{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 25,
  "global_step": 750,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 1.6712856006622314,
      "epoch": 0.1,
      "grad_norm": 0.6304148435592651,
      "learning_rate": 0.0001999990663152786,
      "loss": 2.0491,
      "mean_token_accuracy": 0.6684889650344848,
      "num_tokens": 122580.0,
      "step": 25
    },
    {
      "epoch": 0.1,
      "eval_entropy": 0.9992010481655598,
      "eval_loss": 0.9745309948921204,
      "eval_mean_token_accuracy": 0.815976720303297,
      "eval_num_tokens": 122580.0,
      "eval_runtime": 66.8951,
      "eval_samples_per_second": 29.898,
      "eval_steps_per_second": 0.478,
      "step": 25
    },
    {
      "entropy": 0.9771250176429749,
      "epoch": 0.2,
      "grad_norm": 0.4267440736293793,
      "learning_rate": 0.0001993694918299864,
      "loss": 0.9676,
      "mean_token_accuracy": 0.8179380083084107,
      "num_tokens": 244723.0,
      "step": 50
    },
    {
      "epoch": 0.2,
      "eval_entropy": 0.9376886244863272,
      "eval_loss": 0.940742015838623,
      "eval_mean_token_accuracy": 0.8212726972997189,
      "eval_num_tokens": 244723.0,
      "eval_runtime": 66.3094,
      "eval_samples_per_second": 30.162,
      "eval_steps_per_second": 0.483,
      "step": 50
    },
    {
      "entropy": 0.9467579674720764,
      "epoch": 0.3,
      "grad_norm": 0.39123401045799255,
      "learning_rate": 0.0001975812958575343,
      "loss": 0.9473,
      "mean_token_accuracy": 0.820980327129364,
      "num_tokens": 366803.0,
      "step": 75
    },
    {
      "epoch": 0.3,
      "eval_entropy": 0.930314002558589,
      "eval_loss": 0.9318345785140991,
      "eval_mean_token_accuracy": 0.8218902982771397,
      "eval_num_tokens": 366803.0,
      "eval_runtime": 66.1507,
      "eval_samples_per_second": 30.234,
      "eval_steps_per_second": 0.484,
      "step": 75
    },
    {
      "entropy": 0.9574691414833069,
      "epoch": 0.4,
      "grad_norm": 0.3710257112979889,
      "learning_rate": 0.00019465532828090735,
      "loss": 0.9545,
      "mean_token_accuracy": 0.8186049485206603,
      "num_tokens": 489374.0,
      "step": 100
    },
    {
      "epoch": 0.4,
      "eval_entropy": 0.9250399190932512,
      "eval_loss": 0.927577555179596,
      "eval_mean_token_accuracy": 0.8226032145321369,
      "eval_num_tokens": 489374.0,
      "eval_runtime": 66.6912,
      "eval_samples_per_second": 29.989,
      "eval_steps_per_second": 0.48,
      "step": 100
    },
    {
      "entropy": 0.9415939354896545,
      "epoch": 0.5,
      "grad_norm": 0.3609308898448944,
      "learning_rate": 0.00019062570509327992,
      "loss": 0.938,
      "mean_token_accuracy": 0.8219008255004883,
      "num_tokens": 611509.0,
      "step": 125
    },
    {
      "epoch": 0.5,
      "eval_entropy": 0.9134028796106577,
      "eval_loss": 0.9236319661140442,
      "eval_mean_token_accuracy": 0.8233284279704094,
      "eval_num_tokens": 611509.0,
      "eval_runtime": 67.9585,
      "eval_samples_per_second": 29.43,
      "eval_steps_per_second": 0.471,
      "step": 125
    },
    {
      "entropy": 0.9313359212875366,
      "epoch": 0.6,
      "grad_norm": 0.3689591586589813,
      "learning_rate": 0.00018553941061473218,
      "loss": 0.9277,
      "mean_token_accuracy": 0.8244240188598633,
      "num_tokens": 733437.0,
      "step": 150
    },
    {
      "epoch": 0.6,
      "eval_entropy": 0.9121778700500727,
      "eval_loss": 0.9206869006156921,
      "eval_mean_token_accuracy": 0.8243093993514776,
      "eval_num_tokens": 733437.0,
      "eval_runtime": 68.185,
      "eval_samples_per_second": 29.332,
      "eval_steps_per_second": 0.469,
      "step": 150
    },
    {
      "entropy": 0.9311716175079345,
      "epoch": 0.7,
      "grad_norm": 0.3817085027694702,
      "learning_rate": 0.00017945574966774376,
      "loss": 0.9292,
      "mean_token_accuracy": 0.8232095694541931,
      "num_tokens": 855571.0,
      "step": 175
    },
    {
      "epoch": 0.7,
      "eval_entropy": 0.9240057598799467,
      "eval_loss": 0.9190523028373718,
      "eval_mean_token_accuracy": 0.8243635054677725,
      "eval_num_tokens": 855571.0,
      "eval_runtime": 68.1745,
      "eval_samples_per_second": 29.336,
      "eval_steps_per_second": 0.469,
      "step": 175
    },
    {
      "entropy": 0.9368235039710998,
      "epoch": 0.8,
      "grad_norm": 0.343488872051239,
      "learning_rate": 0.00017244565609895074,
      "loss": 0.9328,
      "mean_token_accuracy": 0.822948260307312,
      "num_tokens": 977886.0,
      "step": 200
    },
    {
      "epoch": 0.8,
      "eval_entropy": 0.9235651567578316,
      "eval_loss": 0.9163441061973572,
      "eval_mean_token_accuracy": 0.8248479198664427,
      "eval_num_tokens": 977886.0,
      "eval_runtime": 68.3524,
      "eval_samples_per_second": 29.26,
      "eval_steps_per_second": 0.468,
      "step": 200
    },
    {
      "entropy": 0.9266206288337707,
      "epoch": 0.9,
      "grad_norm": 0.34697458148002625,
      "learning_rate": 0.00016459086570961594,
      "loss": 0.9247,
      "mean_token_accuracy": 0.8257838773727417,
      "num_tokens": 1099649.0,
      "step": 225
    },
    {
      "epoch": 0.9,
      "eval_entropy": 0.9170300494879484,
      "eval_loss": 0.9154583215713501,
      "eval_mean_token_accuracy": 0.8250364065170288,
      "eval_num_tokens": 1099649.0,
      "eval_runtime": 67.9598,
      "eval_samples_per_second": 29.429,
      "eval_steps_per_second": 0.471,
      "step": 225
    },
    {
      "entropy": 0.9486314964294433,
      "epoch": 1.0,
      "grad_norm": 0.38791951537132263,
      "learning_rate": 0.00015598296323822024,
      "loss": 0.9457,
      "mean_token_accuracy": 0.8199513030052185,
      "num_tokens": 1222434.0,
      "step": 250
    },
    {
      "epoch": 1.0,
      "eval_entropy": 0.9278132077306509,
      "eval_loss": 0.9133847951889038,
      "eval_mean_token_accuracy": 0.8252693247050047,
      "eval_num_tokens": 1222434.0,
      "eval_runtime": 67.957,
      "eval_samples_per_second": 29.43,
      "eval_steps_per_second": 0.471,
      "step": 250
    },
    {
      "entropy": 0.8935548496246338,
      "epoch": 1.1,
      "grad_norm": 0.3913302421569824,
      "learning_rate": 0.00014672231450710066,
      "loss": 0.8794,
      "mean_token_accuracy": 0.8301347184181214,
      "num_tokens": 1344769.0,
      "step": 275
    },
    {
      "epoch": 1.1,
      "eval_entropy": 0.8964261263608932,
      "eval_loss": 0.918248176574707,
      "eval_mean_token_accuracy": 0.8246872704476118,
      "eval_num_tokens": 1344769.0,
      "eval_runtime": 67.949,
      "eval_samples_per_second": 29.434,
      "eval_steps_per_second": 0.471,
      "step": 275
    },
    {
      "entropy": 0.8727394843101501,
      "epoch": 1.2,
      "grad_norm": 0.39644864201545715,
      "learning_rate": 0.00013691689618401835,
      "loss": 0.8669,
      "mean_token_accuracy": 0.8310828638076783,
      "num_tokens": 1466893.0,
      "step": 300
    },
    {
      "epoch": 1.2,
      "eval_entropy": 0.8782581854611635,
      "eval_loss": 0.917015552520752,
      "eval_mean_token_accuracy": 0.8250990845263004,
      "eval_num_tokens": 1466893.0,
      "eval_runtime": 67.8451,
      "eval_samples_per_second": 29.479,
      "eval_steps_per_second": 0.472,
      "step": 300
    },
    {
      "entropy": 0.8778977513313293,
      "epoch": 1.3,
      "grad_norm": 0.3395286500453949,
      "learning_rate": 0.00012668103680332012,
      "loss": 0.876,
      "mean_token_accuracy": 0.8312993144989014,
      "num_tokens": 1589094.0,
      "step": 325
    },
    {
      "epoch": 1.3,
      "eval_entropy": 0.8799433764070272,
      "eval_loss": 0.9174618124961853,
      "eval_mean_token_accuracy": 0.8249436803162098,
      "eval_num_tokens": 1589094.0,
      "eval_runtime": 68.1596,
      "eval_samples_per_second": 29.343,
      "eval_steps_per_second": 0.469,
      "step": 325
    },
    {
      "entropy": 0.8876111268997192,
      "epoch": 1.4,
      "grad_norm": 0.364728182554245,
      "learning_rate": 0.00011613408372604825,
      "loss": 0.8773,
      "mean_token_accuracy": 0.8295953154563904,
      "num_tokens": 1711583.0,
      "step": 350
    },
    {
      "epoch": 1.4,
      "eval_entropy": 0.85817476734519,
      "eval_loss": 0.9163104295730591,
      "eval_mean_token_accuracy": 0.8252090867608786,
      "eval_num_tokens": 1711583.0,
      "eval_runtime": 68.1812,
      "eval_samples_per_second": 29.334,
      "eval_steps_per_second": 0.469,
      "step": 350
    },
    {
      "entropy": 0.88423011302948,
      "epoch": 1.5,
      "grad_norm": 0.38693588972091675,
      "learning_rate": 0.00010539901158188398,
      "loss": 0.883,
      "mean_token_accuracy": 0.8286035037040711,
      "num_tokens": 1834553.0,
      "step": 375
    },
    {
      "epoch": 1.5,
      "eval_entropy": 0.873553803190589,
      "eval_loss": 0.9151167273521423,
      "eval_mean_token_accuracy": 0.8254394326359034,
      "eval_num_tokens": 1834553.0,
      "eval_runtime": 68.1992,
      "eval_samples_per_second": 29.326,
      "eval_steps_per_second": 0.469,
      "step": 375
    },
    {
      "entropy": 0.8715341877937317,
      "epoch": 1.6,
      "grad_norm": 0.3770897686481476,
      "learning_rate": 9.460098841811601e-05,
      "loss": 0.87,
      "mean_token_accuracy": 0.8320170021057129,
      "num_tokens": 1956574.0,
      "step": 400
    },
    {
      "epoch": 1.6,
      "eval_entropy": 0.87935302965343,
      "eval_loss": 0.9151723980903625,
      "eval_mean_token_accuracy": 0.8255144450813532,
      "eval_num_tokens": 1956574.0,
      "eval_runtime": 67.9177,
      "eval_samples_per_second": 29.447,
      "eval_steps_per_second": 0.471,
      "step": 400
    },
    {
      "entropy": 0.8710586881637573,
      "epoch": 1.7,
      "grad_norm": 0.3885301351547241,
      "learning_rate": 8.386591627395173e-05,
      "loss": 0.8672,
      "mean_token_accuracy": 0.8299804759025574,
      "num_tokens": 2078811.0,
      "step": 425
    },
    {
      "epoch": 1.7,
      "eval_entropy": 0.8739563841372728,
      "eval_loss": 0.9149895906448364,
      "eval_mean_token_accuracy": 0.8253704849630594,
      "eval_num_tokens": 2078811.0,
      "eval_runtime": 66.7099,
      "eval_samples_per_second": 29.981,
      "eval_steps_per_second": 0.48,
      "step": 425
    },
    {
      "entropy": 0.8689111661911011,
      "epoch": 1.8,
      "grad_norm": 0.38322678208351135,
      "learning_rate": 7.33189631966799e-05,
      "loss": 0.865,
      "mean_token_accuracy": 0.8322654938697815,
      "num_tokens": 2200571.0,
      "step": 450
    },
    {
      "epoch": 1.8,
      "eval_entropy": 0.8741269316524267,
      "eval_loss": 0.9147102236747742,
      "eval_mean_token_accuracy": 0.8253954574465752,
      "eval_num_tokens": 2200571.0,
      "eval_runtime": 66.1144,
      "eval_samples_per_second": 30.251,
      "eval_steps_per_second": 0.484,
      "step": 450
    },
    {
      "entropy": 0.8786772465705872,
      "epoch": 1.9,
      "grad_norm": 0.3754963278770447,
      "learning_rate": 6.308310381598168e-05,
      "loss": 0.8755,
      "mean_token_accuracy": 0.8297365355491638,
      "num_tokens": 2322788.0,
      "step": 475
    },
    {
      "epoch": 1.9,
      "eval_entropy": 0.8773031029850245,
      "eval_loss": 0.9134438633918762,
      "eval_mean_token_accuracy": 0.8253685813397169,
      "eval_num_tokens": 2322788.0,
      "eval_runtime": 66.1468,
      "eval_samples_per_second": 30.236,
      "eval_steps_per_second": 0.484,
      "step": 475
    },
    {
      "entropy": 0.8698205542564392,
      "epoch": 2.0,
      "grad_norm": 0.37209320068359375,
      "learning_rate": 5.327768549289934e-05,
      "loss": 0.8654,
      "mean_token_accuracy": 0.8314353656768799,
      "num_tokens": 2444868.0,
      "step": 500
    },
    {
      "epoch": 2.0,
      "eval_entropy": 0.874303799122572,
      "eval_loss": 0.9136784076690674,
      "eval_mean_token_accuracy": 0.8255054354667664,
      "eval_num_tokens": 2444868.0,
      "eval_runtime": 65.7882,
      "eval_samples_per_second": 30.401,
      "eval_steps_per_second": 0.486,
      "step": 500
    },
    {
      "entropy": 0.8443045258522034,
      "epoch": 2.1,
      "grad_norm": 0.3855106830596924,
      "learning_rate": 4.4017036761779787e-05,
      "loss": 0.8248,
      "mean_token_accuracy": 0.8369424676895142,
      "num_tokens": 2567342.0,
      "step": 525
    },
    {
      "epoch": 2.1,
      "eval_entropy": 0.8267646413296461,
      "eval_loss": 0.9248229265213013,
      "eval_mean_token_accuracy": 0.8247568681836128,
      "eval_num_tokens": 2567342.0,
      "eval_runtime": 66.4418,
      "eval_samples_per_second": 30.102,
      "eval_steps_per_second": 0.482,
      "step": 525
    },
    {
      "entropy": 0.8175929665565491,
      "epoch": 2.2,
      "grad_norm": 0.3969684839248657,
      "learning_rate": 3.540913429038407e-05,
      "loss": 0.8159,
      "mean_token_accuracy": 0.8389404201507569,
      "num_tokens": 2689296.0,
      "step": 550
    },
    {
      "epoch": 2.2,
      "eval_entropy": 0.8400543294847012,
      "eval_loss": 0.9235786199569702,
      "eval_mean_token_accuracy": 0.8250540643930435,
      "eval_num_tokens": 2689296.0,
      "eval_runtime": 66.682,
      "eval_samples_per_second": 29.993,
      "eval_steps_per_second": 0.48,
      "step": 550
    },
    {
      "entropy": 0.8221414375305176,
      "epoch": 2.3,
      "grad_norm": 0.4103279709815979,
      "learning_rate": 2.7554343901049294e-05,
      "loss": 0.8108,
      "mean_token_accuracy": 0.8397244310379028,
      "num_tokens": 2811102.0,
      "step": 575
    },
    {
      "epoch": 2.3,
      "eval_entropy": 0.8320044614374638,
      "eval_loss": 0.9250128269195557,
      "eval_mean_token_accuracy": 0.8249792046844959,
      "eval_num_tokens": 2811102.0,
      "eval_runtime": 68.2399,
      "eval_samples_per_second": 29.308,
      "eval_steps_per_second": 0.469,
      "step": 575
    },
    {
      "entropy": 0.8239590454101563,
      "epoch": 2.4,
      "grad_norm": 0.4046800434589386,
      "learning_rate": 2.0544250332256276e-05,
      "loss": 0.8193,
      "mean_token_accuracy": 0.8385440015792847,
      "num_tokens": 2933467.0,
      "step": 600
    },
    {
      "epoch": 2.4,
      "eval_entropy": 0.8278196454048157,
      "eval_loss": 0.9243917465209961,
      "eval_mean_token_accuracy": 0.82486542314291,
      "eval_num_tokens": 2933467.0,
      "eval_runtime": 68.1284,
      "eval_samples_per_second": 29.356,
      "eval_steps_per_second": 0.47,
      "step": 600
    },
    {
      "entropy": 0.8189333271980286,
      "epoch": 2.5,
      "grad_norm": 0.38442716002464294,
      "learning_rate": 1.4460589385267842e-05,
      "loss": 0.8178,
      "mean_token_accuracy": 0.8380931878089904,
      "num_tokens": 3055276.0,
      "step": 625
    },
    {
      "epoch": 2.5,
      "eval_entropy": 0.8392351847141981,
      "eval_loss": 0.9240784645080566,
      "eval_mean_token_accuracy": 0.8248794414103031,
      "eval_num_tokens": 3055276.0,
      "eval_runtime": 68.2247,
      "eval_samples_per_second": 29.315,
      "eval_steps_per_second": 0.469,
      "step": 625
    },
    {
      "entropy": 0.8184775829315185,
      "epoch": 2.6,
      "grad_norm": 0.42365893721580505,
      "learning_rate": 9.374294906720082e-06,
      "loss": 0.8096,
      "mean_token_accuracy": 0.8411558413505554,
      "num_tokens": 3176654.0,
      "step": 650
    },
    {
      "epoch": 2.6,
      "eval_entropy": 0.8347331862896681,
      "eval_loss": 0.9245245456695557,
      "eval_mean_token_accuracy": 0.8248291350901127,
      "eval_num_tokens": 3176654.0,
      "eval_runtime": 67.8139,
      "eval_samples_per_second": 29.492,
      "eval_steps_per_second": 0.472,
      "step": 650
    },
    {
      "entropy": 0.8422538375854492,
      "epoch": 2.7,
      "grad_norm": 0.410686194896698,
      "learning_rate": 5.344671719092664e-06,
      "loss": 0.8453,
      "mean_token_accuracy": 0.832388162612915,
      "num_tokens": 3299985.0,
      "step": 675
    },
    {
      "epoch": 2.7,
      "eval_entropy": 0.8364100754261017,
      "eval_loss": 0.9244425296783447,
      "eval_mean_token_accuracy": 0.8246717043220997,
      "eval_num_tokens": 3299985.0,
      "eval_runtime": 67.9993,
      "eval_samples_per_second": 29.412,
      "eval_steps_per_second": 0.471,
      "step": 675
    },
    {
      "entropy": 0.8317058253288269,
      "epoch": 2.8,
      "grad_norm": 0.45394909381866455,
      "learning_rate": 2.418704142465722e-06,
      "loss": 0.8301,
      "mean_token_accuracy": 0.8355569648742676,
      "num_tokens": 3422725.0,
      "step": 700
    },
    {
      "epoch": 2.8,
      "eval_entropy": 0.8373962622135878,
      "eval_loss": 0.924216091632843,
      "eval_mean_token_accuracy": 0.8247099295258522,
      "eval_num_tokens": 3422725.0,
      "eval_runtime": 67.6852,
      "eval_samples_per_second": 29.549,
      "eval_steps_per_second": 0.473,
      "step": 700
    },
    {
      "entropy": 0.8333729934692383,
      "epoch": 2.9,
      "grad_norm": 0.4409957230091095,
      "learning_rate": 6.305081700136328e-07,
      "loss": 0.8241,
      "mean_token_accuracy": 0.83688227891922,
      "num_tokens": 3545132.0,
      "step": 725
    },
    {
      "epoch": 2.9,
      "eval_entropy": 0.8377762027084827,
      "eval_loss": 0.924144446849823,
      "eval_mean_token_accuracy": 0.824684476479888,
      "eval_num_tokens": 3545132.0,
      "eval_runtime": 68.0427,
      "eval_samples_per_second": 29.393,
      "eval_steps_per_second": 0.47,
      "step": 725
    },
    {
      "entropy": 0.8257460308074951,
      "epoch": 3.0,
      "grad_norm": 0.40077298879623413,
      "learning_rate": 9.336847214269639e-10,
      "loss": 0.8146,
      "mean_token_accuracy": 0.8383808851242065,
      "num_tokens": 3667302.0,
      "step": 750
    },
    {
      "epoch": 3.0,
      "eval_entropy": 0.8377156592905521,
      "eval_loss": 0.9241430759429932,
      "eval_mean_token_accuracy": 0.8246781621128321,
      "eval_num_tokens": 3667302.0,
      "eval_runtime": 67.5858,
      "eval_samples_per_second": 29.592,
      "eval_steps_per_second": 0.473,
      "step": 750
    }
  ],
  "logging_steps": 25,
  "max_steps": 750,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 25,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 4.177111226528563e+16,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}