059423208236694,
      "learning_rate": 0.01,
      "loss": 4.488921642303467,
      "num_input_tokens_seen": 7115352,
      "step": 1792,
      "train_runtime": 1456.0767,
      "train_tokens_per_second": 4886.66
    },
    {
      "epoch": 137.84615384615384,
      "num_input_tokens_seen": 7115352,
      "step": 1792,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 137.92307692307693,
      "grad_norm": 0.2996807396411896,
      "learning_rate": 0.01,
      "loss": 4.529014587402344,
      "num_input_tokens_seen": 7119448,
      "step": 1793,
      "train_runtime": 1456.8846,
      "train_tokens_per_second": 4886.762
    },
    {
      "epoch": 137.92307692307693,
      "num_input_tokens_seen": 7119448,
      "step": 1793,
      "train_accuracy": 0.013029315508902073
    },
    {
      "epoch": 138.0,
      "grad_norm": 0.34790003299713135,
      "learning_rate": 0.01,
      "loss": 4.576231479644775,
      "num_input_tokens_seen": 7121904,
      "step": 1794,
      "train_runtime": 1457.3992,
      "train_tokens_per_second": 4886.721
    },
    {
      "epoch": 138.0,
      "num_input_tokens_seen": 7121904,
      "step": 1794,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 138.07692307692307,
      "grad_norm": 0.42401421070098877,
      "learning_rate": 0.01,
      "loss": 4.517054557800293,
      "num_input_tokens_seen": 7126000,
      "step": 1795,
      "train_runtime": 1458.2251,
      "train_tokens_per_second": 4886.763
    },
    {
      "epoch": 138.07692307692307,
      "num_input_tokens_seen": 7126000,
      "step": 1795,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 138.15384615384616,
      "grad_norm": 0.24705839157104492,
      "learning_rate": 0.01,
      "loss": 4.490913391113281,
      "num_input_tokens_seen": 7130096,
      "step": 1796,
      "train_runtime": 1459.0385,
      "train_tokens_per_second": 4886.846
    },
    {
      "epoch": 138.15384615384616,
      "num_input_tokens_seen": 7130096,
      "step": 1796,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 138.23076923076923,
      "grad_norm": 0.40457087755203247,
      "learning_rate": 0.01,
      "loss": 4.516879081726074,
      "num_input_tokens_seen": 7134192,
      "step": 1797,
      "train_runtime": 1459.8552,
      "train_tokens_per_second": 4886.918
    },
    {
      "epoch": 138.23076923076923,
      "num_input_tokens_seen": 7134192,
      "step": 1797,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 138.30769230769232,
      "grad_norm": 0.22481785714626312,
      "learning_rate": 0.01,
      "loss": 4.499124050140381,
      "num_input_tokens_seen": 7138288,
      "step": 1798,
      "train_runtime": 1460.6694,
      "train_tokens_per_second": 4886.998
    },
    {
      "epoch": 138.30769230769232,
      "num_input_tokens_seen": 7138288,
      "step": 1798,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 138.3846153846154,
      "grad_norm": 0.24027518928050995,
      "learning_rate": 0.01,
      "loss": 4.458518028259277,
      "num_input_tokens_seen": 7142384,
      "step": 1799,
      "train_runtime": 1461.482,
      "train_tokens_per_second": 4887.083
    },
    {
      "epoch": 138.3846153846154,
      "num_input_tokens_seen": 7142384,
      "step": 1799,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 138.46153846153845,
      "grad_norm": 0.7147669792175293,
      "learning_rate": 0.01,
      "loss": 4.481321334838867,
      "num_input_tokens_seen": 7146480,
      "step": 1800,
      "train_runtime": 1462.2949,
      "train_tokens_per_second": 4887.167
    },
    {
      "epoch": 138.46153846153845,
      "eval_CMD_3_branch_eval_accuracy": 0.0108499095840868,
      "eval_CMD_3_branch_eval_loss": 4.5855488777160645,
      "eval_CMD_3_branch_eval_runtime": 1.1503,
      "eval_CMD_3_branch_eval_samples_per_second": 2403.647,
      "eval_CMD_3_branch_eval_steps_per_second": 5.216,
      "num_input_tokens_seen": 7146480,
      "step": 1800
    },
    {
      "epoch": 138.46153846153845,
      "num_input_tokens_seen": 7146480,
      "step": 1800,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 138.53846153846155,
      "grad_norm": 0.4119090735912323,
      "learning_rate": 0.01,
      "loss": 4.541964054107666,
      "num_input_tokens_seen": 7150576,
      "step": 1801,
      "train_runtime": 1464.262,
      "train_tokens_per_second": 4883.399
    },
    {
      "epoch": 138.53846153846155,
      "num_input_tokens_seen": 7150576,
      "step": 1801,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 138.6153846153846,
      "grad_norm": 0.42758357524871826,
      "learning_rate": 0.01,
      "loss": 4.5123772621154785,
      "num_input_tokens_seen": 7154672,
      "step": 1802,
      "train_runtime": 1465.076,
      "train_tokens_per_second": 4883.482
    },
    {
      "epoch": 138.6153846153846,
      "num_input_tokens_seen": 7154672,
      "step": 1802,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 138.69230769230768,
      "grad_norm": 0.47022491693496704,
      "learning_rate": 0.01,
      "loss": 4.535153388977051,
      "num_input_tokens_seen": 7158768,
      "step": 1803,
      "train_runtime": 1465.8945,
      "train_tokens_per_second": 4883.549
    },
    {
      "epoch": 138.69230769230768,
      "num_input_tokens_seen": 7158768,
      "step": 1803,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 138.76923076923077,
      "grad_norm": 0.2832880914211273,
      "learning_rate": 0.01,
      "loss": 4.516087532043457,
      "num_input_tokens_seen": 7162864,
      "step": 1804,
      "train_runtime": 1466.7075,
      "train_tokens_per_second": 4883.635
    },
    {
      "epoch": 138.76923076923077,
      "num_input_tokens_seen": 7162864,
      "step": 1804,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 138.84615384615384,
      "grad_norm": 0.37242209911346436,
      "learning_rate": 0.01,
      "loss": 4.56517219543457,
      "num_input_tokens_seen": 7166960,
      "step": 1805,
      "train_runtime": 1467.5209,
      "train_tokens_per_second": 4883.719
    },
    {
      "epoch": 138.84615384615384,
      "num_input_tokens_seen": 7166960,
      "step": 1805,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 138.92307692307693,
      "grad_norm": 0.8365891575813293,
      "learning_rate": 0.01,
      "loss": 4.533134460449219,
      "num_input_tokens_seen": 7171056,
      "step": 1806,
      "train_runtime": 1468.3307,
      "train_tokens_per_second": 4883.815
    },
    {
      "epoch": 138.92307692307693,
      "num_input_tokens_seen": 7171056,
      "step": 1806,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 139.0,
      "grad_norm": 0.3855971693992615,
      "learning_rate": 0.01,
      "loss": 4.516770362854004,
      "num_input_tokens_seen": 7173512,
      "step": 1807,
      "train_runtime": 1468.8459,
      "train_tokens_per_second": 4883.774
    },
    {
      "epoch": 139.0,
      "num_input_tokens_seen": 7173512,
      "step": 1807,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 139.07692307692307,
      "grad_norm": 0.24986007809638977,
      "learning_rate": 0.01,
      "loss": 4.495937347412109,
      "num_input_tokens_seen": 7177608,
      "step": 1808,
      "train_runtime": 1469.6755,
      "train_tokens_per_second": 4883.805
    },
    {
      "epoch": 139.07692307692307,
      "num_input_tokens_seen": 7177608,
      "step": 1808,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 139.15384615384616,
      "grad_norm": 0.25552403926849365,
      "learning_rate": 0.01,
      "loss": 4.464908599853516,
      "num_input_tokens_seen": 7181704,
      "step": 1809,
      "train_runtime": 1470.4878,
      "train_tokens_per_second": 4883.892
    },
    {
      "epoch": 139.15384615384616,
      "num_input_tokens_seen": 7181704,
      "step": 1809,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 139.23076923076923,
      "grad_norm": 0.3772503137588501,
      "learning_rate": 0.01,
      "loss": 4.537103176116943,
      "num_input_tokens_seen": 7185800,
      "step": 1810,
      "train_runtime": 1471.3026,
      "train_tokens_per_second": 4883.971
    },
    {
      "epoch": 139.23076923076923,
      "num_input_tokens_seen": 7185800,
      "step": 1810,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 139.30769230769232,
      "grad_norm": 0.5765798091888428,
      "learning_rate": 0.01,
      "loss": 4.488556861877441,
      "num_input_tokens_seen": 7189896,
      "step": 1811,
      "train_runtime": 1472.117,
      "train_tokens_per_second": 4884.052
    },
    {
      "epoch": 139.30769230769232,
      "num_input_tokens_seen": 7189896,
      "step": 1811,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 139.3846153846154,
      "grad_norm": 0.5421313643455505,
      "learning_rate": 0.01,
      "loss": 4.506270885467529,
      "num_input_tokens_seen": 7193992,
      "step": 1812,
      "train_runtime": 1472.9323,
      "train_tokens_per_second": 4884.129
    },
    {
      "epoch": 139.3846153846154,
      "num_input_tokens_seen": 7193992,
      "step": 1812,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 139.46153846153845,
      "grad_norm": 0.34154364466667175,
      "learning_rate": 0.01,
      "loss": 4.489664077758789,
      "num_input_tokens_seen": 7198088,
      "step": 1813,
      "train_runtime": 1473.746,
      "train_tokens_per_second": 4884.212
    },
    {
      "epoch": 139.46153846153845,
      "num_input_tokens_seen": 7198088,
      "step": 1813,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 139.53846153846155,
      "grad_norm": 0.35041114687919617,
      "learning_rate": 0.01,
      "loss": 4.497374534606934,
      "num_input_tokens_seen": 7202184,
      "step": 1814,
      "train_runtime": 1474.5596,
      "train_tokens_per_second": 4884.295
    },
    {
      "epoch": 139.53846153846155,
      "num_input_tokens_seen": 7202184,
      "step": 1814,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 139.6153846153846,
      "grad_norm": 0.3379485309123993,
      "learning_rate": 0.01,
      "loss": 4.500453472137451,
      "num_input_tokens_seen": 7206280,
      "step": 1815,
      "train_runtime": 1475.3731,
      "train_tokens_per_second": 4884.378
    },
    {
      "epoch": 139.6153846153846,
      "num_input_tokens_seen": 7206280,
      "step": 1815,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 139.69230769230768,
      "grad_norm": 0.3296959400177002,
      "learning_rate": 0.01,
      "loss": 4.494850158691406,
      "num_input_tokens_seen": 7210376,
      "step": 1816,
      "train_runtime": 1476.1901,
      "train_tokens_per_second": 4884.449
    },
    {
      "epoch": 139.69230769230768,
      "num_input_tokens_seen": 7210376,
      "step": 1816,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 139.76923076923077,
      "grad_norm": 0.23183205723762512,
      "learning_rate": 0.01,
      "loss": 4.489829063415527,
      "num_input_tokens_seen": 7214472,
      "step": 1817,
      "train_runtime": 1477.0032,
      "train_tokens_per_second": 4884.534
    },
    {
      "epoch": 139.76923076923077,
      "num_input_tokens_seen": 7214472,
      "step": 1817,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 139.84615384615384,
      "grad_norm": 0.3803265392780304,
      "learning_rate": 0.01,
      "loss": 4.482579231262207,
      "num_input_tokens_seen": 7218568,
      "step": 1818,
      "train_runtime": 1477.8179,
      "train_tokens_per_second": 4884.613
    },
    {
      "epoch": 139.84615384615384,
      "num_input_tokens_seen": 7218568,
      "step": 1818,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 139.92307692307693,
      "grad_norm": 0.4867537021636963,
      "learning_rate": 0.01,
      "loss": 4.499177932739258,
      "num_input_tokens_seen": 7222664,
      "step": 1819,
      "train_runtime": 1478.627,
      "train_tokens_per_second": 4884.71
    },
    {
      "epoch": 139.92307692307693,
      "num_input_tokens_seen": 7222664,
      "step": 1819,
      "train_accuracy": 0.016286645084619522
    },
    {
      "epoch": 140.0,
      "grad_norm": 0.4574130177497864,
      "learning_rate": 0.01,
      "loss": 4.492845058441162,
      "num_input_tokens_seen": 7225120,
      "step": 1820,
      "train_runtime": 1479.1424,
      "train_tokens_per_second": 4884.668
    },
    {
      "epoch": 140.0,
      "num_input_tokens_seen": 7225120,
      "step": 1820,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 140.07692307692307,
      "grad_norm": 0.23928667604923248,
      "learning_rate": 0.01,
      "loss": 4.489180088043213,
      "num_input_tokens_seen": 7229216,
      "step": 1821,
      "train_runtime": 1479.9686,
      "train_tokens_per_second": 4884.709
    },
    {
      "epoch": 140.07692307692307,
      "num_input_tokens_seen": 7229216,
      "step": 1821,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 140.15384615384616,
      "grad_norm": 0.9181494116783142,
      "learning_rate": 0.01,
      "loss": 4.458590507507324,
      "num_input_tokens_seen": 7233312,
      "step": 1822,
      "train_runtime": 1480.7825,
      "train_tokens_per_second": 4884.79
    },
    {
      "epoch": 140.15384615384616,
      "num_input_tokens_seen": 7233312,
      "step": 1822,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 140.23076923076923,
      "grad_norm": 0.28248587250709534,
      "learning_rate": 0.01,
      "loss": 4.466012477874756,
      "num_input_tokens_seen": 7237408,
      "step": 1823,
      "train_runtime": 1481.5969,
      "train_tokens_per_second": 4884.87
    },
    {
      "epoch": 140.23076923076923,
      "num_input_tokens_seen": 7237408,
      "step": 1823,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 140.30769230769232,
      "grad_norm": 0.9355012774467468,
      "learning_rate": 0.01,
      "loss": 4.448394298553467,
      "num_input_tokens_seen": 7241504,
      "step": 1824,
      "train_runtime": 1482.4101,
      "train_tokens_per_second": 4884.953
    },
    {
      "epoch": 140.30769230769232,
      "num_input_tokens_seen": 7241504,
      "step": 1824,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 140.3846153846154,
      "grad_norm": 0.3071248233318329,
      "learning_rate": 0.01,
      "loss": 4.461729526519775,
      "num_input_tokens_seen": 7245600,
      "step": 1825,
      "train_runtime": 1483.224,
      "train_tokens_per_second": 4885.034
    },
    {
      "epoch": 140.3846153846154,
      "num_input_tokens_seen": 7245600,
      "step": 1825,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 140.46153846153845,
      "grad_norm": 0.470536470413208,
      "learning_rate": 0.01,
      "loss": 4.476644515991211,
      "num_input_tokens_seen": 7249696,
      "step": 1826,
      "train_runtime": 1484.0388,
      "train_tokens_per_second": 4885.112
    },
    {
      "epoch": 140.46153846153845,
      "num_input_tokens_seen": 7249696,
      "step": 1826,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 140.53846153846155,
      "grad_norm": 0.7882267236709595,
      "learning_rate": 0.01,
      "loss": 4.481664180755615,
      "num_input_tokens_seen": 7253792,
      "step": 1827,
      "train_runtime": 1484.8533,
      "train_tokens_per_second": 4885.191
    },
    {
      "epoch": 140.53846153846155,
      "num_input_tokens_seen": 7253792,
      "step": 1827,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 140.6153846153846,
      "grad_norm": 0.32621821761131287,
      "learning_rate": 0.01,
      "loss": 4.481396198272705,
      "num_input_tokens_seen": 7257888,
      "step": 1828,
      "train_runtime": 1485.6684,
      "train_tokens_per_second": 4885.268
    },
    {
      "epoch": 140.6153846153846,
      "num_input_tokens_seen": 7257888,
      "step": 1828,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 140.69230769230768,
      "grad_norm": 0.3387461006641388,
      "learning_rate": 0.01,
      "loss": 4.483463287353516,
      "num_input_tokens_seen": 7261984,
      "step": 1829,
      "train_runtime": 1486.4823,
      "train_tokens_per_second": 4885.348
    },
    {
      "epoch": 140.69230769230768,
      "num_input_tokens_seen": 7261984,
      "step": 1829,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 140.76923076923077,
      "grad_norm": 0.2903681993484497,
      "learning_rate": 0.01,
      "loss": 4.483816146850586,
      "num_input_tokens_seen": 7266080,
      "step": 1830,
      "train_runtime": 1487.2961,
      "train_tokens_per_second": 4885.429
    },
    {
      "epoch": 140.76923076923077,
      "num_input_tokens_seen": 7266080,
      "step": 1830,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 140.84615384615384,
      "grad_norm": 22.36224937438965,
      "learning_rate": 0.01,
      "loss": 4.496723175048828,
      "num_input_tokens_seen": 7270176,
      "step": 1831,
      "train_runtime": 1488.1108,
      "train_tokens_per_second": 4885.507
    },
    {
      "epoch": 140.84615384615384,
      "num_input_tokens_seen": 7270176,
      "step": 1831,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 140.92307692307693,
      "grad_norm": 0.5034341216087341,
      "learning_rate": 0.01,
      "loss": 4.457123279571533,
      "num_input_tokens_seen": 7274272,
      "step": 1832,
      "train_runtime": 1488.9209,
      "train_tokens_per_second": 4885.6
    },
    {
      "epoch": 140.92307692307693,
      "num_input_tokens_seen": 7274272,
      "step": 1832,
      "train_accuracy": 0.019543973729014397
    },
    {
      "epoch": 141.0,
      "grad_norm": 0.33878347277641296,
      "learning_rate": 0.01,
      "loss": 4.451435089111328,
      "num_input_tokens_seen": 7276728,
      "step": 1833,
      "train_runtime": 1489.4364,
      "train_tokens_per_second": 4885.558
    },
    {
      "epoch": 141.0,
      "num_input_tokens_seen": 7276728,
      "step": 1833,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 141.07692307692307,
      "grad_norm": 3.739004611968994,
      "learning_rate": 0.01,
      "loss": 4.433147430419922,
      "num_input_tokens_seen": 7280824,
      "step": 1834,
      "train_runtime": 1490.2632,
      "train_tokens_per_second": 4885.596
    },
    {
      "epoch": 141.07692307692307,
      "num_input_tokens_seen": 7280824,
      "step": 1834,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 141.15384615384616,
      "grad_norm": 0.5243370532989502,
      "learning_rate": 0.01,
      "loss": 4.409680366516113,
      "num_input_tokens_seen": 7284920,
      "step": 1835,
      "train_runtime": 1491.0781,
      "train_tokens_per_second": 4885.673
    },
    {
      "epoch": 141.15384615384616,
      "num_input_tokens_seen": 7284920,
      "step": 1835,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 141.23076923076923,
      "grad_norm": 0.9010217785835266,
      "learning_rate": 0.01,
      "loss": 4.446295261383057,
      "num_input_tokens_seen": 7289016,
      "step": 1836,
      "train_runtime": 1491.8915,
      "train_tokens_per_second": 4885.755
    },
    {
      "epoch": 141.23076923076923,
      "num_input_tokens_seen": 7289016,
      "step": 1836,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 141.30769230769232,
      "grad_norm": 0.6142174601554871,
      "learning_rate": 0.01,
      "loss": 4.417339324951172,
      "num_input_tokens_seen": 7293112,
      "step": 1837,
      "train_runtime": 1492.7059,
      "train_tokens_per_second": 4885.833
    },
    {
      "epoch": 141.30769230769232,
      "num_input_tokens_seen": 7293112,
      "step": 1837,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 141.3846153846154,
      "grad_norm": 0.3696889877319336,
      "learning_rate": 0.01,
      "loss": 4.461832523345947,
      "num_input_tokens_seen": 7297208,
      "step": 1838,
      "train_runtime": 1493.5192,
      "train_tokens_per_second": 4885.915
    },
    {
      "epoch": 141.3846153846154,
      "num_input_tokens_seen": 7297208,
      "step": 1838,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 141.46153846153845,
      "grad_norm": 0.4366854429244995,
      "learning_rate": 0.01,
      "loss": 4.467984199523926,
      "num_input_tokens_seen": 7301304,
      "step": 1839,
      "train_runtime": 1494.334,
      "train_tokens_per_second": 4885.992
    },
    {
      "epoch": 141.46153846153845,
      "num_input_tokens_seen": 7301304,
      "step": 1839,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 141.53846153846155,
      "grad_norm": 1.5263376235961914,
      "learning_rate": 0.01,
      "loss": 4.44230318069458,
      "num_input_tokens_seen": 7305400,
      "step": 1840,
      "train_runtime": 1495.1485,
      "train_tokens_per_second": 4886.07
    },
    {
      "epoch": 141.53846153846155,
      "num_input_tokens_seen": 7305400,
      "step": 1840,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 141.6153846153846,
      "grad_norm": 0.508631706237793,
      "learning_rate": 0.01,
      "loss": 4.566743850708008,
      "num_input_tokens_seen": 7309496,
      "step": 1841,
      "train_runtime": 1495.965,
      "train_tokens_per_second": 4886.141
    },
    {
      "epoch": 141.6153846153846,
      "num_input_tokens_seen": 7309496,
      "step": 1841,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 141.69230769230768,
      "grad_norm": 0.3965621888637543,
      "learning_rate": 0.01,
      "loss": 4.438959121704102,
      "num_input_tokens_seen": 7313592,
      "step": 1842,
      "train_runtime": 1496.7792,
      "train_tokens_per_second": 4886.22
    },
    {
      "epoch": 141.69230769230768,
      "num_input_tokens_seen": 7313592,
      "step": 1842,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 141.76923076923077,
      "grad_norm": 0.2859092652797699,
      "learning_rate": 0.01,
      "loss": 4.481933116912842,
      "num_input_tokens_seen": 7317688,
      "step": 1843,
      "train_runtime": 1497.5936,
      "train_tokens_per_second": 4886.298
    },
    {
      "epoch": 141.76923076923077,
      "num_input_tokens_seen": 7317688,
      "step": 1843,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 141.84615384615384,
      "grad_norm": 0.3808438777923584,
      "learning_rate": 0.01,
      "loss": 4.477018356323242,
      "num_input_tokens_seen": 7321784,
      "step": 1844,
      "train_runtime": 1498.4079,
      "train_tokens_per_second": 4886.376
    },
    {
      "epoch": 141.84615384615384,
      "num_input_tokens_seen": 7321784,
      "step": 1844,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 141.92307692307693,
      "grad_norm": 0.37763315439224243,
      "learning_rate": 0.01,
      "loss": 4.482439041137695,
      "num_input_tokens_seen": 7325880,
      "step": 1845,
      "train_runtime": 1499.2177,
      "train_tokens_per_second": 4886.468
    },
    {
      "epoch": 141.92307692307693,
      "num_input_tokens_seen": 7325880,
      "step": 1845,
      "train_accuracy": 0.0032573288772255182
    },
    {
      "epoch": 142.0,
      "grad_norm": 0.4352462589740753,
      "learning_rate": 0.01,
      "loss": 4.516756534576416,
      "num_input_tokens_seen": 7328336,
      "step": 1846,
      "train_runtime": 1499.7329,
      "train_tokens_per_second": 4886.427
    },
    {
      "epoch": 142.0,
      "num_input_tokens_seen": 7328336,
      "step": 1846,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 142.07692307692307,
      "grad_norm": 2.7508509159088135,
      "learning_rate": 0.01,
      "loss": 4.431884288787842,
      "num_input_tokens_seen": 7332432,
      "step": 1847,
      "train_runtime": 1500.5597,
      "train_tokens_per_second": 4886.465
    },
    {
      "epoch": 142.07692307692307,
      "num_input_tokens_seen": 7332432,
      "step": 1847,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 142.15384615384616,
      "grad_norm": 0.295866459608078,
      "learning_rate": 0.01,
      "loss": 4.467733383178711,
      "num_input_tokens_seen": 7336528,
      "step": 1848,
      "train_runtime": 1501.3742,
      "train_tokens_per_second": 4886.542
    },
    {
      "epoch": 142.15384615384616,
      "num_input_tokens_seen": 7336528,
      "step": 1848,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 142.23076923076923,
      "grad_norm": 1.1448103189468384,
      "learning_rate": 0.01,
      "loss": 4.491855144500732,
      "num_input_tokens_seen": 7340624,
      "step": 1849,
      "train_runtime": 1502.1883,
      "train_tokens_per_second": 4886.621
    },
    {
      "epoch": 142.23076923076923,
      "num_input_tokens_seen": 7340624,
      "step": 1849,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 142.30769230769232,
      "grad_norm": 0.5820721387863159,
      "learning_rate": 0.01,
      "loss": 4.487109184265137,
      "num_input_tokens_seen": 7344720,
      "step": 1850,
      "train_runtime": 1503.0034,
      "train_tokens_per_second": 4886.695
    },
    {
      "epoch": 142.30769230769232,
      "num_input_tokens_seen": 7344720,
      "step": 1850,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 142.3846153846154,
      "grad_norm": 0.3473348319530487,
      "learning_rate": 0.01,
      "loss": 4.461572647094727,
      "num_input_tokens_seen": 7348816,
      "step": 1851,
      "train_runtime": 1504.1069,
      "train_tokens_per_second": 4885.834
    },
    {
      "epoch": 142.3846153846154,
      "num_input_tokens_seen": 7348816,
      "step": 1851,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 142.46153846153845,
      "grad_norm": 0.3015472888946533,
      "learning_rate": 0.01,
      "loss": 4.4767165184021,
      "num_input_tokens_seen": 7352912,
      "step": 1852,
      "train_runtime": 1504.9196,
      "train_tokens_per_second": 4885.917
    },
    {
      "epoch": 142.46153846153845,
      "num_input_tokens_seen": 7352912,
      "step": 1852,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 142.53846153846155,
      "grad_norm": 0.26307594776153564,
      "learning_rate": 0.01,
      "loss": 4.494818687438965,
      "num_input_tokens_seen": 7357008,
      "step": 1853,
      "train_runtime": 1505.734,
      "train_tokens_per_second": 4885.995
    },
    {
      "epoch": 142.53846153846155,
      "num_input_tokens_seen": 7357008,
      "step": 1853,
      "train_accuracy": 0.03125
    },
    {
      "epoch": 142.6153846153846,
      "grad_norm": 0.7070944309234619,
      "learning_rate": 0.01,
      "loss": 4.474040508270264,
      "num_input_tokens_seen": 7361104,
      "step": 1854,
      "train_runtime": 1506.5515,
      "train_tokens_per_second": 4886.062
    },
    {
      "epoch": 142.6153846153846,
      "num_input_tokens_seen": 7361104,
      "step": 1854,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 142.69230769230768,
      "grad_norm": 0.29785415530204773,
      "learning_rate": 0.01,
      "loss": 4.507839202880859,
      "num_input_tokens_seen": 7365200,
      "step": 1855,
      "train_runtime": 1507.3658,
      "train_tokens_per_second": 4886.14
    },
    {
      "epoch": 142.69230769230768,
      "num_input_tokens_seen": 7365200,
      "step": 1855,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 142.76923076923077,
      "grad_norm": 0.325640469789505,
      "learning_rate": 0.01,
      "loss": 4.453739166259766,
      "num_input_tokens_seen": 7369296,
      "step": 1856,
      "train_runtime": 1508.1782,
      "train_tokens_per_second": 4886.224
    },
    {
      "epoch": 142.76923076923077,
      "num_input_tokens_seen": 7369296,
      "step": 1856,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 142.84615384615384,
      "grad_norm": 0.3849063813686371,
      "learning_rate": 0.01,
      "loss": 4.478244781494141,
      "num_input_tokens_seen": 7373392,
      "step": 1857,
      "train_runtime": 1508.9918,
      "train_tokens_per_second": 4886.304
    },
    {
      "epoch": 142.84615384615384,
      "num_input_tokens_seen": 7373392,
      "step": 1857,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 142.92307692307693,
      "grad_norm": 0.7774139642715454,
      "learning_rate": 0.01,
      "loss": 4.486651420593262,
      "num_input_tokens_seen": 7377488,
      "step": 1858,
      "train_runtime": 1509.8016,
      "train_tokens_per_second": 4886.396
    },
    {
      "epoch": 142.92307692307693,
      "num_input_tokens_seen": 7377488,
      "step": 1858,
      "train_accuracy": 0.0065146577544510365
    },
    {
      "epoch": 143.0,
      "grad_norm": 0.4661296010017395,
      "learning_rate": 0.01,
      "loss": 4.5151286125183105,
      "num_input_tokens_seen": 7379944,
      "step": 1859,
      "train_runtime": 1510.3161,
      "train_tokens_per_second": 4886.357
    },
    {
      "epoch": 143.0,
      "num_input_tokens_seen": 7379944,
      "step": 1859,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 143.07692307692307,
      "grad_norm": 0.27720293402671814,
      "learning_rate": 0.01,
      "loss": 4.443074703216553,
      "num_input_tokens_seen": 7384040,
      "step": 1860,
      "train_runtime": 1511.1415,
      "train_tokens_per_second": 4886.399
    },
    {
      "epoch": 143.07692307692307,
      "num_input_tokens_seen": 7384040,
      "step": 1860,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 143.15384615384616,
      "grad_norm": 0.24767735600471497,
      "learning_rate": 0.01,
      "loss": 4.409014701843262,
      "num_input_tokens_seen": 7388136,
      "step": 1861,
      "train_runtime": 1511.9548,
      "train_tokens_per_second": 4886.479
    },
    {
      "epoch": 143.15384615384616,
      "num_input_tokens_seen": 7388136,
      "step": 1861,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 143.23076923076923,
      "grad_norm": 0.7404224276542664,
      "learning_rate": 0.01,
      "loss": 4.457136154174805,
      "num_input_tokens_seen": 7392232,
      "step": 1862,
      "train_runtime": 1512.7678,
      "train_tokens_per_second": 4886.561
    },
    {
      "epoch": 143.23076923076923,
      "num_input_tokens_seen": 7392232,
      "step": 1862,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 143.30769230769232,
      "grad_norm": 0.29068946838378906,
      "learning_rate": 0.01,
      "loss": 4.466084003448486,
      "num_input_tokens_seen": 7396328,
      "step": 1863,
      "train_runtime": 1513.5829,
      "train_tokens_per_second": 4886.636
    },
    {
      "epoch": 143.30769230769232,
      "num_input_tokens_seen": 7396328,
      "step": 1863,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 143.3846153846154,
      "grad_norm": 0.3846644163131714,
      "learning_rate": 0.01,
      "loss": 4.4552001953125,
      "num_input_tokens_seen": 7400424,
      "step": 1864,
      "train_runtime": 1514.401,
      "train_tokens_per_second": 4886.701
    },
    {
      "epoch": 143.3846153846154,
      "num_input_tokens_seen": 7400424,
      "step": 1864,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 143.46153846153845,
      "grad_norm": 0.726744532585144,
      "learning_rate": 0.01,
      "loss": 4.420228958129883,
      "num_input_tokens_seen": 7404520,
      "step": 1865,
      "train_runtime": 1515.2135,
      "train_tokens_per_second": 4886.783
    },
    {
      "epoch": 143.46153846153845,
      "num_input_tokens_seen": 7404520,
      "step": 1865,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 143.53846153846155,
      "grad_norm": 0.26240605115890503,
      "learning_rate": 0.01,
      "loss": 4.4455461502075195,
      "num_input_tokens_seen": 7408616,
      "step": 1866,
      "train_runtime": 1516.0269,
      "train_tokens_per_second": 4886.863
    },
    {
      "epoch": 143.53846153846155,
      "num_input_tokens_seen": 7408616,
      "step": 1866,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 143.6153846153846,
      "grad_norm": 0.7446892857551575,
      "learning_rate": 0.01,
      "loss": 4.444561004638672,
      "num_input_tokens_seen": 7412712,
      "step": 1867,
      "train_runtime": 1516.8401,
      "train_tokens_per_second": 4886.944
    },
    {
      "epoch": 143.6153846153846,
      "num_input_tokens_seen": 7412712,
      "step": 1867,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 143.69230769230768,
      "grad_norm": 0.29408547282218933,
      "learning_rate": 0.01,
      "loss": 4.4721174240112305,
      "num_input_tokens_seen": 7416808,
      "step": 1868,
      "train_runtime": 1517.6542,
      "train_tokens_per_second": 4887.021
    },
    {
      "epoch": 143.69230769230768,
      "num_input_tokens_seen": 7416808,
      "step": 1868,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 143.76923076923077,
      "grad_norm": 0.22807541489601135,
      "learning_rate": 0.01,
      "loss": 4.449373245239258,
      "num_input_tokens_seen": 7420904,
      "step": 1869,
      "train_runtime": 1518.4691,
      "train_tokens_per_second": 4887.096
    },
    {
      "epoch": 143.76923076923077,
      "num_input_tokens_seen": 7420904,
      "step": 1869,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 143.84615384615384,
      "grad_norm": 0.357966810464859,
      "learning_rate": 0.01,
      "loss": 4.463508605957031,
      "num_input_tokens_seen": 7425000,
      "step": 1870,
      "train_runtime": 1519.2833,
      "train_tokens_per_second": 4887.173
    },
    {
      "epoch": 143.84615384615384,
      "num_input_tokens_seen": 7425000,
      "step": 1870,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 143.92307692307693,
      "grad_norm": 0.27506953477859497,
      "learning_rate": 0.01,
      "loss": 4.448062896728516,
      "num_input_tokens_seen": 7429096,
      "step": 1871,
      "train_runtime": 1520.0926,
      "train_tokens_per_second": 4887.265
    },
    {
      "epoch": 143.92307692307693,
      "num_input_tokens_seen": 7429096,
      "step": 1871,
      "train_accuracy": 0.026058631017804146
    },
    {
      "epoch": 144.0,
      "grad_norm": 0.5646443367004395,
      "learning_rate": 0.01,
      "loss": 4.415185928344727,
      "num_input_tokens_seen": 7431552,
      "step": 1872,
      "train_runtime": 1520.6075,
      "train_tokens_per_second": 4887.226
    },
    {
      "epoch": 144.0,
      "num_input_tokens_seen": 7431552,
      "step": 1872,
      "train_accuracy": 0.02734375
    },
    {
      "epoch": 144.07692307692307,
      "grad_norm": 0.3606163263320923,
      "learning_rate": 0.01,
      "loss": 4.352210998535156,
      "num_input_tokens_seen": 7435648,
      "step": 1873,
      "train_runtime": 1521.4334,
      "train_tokens_per_second": 4887.265
    },
    {
      "epoch": 144.07692307692307,
      "num_input_tokens_seen": 7435648,
      "step": 1873,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 144.15384615384616,
      "grad_norm": 0.5210484266281128,
      "learning_rate": 0.01,
      "loss": 4.4506659507751465,
      "num_input_tokens_seen": 7439744,
      "step": 1874,
      "train_runtime": 1522.2486,
      "train_tokens_per_second": 4887.338
    },
    {
      "epoch": 144.15384615384616,
      "num_input_tokens_seen": 7439744,
      "step": 1874,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 144.23076923076923,
      "grad_norm": 1.2244443893432617,
      "learning_rate": 0.01,
      "loss": 4.372976303100586,
      "num_input_tokens_seen": 7443840,
      "step": 1875,
      "train_runtime": 1523.0628,
      "train_tokens_per_second": 4887.415
    },
    {
      "epoch": 144.23076923076923,
      "num_input_tokens_seen": 7443840,
      "step": 1875,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 144.30769230769232,
      "grad_norm": 0.552419900894165,
      "learning_rate": 0.01,
      "loss": 4.419890403747559,
      "num_input_tokens_seen": 7447936,
      "step": 1876,
      "train_runtime": 1523.8779,
      "train_tokens_per_second": 4887.489
    },
    {
      "epoch": 144.30769230769232,
      "num_input_tokens_seen": 7447936,
      "step": 1876,
      "train_accuracy": 0.03515625
    },
    {
      "epoch": 144.3846153846154,
      "grad_norm": 0.389026015996933,
      "learning_rate": 0.01,
      "loss": 4.421940803527832,
      "num_input_tokens_seen": 7452032,
      "step": 1877,
      "train_runtime": 1524.6914,
      "train_tokens_per_second": 4887.568
    },
    {
      "epoch": 144.3846153846154,
      "num_input_tokens_seen": 7452032,
      "step": 1877,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 144.46153846153845,
      "grad_norm": 0.4070936441421509,
      "learning_rate": 0.01,
      "loss": 4.416196346282959,
      "num_input_tokens_seen": 7456128,
      "step": 1878,
      "train_runtime": 1525.505,
      "train_tokens_per_second": 4887.646
    },
    {
      "epoch": 144.46153846153845,
      "num_input_tokens_seen": 7456128,
      "step": 1878,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 144.53846153846155,
      "grad_norm": 0.5539480447769165,
      "learning_rate": 0.01,
      "loss": 4.417213439941406,
      "num_input_tokens_seen": 7460224,
      "step": 1879,
      "train_runtime": 1526.319,
      "train_tokens_per_second": 4887.723
    },
    {
      "epoch": 144.53846153846155,
      "num_input_tokens_seen": 7460224,
      "step": 1879,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 144.6153846153846,
      "grad_norm": 0.509876549243927,
      "learning_rate": 0.01,
      "loss": 4.399030685424805,
      "num_input_tokens_seen": 7464320,
      "step": 1880,
      "train_runtime": 1527.1336,
      "train_tokens_per_second": 4887.798
    },
    {
      "epoch": 144.6153846153846,
      "num_input_tokens_seen": 7464320,
      "step": 1880,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 144.69230769230768,
      "grad_norm": 0.31289342045783997,
      "learning_rate": 0.01,
      "loss": 4.391263008117676,
      "num_input_tokens_seen": 7468416,
      "step": 1881,
      "train_runtime": 1527.9479,
      "train_tokens_per_second": 4887.873
    },
    {
      "epoch": 144.69230769230768,
      "num_input_tokens_seen": 7468416,
      "step": 1881,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 144.76923076923077,
      "grad_norm": 0.9623465538024902,
      "learning_rate": 0.01,
      "loss": 4.368143081665039,
      "num_input_tokens_seen": 7472512,
      "step": 1882,
      "train_runtime": 1528.7631,
      "train_tokens_per_second": 4887.946
    },
    {
      "epoch": 144.76923076923077,
      "num_input_tokens_seen": 7472512,
      "step": 1882,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 144.84615384615384,
      "grad_norm": 1.9266918897628784,
      "learning_rate": 0.01,
      "loss": 4.450263977050781,
      "num_input_tokens_seen": 7476608,
      "step": 1883,
      "train_runtime": 1529.577,
      "train_tokens_per_second": 4888.023
    },
    {
      "epoch": 144.84615384615384,
      "num_input_tokens_seen": 7476608,
      "step": 1883,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 144.92307692307693,
      "grad_norm": 0.6616586446762085,
      "learning_rate": 0.01,
      "loss": 4.446620464324951,
      "num_input_tokens_seen": 7480704,
      "step": 1884,
      "train_runtime": 1530.3873,
      "train_tokens_per_second": 4888.112
    },
    {
      "epoch": 144.92307692307693,
      "num_input_tokens_seen": 7480704,
      "step": 1884,
      "train_accuracy": 0.013029315508902073
    },
    {
      "epoch": 145.0,
      "grad_norm": 0.4342552125453949,
      "learning_rate": 0.01,
      "loss": 4.474233150482178,
      "num_input_tokens_seen": 7483160,
      "step": 1885,
      "train_runtime": 1530.904,
      "train_tokens_per_second": 4888.066
    },
    {
      "epoch": 145.0,
      "num_input_tokens_seen": 7483160,
      "step": 1885,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 145.07692307692307,
      "grad_norm": 0.269889771938324,
      "learning_rate": 0.01,
      "loss": 4.450728416442871,
      "num_input_tokens_seen": 7487256,
      "step": 1886,
      "train_runtime": 1531.7306,
      "train_tokens_per_second": 4888.102
    },
    {
      "epoch": 145.07692307692307,
      "num_input_tokens_seen": 7487256,
      "step": 1886,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 145.15384615384616,
      "grad_norm": 0.32820841670036316,
      "learning_rate": 0.01,
      "loss": 4.453962326049805,
      "num_input_tokens_seen": 7491352,
      "step": 1887,
      "train_runtime": 1532.5447,
      "train_tokens_per_second": 4888.178
    },
    {
      "epoch": 145.15384615384616,
      "num_input_tokens_seen": 7491352,
      "step": 1887,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 145.23076923076923,
      "grad_norm": 3.949856758117676,
      "learning_rate": 0.01,
      "loss": 4.493199348449707,
      "num_input_tokens_seen": 7495448,
      "step": 1888,
      "train_runtime": 1533.3602,
      "train_tokens_per_second": 4888.25
    },
    {
      "epoch": 145.23076923076923,
      "num_input_tokens_seen": 7495448,
      "step": 1888,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 145.30769230769232,
      "grad_norm": 0.3343822956085205,
      "learning_rate": 0.01,
      "loss": 4.399847030639648,
      "num_input_tokens_seen": 7499544,
      "step": 1889,
      "train_runtime": 1534.1748,
      "train_tokens_per_second": 4888.324
    },
    {
      "epoch": 145.30769230769232,
      "num_input_tokens_seen": 7499544,
      "step": 1889,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 145.3846153846154,
      "grad_norm": 0.39394888281822205,
      "learning_rate": 0.01,
      "loss": 4.4716997146606445,
      "num_input_tokens_seen": 7503640,
      "step": 1890,
      "train_runtime": 1534.9893,
      "train_tokens_per_second": 4888.399
    },
    {
      "epoch": 145.3846153846154,
      "num_input_tokens_seen": 7503640,
      "step": 1890,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 145.46153846153845,
      "grad_norm": 0.32070469856262207,
      "learning_rate": 0.01,
      "loss": 4.444943428039551,
      "num_input_tokens_seen": 7507736,
      "step": 1891,
      "train_runtime": 1535.8039,
      "train_tokens_per_second": 4888.473
    },
    {
      "epoch": 145.46153846153845,
      "num_input_tokens_seen": 7507736,
      "step": 1891,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 145.53846153846155,
      "grad_norm": 2.3454792499542236,
      "learning_rate": 0.01,
      "loss": 4.485935211181641,
      "num_input_tokens_seen": 7511832,
      "step": 1892,
      "train_runtime": 1536.6191,
      "train_tokens_per_second": 4888.545
    },
    {
      "epoch": 145.53846153846155,
      "num_input_tokens_seen": 7511832,
      "step": 1892,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 145.6153846153846,
      "grad_norm": 0.4269157648086548,
      "learning_rate": 0.01,
      "loss": 4.4301228523254395,
      "num_input_tokens_seen": 7515928,
      "step": 1893,
      "train_runtime": 1537.4332,
      "train_tokens_per_second": 4888.621
    },
    {
      "epoch": 145.6153846153846,
      "num_input_tokens_seen": 7515928,
      "step": 1893,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 145.69230769230768,
      "grad_norm": 0.8848997354507446,
      "learning_rate": 0.01,
      "loss": 4.462418556213379,
      "num_input_tokens_seen": 7520024,
      "step": 1894,
      "train_runtime": 1538.2493,
      "train_tokens_per_second": 4888.69
    },
    {
      "epoch": 145.69230769230768,
      "num_input_tokens_seen": 7520024,
      "step": 1894,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 145.76923076923077,
      "grad_norm": 0.4912108778953552,
      "learning_rate": 0.01,
      "loss": 4.469257831573486,
      "num_input_tokens_seen": 7524120,
      "step": 1895,
      "train_runtime": 1539.0637,
      "train_tokens_per_second": 4888.765
    },
    {
      "epoch": 145.76923076923077,
      "num_input_tokens_seen": 7524120,
      "step": 1895,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 145.84615384615384,
      "grad_norm": 1.5754284858703613,
      "learning_rate": 0.01,
      "loss": 4.464353084564209,
      "num_input_tokens_seen": 7528216,
      "step": 1896,
      "train_runtime": 1539.8779,
      "train_tokens_per_second": 4888.839
    },
    {
      "epoch": 145.84615384615384,
      "num_input_tokens_seen": 7528216,
      "step": 1896,
      "train_accuracy": 0.029296875
    },
    {
      "epoch": 145.92307692307693,
      "grad_norm": 4.208193302154541,
      "learning_rate": 0.01,
      "loss": 4.438633918762207,
      "num_input_tokens_seen": 7532312,
      "step": 1897,
      "train_runtime": 1540.6878,
      "train_tokens_per_second": 4888.928
    },
    {
      "epoch": 145.92307692307693,
      "num_input_tokens_seen": 7532312,
      "step": 1897,
      "train_accuracy": 0.0065146577544510365
    },
    {
      "epoch": 146.0,
      "grad_norm": 0.3978414237499237,
      "learning_rate": 0.01,
      "loss": 4.482831001281738,
      "num_input_tokens_seen": 7534768,
      "step": 1898,
      "train_runtime": 1541.2029,
      "train_tokens_per_second": 4888.888
    },
    {
      "epoch": 146.0,
      "num_input_tokens_seen": 7534768,
      "step": 1898,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 146.07692307692307,
      "grad_norm": 0.5319635272026062,
      "learning_rate": 0.01,
      "loss": 4.4385271072387695,
      "num_input_tokens_seen": 7538864,
      "step": 1899,
      "train_runtime": 1542.0326,
      "train_tokens_per_second": 4888.913
    },
    {
      "epoch": 146.07692307692307,
      "num_input_tokens_seen": 7538864,
      "step": 1899,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 146.15384615384616,
      "grad_norm": 0.4587706923484802,
      "learning_rate": 0.01,
      "loss": 4.444916725158691,
      "num_input_tokens_seen": 7542960,
      "step": 1900,
      "train_runtime": 1542.851,
      "train_tokens_per_second": 4888.975
    },
    {
      "epoch": 146.15384615384616,
      "eval_CMD_3_branch_eval_accuracy": 0.011934900542495479,
      "eval_CMD_3_branch_eval_loss": 4.494749546051025,
      "eval_CMD_3_branch_eval_runtime": 1.1543,
      "eval_CMD_3_branch_eval_samples_per_second": 2395.338,
      "eval_CMD_3_branch_eval_steps_per_second": 5.198,
      "num_input_tokens_seen": 7542960,
      "step": 1900
    },
    {
      "epoch": 146.15384615384616,
      "num_input_tokens_seen": 7542960,
      "step": 1900,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 146.23076923076923,
      "grad_norm": 0.8114116787910461,
      "learning_rate": 0.01,
      "loss": 4.4554595947265625,
      "num_input_tokens_seen": 7547056,
      "step": 1901,
      "train_runtime": 1544.8254,
      "train_tokens_per_second": 4885.378
    },
    {
      "epoch": 146.23076923076923,
      "num_input_tokens_seen": 7547056,
      "step": 1901,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 146.30769230769232,
      "grad_norm": 0.2652541399002075,
      "learning_rate": 0.01,
      "loss": 4.4496169090271,
      "num_input_tokens_seen": 7551152,
      "step": 1902,
      "train_runtime": 1545.6419,
      "train_tokens_per_second": 4885.447
    },
    {
      "epoch": 146.30769230769232,
      "num_input_tokens_seen": 7551152,
      "step": 1902,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 146.3846153846154,
      "grad_norm": 2.1605048179626465,
      "learning_rate": 0.01,
      "loss": 4.47141695022583,
      "num_input_tokens_seen": 7555248,
      "step": 1903,
      "train_runtime": 1546.4649,
      "train_tokens_per_second": 4885.496
    },
    {
      "epoch": 146.3846153846154,
      "num_input_tokens_seen": 7555248,
      "step": 1903,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 146.46153846153845,
      "grad_norm": 0.5354732871055603,
      "learning_rate": 0.01,
      "loss": 4.465007781982422,
      "num_input_tokens_seen": 7559344,
      "step": 1904,
      "train_runtime": 1547.2808,
      "train_tokens_per_second": 4885.567
    },
    {
      "epoch": 146.46153846153845,
      "num_input_tokens_seen": 7559344,
      "step": 1904,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 146.53846153846155,
      "grad_norm": 0.6656928658485413,
      "learning_rate": 0.01,
      "loss": 4.482838153839111,
      "num_input_tokens_seen": 7563440,
      "step": 1905,
      "train_runtime": 1548.0968,
      "train_tokens_per_second": 4885.638
    },
    {
      "epoch": 146.53846153846155,
      "num_input_tokens_seen": 7563440,
      "step": 1905,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 146.6153846153846,
      "grad_norm": 2.222104072570801,
      "learning_rate": 0.01,
      "loss": 4.4592742919921875,
      "num_input_tokens_seen": 7567536,
      "step": 1906,
      "train_runtime": 1548.9114,
      "train_tokens_per_second": 4885.713
    },
    {
      "epoch": 146.6153846153846,
      "num_input_tokens_seen": 7567536,
      "step": 1906,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 146.69230769230768,
      "grad_norm": 0.7789755463600159,
      "learning_rate": 0.01,
      "loss": 4.525387763977051,
      "num_input_tokens_seen": 7571632,
      "step": 1907,
      "train_runtime": 1549.7265,
      "train_tokens_per_second": 4885.786
    },
    {
      "epoch": 146.69230769230768,
      "num_input_tokens_seen": 7571632,
      "step": 1907,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 146.76923076923077,
      "grad_norm": 0.37413662672042847,
      "learning_rate": 0.01,
      "loss": 4.4984517097473145,
      "num_input_tokens_seen": 7575728,
      "step": 1908,
      "train_runtime": 1550.5421,
      "train_tokens_per_second": 4885.858
    },
    {
      "epoch": 146.76923076923077,
      "num_input_tokens_seen": 7575728,
      "step": 1908,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 146.84615384615384,
      "grad_norm": 0.33064937591552734,
      "learning_rate": 0.01,
      "loss": 4.53651237487793,
      "num_input_tokens_seen": 7579824,
      "step": 1909,
      "train_runtime": 1551.3579,
      "train_tokens_per_second": 4885.929
    },
    {
      "epoch": 146.84615384615384,
      "num_input_tokens_seen": 7579824,
      "step": 1909,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 146.92307692307693,
      "grad_norm": 0.8711787462234497,
      "learning_rate": 0.01,
      "loss": 4.535449028015137,
      "num_input_tokens_seen": 7583920,
      "step": 1910,
      "train_runtime": 1552.1685,
      "train_tokens_per_second": 4886.016
    },
    {
      "epoch": 146.92307692307693,
      "num_input_tokens_seen": 7583920,
      "step": 1910,
      "train_accuracy": 0.013029315508902073
    },
    {
      "epoch": 147.0,
      "grad_norm": 0.39004987478256226,
      "learning_rate": 0.01,
      "loss": 4.533785343170166,
      "num_input_tokens_seen": 7586376,
      "step": 1911,
      "train_runtime": 1552.684,
      "train_tokens_per_second": 4885.975
    },
    {
      "epoch": 147.0,
      "num_input_tokens_seen": 7586376,
      "step": 1911,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 147.07692307692307,
      "grad_norm": 0.3635663688182831,
      "learning_rate": 0.01,
      "loss": 4.534843921661377,
      "num_input_tokens_seen": 7590472,
      "step": 1912,
      "train_runtime": 1553.511,
      "train_tokens_per_second": 4886.011
    },
    {
      "epoch": 147.07692307692307,
      "num_input_tokens_seen": 7590472,
      "step": 1912,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 147.15384615384616,
      "grad_norm": 0.2025507241487503,
      "learning_rate": 0.01,
      "loss": 4.535642623901367,
      "num_input_tokens_seen": 7594568,
      "step": 1913,
      "train_runtime": 1554.3274,
      "train_tokens_per_second": 4886.08
    },
    {
      "epoch": 147.15384615384616,
      "num_input_tokens_seen": 7594568,
      "step": 1913,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 147.23076923076923,
      "grad_norm": 0.2537078857421875,
      "learning_rate": 0.01,
      "loss": 4.543508529663086,
      "num_input_tokens_seen": 7598664,
      "step": 1914,
      "train_runtime": 1555.1428,
      "train_tokens_per_second": 4886.152
    },
    {
      "epoch": 147.23076923076923,
      "num_input_tokens_seen": 7598664,
      "step": 1914,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 147.30769230769232,
      "grad_norm": 0.25231072306632996,
      "learning_rate": 0.01,
      "loss": 4.54705810546875,
      "num_input_tokens_seen": 7602760,
      "step": 1915,
      "train_runtime": 1555.9582,
      "train_tokens_per_second": 4886.224
    },
    {
      "epoch": 147.30769230769232,
      "num_input_tokens_seen": 7602760,
      "step": 1915,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 147.3846153846154,
      "grad_norm": 0.2876351773738861,
      "learning_rate": 0.01,
      "loss": 4.552492141723633,
      "num_input_tokens_seen": 7606856,
      "step": 1916,
      "train_runtime": 1556.7733,
      "train_tokens_per_second": 4886.296
    },
    {
      "epoch": 147.3846153846154,
      "num_input_tokens_seen": 7606856,
      "step": 1916,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 147.46153846153845,
      "grad_norm": 0.2483632117509842,
      "learning_rate": 0.01,
      "loss": 4.564187049865723,
      "num_input_tokens_seen": 7610952,
      "step": 1917,
      "train_runtime": 1557.5879,
      "train_tokens_per_second": 4886.371
    },
    {
      "epoch": 147.46153846153845,
      "num_input_tokens_seen": 7610952,
      "step": 1917,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 147.53846153846155,
      "grad_norm": 0.2530250549316406,
      "learning_rate": 0.01,
      "loss": 4.551231384277344,
      "num_input_tokens_seen": 7615048,
      "step": 1918,
      "train_runtime": 1558.4031,
      "train_tokens_per_second": 4886.443
    },
    {
      "epoch": 147.53846153846155,
      "num_input_tokens_seen": 7615048,
      "step": 1918,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 147.6153846153846,
      "grad_norm": 0.21836470067501068,
      "learning_rate": 0.01,
      "loss": 4.567032814025879,
      "num_input_tokens_seen": 7619144,
      "step": 1919,
      "train_runtime": 1559.2185,
      "train_tokens_per_second": 4886.515
    },
    {
      "epoch": 147.6153846153846,
      "num_input_tokens_seen": 7619144,
      "step": 1919,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 147.69230769230768,
      "grad_norm": 0.30134090781211853,
      "learning_rate": 0.01,
      "loss": 4.554921627044678,
      "num_input_tokens_seen": 7623240,
      "step": 1920,
      "train_runtime": 1560.0346,
      "train_tokens_per_second": 4886.584
    },
    {
      "epoch": 147.69230769230768,
      "num_input_tokens_seen": 7623240,
      "step": 1920,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 147.76923076923077,
      "grad_norm": 0.39850473403930664,
      "learning_rate": 0.01,
      "loss": 4.575058937072754,
      "num_input_tokens_seen": 7627336,
      "step": 1921,
      "train_runtime": 1560.8497,
      "train_tokens_per_second": 4886.656
    },
    {
      "epoch": 147.76923076923077,
      "num_input_tokens_seen": 7627336,
      "step": 1921,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 147.84615384615384,
      "grad_norm": 0.4197423458099365,
      "learning_rate": 0.01,
      "loss": 4.550651550292969,
      "num_input_tokens_seen": 7631432,
      "step": 1922,
      "train_runtime": 1561.6643,
      "train_tokens_per_second": 4886.73
    },
    {
      "epoch": 147.84615384615384,
      "num_input_tokens_seen": 7631432,
      "step": 1922,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 147.92307692307693,
      "grad_norm": 0.2957670986652374,
      "learning_rate": 0.01,
      "loss": 4.577169418334961,
      "num_input_tokens_seen": 7635528,
      "step": 1923,
      "train_runtime": 1562.4753,
      "train_tokens_per_second": 4886.815
    },
    {
      "epoch": 147.92307692307693,
      "num_input_tokens_seen": 7635528,
      "step": 1923,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 148.0,
      "grad_norm": 0.34968140721321106,
      "learning_rate": 0.01,
      "loss": 4.567928314208984,
      "num_input_tokens_seen": 7637984,
      "step": 1924,
      "train_runtime": 1562.9904,
      "train_tokens_per_second": 4886.776
    },
    {
      "epoch": 148.0,
      "num_input_tokens_seen": 7637984,
      "step": 1924,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 148.07692307692307,
      "grad_norm": 0.21659724414348602,
      "learning_rate": 0.01,
      "loss": 4.5482177734375,
      "num_input_tokens_seen": 7642080,
      "step": 1925,
      "train_runtime": 1563.8208,
      "train_tokens_per_second": 4886.8
    },
    {
      "epoch": 148.07692307692307,
      "num_input_tokens_seen": 7642080,
      "step": 1925,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 148.15384615384616,
      "grad_norm": 0.5055199265480042,
      "learning_rate": 0.01,
      "loss": 4.555220603942871,
      "num_input_tokens_seen": 7646176,
      "step": 1926,
      "train_runtime": 1564.6357,
      "train_tokens_per_second": 4886.873
    },
    {
      "epoch": 148.15384615384616,
      "num_input_tokens_seen": 7646176,
      "step": 1926,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 148.23076923076923,
      "grad_norm": 0.39798504114151,
      "learning_rate": 0.01,
      "loss": 4.528914928436279,
      "num_input_tokens_seen": 7650272,
      "step": 1927,
      "train_runtime": 1565.4526,
      "train_tokens_per_second": 4886.939
    },
    {
      "epoch": 148.23076923076923,
      "num_input_tokens_seen": 7650272,
      "step": 1927,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 148.30769230769232,
      "grad_norm": 0.267161101102829,
      "learning_rate": 0.01,
      "loss": 4.556334018707275,
      "num_input_tokens_seen": 7654368,
      "step": 1928,
      "train_runtime": 1566.2679,
      "train_tokens_per_second": 4887.011
    },
    {
      "epoch": 148.30769230769232,
      "num_input_tokens_seen": 7654368,
      "step": 1928,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 148.3846153846154,
      "grad_norm": 0.337952196598053,
      "learning_rate": 0.01,
      "loss": 4.570736885070801,
      "num_input_tokens_seen": 7658464,
      "step": 1929,
      "train_runtime": 1567.0834,
      "train_tokens_per_second": 4887.081
    },
    {
      "epoch": 148.3846153846154,
      "num_input_tokens_seen": 7658464,
      "step": 1929,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 148.46153846153845,
      "grad_norm": 0.2820669412612915,
      "learning_rate": 0.01,
      "loss": 4.561923980712891,
      "num_input_tokens_seen": 7662560,
      "step": 1930,
      "train_runtime": 1567.8978,
      "train_tokens_per_second": 4887.155
    },
    {
      "epoch": 148.46153846153845,
      "num_input_tokens_seen": 7662560,
      "step": 1930,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 148.53846153846155,
      "grad_norm": 0.2407274842262268,
      "learning_rate": 0.01,
      "loss": 4.524092674255371,
      "num_input_tokens_seen": 7666656,
      "step": 1931,
      "train_runtime": 1568.714,
      "train_tokens_per_second": 4887.223
    },
    {
      "epoch": 148.53846153846155,
      "num_input_tokens_seen": 7666656,
      "step": 1931,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 148.6153846153846,
      "grad_norm": 0.24104392528533936,
      "learning_rate": 0.01,
      "loss": 4.526534080505371,
      "num_input_tokens_seen": 7670752,
      "step": 1932,
      "train_runtime": 1569.5302,
      "train_tokens_per_second": 4887.292
    },
    {
      "epoch": 148.6153846153846,
      "num_input_tokens_seen": 7670752,
      "step": 1932,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 148.69230769230768,
      "grad_norm": 0.31010931730270386,
      "learning_rate": 0.01,
      "loss": 4.554614067077637,
      "num_input_tokens_seen": 7674848,
      "step": 1933,
      "train_runtime": 1570.3486,
      "train_tokens_per_second": 4887.353
    },
    {
      "epoch": 148.69230769230768,
      "num_input_tokens_seen": 7674848,
      "step": 1933,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 148.76923076923077,
      "grad_norm": 0.1968391090631485,
      "learning_rate": 0.01,
      "loss": 4.527078628540039,
      "num_input_tokens_seen": 7678944,
      "step": 1934,
      "train_runtime": 1571.1653,
      "train_tokens_per_second": 4887.419
    },
    {
      "epoch": 148.76923076923077,
      "num_input_tokens_seen": 7678944,
      "step": 1934,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 148.84615384615384,
      "grad_norm": 0.263007789850235,
      "learning_rate": 0.01,
      "loss": 4.539289951324463,
      "num_input_tokens_seen": 7683040,
      "step": 1935,
      "train_runtime": 1571.9799,
      "train_tokens_per_second": 4887.493
    },
    {
      "epoch": 148.84615384615384,
      "num_input_tokens_seen": 7683040,
      "step": 1935,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 148.92307692307693,
      "grad_norm": 0.2502298951148987,
      "learning_rate": 0.01,
      "loss": 4.544886589050293,
      "num_input_tokens_seen": 7687136,
      "step": 1936,
      "train_runtime": 1572.7907,
      "train_tokens_per_second": 4887.577
    },
    {
      "epoch": 148.92307692307693,
      "num_input_tokens_seen": 7687136,
      "step": 1936,
      "train_accuracy": 0.013029315508902073
    },
    {
      "epoch": 149.0,
      "grad_norm": 0.3240714371204376,
      "learning_rate": 0.01,
      "loss": 4.544671535491943,
      "num_input_tokens_seen": 7689592,
      "step": 1937,
      "train_runtime": 1573.3074,
      "train_tokens_per_second": 4887.533
    },
    {
      "epoch": 149.0,
      "num_input_tokens_seen": 7689592,
      "step": 1937,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 149.07692307692307,
      "grad_norm": 0.2039574384689331,
      "learning_rate": 0.01,
      "loss": 4.522834777832031,
      "num_input_tokens_seen": 7693688,
      "step": 1938,
      "train_runtime": 1574.1368,
      "train_tokens_per_second": 4887.56
    },
    {
      "epoch": 149.07692307692307,
      "num_input_tokens_seen": 7693688,
      "step": 1938,
      "train_accuracy": 0.01953125
    },
    {
      "epoch": 149.15384615384616,
      "grad_norm": 1.0362766981124878,
      "learning_rate": 0.01,
      "loss": 4.514573097229004,
      "num_input_tokens_seen": 7697784,
      "step": 1939,
      "train_runtime": 1574.9517,
      "train_tokens_per_second": 4887.632
    },
    {
      "epoch": 149.15384615384616,
      "num_input_tokens_seen": 7697784,
      "step": 1939,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 149.23076923076923,
      "grad_norm": 0.29704248905181885,
      "learning_rate": 0.01,
      "loss": 4.513667106628418,
      "num_input_tokens_seen": 7701880,
      "step": 1940,
      "train_runtime": 1575.7694,
      "train_tokens_per_second": 4887.695
    },
    {
      "epoch": 149.23076923076923,
      "num_input_tokens_seen": 7701880,
      "step": 1940,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 149.30769230769232,
      "grad_norm": 0.4148108959197998,
      "learning_rate": 0.01,
      "loss": 4.529236316680908,
      "num_input_tokens_seen": 7705976,
      "step": 1941,
      "train_runtime": 1576.5853,
      "train_tokens_per_second": 4887.764
    },
    {
      "epoch": 149.30769230769232,
      "num_input_tokens_seen": 7705976,
      "step": 1941,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 149.3846153846154,
      "grad_norm": 0.3442012369632721,
      "learning_rate": 0.01,
      "loss": 4.504043102264404,
      "num_input_tokens_seen": 7710072,
      "step": 1942,
      "train_runtime": 1577.3992,
      "train_tokens_per_second": 4887.838
    },
    {
      "epoch": 149.3846153846154,
      "num_input_tokens_seen": 7710072,
      "step": 1942,
      "train_accuracy": 0.005859375
    },
    {
      "epoch": 149.46153846153845,
      "grad_norm": 0.2795056700706482,
      "learning_rate": 0.01,
      "loss": 4.526984691619873,
      "num_input_tokens_seen": 7714168,
      "step": 1943,
      "train_runtime": 1578.2143,
      "train_tokens_per_second": 4887.909
    },
    {
      "epoch": 149.46153846153845,
      "num_input_tokens_seen": 7714168,
      "step": 1943,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 149.53846153846155,
      "grad_norm": 0.25076037645339966,
      "learning_rate": 0.01,
      "loss": 4.546579360961914,
      "num_input_tokens_seen": 7718264,
      "step": 1944,
      "train_runtime": 1579.0301,
      "train_tokens_per_second": 4887.978
    },
    {
      "epoch": 149.53846153846155,
      "num_input_tokens_seen": 7718264,
      "step": 1944,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 149.6153846153846,
      "grad_norm": 0.25393232703208923,
      "learning_rate": 0.01,
      "loss": 4.53092622756958,
      "num_input_tokens_seen": 7722360,
      "step": 1945,
      "train_runtime": 1579.8478,
      "train_tokens_per_second": 4888.041
    },
    {
      "epoch": 149.6153846153846,
      "num_input_tokens_seen": 7722360,
      "step": 1945,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 149.69230769230768,
      "grad_norm": 0.3227154314517975,
      "learning_rate": 0.01,
      "loss": 4.532125473022461,
      "num_input_tokens_seen": 7726456,
      "step": 1946,
      "train_runtime": 1580.6623,
      "train_tokens_per_second": 4888.113
    },
    {
      "epoch": 149.69230769230768,
      "num_input_tokens_seen": 7726456,
      "step": 1946,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 149.76923076923077,
      "grad_norm": 1.8587673902511597,
      "learning_rate": 0.01,
      "loss": 4.552395343780518,
      "num_input_tokens_seen": 7730552,
      "step": 1947,
      "train_runtime": 1581.4791,
      "train_tokens_per_second": 4888.178
    },
    {
      "epoch": 149.76923076923077,
      "num_input_tokens_seen": 7730552,
      "step": 1947,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 149.84615384615384,
      "grad_norm": 0.2996019124984741,
      "learning_rate": 0.01,
      "loss": 4.536066055297852,
      "num_input_tokens_seen": 7734648,
      "step": 1948,
      "train_runtime": 1582.296,
      "train_tokens_per_second": 4888.243
    },
    {
      "epoch": 149.84615384615384,
      "num_input_tokens_seen": 7734648,
      "step": 1948,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 149.92307692307693,
      "grad_norm": 0.2963835895061493,
      "learning_rate": 0.01,
      "loss": 4.584995269775391,
      "num_input_tokens_seen": 7738744,
      "step": 1949,
      "train_runtime": 1583.1078,
      "train_tokens_per_second": 4888.324
    },
    {
      "epoch": 149.92307692307693,
      "num_input_tokens_seen": 7738744,
      "step": 1949,
      "train_accuracy": 0.009771986864507198
    },
    {
      "epoch": 150.0,
      "grad_norm": 0.5054681897163391,
      "learning_rate": 0.01,
      "loss": 4.556207656860352,
      "num_input_tokens_seen": 7741200,
      "step": 1950,
      "train_runtime": 1583.6243,
      "train_tokens_per_second": 4888.281
    },
    {
      "epoch": 150.0,
      "num_input_tokens_seen": 7741200,
      "step": 1950,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 150.07692307692307,
      "grad_norm": 0.3328983783721924,
      "learning_rate": 0.01,
      "loss": 4.530587196350098,
      "num_input_tokens_seen": 7745296,
      "step": 1951,
      "train_runtime": 1584.454,
      "train_tokens_per_second": 4888.306
    },
    {
      "epoch": 150.07692307692307,
      "num_input_tokens_seen": 7745296,
      "step": 1951,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 150.15384615384616,
      "grad_norm": 0.38532009720802307,
      "learning_rate": 0.01,
      "loss": 4.567971706390381,
      "num_input_tokens_seen": 7749392,
      "step": 1952,
      "train_runtime": 1585.2715,
      "train_tokens_per_second": 4888.369
    },
    {
      "epoch": 150.15384615384616,
      "num_input_tokens_seen": 7749392,
      "step": 1952,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 150.23076923076923,
      "grad_norm": 0.427386611700058,
      "learning_rate": 0.01,
      "loss": 4.555583953857422,
      "num_input_tokens_seen": 7753488,
      "step": 1953,
      "train_runtime": 1586.087,
      "train_tokens_per_second": 4888.438
    },
    {
      "epoch": 150.23076923076923,
      "num_input_tokens_seen": 7753488,
      "step": 1953,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 150.30769230769232,
      "grad_norm": 0.28484898805618286,
      "learning_rate": 0.01,
      "loss": 4.55283260345459,
      "num_input_tokens_seen": 7757584,
      "step": 1954,
      "train_runtime": 1586.9021,
      "train_tokens_per_second": 4888.508
    },
    {
      "epoch": 150.30769230769232,
      "num_input_tokens_seen": 7757584,
      "step": 1954,
      "train_accuracy": 0.0234375
    },
    {
      "epoch": 150.3846153846154,
      "grad_norm": 0.27168798446655273,
      "learning_rate": 0.01,
      "loss": 4.511292457580566,
      "num_input_tokens_seen": 7761680,
      "step": 1955,
      "train_runtime": 1587.7182,
      "train_tokens_per_second": 4888.575
    },
    {
      "epoch": 150.3846153846154,
      "num_input_tokens_seen": 7761680,
      "step": 1955,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 150.46153846153845,
      "grad_norm": 0.275595486164093,
      "learning_rate": 0.01,
      "loss": 4.548511505126953,
      "num_input_tokens_seen": 7765776,
      "step": 1956,
      "train_runtime": 1588.5354,
      "train_tokens_per_second": 4888.639
    },
    {
      "epoch": 150.46153846153845,
      "num_input_tokens_seen": 7765776,
      "step": 1956,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 150.53846153846155,
      "grad_norm": 0.4258243441581726,
      "learning_rate": 0.01,
      "loss": 4.549992084503174,
      "num_input_tokens_seen": 7769872,
      "step": 1957,
      "train_runtime": 1589.3506,
      "train_tokens_per_second": 4888.709
    },
    {
      "epoch": 150.53846153846155,
      "num_input_tokens_seen": 7769872,
      "step": 1957,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 150.6153846153846,
      "grad_norm": 0.3259783089160919,
      "learning_rate": 0.01,
      "loss": 4.542173385620117,
      "num_input_tokens_seen": 7773968,
      "step": 1958,
      "train_runtime": 1590.1668,
      "train_tokens_per_second": 4888.775
    },
    {
      "epoch": 150.6153846153846,
      "num_input_tokens_seen": 7773968,
      "step": 1958,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 150.69230769230768,
      "grad_norm": 1.607349157333374,
      "learning_rate": 0.01,
      "loss": 4.520785331726074,
      "num_input_tokens_seen": 7778064,
      "step": 1959,
      "train_runtime": 1590.9819,
      "train_tokens_per_second": 4888.845
    },
    {
      "epoch": 150.69230769230768,
      "num_input_tokens_seen": 7778064,
      "step": 1959,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 150.76923076923077,
      "grad_norm": 0.285757839679718,
      "learning_rate": 0.01,
      "loss": 4.542995452880859,
      "num_input_tokens_seen": 7782160,
      "step": 1960,
      "train_runtime": 1591.7968,
      "train_tokens_per_second": 4888.915
    },
    {
      "epoch": 150.76923076923077,
      "num_input_tokens_seen": 7782160,
      "step": 1960,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 150.84615384615384,
      "grad_norm": 0.7893608808517456,
      "learning_rate": 0.01,
      "loss": 4.516138076782227,
      "num_input_tokens_seen": 7786256,
      "step": 1961,
      "train_runtime": 1592.6111,
      "train_tokens_per_second": 4888.988
    },
    {
      "epoch": 150.84615384615384,
      "num_input_tokens_seen": 7786256,
      "step": 1961,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 150.92307692307693,
      "grad_norm": 0.37208184599876404,
      "learning_rate": 0.01,
      "loss": 4.5633158683776855,
      "num_input_tokens_seen": 7790352,
      "step": 1962,
      "train_runtime": 1593.4234,
      "train_tokens_per_second": 4889.066
    },
    {
      "epoch": 150.92307692307693,
      "num_input_tokens_seen": 7790352,
      "step": 1962,
      "train_accuracy": 0.016286645084619522
    },
    {
      "epoch": 151.0,
      "grad_norm": 0.6915590167045593,
      "learning_rate": 0.01,
      "loss": 4.537909984588623,
      "num_input_tokens_seen": 7792808,
      "step": 1963,
      "train_runtime": 1593.9393,
      "train_tokens_per_second": 4889.024
    },
    {
      "epoch": 151.0,
      "num_input_tokens_seen": 7792808,
      "step": 1963,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 151.07692307692307,
      "grad_norm": 2.010768175125122,
      "learning_rate": 0.01,
      "loss": 4.547728538513184,
      "num_input_tokens_seen": 7796904,
      "step": 1964,
      "train_runtime": 1594.7667,
      "train_tokens_per_second": 4889.056
    },
    {
      "epoch": 151.07692307692307,
      "num_input_tokens_seen": 7796904,
      "step": 1964,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 151.15384615384616,
      "grad_norm": 0.3066599667072296,
      "learning_rate": 0.01,
      "loss": 4.548813819885254,
      "num_input_tokens_seen": 7801000,
      "step": 1965,
      "train_runtime": 1595.5812,
      "train_tokens_per_second": 4889.128
    },
    {
      "epoch": 151.15384615384616,
      "num_input_tokens_seen": 7801000,
      "step": 1965,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 151.23076923076923,
      "grad_norm": 0.2867196798324585,
      "learning_rate": 0.01,
      "loss": 4.550006866455078,
      "num_input_tokens_seen": 7805096,
      "step": 1966,
      "train_runtime": 1596.3964,
      "train_tokens_per_second": 4889.197
    },
    {
      "epoch": 151.23076923076923,
      "num_input_tokens_seen": 7805096,
      "step": 1966,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 151.30769230769232,
      "grad_norm": 0.27043581008911133,
      "learning_rate": 0.01,
      "loss": 4.536755561828613,
      "num_input_tokens_seen": 7809192,
      "step": 1967,
      "train_runtime": 1597.2101,
      "train_tokens_per_second": 4889.27
    },
    {
      "epoch": 151.30769230769232,
      "num_input_tokens_seen": 7809192,
      "step": 1967,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 151.3846153846154,
      "grad_norm": 0.2750074863433838,
      "learning_rate": 0.01,
      "loss": 4.558106422424316,
      "num_input_tokens_seen": 7813288,
      "step": 1968,
      "train_runtime": 1598.0251,
      "train_tokens_per_second": 4889.34
    },
    {
      "epoch": 151.3846153846154,
      "num_input_tokens_seen": 7813288,
      "step": 1968,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 151.46153846153845,
      "grad_norm": 0.3051276206970215,
      "learning_rate": 0.01,
      "loss": 4.566568374633789,
      "num_input_tokens_seen": 7817384,
      "step": 1969,
      "train_runtime": 1598.8393,
      "train_tokens_per_second": 4889.412
    },
    {
      "epoch": 151.46153846153845,
      "num_input_tokens_seen": 7817384,
      "step": 1969,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 151.53846153846155,
      "grad_norm": 0.26743951439857483,
      "learning_rate": 0.01,
      "loss": 4.548969745635986,
      "num_input_tokens_seen": 7821480,
      "step": 1970,
      "train_runtime": 1599.6544,
      "train_tokens_per_second": 4889.481
    },
    {
      "epoch": 151.53846153846155,
      "num_input_tokens_seen": 7821480,
      "step": 1970,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 151.6153846153846,
      "grad_norm": 0.3997062146663666,
      "learning_rate": 0.01,
      "loss": 4.54901123046875,
      "num_input_tokens_seen": 7825576,
      "step": 1971,
      "train_runtime": 1600.4709,
      "train_tokens_per_second": 4889.546
    },
    {
      "epoch": 151.6153846153846,
      "num_input_tokens_seen": 7825576,
      "step": 1971,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 151.69230769230768,
      "grad_norm": 1.5198040008544922,
      "learning_rate": 0.01,
      "loss": 4.561570167541504,
      "num_input_tokens_seen": 7829672,
      "step": 1972,
      "train_runtime": 1601.2849,
      "train_tokens_per_second": 4889.618
    },
    {
      "epoch": 151.69230769230768,
      "num_input_tokens_seen": 7829672,
      "step": 1972,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 151.76923076923077,
      "grad_norm": 0.2400762289762497,
      "learning_rate": 0.01,
      "loss": 4.562237739562988,
      "num_input_tokens_seen": 7833768,
      "step": 1973,
      "train_runtime": 1602.0994,
      "train_tokens_per_second": 4889.689
    },
    {
      "epoch": 151.76923076923077,
      "num_input_tokens_seen": 7833768,
      "step": 1973,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 151.84615384615384,
      "grad_norm": 0.38551050424575806,
      "learning_rate": 0.01,
      "loss": 4.566306114196777,
      "num_input_tokens_seen": 7837864,
      "step": 1974,
      "train_runtime": 1602.9143,
      "train_tokens_per_second": 4889.759
    },
    {
      "epoch": 151.84615384615384,
      "num_input_tokens_seen": 7837864,
      "step": 1974,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 151.92307692307693,
      "grad_norm": 0.2782340347766876,
      "learning_rate": 0.01,
      "loss": 4.544930458068848,
      "num_input_tokens_seen": 7841960,
      "step": 1975,
      "train_runtime": 1603.7244,
      "train_tokens_per_second": 4889.843
    },
    {
      "epoch": 151.92307692307693,
      "num_input_tokens_seen": 7841960,
      "step": 1975,
      "train_accuracy": 0.016286645084619522
    },
    {
      "epoch": 152.0,
      "grad_norm": 0.3305472433567047,
      "learning_rate": 0.01,
      "loss": 4.587853908538818,
      "num_input_tokens_seen": 7844416,
      "step": 1976,
      "train_runtime": 1604.2395,
      "train_tokens_per_second": 4889.804
    },
    {
      "epoch": 152.0,
      "num_input_tokens_seen": 7844416,
      "step": 1976,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 152.07692307692307,
      "grad_norm": 0.5381389856338501,
      "learning_rate": 0.01,
      "loss": 4.540996551513672,
      "num_input_tokens_seen": 7848512,
      "step": 1977,
      "train_runtime": 1605.0668,
      "train_tokens_per_second": 4889.835
    },
    {
      "epoch": 152.07692307692307,
      "num_input_tokens_seen": 7848512,
      "step": 1977,
      "train_accuracy": 0.021484375
    },
    {
      "epoch": 152.15384615384616,
      "grad_norm": 0.2174607664346695,
      "learning_rate": 0.01,
      "loss": 4.534534931182861,
      "num_input_tokens_seen": 7852608,
      "step": 1978,
      "train_runtime": 1605.8822,
      "train_tokens_per_second": 4889.903
    },
    {
      "epoch": 152.15384615384616,
      "num_input_tokens_seen": 7852608,
      "step": 1978,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 152.23076923076923,
      "grad_norm": 0.26913097500801086,
      "learning_rate": 0.01,
      "loss": 4.543195724487305,
      "num_input_tokens_seen": 7856704,
      "step": 1979,
      "train_runtime": 1606.6973,
      "train_tokens_per_second": 4889.971
    },
    {
      "epoch": 152.23076923076923,
      "num_input_tokens_seen": 7856704,
      "step": 1979,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 152.30769230769232,
      "grad_norm": 0.28455042839050293,
      "learning_rate": 0.01,
      "loss": 4.569001197814941,
      "num_input_tokens_seen": 7860800,
      "step": 1980,
      "train_runtime": 1607.5115,
      "train_tokens_per_second": 4890.043
    },
    {
      "epoch": 152.30769230769232,
      "num_input_tokens_seen": 7860800,
      "step": 1980,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 152.3846153846154,
      "grad_norm": 0.37407955527305603,
      "learning_rate": 0.01,
      "loss": 4.567748069763184,
      "num_input_tokens_seen": 7864896,
      "step": 1981,
      "train_runtime": 1608.326,
      "train_tokens_per_second": 4890.113
    },
    {
      "epoch": 152.3846153846154,
      "num_input_tokens_seen": 7864896,
      "step": 1981,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 152.46153846153845,
      "grad_norm": 0.40064266324043274,
      "learning_rate": 0.01,
      "loss": 4.55222225189209,
      "num_input_tokens_seen": 7868992,
      "step": 1982,
      "train_runtime": 1609.1415,
      "train_tokens_per_second": 4890.18
    },
    {
      "epoch": 152.46153846153845,
      "num_input_tokens_seen": 7868992,
      "step": 1982,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 152.53846153846155,
      "grad_norm": 0.25044623017311096,
      "learning_rate": 0.01,
      "loss": 4.574286460876465,
      "num_input_tokens_seen": 7873088,
      "step": 1983,
      "train_runtime": 1609.9555,
      "train_tokens_per_second": 4890.252
    },
    {
      "epoch": 152.53846153846155,
      "num_input_tokens_seen": 7873088,
      "step": 1983,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 152.6153846153846,
      "grad_norm": 0.2382778376340866,
      "learning_rate": 0.01,
      "loss": 4.558467864990234,
      "num_input_tokens_seen": 7877184,
      "step": 1984,
      "train_runtime": 1610.7705,
      "train_tokens_per_second": 4890.32
    },
    {
      "epoch": 152.6153846153846,
      "num_input_tokens_seen": 7877184,
      "step": 1984,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 152.69230769230768,
      "grad_norm": 0.2806725800037384,
      "learning_rate": 0.01,
      "loss": 4.548799514770508,
      "num_input_tokens_seen": 7881280,
      "step": 1985,
      "train_runtime": 1611.5854,
      "train_tokens_per_second": 4890.389
    },
    {
      "epoch": 152.69230769230768,
      "num_input_tokens_seen": 7881280,
      "step": 1985,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 152.76923076923077,
      "grad_norm": 0.24032214283943176,
      "learning_rate": 0.01,
      "loss": 4.5399017333984375,
      "num_input_tokens_seen": 7885376,
      "step": 1986,
      "train_runtime": 1612.3997,
      "train_tokens_per_second": 4890.46
    },
    {
      "epoch": 152.76923076923077,
      "num_input_tokens_seen": 7885376,
      "step": 1986,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 152.84615384615384,
      "grad_norm": 0.2893168032169342,
      "learning_rate": 0.01,
      "loss": 4.548337936401367,
      "num_input_tokens_seen": 7889472,
      "step": 1987,
      "train_runtime": 1613.2143,
      "train_tokens_per_second": 4890.529
    },
    {
      "epoch": 152.84615384615384,
      "num_input_tokens_seen": 7889472,
      "step": 1987,
      "train_accuracy": 0.00390625
    },
    {
      "epoch": 152.92307692307693,
      "grad_norm": 0.34196656942367554,
      "learning_rate": 0.01,
      "loss": 4.555220127105713,
      "num_input_tokens_seen": 7893568,
      "step": 1988,
      "train_runtime": 1614.0252,
      "train_tokens_per_second": 4890.61
    },
    {
      "epoch": 152.92307692307693,
      "num_input_tokens_seen": 7893568,
      "step": 1988,
      "train_accuracy": 0.0065146577544510365
    },
    {
      "epoch": 153.0,
      "grad_norm": 0.3127935826778412,
      "learning_rate": 0.01,
      "loss": 4.5469441413879395,
      "num_input_tokens_seen": 7896024,
      "step": 1989,
      "train_runtime": 1614.5406,
      "train_tokens_per_second": 4890.57
    },
    {
      "epoch": 153.0,
      "num_input_tokens_seen": 7896024,
      "step": 1989,
      "train_accuracy": 0.025390625
    },
    {
      "epoch": 153.07692307692307,
      "grad_norm": 0.2148197442293167,
      "learning_rate": 0.01,
      "loss": 4.529175758361816,
      "num_input_tokens_seen": 7900120,
      "step": 1990,
      "train_runtime": 1615.369,
      "train_tokens_per_second": 4890.598
    },
    {
      "epoch": 153.07692307692307,
      "num_input_tokens_seen": 7900120,
      "step": 1990,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 153.15384615384616,
      "grad_norm": 0.2560349702835083,
      "learning_rate": 0.01,
      "loss": 4.532726287841797,
      "num_input_tokens_seen": 7904216,
      "step": 1991,
      "train_runtime": 1616.1844,
      "train_tokens_per_second": 4890.665
    },
    {
      "epoch": 153.15384615384616,
      "num_input_tokens_seen": 7904216,
      "step": 1991,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 153.23076923076923,
      "grad_norm": 0.19809867441654205,
      "learning_rate": 0.01,
      "loss": 4.539192199707031,
      "num_input_tokens_seen": 7908312,
      "step": 1992,
      "train_runtime": 1616.9998,
      "train_tokens_per_second": 4890.732
    },
    {
      "epoch": 153.23076923076923,
      "num_input_tokens_seen": 7908312,
      "step": 1992,
      "train_accuracy": 0.013671875
    },
    {
      "epoch": 153.30769230769232,
      "grad_norm": 0.29816627502441406,
      "learning_rate": 0.01,
      "loss": 4.561485767364502,
      "num_input_tokens_seen": 7912408,
      "step": 1993,
      "train_runtime": 1617.8159,
      "train_tokens_per_second": 4890.796
    },
    {
      "epoch": 153.30769230769232,
      "num_input_tokens_seen": 7912408,
      "step": 1993,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 153.3846153846154,
      "grad_norm": 0.26226750016212463,
      "learning_rate": 0.01,
      "loss": 4.529763221740723,
      "num_input_tokens_seen": 7916504,
      "step": 1994,
      "train_runtime": 1618.6308,
      "train_tokens_per_second": 4890.864
    },
    {
      "epoch": 153.3846153846154,
      "num_input_tokens_seen": 7916504,
      "step": 1994,
      "train_accuracy": 0.009765625
    },
    {
      "epoch": 153.46153846153845,
      "grad_norm": 0.25887763500213623,
      "learning_rate": 0.01,
      "loss": 4.545772552490234,
      "num_input_tokens_seen": 7920600,
      "step": 1995,
      "train_runtime": 1619.4517,
      "train_tokens_per_second": 4890.915
    },
    {
      "epoch": 153.46153846153845,
      "num_input_tokens_seen": 7920600,
      "step": 1995,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 153.53846153846155,
      "grad_norm": 0.24705550074577332,
      "learning_rate": 0.01,
      "loss": 4.547776699066162,
      "num_input_tokens_seen": 7924696,
      "step": 1996,
      "train_runtime": 1620.2668,
      "train_tokens_per_second": 4890.982
    },
    {
      "epoch": 153.53846153846155,
      "num_input_tokens_seen": 7924696,
      "step": 1996,
      "train_accuracy": 0.017578125
    },
    {
      "epoch": 153.6153846153846,
      "grad_norm": 0.29341304302215576,
      "learning_rate": 0.01,
      "loss": 4.559391021728516,
      "num_input_tokens_seen": 7928792,
      "step": 1997,
      "train_runtime": 1621.083,
      "train_tokens_per_second": 4891.046
    },
    {
      "epoch": 153.6153846153846,
      "num_input_tokens_seen": 7928792,
      "step": 1997,
      "train_accuracy": 0.01171875
    },
    {
      "epoch": 153.69230769230768,
      "grad_norm": 0.276309609413147,
      "learning_rate": 0.01,
      "loss": 4.530332565307617,
      "num_input_tokens_seen": 7932888,
      "step": 1998,
      "train_runtime": 1621.8971,
      "train_tokens_per_second": 4891.117
    },
    {
      "epoch": 153.69230769230768,
      "num_input_tokens_seen": 7932888,
      "step": 1998,
      "train_accuracy": 0.0078125
    },
    {
      "epoch": 153.76923076923077,
      "grad_norm": 0.20727954804897308,
      "learning_rate": 0.01,
      "loss": 4.541828155517578,
      "num_input_tokens_seen": 7936984,
      "step": 1999,
      "train_runtime": 1622.7117,
      "train_tokens_per_second": 4891.186
    },
    {
      "epoch": 153.76923076923077,
      "num_input_tokens_seen": 7936984,
      "step": 1999,
      "train_accuracy": 0.015625
    },
    {
      "epoch": 153.84615384615384,
      "grad_norm": 0.2459501475095749,
      "learning_rate": 0.01,
      "loss": 4.527507781982422,
      "num_input_tokens_seen": 7941080,
      "step": 2000,
      "train_runtime": 1623.527,
      "train_tokens_per_second": 4891.252
    },
    {
      "epoch": 153.84615384615384,
      "eval_CMD_3_branch_eval_accuracy": 0.011211573236889693,
      "eval_CMD_3_branch_eval_loss": 4.569759368896484,
      "eval_CMD_3_branch_eval_runtime": 1.1547,
      "eval_CMD_3_branch_eval_samples_per_second": 2394.655,
      "eval_CMD_3_branch_eval_steps_per_second": 5.196,
      "num_input_tokens_seen": 7941080,
      "step": 2000
    }
  ],
  "logging_steps": 1,
  "max_steps": 1300000000,
  "num_input_tokens_seen": 7941080,
  "num_train_epochs": 100000000,
  "save_steps": 1000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 4.636698676248576e+16,
  "train_batch_size": 512,
  "trial_name": null,
  "trial_params": null
}