{
  "best_global_step": 700,
  "best_metric": 1.2832741731472197e-06,
  "best_model_checkpoint": "qwen2-7b-instruct-trl-sft-newick_extraction_1/checkpoint-700",
  "epoch": 2.4835555555555557,
  "eval_steps": 10,
  "global_step": 700,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.035555555555555556,
      "grad_norm": 4.419339179992676,
      "learning_rate": 0.0002,
      "loss": 14.4699,
      "mean_token_accuracy": 0.5754725567996501,
      "num_tokens": 570240.0,
      "step": 10
    },
    {
      "epoch": 0.035555555555555556,
      "eval_loss": 1.5543439388275146,
      "eval_mean_token_accuracy": 0.6078731870651245,
      "eval_num_tokens": 570240.0,
      "eval_runtime": 648.0159,
      "eval_samples_per_second": 1.543,
      "eval_steps_per_second": 0.386,
      "step": 10
    },
    {
      "epoch": 0.07111111111111111,
      "grad_norm": 9.82264232635498,
      "learning_rate": 0.0002,
      "loss": 10.3302,
      "mean_token_accuracy": 0.6614481717348099,
      "num_tokens": 1140480.0,
      "step": 20
    },
    {
      "epoch": 0.07111111111111111,
      "eval_loss": 0.9738740921020508,
      "eval_mean_token_accuracy": 0.7448731756210327,
      "eval_num_tokens": 1140480.0,
      "eval_runtime": 630.2188,
      "eval_samples_per_second": 1.587,
      "eval_steps_per_second": 0.397,
      "step": 20
    },
    {
      "epoch": 0.10666666666666667,
      "grad_norm": 11.91273307800293,
      "learning_rate": 0.0002,
      "loss": 5.6105,
      "mean_token_accuracy": 0.8060442075133324,
      "num_tokens": 1710720.0,
      "step": 30
    },
    {
      "epoch": 0.10666666666666667,
      "eval_loss": 0.3900993764400482,
      "eval_mean_token_accuracy": 0.8829268217086792,
      "eval_num_tokens": 1710720.0,
      "eval_runtime": 631.6971,
      "eval_samples_per_second": 1.583,
      "eval_steps_per_second": 0.396,
      "step": 30
    },
    {
      "epoch": 0.14222222222222222,
      "grad_norm": 6.84392786026001,
      "learning_rate": 0.0002,
      "loss": 1.7204,
      "mean_token_accuracy": 0.943353658914566,
      "num_tokens": 2280960.0,
      "step": 40
    },
    {
      "epoch": 0.14222222222222222,
      "eval_loss": 0.06896545737981796,
      "eval_mean_token_accuracy": 0.9878048896789551,
      "eval_num_tokens": 2280960.0,
      "eval_runtime": 627.3952,
      "eval_samples_per_second": 1.594,
      "eval_steps_per_second": 0.398,
      "step": 40
    },
    {
      "epoch": 0.17777777777777778,
      "grad_norm": 0.21437789499759674,
      "learning_rate": 0.0002,
      "loss": 0.3396,
      "mean_token_accuracy": 0.9958536624908447,
      "num_tokens": 2851200.0,
      "step": 50
    },
    {
      "epoch": 0.17777777777777778,
      "eval_loss": 0.03179474174976349,
      "eval_mean_token_accuracy": 0.997560977935791,
      "eval_num_tokens": 2851200.0,
      "eval_runtime": 631.7748,
      "eval_samples_per_second": 1.583,
      "eval_steps_per_second": 0.396,
      "step": 50
    },
    {
      "epoch": 0.21333333333333335,
      "grad_norm": 1.2222466468811035,
      "learning_rate": 0.0002,
      "loss": 0.2231,
      "mean_token_accuracy": 0.997560977935791,
      "num_tokens": 3421440.0,
      "step": 60
    },
    {
      "epoch": 0.21333333333333335,
      "eval_loss": 0.017346156761050224,
      "eval_mean_token_accuracy": 0.997560977935791,
      "eval_num_tokens": 3421440.0,
      "eval_runtime": 627.3871,
      "eval_samples_per_second": 1.594,
      "eval_steps_per_second": 0.398,
      "step": 60
    },
    {
      "epoch": 0.24888888888888888,
      "grad_norm": 1.549639105796814,
      "learning_rate": 0.0002,
      "loss": 0.0636,
      "mean_token_accuracy": 0.9984908550977707,
      "num_tokens": 3991680.0,
      "step": 70
    },
    {
      "epoch": 0.24888888888888888,
      "eval_loss": 0.0006659275386482477,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 3991680.0,
      "eval_runtime": 627.7647,
      "eval_samples_per_second": 1.593,
      "eval_steps_per_second": 0.398,
      "step": 70
    },
    {
      "epoch": 0.28444444444444444,
      "grad_norm": 0.034244876354932785,
      "learning_rate": 0.0002,
      "loss": 0.0017,
      "mean_token_accuracy": 1.0,
      "num_tokens": 4561920.0,
      "step": 80
    },
    {
      "epoch": 0.28444444444444444,
      "eval_loss": 4.906835238216445e-05,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 4561920.0,
      "eval_runtime": 627.866,
      "eval_samples_per_second": 1.593,
      "eval_steps_per_second": 0.398,
      "step": 80
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.004514188971370459,
      "learning_rate": 0.0002,
      "loss": 0.0003,
      "mean_token_accuracy": 1.0,
      "num_tokens": 5132160.0,
      "step": 90
    },
    {
      "epoch": 0.32,
      "eval_loss": 2.5992721930379048e-05,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 5132160.0,
      "eval_runtime": 630.504,
      "eval_samples_per_second": 1.586,
      "eval_steps_per_second": 0.397,
      "step": 90
    },
    {
      "epoch": 0.35555555555555557,
      "grad_norm": 0.0026128594763576984,
      "learning_rate": 0.0002,
      "loss": 0.0002,
      "mean_token_accuracy": 1.0,
      "num_tokens": 5702400.0,
      "step": 100
    },
    {
      "epoch": 0.35555555555555557,
      "eval_loss": 1.878049260994885e-05,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 5702400.0,
      "eval_runtime": 628.0227,
      "eval_samples_per_second": 1.592,
      "eval_steps_per_second": 0.398,
      "step": 100
    },
    {
      "epoch": 0.39111111111111113,
      "grad_norm": 0.0018496178090572357,
      "learning_rate": 0.0002,
      "loss": 0.0001,
      "mean_token_accuracy": 1.0,
      "num_tokens": 6272640.0,
      "step": 110
    },
    {
      "epoch": 0.39111111111111113,
      "eval_loss": 1.4819584976066835e-05,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 6272640.0,
      "eval_runtime": 632.337,
      "eval_samples_per_second": 1.581,
      "eval_steps_per_second": 0.395,
      "step": 110
    },
    {
      "epoch": 0.4266666666666667,
      "grad_norm": 0.001428118092007935,
      "learning_rate": 0.0002,
      "loss": 0.0001,
      "mean_token_accuracy": 1.0,
      "num_tokens": 6842880.0,
      "step": 120
    },
    {
      "epoch": 0.4266666666666667,
      "eval_loss": 1.2422192412486766e-05,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 6842880.0,
      "eval_runtime": 627.3485,
      "eval_samples_per_second": 1.594,
      "eval_steps_per_second": 0.399,
      "step": 120
    },
    {
      "epoch": 0.4622222222222222,
      "grad_norm": 0.0011839906219393015,
      "learning_rate": 0.0002,
      "loss": 0.0001,
      "mean_token_accuracy": 1.0,
      "num_tokens": 7413120.0,
      "step": 130
    },
    {
      "epoch": 0.4622222222222222,
      "eval_loss": 1.0789405678224284e-05,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 7413120.0,
      "eval_runtime": 630.5032,
      "eval_samples_per_second": 1.586,
      "eval_steps_per_second": 0.397,
      "step": 130
    },
    {
      "epoch": 0.49777777777777776,
      "grad_norm": 0.001027945545502007,
      "learning_rate": 0.0002,
      "loss": 0.0001,
      "mean_token_accuracy": 1.0,
      "num_tokens": 7983360.0,
      "step": 140
    },
    {
      "epoch": 0.49777777777777776,
      "eval_loss": 9.804618457565084e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 7983360.0,
      "eval_runtime": 626.6713,
      "eval_samples_per_second": 1.596,
      "eval_steps_per_second": 0.399,
      "step": 140
    },
    {
      "epoch": 0.5333333333333333,
      "grad_norm": 0.0009170864359475672,
      "learning_rate": 0.0002,
      "loss": 0.0001,
      "mean_token_accuracy": 1.0,
      "num_tokens": 8553600.0,
      "step": 150
    },
    {
      "epoch": 0.5333333333333333,
      "eval_loss": 8.696494660398457e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 8553600.0,
      "eval_runtime": 630.4604,
      "eval_samples_per_second": 1.586,
      "eval_steps_per_second": 0.397,
      "step": 150
    },
    {
      "epoch": 0.5688888888888889,
      "grad_norm": 0.0008168203639797866,
      "learning_rate": 0.0002,
      "loss": 0.0001,
      "mean_token_accuracy": 1.0,
      "num_tokens": 9123840.0,
      "step": 160
    },
    {
      "epoch": 0.5688888888888889,
      "eval_loss": 7.992691280378494e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 9123840.0,
      "eval_runtime": 628.5297,
      "eval_samples_per_second": 1.591,
      "eval_steps_per_second": 0.398,
      "step": 160
    },
    {
      "epoch": 0.6044444444444445,
      "grad_norm": 0.0007576971547678113,
      "learning_rate": 0.0002,
      "loss": 0.0001,
      "mean_token_accuracy": 1.0,
      "num_tokens": 9694080.0,
      "step": 170
    },
    {
      "epoch": 0.6044444444444445,
      "eval_loss": 7.28438317310065e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 9694080.0,
      "eval_runtime": 626.7184,
      "eval_samples_per_second": 1.596,
      "eval_steps_per_second": 0.399,
      "step": 170
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.0006879908032715321,
      "learning_rate": 0.0002,
      "loss": 0.0001,
      "mean_token_accuracy": 1.0,
      "num_tokens": 10264320.0,
      "step": 180
    },
    {
      "epoch": 0.64,
      "eval_loss": 6.8836661739624105e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 10264320.0,
      "eval_runtime": 756.3063,
      "eval_samples_per_second": 1.322,
      "eval_steps_per_second": 0.331,
      "step": 180
    },
    {
      "epoch": 0.6755555555555556,
      "grad_norm": 0.0006444182363338768,
      "learning_rate": 0.0002,
      "loss": 0.0001,
      "mean_token_accuracy": 1.0,
      "num_tokens": 10834560.0,
      "step": 190
    },
    {
      "epoch": 0.6755555555555556,
      "eval_loss": 6.4065607148222625e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 10834560.0,
      "eval_runtime": 628.8265,
      "eval_samples_per_second": 1.59,
      "eval_steps_per_second": 0.398,
      "step": 190
    },
    {
      "epoch": 0.7111111111111111,
      "grad_norm": 0.0005997396074235439,
      "learning_rate": 0.0002,
      "loss": 0.0001,
      "mean_token_accuracy": 1.0,
      "num_tokens": 11404800.0,
      "step": 200
    },
    {
      "epoch": 0.7111111111111111,
      "eval_loss": 6.104866770328954e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 11404800.0,
      "eval_runtime": 688.2367,
      "eval_samples_per_second": 1.453,
      "eval_steps_per_second": 0.363,
      "step": 200
    },
    {
      "epoch": 0.7466666666666667,
      "grad_norm": 0.0005526828463189304,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 11975040.0,
      "step": 210
    },
    {
      "epoch": 0.7466666666666667,
      "eval_loss": 5.683682047674665e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 11975040.0,
      "eval_runtime": 754.3494,
      "eval_samples_per_second": 1.326,
      "eval_steps_per_second": 0.331,
      "step": 210
    },
    {
      "epoch": 0.7822222222222223,
      "grad_norm": 0.0005140851717442274,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 12545280.0,
      "step": 220
    },
    {
      "epoch": 0.7822222222222223,
      "eval_loss": 5.339933522918727e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 12545280.0,
      "eval_runtime": 647.6999,
      "eval_samples_per_second": 1.544,
      "eval_steps_per_second": 0.386,
      "step": 220
    },
    {
      "epoch": 0.8177777777777778,
      "grad_norm": 0.0004886530223302543,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 13115520.0,
      "step": 230
    },
    {
      "epoch": 0.8177777777777778,
      "eval_loss": 5.0328567340329755e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 13115520.0,
      "eval_runtime": 631.1227,
      "eval_samples_per_second": 1.584,
      "eval_steps_per_second": 0.396,
      "step": 230
    },
    {
      "epoch": 0.8533333333333334,
      "grad_norm": 0.0004592030309140682,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 13685760.0,
      "step": 240
    },
    {
      "epoch": 0.8533333333333334,
      "eval_loss": 4.829912541026715e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 13685760.0,
      "eval_runtime": 631.5138,
      "eval_samples_per_second": 1.583,
      "eval_steps_per_second": 0.396,
      "step": 240
    },
    {
      "epoch": 0.8888888888888888,
      "grad_norm": 0.0004404531209729612,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 14256000.0,
      "step": 250
    },
    {
      "epoch": 0.8888888888888888,
      "eval_loss": 4.590192020259565e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 14256000.0,
      "eval_runtime": 631.121,
      "eval_samples_per_second": 1.584,
      "eval_steps_per_second": 0.396,
      "step": 250
    },
    {
      "epoch": 0.9244444444444444,
      "grad_norm": 0.00041581166442483664,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 14826240.0,
      "step": 260
    },
    {
      "epoch": 0.9244444444444444,
      "eval_loss": 4.422760412126081e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 14826240.0,
      "eval_runtime": 626.5972,
      "eval_samples_per_second": 1.596,
      "eval_steps_per_second": 0.399,
      "step": 260
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.00039238750468939543,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 15396480.0,
      "step": 270
    },
    {
      "epoch": 0.96,
      "eval_loss": 4.159514901402872e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 15396480.0,
      "eval_runtime": 626.6804,
      "eval_samples_per_second": 1.596,
      "eval_steps_per_second": 0.399,
      "step": 270
    },
    {
      "epoch": 0.9955555555555555,
      "grad_norm": 0.0003751284384634346,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 15966720.0,
      "step": 280
    },
    {
      "epoch": 0.9955555555555555,
      "eval_loss": 4.007378265669104e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 15966720.0,
      "eval_runtime": 626.6082,
      "eval_samples_per_second": 1.596,
      "eval_steps_per_second": 0.399,
      "step": 280
    },
    {
      "epoch": 1.0284444444444445,
      "grad_norm": 0.0003600440395530313,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 16492410.0,
      "step": 290
    },
    {
      "epoch": 1.0284444444444445,
      "eval_loss": 3.86876126867719e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 16492410.0,
      "eval_runtime": 630.0967,
      "eval_samples_per_second": 1.587,
      "eval_steps_per_second": 0.397,
      "step": 290
    },
    {
      "epoch": 1.064,
      "grad_norm": 0.00035800470504909754,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 17062650.0,
      "step": 300
    },
    {
      "epoch": 1.064,
      "eval_loss": 3.728666797542246e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 17062650.0,
      "eval_runtime": 630.6538,
      "eval_samples_per_second": 1.586,
      "eval_steps_per_second": 0.396,
      "step": 300
    },
    {
      "epoch": 1.0995555555555556,
      "grad_norm": 0.0003293608024250716,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 17632890.0,
      "step": 310
    },
    {
      "epoch": 1.0995555555555556,
      "eval_loss": 3.5069003843091195e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 17632890.0,
      "eval_runtime": 626.3062,
      "eval_samples_per_second": 1.597,
      "eval_steps_per_second": 0.399,
      "step": 310
    },
    {
      "epoch": 1.1351111111111112,
      "grad_norm": 0.000320440623909235,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 18203130.0,
      "step": 320
    },
    {
      "epoch": 1.1351111111111112,
      "eval_loss": 3.4313563901378075e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 18203130.0,
      "eval_runtime": 626.8382,
      "eval_samples_per_second": 1.595,
      "eval_steps_per_second": 0.399,
      "step": 320
    },
    {
      "epoch": 1.1706666666666667,
      "grad_norm": 0.00031058763852342963,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 18773370.0,
      "step": 330
    },
    {
      "epoch": 1.1706666666666667,
      "eval_loss": 3.2939326501946198e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 18773370.0,
      "eval_runtime": 631.1457,
      "eval_samples_per_second": 1.584,
      "eval_steps_per_second": 0.396,
      "step": 330
    },
    {
      "epoch": 1.2062222222222223,
      "grad_norm": 0.0002935119846370071,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 19343610.0,
      "step": 340
    },
    {
      "epoch": 1.2062222222222223,
      "eval_loss": 3.1630906960344873e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 19343610.0,
      "eval_runtime": 631.1257,
      "eval_samples_per_second": 1.584,
      "eval_steps_per_second": 0.396,
      "step": 340
    },
    {
      "epoch": 1.2417777777777779,
      "grad_norm": 0.00028090347768738866,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 19913850.0,
      "step": 350
    },
    {
      "epoch": 1.2417777777777779,
      "eval_loss": 3.1076663162821205e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 19913850.0,
      "eval_runtime": 631.1307,
      "eval_samples_per_second": 1.584,
      "eval_steps_per_second": 0.396,
      "step": 350
    },
    {
      "epoch": 1.2773333333333334,
      "grad_norm": 0.000271222204901278,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 20484090.0,
      "step": 360
    },
    {
      "epoch": 1.2773333333333334,
      "eval_loss": 2.9442144295899197e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 20484090.0,
      "eval_runtime": 631.362,
      "eval_samples_per_second": 1.584,
      "eval_steps_per_second": 0.396,
      "step": 360
    },
    {
      "epoch": 1.3128888888888888,
      "grad_norm": 0.0002641026512719691,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 21054330.0,
      "step": 370
    },
    {
      "epoch": 1.3128888888888888,
      "eval_loss": 2.8731856218655594e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 21054330.0,
      "eval_runtime": 626.4927,
      "eval_samples_per_second": 1.596,
      "eval_steps_per_second": 0.399,
      "step": 370
    },
    {
      "epoch": 1.3484444444444446,
      "grad_norm": 0.0002599551226012409,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 21624570.0,
      "step": 380
    },
    {
      "epoch": 1.3484444444444446,
      "eval_loss": 2.7712894734577276e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 21624570.0,
      "eval_runtime": 626.4481,
      "eval_samples_per_second": 1.596,
      "eval_steps_per_second": 0.399,
      "step": 380
    },
    {
      "epoch": 1.384,
      "grad_norm": 0.0002494712534826249,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 22194810.0,
      "step": 390
    },
    {
      "epoch": 1.384,
      "eval_loss": 2.666238515303121e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 22194810.0,
      "eval_runtime": 626.3873,
      "eval_samples_per_second": 1.596,
      "eval_steps_per_second": 0.399,
      "step": 390
    },
    {
      "epoch": 1.4195555555555557,
      "grad_norm": 0.00024175533326342702,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 22765050.0,
      "step": 400
    },
    {
      "epoch": 1.4195555555555557,
      "eval_loss": 2.613794549688464e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 22765050.0,
      "eval_runtime": 625.5663,
      "eval_samples_per_second": 1.599,
      "eval_steps_per_second": 0.4,
      "step": 400
    },
    {
      "epoch": 1.455111111111111,
      "grad_norm": 0.0002301511267432943,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 23335290.0,
      "step": 410
    },
    {
      "epoch": 1.455111111111111,
      "eval_loss": 2.55880513577722e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 23335290.0,
      "eval_runtime": 630.8138,
      "eval_samples_per_second": 1.585,
      "eval_steps_per_second": 0.396,
      "step": 410
    },
    {
      "epoch": 1.4906666666666666,
      "grad_norm": 0.00023040513042360544,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 23905530.0,
      "step": 420
    },
    {
      "epoch": 1.4906666666666666,
      "eval_loss": 2.4679234229552094e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 23905530.0,
      "eval_runtime": 630.4552,
      "eval_samples_per_second": 1.586,
      "eval_steps_per_second": 0.397,
      "step": 420
    },
    {
      "epoch": 1.5262222222222221,
      "grad_norm": 0.0002182941243518144,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 24475770.0,
      "step": 430
    },
    {
      "epoch": 1.5262222222222221,
      "eval_loss": 2.395064711890882e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 24475770.0,
      "eval_runtime": 627.55,
      "eval_samples_per_second": 1.593,
      "eval_steps_per_second": 0.398,
      "step": 430
    },
    {
      "epoch": 1.561777777777778,
      "grad_norm": 0.00021538310102187097,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 25046010.0,
      "step": 440
    },
    {
      "epoch": 1.561777777777778,
      "eval_loss": 2.3233274077938404e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 25046010.0,
      "eval_runtime": 626.3834,
      "eval_samples_per_second": 1.596,
      "eval_steps_per_second": 0.399,
      "step": 440
    },
    {
      "epoch": 1.5973333333333333,
      "grad_norm": 0.000205744436243549,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 25616250.0,
      "step": 450
    },
    {
      "epoch": 1.5973333333333333,
      "eval_loss": 2.2373139927367447e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 25616250.0,
      "eval_runtime": 626.4713,
      "eval_samples_per_second": 1.596,
      "eval_steps_per_second": 0.399,
      "step": 450
    },
    {
      "epoch": 1.6328888888888888,
      "grad_norm": 0.00020622971351258457,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 26186490.0,
      "step": 460
    },
    {
      "epoch": 1.6328888888888888,
      "eval_loss": 2.2288552372629056e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 26186490.0,
      "eval_runtime": 629.9221,
      "eval_samples_per_second": 1.587,
      "eval_steps_per_second": 0.397,
      "step": 460
    },
    {
      "epoch": 1.6684444444444444,
      "grad_norm": 0.0001915807370096445,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 26756730.0,
      "step": 470
    },
    {
      "epoch": 1.6684444444444444,
      "eval_loss": 2.1382338672992773e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 26756730.0,
      "eval_runtime": 631.3677,
      "eval_samples_per_second": 1.584,
      "eval_steps_per_second": 0.396,
      "step": 470
    },
    {
      "epoch": 1.704,
      "grad_norm": 0.0001857055613072589,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 27326970.0,
      "step": 480
    },
    {
      "epoch": 1.704,
      "eval_loss": 2.078363195323618e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 27326970.0,
      "eval_runtime": 626.5474,
      "eval_samples_per_second": 1.596,
      "eval_steps_per_second": 0.399,
      "step": 480
    },
    {
      "epoch": 1.7395555555555555,
      "grad_norm": 0.0001786583598004654,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 27897210.0,
      "step": 490
    },
    {
      "epoch": 1.7395555555555555,
      "eval_loss": 2.0040279196109623e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 27897210.0,
      "eval_runtime": 630.3478,
      "eval_samples_per_second": 1.586,
      "eval_steps_per_second": 0.397,
      "step": 490
    },
    {
      "epoch": 1.775111111111111,
      "grad_norm": 0.00017898347869049758,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 28467450.0,
      "step": 500
    },
    {
      "epoch": 1.775111111111111,
      "eval_loss": 1.946991005752352e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 28467450.0,
      "eval_runtime": 630.0683,
      "eval_samples_per_second": 1.587,
      "eval_steps_per_second": 0.397,
      "step": 500
    },
    {
      "epoch": 1.8106666666666666,
      "grad_norm": 0.00017079812823794782,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 29037690.0,
      "step": 510
    },
    {
      "epoch": 1.8106666666666666,
      "eval_loss": 1.950623527591233e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 29037690.0,
      "eval_runtime": 626.6265,
      "eval_samples_per_second": 1.596,
      "eval_steps_per_second": 0.399,
      "step": 510
    },
    {
      "epoch": 1.8462222222222222,
      "grad_norm": 0.00016901094932109118,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 29607930.0,
      "step": 520
    },
    {
      "epoch": 1.8462222222222222,
      "eval_loss": 1.898369532682409e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 29607930.0,
      "eval_runtime": 626.583,
      "eval_samples_per_second": 1.596,
      "eval_steps_per_second": 0.399,
      "step": 520
    },
    {
      "epoch": 1.8817777777777778,
      "grad_norm": 0.00017019332153722644,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 30178170.0,
      "step": 530
    },
    {
      "epoch": 1.8817777777777778,
      "eval_loss": 1.8598769884192734e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 30178170.0,
      "eval_runtime": 626.4907,
      "eval_samples_per_second": 1.596,
      "eval_steps_per_second": 0.399,
      "step": 530
    },
    {
      "epoch": 1.9173333333333333,
      "grad_norm": 0.0001607046724529937,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 30748410.0,
      "step": 540
    },
    {
      "epoch": 1.9173333333333333,
      "eval_loss": 1.798646053430275e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 30748410.0,
      "eval_runtime": 626.1949,
      "eval_samples_per_second": 1.597,
      "eval_steps_per_second": 0.399,
      "step": 540
    },
    {
      "epoch": 1.952888888888889,
      "grad_norm": 0.00015570198593195528,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 31318650.0,
      "step": 550
    },
    {
      "epoch": 1.952888888888889,
      "eval_loss": 1.7364616269333055e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 31318650.0,
      "eval_runtime": 631.0041,
      "eval_samples_per_second": 1.585,
      "eval_steps_per_second": 0.396,
      "step": 550
    },
    {
      "epoch": 1.9884444444444445,
      "grad_norm": 0.00015243745292536914,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 31888890.0,
      "step": 560
    },
    {
      "epoch": 1.9884444444444445,
      "eval_loss": 1.7059577430700301e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 31888890.0,
      "eval_runtime": 628.2317,
      "eval_samples_per_second": 1.592,
      "eval_steps_per_second": 0.398,
      "step": 560
    },
    {
      "epoch": 2.021333333333333,
      "grad_norm": 0.00014857217320241034,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 32414580.0,
      "step": 570
    },
    {
      "epoch": 2.021333333333333,
      "eval_loss": 1.6856015463417862e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 32414580.0,
      "eval_runtime": 1329.2357,
      "eval_samples_per_second": 0.752,
      "eval_steps_per_second": 0.188,
      "step": 570
    },
    {
      "epoch": 2.056888888888889,
      "grad_norm": 0.00014311449194792658,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 32984820.0,
      "step": 580
    },
    {
      "epoch": 2.056888888888889,
      "eval_loss": 1.6173969470401062e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 32984820.0,
      "eval_runtime": 1323.5911,
      "eval_samples_per_second": 0.756,
      "eval_steps_per_second": 0.189,
      "step": 580
    },
    {
      "epoch": 2.0924444444444443,
      "grad_norm": 0.0001415183360222727,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 33555060.0,
      "step": 590
    },
    {
      "epoch": 2.0924444444444443,
      "eval_loss": 1.6115862990773167e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 33555060.0,
      "eval_runtime": 1365.3901,
      "eval_samples_per_second": 0.732,
      "eval_steps_per_second": 0.183,
      "step": 590
    },
    {
      "epoch": 2.128,
      "grad_norm": 0.0001429204421583563,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 34125300.0,
      "step": 600
    },
    {
      "epoch": 2.128,
      "eval_loss": 1.558861640660325e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 34125300.0,
      "eval_runtime": 1397.439,
      "eval_samples_per_second": 0.716,
      "eval_steps_per_second": 0.179,
      "step": 600
    },
    {
      "epoch": 2.1635555555555555,
      "grad_norm": 0.0001378554297843948,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 34695540.0,
      "step": 610
    },
    {
      "epoch": 2.1635555555555555,
      "eval_loss": 1.5426779782501399e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 34695540.0,
      "eval_runtime": 1377.4259,
      "eval_samples_per_second": 0.726,
      "eval_steps_per_second": 0.181,
      "step": 610
    },
    {
      "epoch": 2.1991111111111112,
      "grad_norm": 0.00013565713015850633,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 35265780.0,
      "step": 620
    },
    {
      "epoch": 2.1991111111111112,
      "eval_loss": 1.5140946061364957e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 35265780.0,
      "eval_runtime": 1382.8385,
      "eval_samples_per_second": 0.723,
      "eval_steps_per_second": 0.181,
      "step": 620
    },
    {
      "epoch": 2.2346666666666666,
      "grad_norm": 0.00012901326408609748,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 35836020.0,
      "step": 630
    },
    {
      "epoch": 2.2346666666666666,
      "eval_loss": 1.4765602145416779e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 35836020.0,
      "eval_runtime": 1383.5639,
      "eval_samples_per_second": 0.723,
      "eval_steps_per_second": 0.181,
      "step": 630
    },
    {
      "epoch": 2.2702222222222224,
      "grad_norm": 0.00012723123654723167,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 36406260.0,
      "step": 640
    },
    {
      "epoch": 2.2702222222222224,
      "eval_loss": 1.4442689462157432e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 36406260.0,
      "eval_runtime": 1327.4282,
      "eval_samples_per_second": 0.753,
      "eval_steps_per_second": 0.188,
      "step": 640
    },
    {
      "epoch": 2.3057777777777777,
      "grad_norm": 0.00012795526708941907,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 36976500.0,
      "step": 650
    },
    {
      "epoch": 2.3057777777777777,
      "eval_loss": 1.4113712722974014e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 36976500.0,
      "eval_runtime": 1027.3957,
      "eval_samples_per_second": 0.973,
      "eval_steps_per_second": 0.243,
      "step": 650
    },
    {
      "epoch": 2.3413333333333335,
      "grad_norm": 0.0001269335625693202,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 37546740.0,
      "step": 660
    },
    {
      "epoch": 2.3413333333333335,
      "eval_loss": 1.3910276948081446e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 37546740.0,
      "eval_runtime": 850.9645,
      "eval_samples_per_second": 1.175,
      "eval_steps_per_second": 0.294,
      "step": 660
    },
    {
      "epoch": 2.376888888888889,
      "grad_norm": 0.00012149213580414653,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 38116980.0,
      "step": 670
    },
    {
      "epoch": 2.376888888888889,
      "eval_loss": 1.3603197430711589e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 38116980.0,
      "eval_runtime": 826.5336,
      "eval_samples_per_second": 1.21,
      "eval_steps_per_second": 0.302,
      "step": 670
    },
    {
      "epoch": 2.4124444444444446,
      "grad_norm": 0.00012022069131489843,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 38687220.0,
      "step": 680
    },
    {
      "epoch": 2.4124444444444446,
      "eval_loss": 1.3281172641654848e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 38687220.0,
      "eval_runtime": 629.2123,
      "eval_samples_per_second": 1.589,
      "eval_steps_per_second": 0.397,
      "step": 680
    },
    {
      "epoch": 2.448,
      "grad_norm": 0.00011441825336078182,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 39257460.0,
      "step": 690
    },
    {
      "epoch": 2.448,
      "eval_loss": 1.289061628995114e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 39257460.0,
      "eval_runtime": 626.5287,
      "eval_samples_per_second": 1.596,
      "eval_steps_per_second": 0.399,
      "step": 690
    },
    {
      "epoch": 2.4835555555555557,
      "grad_norm": 0.00011212020763196051,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 39827700.0,
      "step": 700
    },
    {
      "epoch": 2.4835555555555557,
      "eval_loss": 1.2832741731472197e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 39827700.0,
      "eval_runtime": 626.6259,
      "eval_samples_per_second": 1.596,
      "eval_steps_per_second": 0.399,
      "step": 700
    }
  ],
  "logging_steps": 10,
  "max_steps": 846,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 20,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.8517255159974912e+18,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}