{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 2096,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 2.0116909193992614,
      "epoch": 0.047744091668656,
      "grad_norm": 3.390625,
      "learning_rate": 1.0315789473684213e-05,
      "loss": 2.0623187255859374,
      "mean_token_accuracy": 0.589528747946024,
      "num_tokens": 62415.0,
      "step": 50
    },
    {
      "entropy": 1.7590599358081818,
      "epoch": 0.095488183337312,
      "grad_norm": 3.078125,
      "learning_rate": 1.9999915067429385e-05,
      "loss": 1.754022216796875,
      "mean_token_accuracy": 0.6332476472854615,
      "num_tokens": 126004.0,
      "step": 100
    },
    {
      "entropy": 1.7224605268239974,
      "epoch": 0.14323227500596802,
      "grad_norm": 3.203125,
      "learning_rate": 1.9984525009996967e-05,
      "loss": 1.7351203918457032,
      "mean_token_accuracy": 0.6340418252348899,
      "num_tokens": 188614.0,
      "step": 150
    },
    {
      "entropy": 1.6983880323171616,
      "epoch": 0.190976366674624,
      "grad_norm": 2.984375,
      "learning_rate": 1.9942640420376764e-05,
      "loss": 1.7146597290039063,
      "mean_token_accuracy": 0.6401417210698128,
      "num_tokens": 251062.0,
      "step": 200
    },
    {
      "entropy": 1.6483143484592437,
      "epoch": 0.23872045834328,
      "grad_norm": 2.734375,
      "learning_rate": 1.9874372441823628e-05,
      "loss": 1.658107147216797,
      "mean_token_accuracy": 0.6478794780373573,
      "num_tokens": 313957.0,
      "step": 250
    },
    {
      "entropy": 1.6347757685184479,
      "epoch": 0.28646455001193605,
      "grad_norm": 2.703125,
      "learning_rate": 1.9779902227487397e-05,
      "loss": 1.6550674438476562,
      "mean_token_accuracy": 0.6487274625897408,
      "num_tokens": 376779.0,
      "step": 300
    },
    {
      "entropy": 1.5762060970067977,
      "epoch": 0.334208641680592,
      "grad_norm": 2.34375,
      "learning_rate": 1.9659480459712203e-05,
      "loss": 1.5960990905761718,
      "mean_token_accuracy": 0.6587067684531211,
      "num_tokens": 439472.0,
      "step": 350
    },
    {
      "entropy": 1.534512039422989,
      "epoch": 0.381952733349248,
      "grad_norm": 2.5625,
      "learning_rate": 1.9513426684835922e-05,
      "loss": 1.55486572265625,
      "mean_token_accuracy": 0.6660523465275765,
      "num_tokens": 501279.0,
      "step": 400
    },
    {
      "entropy": 1.5402596753835678,
      "epoch": 0.429696825017904,
      "grad_norm": 2.53125,
      "learning_rate": 1.9342128465254943e-05,
      "loss": 1.5535296630859374,
      "mean_token_accuracy": 0.6673906499147415,
      "num_tokens": 564442.0,
      "step": 450
    },
    {
      "entropy": 1.5164723497629167,
      "epoch": 0.47744091668656,
      "grad_norm": 2.34375,
      "learning_rate": 1.9146040351004225e-05,
      "loss": 1.5373231506347655,
      "mean_token_accuracy": 0.670901304781437,
      "num_tokens": 627102.0,
      "step": 500
    },
    {
      "entropy": 1.4891952139139175,
      "epoch": 0.5251850083552161,
      "grad_norm": 2.296875,
      "learning_rate": 1.8925682673581706e-05,
      "loss": 1.5036320495605469,
      "mean_token_accuracy": 0.6744937270879745,
      "num_tokens": 689421.0,
      "step": 550
    },
    {
      "entropy": 1.488253089785576,
      "epoch": 0.5729291000238721,
      "grad_norm": 2.15625,
      "learning_rate": 1.868164016521766e-05,
      "loss": 1.5044355773925782,
      "mean_token_accuracy": 0.6755094349384307,
      "num_tokens": 752388.0,
      "step": 600
    },
    {
      "entropy": 1.4446531629562378,
      "epoch": 0.6206731916925281,
      "grad_norm": 2.34375,
      "learning_rate": 1.8414560407252874e-05,
      "loss": 1.4617152404785156,
      "mean_token_accuracy": 0.6805451279878616,
      "num_tokens": 815803.0,
      "step": 650
    },
    {
      "entropy": 1.4555100780725478,
      "epoch": 0.668417283361184,
      "grad_norm": 2.109375,
      "learning_rate": 1.8125152111742944e-05,
      "loss": 1.469644775390625,
      "mean_token_accuracy": 0.6796707132458687,
      "num_tokens": 878686.0,
      "step": 700
    },
    {
      "entropy": 1.414481201171875,
      "epoch": 0.71616137502984,
      "grad_norm": 2.140625,
      "learning_rate": 1.7814183240848635e-05,
      "loss": 1.4264332580566406,
      "mean_token_accuracy": 0.6893567860126495,
      "num_tokens": 940629.0,
      "step": 750
    },
    {
      "entropy": 1.3958445823192596,
      "epoch": 0.763905466698496,
      "grad_norm": 2.34375,
      "learning_rate": 1.7482478969002487e-05,
      "loss": 1.4121397399902345,
      "mean_token_accuracy": 0.6902168181538582,
      "num_tokens": 1002810.0,
      "step": 800
    },
    {
      "entropy": 1.3906721925735475,
      "epoch": 0.811649558367152,
      "grad_norm": 2.421875,
      "learning_rate": 1.7130919493259284e-05,
      "loss": 1.3895584106445313,
      "mean_token_accuracy": 0.6943912228941918,
      "num_tokens": 1067041.0,
      "step": 850
    },
    {
      "entropy": 1.3663516825437545,
      "epoch": 0.859393650035808,
      "grad_norm": 2.375,
      "learning_rate": 1.676043769764072e-05,
      "loss": 1.3753732299804688,
      "mean_token_accuracy": 0.6982124423980713,
      "num_tokens": 1130309.0,
      "step": 900
    },
    {
      "entropy": 1.370076591670513,
      "epoch": 0.907137741704464,
      "grad_norm": 1.9296875,
      "learning_rate": 1.637201667767204e-05,
      "loss": 1.3839447021484375,
      "mean_token_accuracy": 0.6986603051424026,
      "num_tokens": 1192924.0,
      "step": 950
    },
    {
      "entropy": 1.3369020774960518,
      "epoch": 0.95488183337312,
      "grad_norm": 1.9609375,
      "learning_rate": 1.5966687131679463e-05,
      "loss": 1.3427734375,
      "mean_token_accuracy": 0.7060227099061013,
      "num_tokens": 1254418.0,
      "step": 1000
    },
    {
      "epoch": 1.0,
      "eval_entropy": 1.2727481353174803,
      "eval_loss": 1.291359543800354,
      "eval_mean_token_accuracy": 0.7115251171210456,
      "eval_num_tokens": 1314012.0,
      "eval_runtime": 52.7795,
      "eval_samples_per_second": 16.711,
      "eval_steps_per_second": 16.711,
      "step": 1048
    },
    {
      "entropy": 1.3094463989819367,
      "epoch": 1.0019097636667462,
      "grad_norm": 2.078125,
      "learning_rate": 1.554552462577077e-05,
      "loss": 1.3102546691894532,
      "mean_token_accuracy": 0.7078821743805397,
      "num_tokens": 1316679.0,
      "step": 1050
    },
    {
      "entropy": 0.9862651464343071,
      "epoch": 1.0496538553354022,
      "grad_norm": 1.9609375,
      "learning_rate": 1.510964673975651e-05,
      "loss": 0.94318603515625,
      "mean_token_accuracy": 0.7765462607145309,
      "num_tokens": 1379361.0,
      "step": 1100
    },
    {
      "entropy": 0.9908152732253075,
      "epoch": 1.0973979470040582,
      "grad_norm": 2.453125,
      "learning_rate": 1.4660210101585367e-05,
      "loss": 0.9593799591064454,
      "mean_token_accuracy": 0.7724591562151909,
      "num_tokens": 1442332.0,
      "step": 1150
    },
    {
      "entropy": 0.9940855634212494,
      "epoch": 1.1451420386727142,
      "grad_norm": 2.21875,
      "learning_rate": 1.4198407318162977e-05,
      "loss": 0.9572904968261718,
      "mean_token_accuracy": 0.7741717839241028,
      "num_tokens": 1506048.0,
      "step": 1200
    },
    {
      "entropy": 0.997409186065197,
      "epoch": 1.1928861303413703,
      "grad_norm": 2.0625,
      "learning_rate": 1.3725463810698418e-05,
      "loss": 0.962403564453125,
      "mean_token_accuracy": 0.7725697284936905,
      "num_tokens": 1569529.0,
      "step": 1250
    },
    {
      "entropy": 0.9752482569217682,
      "epoch": 1.2406302220100263,
      "grad_norm": 2.234375,
      "learning_rate": 1.3242634562975988e-05,
      "loss": 0.9287136840820313,
      "mean_token_accuracy": 0.7795223397016525,
      "num_tokens": 1632744.0,
      "step": 1300
    },
    {
      "entropy": 0.972535012960434,
      "epoch": 1.2883743136786823,
      "grad_norm": 2.234375,
      "learning_rate": 1.2751200791180963e-05,
      "loss": 0.9352114105224609,
      "mean_token_accuracy": 0.7785275393724441,
      "num_tokens": 1695454.0,
      "step": 1350
    },
    {
      "entropy": 0.9388299316167832,
      "epoch": 1.3361184053473383,
      "grad_norm": 2.15625,
      "learning_rate": 1.225246654411605e-05,
      "loss": 0.9005762481689453,
      "mean_token_accuracy": 0.7845277190208435,
      "num_tokens": 1759868.0,
      "step": 1400
    },
    {
      "entropy": 0.9462847051024437,
      "epoch": 1.3838624970159943,
      "grad_norm": 2.234375,
      "learning_rate": 1.1747755242830204e-05,
      "loss": 0.9043065643310547,
      "mean_token_accuracy": 0.782291199862957,
      "num_tokens": 1820879.0,
      "step": 1450
    },
    {
      "entropy": 0.9308868950605392,
      "epoch": 1.4316065886846503,
      "grad_norm": 2.046875,
      "learning_rate": 1.1238406168841981e-05,
      "loss": 0.89322265625,
      "mean_token_accuracy": 0.7880325165390968,
      "num_tokens": 1883307.0,
      "step": 1500
    },
    {
      "entropy": 0.9272586759924889,
      "epoch": 1.4793506803533063,
      "grad_norm": 2.09375,
      "learning_rate": 1.0725770910276218e-05,
      "loss": 0.8859436798095703,
      "mean_token_accuracy": 0.7878314164280892,
      "num_tokens": 1946095.0,
      "step": 1550
    },
    {
      "entropy": 0.9405490574240685,
      "epoch": 1.5270947720219623,
      "grad_norm": 2.09375,
      "learning_rate": 1.0211209775344377e-05,
      "loss": 0.8991239929199218,
      "mean_token_accuracy": 0.787108507156372,
      "num_tokens": 2009039.0,
      "step": 1600
    },
    {
      "entropy": 0.9246599149703979,
      "epoch": 1.5748388636906183,
      "grad_norm": 2.1875,
      "learning_rate": 9.696088182685638e-06,
      "loss": 0.8893150329589844,
      "mean_token_accuracy": 0.787010528743267,
      "num_tokens": 2070866.0,
      "step": 1650
    },
    {
      "entropy": 0.9010607555508614,
      "epoch": 1.6225829553592743,
      "grad_norm": 2.5625,
      "learning_rate": 9.18177303814717e-06,
      "loss": 0.8644073486328125,
      "mean_token_accuracy": 0.7935708364844323,
      "num_tokens": 2133531.0,
      "step": 1700
    },
    {
      "entropy": 0.9290094000101089,
      "epoch": 1.6703270470279303,
      "grad_norm": 1.984375,
      "learning_rate": 8.66962910761798e-06,
      "loss": 0.8792234039306641,
      "mean_token_accuracy": 0.7916052675247193,
      "num_tokens": 2196212.0,
      "step": 1750
    },
    {
      "entropy": 0.9067334878444672,
      "epoch": 1.7180711386965863,
      "grad_norm": 2.0625,
      "learning_rate": 8.161015395541311e-06,
      "loss": 0.8666587829589844,
      "mean_token_accuracy": 0.7931482782959938,
      "num_tokens": 2259117.0,
      "step": 1800
    },
    {
      "entropy": 0.9192495954036712,
      "epoch": 1.7658152303652423,
      "grad_norm": 2.421875,
      "learning_rate": 7.65728153871533e-06,
      "loss": 0.8764252471923828,
      "mean_token_accuracy": 0.792299057841301,
      "num_tokens": 2320595.0,
      "step": 1850
    },
    {
      "entropy": 0.9041965058445931,
      "epoch": 1.8135593220338984,
      "grad_norm": 2.34375,
      "learning_rate": 7.159764224951467e-06,
      "loss": 0.8590383148193359,
      "mean_token_accuracy": 0.7972308859229088,
      "num_tokens": 2384167.0,
      "step": 1900
    },
    {
      "entropy": 0.9144112569093704,
      "epoch": 1.8613034137025544,
      "grad_norm": 2.21875,
      "learning_rate": 6.669783646093635e-06,
      "loss": 0.8813047790527344,
      "mean_token_accuracy": 0.7919116005301475,
      "num_tokens": 2444983.0,
      "step": 1950
    },
    {
      "entropy": 0.8830939248204231,
      "epoch": 1.9090475053712104,
      "grad_norm": 2.15625,
      "learning_rate": 6.188639994810476e-06,
      "loss": 0.8278516387939453,
      "mean_token_accuracy": 0.8022209334373475,
      "num_tokens": 2508413.0,
      "step": 2000
    },
    {
      "entropy": 0.9015234977006912,
      "epoch": 1.9567915970398664,
      "grad_norm": 2.1875,
      "learning_rate": 5.717610014456644e-06,
      "loss": 0.8635023498535156,
      "mean_token_accuracy": 0.7982774981856346,
      "num_tokens": 2570521.0,
      "step": 2050
    },
    {
      "epoch": 2.0,
      "eval_entropy": 1.0000475248545746,
      "eval_loss": 1.178954005241394,
      "eval_mean_token_accuracy": 0.74215727841773,
      "eval_num_tokens": 2628024.0,
      "eval_runtime": 52.7405,
      "eval_samples_per_second": 16.723,
      "eval_steps_per_second": 16.723,
      "step": 2096
    }
  ],
  "logging_steps": 50,
  "max_steps": 3144,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.5924789720612864e+17,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}