{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.011254924029262802,
  "eval_steps": 9,
  "global_step": 100,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.00011254924029262802,
      "grad_norm": NaN,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.0,
      "step": 1
    },
    {
      "epoch": 0.00011254924029262802,
      "eval_loss": NaN,
      "eval_runtime": 880.1072,
      "eval_samples_per_second": 8.502,
      "eval_steps_per_second": 1.064,
      "step": 1
    },
    {
      "epoch": 0.00022509848058525604,
      "grad_norm": NaN,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.0,
      "step": 2
    },
    {
      "epoch": 0.0003376477208778841,
      "grad_norm": NaN,
      "learning_rate": 6e-06,
      "loss": 0.0,
      "step": 3
    },
    {
      "epoch": 0.0004501969611705121,
      "grad_norm": NaN,
      "learning_rate": 8.000000000000001e-06,
      "loss": 0.0,
      "step": 4
    },
    {
      "epoch": 0.0005627462014631402,
      "grad_norm": NaN,
      "learning_rate": 1e-05,
      "loss": 0.0,
      "step": 5
    },
    {
      "epoch": 0.0006752954417557682,
      "grad_norm": NaN,
      "learning_rate": 1.2e-05,
      "loss": 0.0,
      "step": 6
    },
    {
      "epoch": 0.0007878446820483962,
      "grad_norm": NaN,
      "learning_rate": 1.4e-05,
      "loss": 0.0,
      "step": 7
    },
    {
      "epoch": 0.0009003939223410242,
      "grad_norm": NaN,
      "learning_rate": 1.6000000000000003e-05,
      "loss": 0.0,
      "step": 8
    },
    {
      "epoch": 0.0010129431626336523,
      "grad_norm": NaN,
      "learning_rate": 1.8e-05,
      "loss": 0.0,
      "step": 9
    },
    {
      "epoch": 0.0010129431626336523,
      "eval_loss": NaN,
      "eval_runtime": 878.9983,
      "eval_samples_per_second": 8.513,
      "eval_steps_per_second": 1.065,
      "step": 9
    },
    {
      "epoch": 0.0011254924029262803,
      "grad_norm": NaN,
      "learning_rate": 2e-05,
      "loss": 0.0,
      "step": 10
    },
    {
      "epoch": 0.0012380416432189083,
      "grad_norm": NaN,
      "learning_rate": 1.999390827019096e-05,
      "loss": 0.0,
      "step": 11
    },
    {
      "epoch": 0.0013505908835115363,
      "grad_norm": NaN,
      "learning_rate": 1.9975640502598243e-05,
      "loss": 0.0,
      "step": 12
    },
    {
      "epoch": 0.0014631401238041643,
      "grad_norm": NaN,
      "learning_rate": 1.9945218953682736e-05,
      "loss": 0.0,
      "step": 13
    },
    {
      "epoch": 0.0015756893640967923,
      "grad_norm": NaN,
      "learning_rate": 1.9902680687415704e-05,
      "loss": 0.0,
      "step": 14
    },
    {
      "epoch": 0.0016882386043894203,
      "grad_norm": NaN,
      "learning_rate": 1.9848077530122083e-05,
      "loss": 0.0,
      "step": 15
    },
    {
      "epoch": 0.0018007878446820484,
      "grad_norm": NaN,
      "learning_rate": 1.9781476007338058e-05,
      "loss": 0.0,
      "step": 16
    },
    {
      "epoch": 0.0019133370849746764,
      "grad_norm": NaN,
      "learning_rate": 1.9702957262759964e-05,
      "loss": 0.0,
      "step": 17
    },
    {
      "epoch": 0.0020258863252673046,
      "grad_norm": NaN,
      "learning_rate": 1.961261695938319e-05,
      "loss": 0.0,
      "step": 18
    },
    {
      "epoch": 0.0020258863252673046,
      "eval_loss": NaN,
      "eval_runtime": 879.0488,
      "eval_samples_per_second": 8.513,
      "eval_steps_per_second": 1.065,
      "step": 18
    },
    {
      "epoch": 0.0021384355655599324,
      "grad_norm": NaN,
      "learning_rate": 1.9510565162951538e-05,
      "loss": 0.0,
      "step": 19
    },
    {
      "epoch": 0.0022509848058525606,
      "grad_norm": NaN,
      "learning_rate": 1.9396926207859085e-05,
      "loss": 0.0,
      "step": 20
    },
    {
      "epoch": 0.0023635340461451884,
      "grad_norm": NaN,
      "learning_rate": 1.9271838545667876e-05,
      "loss": 0.0,
      "step": 21
    },
    {
      "epoch": 0.0024760832864378166,
      "grad_norm": NaN,
      "learning_rate": 1.913545457642601e-05,
      "loss": 0.0,
      "step": 22
    },
    {
      "epoch": 0.0025886325267304444,
      "grad_norm": NaN,
      "learning_rate": 1.8987940462991673e-05,
      "loss": 0.0,
      "step": 23
    },
    {
      "epoch": 0.0027011817670230726,
      "grad_norm": NaN,
      "learning_rate": 1.8829475928589272e-05,
      "loss": 0.0,
      "step": 24
    },
    {
      "epoch": 0.0028137310073157004,
      "grad_norm": NaN,
      "learning_rate": 1.866025403784439e-05,
      "loss": 0.0,
      "step": 25
    },
    {
      "epoch": 0.0029262802476083287,
      "grad_norm": NaN,
      "learning_rate": 1.848048096156426e-05,
      "loss": 0.0,
      "step": 26
    },
    {
      "epoch": 0.003038829487900957,
      "grad_norm": NaN,
      "learning_rate": 1.8290375725550417e-05,
      "loss": 0.0,
      "step": 27
    },
    {
      "epoch": 0.003038829487900957,
      "eval_loss": NaN,
      "eval_runtime": 879.7653,
      "eval_samples_per_second": 8.506,
      "eval_steps_per_second": 1.064,
      "step": 27
    },
    {
      "epoch": 0.0031513787281935847,
      "grad_norm": NaN,
      "learning_rate": 1.8090169943749477e-05,
      "loss": 0.0,
      "step": 28
    },
    {
      "epoch": 0.003263927968486213,
      "grad_norm": NaN,
      "learning_rate": 1.788010753606722e-05,
      "loss": 0.0,
      "step": 29
    },
    {
      "epoch": 0.0033764772087788407,
      "grad_norm": NaN,
      "learning_rate": 1.766044443118978e-05,
      "loss": 0.0,
      "step": 30
    },
    {
      "epoch": 0.003489026449071469,
      "grad_norm": NaN,
      "learning_rate": 1.7431448254773943e-05,
      "loss": 0.0,
      "step": 31
    },
    {
      "epoch": 0.0036015756893640967,
      "grad_norm": NaN,
      "learning_rate": 1.7193398003386514e-05,
      "loss": 0.0,
      "step": 32
    },
    {
      "epoch": 0.003714124929656725,
      "grad_norm": NaN,
      "learning_rate": 1.6946583704589973e-05,
      "loss": 0.0,
      "step": 33
    },
    {
      "epoch": 0.0038266741699493527,
      "grad_norm": NaN,
      "learning_rate": 1.6691306063588583e-05,
      "loss": 0.0,
      "step": 34
    },
    {
      "epoch": 0.0039392234102419805,
      "grad_norm": NaN,
      "learning_rate": 1.6427876096865394e-05,
      "loss": 0.0,
      "step": 35
    },
    {
      "epoch": 0.004051772650534609,
      "grad_norm": NaN,
      "learning_rate": 1.6156614753256583e-05,
      "loss": 0.0,
      "step": 36
    },
    {
      "epoch": 0.004051772650534609,
      "eval_loss": NaN,
      "eval_runtime": 879.2177,
      "eval_samples_per_second": 8.511,
      "eval_steps_per_second": 1.065,
      "step": 36
    },
    {
      "epoch": 0.004164321890827237,
      "grad_norm": NaN,
      "learning_rate": 1.5877852522924733e-05,
      "loss": 0.0,
      "step": 37
    },
    {
      "epoch": 0.004276871131119865,
      "grad_norm": NaN,
      "learning_rate": 1.5591929034707468e-05,
      "loss": 0.0,
      "step": 38
    },
    {
      "epoch": 0.0043894203714124926,
      "grad_norm": NaN,
      "learning_rate": 1.529919264233205e-05,
      "loss": 0.0,
      "step": 39
    },
    {
      "epoch": 0.004501969611705121,
      "grad_norm": NaN,
      "learning_rate": 1.5000000000000002e-05,
      "loss": 0.0,
      "step": 40
    },
    {
      "epoch": 0.004614518851997749,
      "grad_norm": NaN,
      "learning_rate": 1.469471562785891e-05,
      "loss": 0.0,
      "step": 41
    },
    {
      "epoch": 0.004727068092290377,
      "grad_norm": NaN,
      "learning_rate": 1.4383711467890776e-05,
      "loss": 0.0,
      "step": 42
    },
    {
      "epoch": 0.0048396173325830055,
      "grad_norm": NaN,
      "learning_rate": 1.4067366430758004e-05,
      "loss": 0.0,
      "step": 43
    },
    {
      "epoch": 0.004952166572875633,
      "grad_norm": NaN,
      "learning_rate": 1.3746065934159123e-05,
      "loss": 0.0,
      "step": 44
    },
    {
      "epoch": 0.005064715813168261,
      "grad_norm": NaN,
      "learning_rate": 1.342020143325669e-05,
      "loss": 0.0,
      "step": 45
    },
    {
      "epoch": 0.005064715813168261,
      "eval_loss": NaN,
      "eval_runtime": 879.2121,
      "eval_samples_per_second": 8.511,
      "eval_steps_per_second": 1.065,
      "step": 45
    },
    {
      "epoch": 0.005177265053460889,
      "grad_norm": NaN,
      "learning_rate": 1.3090169943749475e-05,
      "loss": 0.0,
      "step": 46
    },
    {
      "epoch": 0.0052898142937535175,
      "grad_norm": NaN,
      "learning_rate": 1.2756373558169992e-05,
      "loss": 0.0,
      "step": 47
    },
    {
      "epoch": 0.005402363534046145,
      "grad_norm": NaN,
      "learning_rate": 1.2419218955996677e-05,
      "loss": 0.0,
      "step": 48
    },
    {
      "epoch": 0.005514912774338773,
      "grad_norm": NaN,
      "learning_rate": 1.2079116908177592e-05,
      "loss": 0.0,
      "step": 49
    },
    {
      "epoch": 0.005627462014631401,
      "grad_norm": NaN,
      "learning_rate": 1.1736481776669307e-05,
      "loss": 0.0,
      "step": 50
    },
    {
      "epoch": 0.0057400112549240295,
      "grad_norm": NaN,
      "learning_rate": 1.1391731009600655e-05,
      "loss": 0.0,
      "step": 51
    },
    {
      "epoch": 0.005852560495216657,
      "grad_norm": NaN,
      "learning_rate": 1.1045284632676535e-05,
      "loss": 0.0,
      "step": 52
    },
    {
      "epoch": 0.005965109735509285,
      "grad_norm": NaN,
      "learning_rate": 1.0697564737441254e-05,
      "loss": 0.0,
      "step": 53
    },
    {
      "epoch": 0.006077658975801914,
      "grad_norm": NaN,
      "learning_rate": 1.0348994967025012e-05,
      "loss": 0.0,
      "step": 54
    },
    {
      "epoch": 0.006077658975801914,
      "eval_loss": NaN,
      "eval_runtime": 879.2894,
      "eval_samples_per_second": 8.51,
      "eval_steps_per_second": 1.064,
      "step": 54
    },
    {
      "epoch": 0.0061902082160945416,
      "grad_norm": NaN,
      "learning_rate": 1e-05,
      "loss": 0.0,
      "step": 55
    },
    {
      "epoch": 0.006302757456387169,
      "grad_norm": NaN,
      "learning_rate": 9.651005032974994e-06,
      "loss": 0.0,
      "step": 56
    },
    {
      "epoch": 0.006415306696679797,
      "grad_norm": NaN,
      "learning_rate": 9.302435262558748e-06,
      "loss": 0.0,
      "step": 57
    },
    {
      "epoch": 0.006527855936972426,
      "grad_norm": NaN,
      "learning_rate": 8.954715367323468e-06,
      "loss": 0.0,
      "step": 58
    },
    {
      "epoch": 0.006640405177265054,
      "grad_norm": NaN,
      "learning_rate": 8.60826899039935e-06,
      "loss": 0.0,
      "step": 59
    },
    {
      "epoch": 0.006752954417557681,
      "grad_norm": NaN,
      "learning_rate": 8.263518223330698e-06,
      "loss": 0.0,
      "step": 60
    },
    {
      "epoch": 0.006865503657850309,
      "grad_norm": NaN,
      "learning_rate": 7.92088309182241e-06,
      "loss": 0.0,
      "step": 61
    },
    {
      "epoch": 0.006978052898142938,
      "grad_norm": NaN,
      "learning_rate": 7.580781044003324e-06,
      "loss": 0.0,
      "step": 62
    },
    {
      "epoch": 0.007090602138435566,
      "grad_norm": NaN,
      "learning_rate": 7.243626441830009e-06,
      "loss": 0.0,
      "step": 63
    },
    {
      "epoch": 0.007090602138435566,
      "eval_loss": NaN,
      "eval_runtime": 879.3072,
      "eval_samples_per_second": 8.51,
      "eval_steps_per_second": 1.064,
      "step": 63
    },
    {
      "epoch": 0.007203151378728193,
      "grad_norm": NaN,
      "learning_rate": 6.909830056250527e-06,
      "loss": 0.0,
      "step": 64
    },
    {
      "epoch": 0.007315700619020821,
      "grad_norm": NaN,
      "learning_rate": 6.579798566743314e-06,
      "loss": 0.0,
      "step": 65
    },
    {
      "epoch": 0.00742824985931345,
      "grad_norm": NaN,
      "learning_rate": 6.25393406584088e-06,
      "loss": 0.0,
      "step": 66
    },
    {
      "epoch": 0.007540799099606078,
      "grad_norm": NaN,
      "learning_rate": 5.932633569242e-06,
      "loss": 0.0,
      "step": 67
    },
    {
      "epoch": 0.0076533483398987055,
      "grad_norm": NaN,
      "learning_rate": 5.616288532109225e-06,
      "loss": 0.0,
      "step": 68
    },
    {
      "epoch": 0.007765897580191334,
      "grad_norm": NaN,
      "learning_rate": 5.305284372141095e-06,
      "loss": 0.0,
      "step": 69
    },
    {
      "epoch": 0.007878446820483961,
      "grad_norm": NaN,
      "learning_rate": 5.000000000000003e-06,
      "loss": 0.0,
      "step": 70
    },
    {
      "epoch": 0.00799099606077659,
      "grad_norm": NaN,
      "learning_rate": 4.700807357667953e-06,
      "loss": 0.0,
      "step": 71
    },
    {
      "epoch": 0.008103545301069218,
      "grad_norm": NaN,
      "learning_rate": 4.408070965292534e-06,
      "loss": 0.0,
      "step": 72
    },
    {
      "epoch": 0.008103545301069218,
      "eval_loss": NaN,
      "eval_runtime": 879.1927,
      "eval_samples_per_second": 8.511,
      "eval_steps_per_second": 1.065,
      "step": 72
    },
    {
      "epoch": 0.008216094541361846,
      "grad_norm": NaN,
      "learning_rate": 4.12214747707527e-06,
      "loss": 0.0,
      "step": 73
    },
    {
      "epoch": 0.008328643781654474,
      "grad_norm": NaN,
      "learning_rate": 3.8433852467434175e-06,
      "loss": 0.0,
      "step": 74
    },
    {
      "epoch": 0.008441193021947102,
      "grad_norm": NaN,
      "learning_rate": 3.5721239031346067e-06,
      "loss": 0.0,
      "step": 75
    },
    {
      "epoch": 0.00855374226223973,
      "grad_norm": NaN,
      "learning_rate": 3.308693936411421e-06,
      "loss": 0.0,
      "step": 76
    },
    {
      "epoch": 0.008666291502532357,
      "grad_norm": NaN,
      "learning_rate": 3.0534162954100264e-06,
      "loss": 0.0,
      "step": 77
    },
    {
      "epoch": 0.008778840742824985,
      "grad_norm": NaN,
      "learning_rate": 2.8066019966134907e-06,
      "loss": 0.0,
      "step": 78
    },
    {
      "epoch": 0.008891389983117615,
      "grad_norm": NaN,
      "learning_rate": 2.5685517452260566e-06,
      "loss": 0.0,
      "step": 79
    },
    {
      "epoch": 0.009003939223410242,
      "grad_norm": NaN,
      "learning_rate": 2.339555568810221e-06,
      "loss": 0.0,
      "step": 80
    },
    {
      "epoch": 0.00911648846370287,
      "grad_norm": NaN,
      "learning_rate": 2.119892463932781e-06,
      "loss": 0.0,
      "step": 81
    },
    {
      "epoch": 0.00911648846370287,
      "eval_loss": NaN,
      "eval_runtime": 879.305,
      "eval_samples_per_second": 8.51,
      "eval_steps_per_second": 1.064,
      "step": 81
    },
    {
      "epoch": 0.009229037703995498,
      "grad_norm": NaN,
      "learning_rate": 1.9098300562505266e-06,
      "loss": 0.0,
      "step": 82
    },
    {
      "epoch": 0.009341586944288126,
      "grad_norm": NaN,
      "learning_rate": 1.709624274449584e-06,
      "loss": 0.0,
      "step": 83
    },
    {
      "epoch": 0.009454136184580754,
      "grad_norm": NaN,
      "learning_rate": 1.5195190384357405e-06,
      "loss": 0.0,
      "step": 84
    },
    {
      "epoch": 0.009566685424873381,
      "grad_norm": NaN,
      "learning_rate": 1.339745962155613e-06,
      "loss": 0.0,
      "step": 85
    },
    {
      "epoch": 0.009679234665166011,
      "grad_norm": NaN,
      "learning_rate": 1.1705240714107301e-06,
      "loss": 0.0,
      "step": 86
    },
    {
      "epoch": 0.009791783905458639,
      "grad_norm": NaN,
      "learning_rate": 1.012059537008332e-06,
      "loss": 0.0,
      "step": 87
    },
    {
      "epoch": 0.009904333145751266,
      "grad_norm": NaN,
      "learning_rate": 8.645454235739903e-07,
      "loss": 0.0,
      "step": 88
    },
    {
      "epoch": 0.010016882386043894,
      "grad_norm": NaN,
      "learning_rate": 7.281614543321269e-07,
      "loss": 0.0,
      "step": 89
    },
    {
      "epoch": 0.010129431626336522,
      "grad_norm": NaN,
      "learning_rate": 6.030737921409169e-07,
      "loss": 0.0,
      "step": 90
    },
    {
      "epoch": 0.010129431626336522,
      "eval_loss": NaN,
      "eval_runtime": 879.5002,
      "eval_samples_per_second": 8.508,
      "eval_steps_per_second": 1.064,
      "step": 90
    },
    {
      "epoch": 0.01024198086662915,
      "grad_norm": NaN,
      "learning_rate": 4.894348370484648e-07,
      "loss": 0.0,
      "step": 91
    },
    {
      "epoch": 0.010354530106921778,
      "grad_norm": NaN,
      "learning_rate": 3.8738304061681107e-07,
      "loss": 0.0,
      "step": 92
    },
    {
      "epoch": 0.010467079347214405,
      "grad_norm": NaN,
      "learning_rate": 2.970427372400353e-07,
      "loss": 0.0,
      "step": 93
    },
    {
      "epoch": 0.010579628587507035,
      "grad_norm": NaN,
      "learning_rate": 2.1852399266194312e-07,
      "loss": 0.0,
      "step": 94
    },
    {
      "epoch": 0.010692177827799663,
      "grad_norm": NaN,
      "learning_rate": 1.519224698779198e-07,
      "loss": 0.0,
      "step": 95
    },
    {
      "epoch": 0.01080472706809229,
      "grad_norm": NaN,
      "learning_rate": 9.731931258429638e-08,
      "loss": 0.0,
      "step": 96
    },
    {
      "epoch": 0.010917276308384918,
      "grad_norm": NaN,
      "learning_rate": 5.4781046317267103e-08,
      "loss": 0.0,
      "step": 97
    },
    {
      "epoch": 0.011029825548677546,
      "grad_norm": NaN,
      "learning_rate": 2.4359497401758026e-08,
      "loss": 0.0,
      "step": 98
    },
    {
      "epoch": 0.011142374788970174,
      "grad_norm": NaN,
      "learning_rate": 6.091729809042379e-09,
      "loss": 0.0,
      "step": 99
    },
    {
      "epoch": 0.011142374788970174,
      "eval_loss": NaN,
      "eval_runtime": 879.4643,
      "eval_samples_per_second": 8.509,
      "eval_steps_per_second": 1.064,
      "step": 99
    },
    {
      "epoch": 0.011254924029262802,
      "grad_norm": NaN,
      "learning_rate": 0.0,
      "loss": 0.0,
      "step": 100
    }
  ],
  "logging_steps": 1,
  "max_steps": 100,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "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": 7.901099569918771e+16,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}