{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 100,
  "global_step": 354,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 1.3227437630295753,
      "epoch": 0.08501594048884166,
      "grad_norm": 1.4296875,
      "learning_rate": 0.00016363636363636366,
      "loss": 1.518,
      "mean_token_accuracy": 0.7175316952168942,
      "num_tokens": 48910.0,
      "step": 10
    },
    {
      "entropy": 0.7062680356204509,
      "epoch": 0.17003188097768332,
      "grad_norm": 0.439453125,
      "learning_rate": 0.00019973167127614215,
      "loss": 0.7921,
      "mean_token_accuracy": 0.8361715286970138,
      "num_tokens": 96994.0,
      "step": 20
    },
    {
      "entropy": 0.6654534563422203,
      "epoch": 0.255047821466525,
      "grad_norm": 0.46875,
      "learning_rate": 0.00019864405334538517,
      "loss": 0.6723,
      "mean_token_accuracy": 0.848845011740923,
      "num_tokens": 144332.0,
      "step": 30
    },
    {
      "entropy": 0.5781836319714785,
      "epoch": 0.34006376195536664,
      "grad_norm": 0.447265625,
      "learning_rate": 0.00019672948630390294,
      "loss": 0.6044,
      "mean_token_accuracy": 0.8551476895809174,
      "num_tokens": 192719.0,
      "step": 40
    },
    {
      "entropy": 0.5426878597587347,
      "epoch": 0.4250797024442083,
      "grad_norm": 0.466796875,
      "learning_rate": 0.00019400402027796955,
      "loss": 0.5788,
      "mean_token_accuracy": 0.8572557404637337,
      "num_tokens": 241632.0,
      "step": 50
    },
    {
      "entropy": 0.5355116073042154,
      "epoch": 0.51009564293305,
      "grad_norm": 0.5625,
      "learning_rate": 0.00019049050329166778,
      "loss": 0.5629,
      "mean_token_accuracy": 0.862634701281786,
      "num_tokens": 291116.0,
      "step": 60
    },
    {
      "entropy": 0.5224194105714559,
      "epoch": 0.5951115834218916,
      "grad_norm": 0.55078125,
      "learning_rate": 0.00018621838972819458,
      "loss": 0.551,
      "mean_token_accuracy": 0.8659204967319966,
      "num_tokens": 340898.0,
      "step": 70
    },
    {
      "entropy": 0.4716725114732981,
      "epoch": 0.6801275239107333,
      "grad_norm": 0.52734375,
      "learning_rate": 0.00018122349340898595,
      "loss": 0.5116,
      "mean_token_accuracy": 0.8745670169591904,
      "num_tokens": 386799.0,
      "step": 80
    },
    {
      "entropy": 0.4831196520477533,
      "epoch": 0.7651434643995749,
      "grad_norm": 0.54296875,
      "learning_rate": 0.00017554768736063859,
      "loss": 0.5071,
      "mean_token_accuracy": 0.8737691469490528,
      "num_tokens": 436798.0,
      "step": 90
    },
    {
      "entropy": 0.44570639468729495,
      "epoch": 0.8501594048884166,
      "grad_norm": 0.515625,
      "learning_rate": 0.00016923855278653114,
      "loss": 0.465,
      "mean_token_accuracy": 0.8849786855280399,
      "num_tokens": 485573.0,
      "step": 100
    },
    {
      "epoch": 0.8501594048884166,
      "eval_entropy": 0.48599471364702496,
      "eval_loss": 0.475477933883667,
      "eval_mean_token_accuracy": 0.8784565641766503,
      "eval_num_tokens": 485573.0,
      "eval_runtime": 17.16,
      "eval_samples_per_second": 9.732,
      "eval_steps_per_second": 4.895,
      "step": 100
    },
    {
      "entropy": 0.43454019390046594,
      "epoch": 0.9351753453772582,
      "grad_norm": 0.62890625,
      "learning_rate": 0.00016234898018587337,
      "loss": 0.4552,
      "mean_token_accuracy": 0.8844782762229443,
      "num_tokens": 530654.0,
      "step": 110
    },
    {
      "entropy": 0.41183759620437377,
      "epoch": 1.0170031880977684,
      "grad_norm": 0.5078125,
      "learning_rate": 0.00015493672596406598,
      "loss": 0.4096,
      "mean_token_accuracy": 0.8991486182460537,
      "num_tokens": 574892.0,
      "step": 120
    },
    {
      "entropy": 0.35238806419074536,
      "epoch": 1.1020191285866099,
      "grad_norm": 0.5546875,
      "learning_rate": 0.00014706392825137964,
      "loss": 0.3574,
      "mean_token_accuracy": 0.9064381532371044,
      "num_tokens": 620523.0,
      "step": 130
    },
    {
      "entropy": 0.3248062996193767,
      "epoch": 1.1870350690754516,
      "grad_norm": 0.69921875,
      "learning_rate": 0.00013879658598892254,
      "loss": 0.3389,
      "mean_token_accuracy": 0.9107168510556221,
      "num_tokens": 670456.0,
      "step": 140
    },
    {
      "entropy": 0.3360298121348023,
      "epoch": 1.2720510095642932,
      "grad_norm": 0.65234375,
      "learning_rate": 0.0001302040056488047,
      "loss": 0.3444,
      "mean_token_accuracy": 0.9091025076806545,
      "num_tokens": 718444.0,
      "step": 150
    },
    {
      "entropy": 0.3400493461638689,
      "epoch": 1.357066950053135,
      "grad_norm": 0.58203125,
      "learning_rate": 0.00012135822022673263,
      "loss": 0.3558,
      "mean_token_accuracy": 0.910928662866354,
      "num_tokens": 767549.0,
      "step": 160
    },
    {
      "entropy": 0.3092759184539318,
      "epoch": 1.4420828905419767,
      "grad_norm": 0.625,
      "learning_rate": 0.00011233338537771407,
      "loss": 0.3089,
      "mean_token_accuracy": 0.9161648266017437,
      "num_tokens": 818123.0,
      "step": 170
    },
    {
      "entropy": 0.3046819636598229,
      "epoch": 1.5270988310308182,
      "grad_norm": 0.55859375,
      "learning_rate": 0.00010320515775716555,
      "loss": 0.3032,
      "mean_token_accuracy": 0.9219753406941891,
      "num_tokens": 867721.0,
      "step": 180
    },
    {
      "entropy": 0.2906474955379963,
      "epoch": 1.61211477151966,
      "grad_norm": 0.5,
      "learning_rate": 9.405006077888954e-05,
      "loss": 0.2992,
      "mean_token_accuracy": 0.9225860193371773,
      "num_tokens": 917153.0,
      "step": 190
    },
    {
      "entropy": 0.30296949967741965,
      "epoch": 1.6971307120085015,
      "grad_norm": 0.546875,
      "learning_rate": 8.494484310687581e-05,
      "loss": 0.3053,
      "mean_token_accuracy": 0.9186053372919559,
      "num_tokens": 964600.0,
      "step": 200
    },
    {
      "epoch": 1.6971307120085015,
      "eval_entropy": 0.35429788788869265,
      "eval_loss": 0.404356449842453,
      "eval_mean_token_accuracy": 0.8999107018822715,
      "eval_num_tokens": 964600.0,
      "eval_runtime": 17.2396,
      "eval_samples_per_second": 9.687,
      "eval_steps_per_second": 4.872,
      "step": 200
    },
    {
      "entropy": 0.30462323110550643,
      "epoch": 1.7821466524973433,
      "grad_norm": 0.59375,
      "learning_rate": 7.596583525879344e-05,
      "loss": 0.308,
      "mean_token_accuracy": 0.9176651492714882,
      "num_tokens": 1010401.0,
      "step": 210
    },
    {
      "entropy": 0.29319654731079936,
      "epoch": 1.867162592986185,
      "grad_norm": 0.64453125,
      "learning_rate": 6.718830971487165e-05,
      "loss": 0.2912,
      "mean_token_accuracy": 0.9217840351164341,
      "num_tokens": 1058672.0,
      "step": 220
    },
    {
      "entropy": 0.28096328396350145,
      "epoch": 1.9521785334750266,
      "grad_norm": 0.6796875,
      "learning_rate": 5.868584989647994e-05,
      "loss": 0.2825,
      "mean_token_accuracy": 0.9237303651869297,
      "num_tokens": 1106102.0,
      "step": 230
    },
    {
      "entropy": 0.27870092405514285,
      "epoch": 2.0340063761955367,
      "grad_norm": 0.4296875,
      "learning_rate": 5.0529733304363145e-05,
      "loss": 0.2657,
      "mean_token_accuracy": 0.9302684858247832,
      "num_tokens": 1151178.0,
      "step": 240
    },
    {
      "entropy": 0.2116462522186339,
      "epoch": 2.1190223166843785,
      "grad_norm": 0.625,
      "learning_rate": 4.278833398778306e-05,
      "loss": 0.2033,
      "mean_token_accuracy": 0.9440436236560344,
      "num_tokens": 1198945.0,
      "step": 250
    },
    {
      "entropy": 0.21235913261771203,
      "epoch": 2.2040382571732198,
      "grad_norm": 0.5,
      "learning_rate": 3.5526549353765296e-05,
      "loss": 0.2133,
      "mean_token_accuracy": 0.9405460350215435,
      "num_tokens": 1246732.0,
      "step": 260
    },
    {
      "entropy": 0.22362984605133535,
      "epoch": 2.2890541976620615,
      "grad_norm": 0.494140625,
      "learning_rate": 2.8805256121602398e-05,
      "loss": 0.2225,
      "mean_token_accuracy": 0.9388925828039646,
      "num_tokens": 1291061.0,
      "step": 270
    },
    {
      "entropy": 0.22003474170342088,
      "epoch": 2.3740701381509033,
      "grad_norm": 0.53125,
      "learning_rate": 2.268079998343453e-05,
      "loss": 0.2244,
      "mean_token_accuracy": 0.9388979405164719,
      "num_tokens": 1340506.0,
      "step": 280
    },
    {
      "entropy": 0.20420914590358735,
      "epoch": 2.459086078639745,
      "grad_norm": 0.59765625,
      "learning_rate": 1.720452324916656e-05,
      "loss": 0.2062,
      "mean_token_accuracy": 0.941665968298912,
      "num_tokens": 1389105.0,
      "step": 290
    },
    {
      "entropy": 0.21224667700007557,
      "epoch": 2.5441020191285864,
      "grad_norm": 0.470703125,
      "learning_rate": 1.2422334435550509e-05,
      "loss": 0.2076,
      "mean_token_accuracy": 0.9419615142047405,
      "num_tokens": 1436911.0,
      "step": 300
    },
    {
      "epoch": 2.5441020191285864,
      "eval_entropy": 0.2823823473105828,
      "eval_loss": 0.3997238576412201,
      "eval_mean_token_accuracy": 0.906679209499132,
      "eval_num_tokens": 1436911.0,
      "eval_runtime": 17.0522,
      "eval_samples_per_second": 9.793,
      "eval_steps_per_second": 4.926,
      "step": 300
    },
    {
      "entropy": 0.21214768588542937,
      "epoch": 2.629117959617428,
      "grad_norm": 0.53125,
      "learning_rate": 8.374323407637742e-06,
      "loss": 0.1961,
      "mean_token_accuracy": 0.9436013922095299,
      "num_tokens": 1485939.0,
      "step": 310
    },
    {
      "entropy": 0.1970360592007637,
      "epoch": 2.71413390010627,
      "grad_norm": 0.63671875,
      "learning_rate": 5.094425298933136e-06,
      "loss": 0.1917,
      "mean_token_accuracy": 0.9456975102424622,
      "num_tokens": 1534218.0,
      "step": 320
    },
    {
      "entropy": 0.2043833415955305,
      "epoch": 2.7991498405951116,
      "grad_norm": 0.51171875,
      "learning_rate": 2.6101360276626798e-06,
      "loss": 0.1931,
      "mean_token_accuracy": 0.9451199650764466,
      "num_tokens": 1583541.0,
      "step": 330
    },
    {
      "entropy": 0.20932846572250127,
      "epoch": 2.8841657810839534,
      "grad_norm": 0.65234375,
      "learning_rate": 9.422817940278772e-07,
      "loss": 0.2116,
      "mean_token_accuracy": 0.9425160974264145,
      "num_tokens": 1632537.0,
      "step": 340
    },
    {
      "entropy": 0.21047836663201452,
      "epoch": 2.969181721572795,
      "grad_norm": 0.54296875,
      "learning_rate": 1.0484449078912439e-07,
      "loss": 0.21,
      "mean_token_accuracy": 0.9405752278864383,
      "num_tokens": 1681400.0,
      "step": 350
    }
  ],
  "logging_steps": 10,
  "max_steps": 354,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.1052763677773005e+17,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}