{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 15.0,
  "eval_steps": 200,
  "global_step": 6210,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 2.5034248107671737,
      "epoch": 0.4839685420447671,
      "grad_norm": 0.609375,
      "learning_rate": 1.9900000000000003e-05,
      "loss": 3.8386904907226564,
      "mean_token_accuracy": 0.4763091333582997,
      "num_tokens": 258555.0,
      "step": 200
    },
    {
      "epoch": 0.4839685420447671,
      "eval_entropy": 1.89908932659724,
      "eval_loss": 1.6288481950759888,
      "eval_mean_token_accuracy": 0.7560858702006406,
      "eval_num_tokens": 258555.0,
      "eval_runtime": 48.9754,
      "eval_samples_per_second": 5.921,
      "eval_steps_per_second": 1.491,
      "step": 200
    },
    {
      "entropy": 0.9987735440582037,
      "epoch": 0.9679370840895342,
      "grad_norm": 0.1767578125,
      "learning_rate": 3.99e-05,
      "loss": 0.8650601196289063,
      "mean_token_accuracy": 0.8588662607967854,
      "num_tokens": 516030.0,
      "step": 400
    },
    {
      "epoch": 0.9679370840895342,
      "eval_entropy": 0.7666851747525881,
      "eval_loss": 0.7402243614196777,
      "eval_mean_token_accuracy": 0.8760081317326794,
      "eval_num_tokens": 516030.0,
      "eval_runtime": 46.9,
      "eval_samples_per_second": 6.183,
      "eval_steps_per_second": 1.557,
      "step": 400
    },
    {
      "entropy": 0.6914621311119539,
      "epoch": 1.4500907441016335,
      "grad_norm": 0.265625,
      "learning_rate": 5.99e-05,
      "loss": 0.6239362335205079,
      "mean_token_accuracy": 0.8933641332455232,
      "num_tokens": 773384.0,
      "step": 600
    },
    {
      "epoch": 1.4500907441016335,
      "eval_entropy": 0.6780054618234503,
      "eval_loss": 0.6326681971549988,
      "eval_mean_token_accuracy": 0.8811007974898979,
      "eval_num_tokens": 773384.0,
      "eval_runtime": 47.2911,
      "eval_samples_per_second": 6.132,
      "eval_steps_per_second": 1.544,
      "step": 600
    },
    {
      "entropy": 0.5927861626073718,
      "epoch": 1.9340592861464003,
      "grad_norm": 0.240234375,
      "learning_rate": 7.99e-05,
      "loss": 0.5069254684448242,
      "mean_token_accuracy": 0.9013223663717508,
      "num_tokens": 1031174.0,
      "step": 800
    },
    {
      "epoch": 1.9340592861464003,
      "eval_entropy": 0.6229167531614435,
      "eval_loss": 0.5980483293533325,
      "eval_mean_token_accuracy": 0.893642638644127,
      "eval_num_tokens": 1031174.0,
      "eval_runtime": 47.2124,
      "eval_samples_per_second": 6.142,
      "eval_steps_per_second": 1.546,
      "step": 800
    },
    {
      "entropy": 0.5391529030452857,
      "epoch": 2.4162129461585,
      "grad_norm": 0.2080078125,
      "learning_rate": 9.99e-05,
      "loss": 0.4552752685546875,
      "mean_token_accuracy": 0.913724020927434,
      "num_tokens": 1287528.0,
      "step": 1000
    },
    {
      "epoch": 2.4162129461585,
      "eval_entropy": 0.5597588991465634,
      "eval_loss": 0.5732213258743286,
      "eval_mean_token_accuracy": 0.8946846454110864,
      "eval_num_tokens": 1287528.0,
      "eval_runtime": 46.718,
      "eval_samples_per_second": 6.207,
      "eval_steps_per_second": 1.563,
      "step": 1000
    },
    {
      "entropy": 0.48732462897896767,
      "epoch": 2.900181488203267,
      "grad_norm": 0.1865234375,
      "learning_rate": 9.964045828736265e-05,
      "loss": 0.43026100158691405,
      "mean_token_accuracy": 0.9152925322949886,
      "num_tokens": 1545979.0,
      "step": 1200
    },
    {
      "epoch": 2.900181488203267,
      "eval_entropy": 0.5305672479818945,
      "eval_loss": 0.571603000164032,
      "eval_mean_token_accuracy": 0.8954706118531424,
      "eval_num_tokens": 1545979.0,
      "eval_runtime": 87.5618,
      "eval_samples_per_second": 3.312,
      "eval_steps_per_second": 0.834,
      "step": 1200
    },
    {
      "entropy": 0.4696483660297681,
      "epoch": 3.382335148215366,
      "grad_norm": 0.1953125,
      "learning_rate": 9.855982874039193e-05,
      "loss": 0.41300594329833984,
      "mean_token_accuracy": 0.9177645573502353,
      "num_tokens": 1802155.0,
      "step": 1400
    },
    {
      "epoch": 3.382335148215366,
      "eval_entropy": 0.5139341966746604,
      "eval_loss": 0.580710232257843,
      "eval_mean_token_accuracy": 0.8955999376022652,
      "eval_num_tokens": 1802155.0,
      "eval_runtime": 87.6062,
      "eval_samples_per_second": 3.31,
      "eval_steps_per_second": 0.833,
      "step": 1400
    },
    {
      "entropy": 0.4651796716451645,
      "epoch": 3.866303690260133,
      "grad_norm": 0.2001953125,
      "learning_rate": 9.677379992114693e-05,
      "loss": 0.40818458557128906,
      "mean_token_accuracy": 0.9180439539253712,
      "num_tokens": 2060871.0,
      "step": 1600
    },
    {
      "epoch": 3.866303690260133,
      "eval_entropy": 0.5178806251042509,
      "eval_loss": 0.5665879845619202,
      "eval_mean_token_accuracy": 0.8960714160579525,
      "eval_num_tokens": 2060871.0,
      "eval_runtime": 87.2236,
      "eval_samples_per_second": 3.325,
      "eval_steps_per_second": 0.837,
      "step": 1600
    },
    {
      "entropy": 0.44904655928593806,
      "epoch": 4.348457350272232,
      "grad_norm": 0.208984375,
      "learning_rate": 9.430831639043096e-05,
      "loss": 0.39284423828125,
      "mean_token_accuracy": 0.9206457537420122,
      "num_tokens": 2317399.0,
      "step": 1800
    },
    {
      "epoch": 4.348457350272232,
      "eval_entropy": 0.5153208285161893,
      "eval_loss": 0.5782919526100159,
      "eval_mean_token_accuracy": 0.8959851983475359,
      "eval_num_tokens": 2317399.0,
      "eval_runtime": 47.2746,
      "eval_samples_per_second": 6.134,
      "eval_steps_per_second": 1.544,
      "step": 1800
    },
    {
      "entropy": 0.44385369334369895,
      "epoch": 4.832425892317,
      "grad_norm": 0.19921875,
      "learning_rate": 9.119919273683477e-05,
      "loss": 0.3875685501098633,
      "mean_token_accuracy": 0.9209027971327305,
      "num_tokens": 2575616.0,
      "step": 2000
    },
    {
      "epoch": 4.832425892317,
      "eval_entropy": 0.4954172817811574,
      "eval_loss": 0.5729158520698547,
      "eval_mean_token_accuracy": 0.895403119799209,
      "eval_num_tokens": 2575616.0,
      "eval_runtime": 46.8739,
      "eval_samples_per_second": 6.187,
      "eval_steps_per_second": 1.557,
      "step": 2000
    },
    {
      "entropy": 0.43330108902535147,
      "epoch": 5.314579552329098,
      "grad_norm": 0.201171875,
      "learning_rate": 8.749159331986178e-05,
      "loss": 0.37568603515625,
      "mean_token_accuracy": 0.9220367068029856,
      "num_tokens": 2832652.0,
      "step": 2200
    },
    {
      "epoch": 5.314579552329098,
      "eval_entropy": 0.4961876812046521,
      "eval_loss": 0.5800752639770508,
      "eval_mean_token_accuracy": 0.8957706052963048,
      "eval_num_tokens": 2832652.0,
      "eval_runtime": 47.319,
      "eval_samples_per_second": 6.129,
      "eval_steps_per_second": 1.543,
      "step": 2200
    },
    {
      "entropy": 0.430719858892262,
      "epoch": 5.798548094373865,
      "grad_norm": 0.26171875,
      "learning_rate": 8.323937619458462e-05,
      "loss": 0.3715415191650391,
      "mean_token_accuracy": 0.9230316735059023,
      "num_tokens": 3090576.0,
      "step": 2400
    },
    {
      "epoch": 5.798548094373865,
      "eval_entropy": 0.4881298513445136,
      "eval_loss": 0.5789266228675842,
      "eval_mean_token_accuracy": 0.8953908681869507,
      "eval_num_tokens": 3090576.0,
      "eval_runtime": 47.2341,
      "eval_samples_per_second": 6.14,
      "eval_steps_per_second": 1.545,
      "step": 2400
    },
    {
      "entropy": 0.4203071318650934,
      "epoch": 6.280701754385965,
      "grad_norm": 0.2294921875,
      "learning_rate": 7.850431074824228e-05,
      "loss": 0.3627183151245117,
      "mean_token_accuracy": 0.9248409140364289,
      "num_tokens": 3346603.0,
      "step": 2600
    },
    {
      "epoch": 6.280701754385965,
      "eval_entropy": 0.479939534647824,
      "eval_loss": 0.5780320167541504,
      "eval_mean_token_accuracy": 0.8957429552731448,
      "eval_num_tokens": 3346603.0,
      "eval_runtime": 46.6552,
      "eval_samples_per_second": 6.216,
      "eval_steps_per_second": 1.565,
      "step": 2600
    },
    {
      "entropy": 0.4207753050327301,
      "epoch": 6.764670296430732,
      "grad_norm": 0.2109375,
      "learning_rate": 7.335518041369806e-05,
      "loss": 0.36228099822998044,
      "mean_token_accuracy": 0.924303879737854,
      "num_tokens": 3605206.0,
      "step": 2800
    },
    {
      "epoch": 6.764670296430732,
      "eval_entropy": 0.4856129292755911,
      "eval_loss": 0.5810655355453491,
      "eval_mean_token_accuracy": 0.8950460528674191,
      "eval_num_tokens": 3605206.0,
      "eval_runtime": 47.2855,
      "eval_samples_per_second": 6.133,
      "eval_steps_per_second": 1.544,
      "step": 2800
    },
    {
      "entropy": 0.4141107780421842,
      "epoch": 7.246823956442832,
      "grad_norm": 0.240234375,
      "learning_rate": 6.786678349409785e-05,
      "loss": 0.353370361328125,
      "mean_token_accuracy": 0.9260965282076425,
      "num_tokens": 3862259.0,
      "step": 3000
    },
    {
      "epoch": 7.246823956442832,
      "eval_entropy": 0.47775410626032583,
      "eval_loss": 0.5862641930580139,
      "eval_mean_token_accuracy": 0.8956269344238386,
      "eval_num_tokens": 3862259.0,
      "eval_runtime": 47.1135,
      "eval_samples_per_second": 6.155,
      "eval_steps_per_second": 1.549,
      "step": 3000
    },
    {
      "entropy": 0.4134879305586219,
      "epoch": 7.730792498487598,
      "grad_norm": 0.212890625,
      "learning_rate": 6.211884661314404e-05,
      "loss": 0.3520899963378906,
      "mean_token_accuracy": 0.9260335348546505,
      "num_tokens": 4120020.0,
      "step": 3200
    },
    {
      "epoch": 7.730792498487598,
      "eval_entropy": 0.47687137167747706,
      "eval_loss": 0.5826675295829773,
      "eval_mean_token_accuracy": 0.8959454404164667,
      "eval_num_tokens": 4120020.0,
      "eval_runtime": 46.9524,
      "eval_samples_per_second": 6.176,
      "eval_steps_per_second": 1.555,
      "step": 3200
    },
    {
      "entropy": 0.41237180912808763,
      "epoch": 8.212946158499697,
      "grad_norm": 0.23828125,
      "learning_rate": 5.6194866574636726e-05,
      "loss": 0.3477163314819336,
      "mean_token_accuracy": 0.9267811967408194,
      "num_tokens": 4377139.0,
      "step": 3400
    },
    {
      "epoch": 8.212946158499697,
      "eval_entropy": 0.47165755008997984,
      "eval_loss": 0.5873588919639587,
      "eval_mean_token_accuracy": 0.8950836046101296,
      "eval_num_tokens": 4377139.0,
      "eval_runtime": 54.7647,
      "eval_samples_per_second": 5.295,
      "eval_steps_per_second": 1.333,
      "step": 3400
    },
    {
      "entropy": 0.40712683595716953,
      "epoch": 8.696914700544465,
      "grad_norm": 0.232421875,
      "learning_rate": 5.018089745488805e-05,
      "loss": 0.3462469863891602,
      "mean_token_accuracy": 0.9266008900105953,
      "num_tokens": 4635344.0,
      "step": 3600
    },
    {
      "epoch": 8.696914700544465,
      "eval_entropy": 0.4733920579087244,
      "eval_loss": 0.5913234353065491,
      "eval_mean_token_accuracy": 0.895441471713863,
      "eval_num_tokens": 4635344.0,
      "eval_runtime": 98.6499,
      "eval_samples_per_second": 2.94,
      "eval_steps_per_second": 0.74,
      "step": 3600
    },
    {
      "entropy": 0.40435460647346083,
      "epoch": 9.179068360556563,
      "grad_norm": 0.2392578125,
      "learning_rate": 4.416430054718493e-05,
      "loss": 0.33988548278808595,
      "mean_token_accuracy": 0.9284380353529151,
      "num_tokens": 4891959.0,
      "step": 3800
    },
    {
      "epoch": 9.179068360556563,
      "eval_entropy": 0.4687396479796057,
      "eval_loss": 0.5922302603721619,
      "eval_mean_token_accuracy": 0.8952091801656435,
      "eval_num_tokens": 4891959.0,
      "eval_runtime": 98.6504,
      "eval_samples_per_second": 2.94,
      "eval_steps_per_second": 0.74,
      "step": 3800
    },
    {
      "entropy": 0.40334540244191885,
      "epoch": 9.66303690260133,
      "grad_norm": 0.25,
      "learning_rate": 3.823247531710108e-05,
      "loss": 0.34033252716064455,
      "mean_token_accuracy": 0.9283645600825549,
      "num_tokens": 5150117.0,
      "step": 4000
    },
    {
      "epoch": 9.66303690260133,
      "eval_entropy": 0.46922515271461174,
      "eval_loss": 0.5946717262268066,
      "eval_mean_token_accuracy": 0.8953593046697852,
      "eval_num_tokens": 5150117.0,
      "eval_runtime": 98.7836,
      "eval_samples_per_second": 2.936,
      "eval_steps_per_second": 0.739,
      "step": 4000
    },
    {
      "entropy": 0.39993977449469764,
      "epoch": 10.14519056261343,
      "grad_norm": 0.232421875,
      "learning_rate": 3.247158980330338e-05,
      "loss": 0.33406734466552734,
      "mean_token_accuracy": 0.9284365516684137,
      "num_tokens": 5406258.0,
      "step": 4200
    },
    {
      "epoch": 10.14519056261343,
      "eval_entropy": 0.4659598641199608,
      "eval_loss": 0.59522944688797,
      "eval_mean_token_accuracy": 0.894354227471025,
      "eval_num_tokens": 5406258.0,
      "eval_runtime": 98.2511,
      "eval_samples_per_second": 2.952,
      "eval_steps_per_second": 0.743,
      "step": 4200
    },
    {
      "entropy": 0.4000014416128397,
      "epoch": 10.629159104658196,
      "grad_norm": 0.28125,
      "learning_rate": 2.6965328906552368e-05,
      "loss": 0.33594528198242185,
      "mean_token_accuracy": 0.928432078063488,
      "num_tokens": 5665214.0,
      "step": 4400
    },
    {
      "epoch": 10.629159104658196,
      "eval_entropy": 0.4656533133493711,
      "eval_loss": 0.5966197848320007,
      "eval_mean_token_accuracy": 0.8958511099423447,
      "eval_num_tokens": 5665214.0,
      "eval_runtime": 97.708,
      "eval_samples_per_second": 2.968,
      "eval_steps_per_second": 0.747,
      "step": 4400
    },
    {
      "entropy": 0.39988665288335257,
      "epoch": 11.111312764670297,
      "grad_norm": 0.255859375,
      "learning_rate": 2.1793678749745773e-05,
      "loss": 0.33592803955078127,
      "mean_token_accuracy": 0.9285333580623756,
      "num_tokens": 5921976.0,
      "step": 4600
    },
    {
      "epoch": 11.111312764670297,
      "eval_entropy": 0.4668905788088498,
      "eval_loss": 0.595775306224823,
      "eval_mean_token_accuracy": 0.895175910975835,
      "eval_num_tokens": 5921976.0,
      "eval_runtime": 98.0497,
      "eval_samples_per_second": 2.958,
      "eval_steps_per_second": 0.745,
      "step": 4600
    },
    {
      "entropy": 0.39747313819825647,
      "epoch": 11.595281306715064,
      "grad_norm": 0.265625,
      "learning_rate": 1.703176476787164e-05,
      "loss": 0.33493621826171877,
      "mean_token_accuracy": 0.9288852622359991,
      "num_tokens": 6180319.0,
      "step": 4800
    },
    {
      "epoch": 11.595281306715064,
      "eval_entropy": 0.46607412214148536,
      "eval_loss": 0.5940157175064087,
      "eval_mean_token_accuracy": 0.8953437478574988,
      "eval_num_tokens": 6180319.0,
      "eval_runtime": 98.3503,
      "eval_samples_per_second": 2.949,
      "eval_steps_per_second": 0.742,
      "step": 4800
    },
    {
      "entropy": 0.3974911795655637,
      "epoch": 12.077434966727163,
      "grad_norm": 0.2333984375,
      "learning_rate": 1.2748760406236098e-05,
      "loss": 0.3305249786376953,
      "mean_token_accuracy": 0.9298606089396937,
      "num_tokens": 6436530.0,
      "step": 5000
    },
    {
      "epoch": 12.077434966727163,
      "eval_entropy": 0.46489216652635024,
      "eval_loss": 0.5976075530052185,
      "eval_mean_token_accuracy": 0.8952453854965837,
      "eval_num_tokens": 6436530.0,
      "eval_runtime": 98.1183,
      "eval_samples_per_second": 2.956,
      "eval_steps_per_second": 0.744,
      "step": 5000
    },
    {
      "entropy": 0.39817900501191617,
      "epoch": 12.56140350877193,
      "grad_norm": 0.271484375,
      "learning_rate": 9.006882279646356e-06,
      "loss": 0.3339969635009766,
      "mean_token_accuracy": 0.9290466893464326,
      "num_tokens": 6694685.0,
      "step": 5200
    },
    {
      "epoch": 12.56140350877193,
      "eval_entropy": 0.4657470506348022,
      "eval_loss": 0.597287118434906,
      "eval_mean_token_accuracy": 0.8953335562797442,
      "eval_num_tokens": 6694685.0,
      "eval_runtime": 47.8386,
      "eval_samples_per_second": 6.062,
      "eval_steps_per_second": 1.526,
      "step": 5200
    },
    {
      "entropy": 0.39675679298087374,
      "epoch": 13.043557168784028,
      "grad_norm": 0.25390625,
      "learning_rate": 5.860486389264325e-06,
      "loss": 0.3299782180786133,
      "mean_token_accuracy": 0.9294995974417463,
      "num_tokens": 6951487.0,
      "step": 5400
    },
    {
      "epoch": 13.043557168784028,
      "eval_entropy": 0.4653073390052743,
      "eval_loss": 0.5966233611106873,
      "eval_mean_token_accuracy": 0.8951566096854536,
      "eval_num_tokens": 6951487.0,
      "eval_runtime": 98.2374,
      "eval_samples_per_second": 2.952,
      "eval_steps_per_second": 0.743,
      "step": 5400
    },
    {
      "entropy": 0.3993960711359978,
      "epoch": 13.527525710828796,
      "grad_norm": 0.26171875,
      "learning_rate": 3.355278525842043e-06,
      "loss": 0.33360870361328127,
      "mean_token_accuracy": 0.9292359718680382,
      "num_tokens": 7209994.0,
      "step": 5600
    },
    {
      "epoch": 13.527525710828796,
      "eval_entropy": 0.46439201129625923,
      "eval_loss": 0.596992552280426,
      "eval_mean_token_accuracy": 0.8953333423562246,
      "eval_num_tokens": 7209994.0,
      "eval_runtime": 97.8986,
      "eval_samples_per_second": 2.962,
      "eval_steps_per_second": 0.746,
      "step": 5600
    },
    {
      "entropy": 0.3950923753053959,
      "epoch": 14.009679370840896,
      "grad_norm": 0.24609375,
      "learning_rate": 1.5276503293336187e-06,
      "loss": 0.3306077194213867,
      "mean_token_accuracy": 0.9297496843218355,
      "num_tokens": 7466401.0,
      "step": 5800
    },
    {
      "epoch": 14.009679370840896,
      "eval_entropy": 0.4649810607302679,
      "eval_loss": 0.5969449877738953,
      "eval_mean_token_accuracy": 0.8953231295494184,
      "eval_num_tokens": 7466401.0,
      "eval_runtime": 97.4653,
      "eval_samples_per_second": 2.975,
      "eval_steps_per_second": 0.749,
      "step": 5800
    },
    {
      "entropy": 0.39669271767139436,
      "epoch": 14.493647912885663,
      "grad_norm": 0.2412109375,
      "learning_rate": 4.0415064954446023e-07,
      "loss": 0.33266185760498046,
      "mean_token_accuracy": 0.929519795998931,
      "num_tokens": 7724491.0,
      "step": 6000
    },
    {
      "epoch": 14.493647912885663,
      "eval_entropy": 0.46521988633560807,
      "eval_loss": 0.5968044400215149,
      "eval_mean_token_accuracy": 0.8951437881548111,
      "eval_num_tokens": 7724491.0,
      "eval_runtime": 97.4586,
      "eval_samples_per_second": 2.976,
      "eval_steps_per_second": 0.749,
      "step": 6000
    },
    {
      "entropy": 0.3974524886161089,
      "epoch": 14.977616454930429,
      "grad_norm": 0.310546875,
      "learning_rate": 1.099887041894343e-09,
      "loss": 0.3300043487548828,
      "mean_token_accuracy": 0.9291912470757961,
      "num_tokens": 7982222.0,
      "step": 6200
    },
    {
      "epoch": 14.977616454930429,
      "eval_entropy": 0.4640667352774372,
      "eval_loss": 0.5965780019760132,
      "eval_mean_token_accuracy": 0.8952341332827529,
      "eval_num_tokens": 7982222.0,
      "eval_runtime": 98.1877,
      "eval_samples_per_second": 2.954,
      "eval_steps_per_second": 0.743,
      "step": 6200
    },
    {
      "epoch": 15.0,
      "eval_entropy": 0.4640667352774372,
      "eval_loss": 0.5965780019760132,
      "eval_mean_token_accuracy": 0.8952341332827529,
      "eval_num_tokens": 7994175.0,
      "eval_runtime": 97.9356,
      "eval_samples_per_second": 2.961,
      "eval_steps_per_second": 0.745,
      "step": 6210
    }
  ],
  "logging_steps": 200,
  "max_steps": 6210,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 15,
  "save_steps": 200,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.1728842516559872e+17,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}