{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.0,
  "eval_steps": 1024,
  "global_step": 8100,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.06320987654320988,
      "grad_norm": 1.6552858352661133,
      "learning_rate": 0.000248046875,
      "loss": 3.450495481491089,
      "step": 128
    },
    {
      "epoch": 0.12641975308641976,
      "grad_norm": 1.9175924062728882,
      "learning_rate": 0.000498046875,
      "loss": 1.3704386949539185,
      "step": 256
    },
    {
      "epoch": 0.18962962962962962,
      "grad_norm": 1.2898952960968018,
      "learning_rate": 0.000748046875,
      "loss": 1.171890139579773,
      "step": 384
    },
    {
      "epoch": 0.2528395061728395,
      "grad_norm": 0.7517929673194885,
      "learning_rate": 0.000998046875,
      "loss": 1.1315792798995972,
      "step": 512
    },
    {
      "epoch": 0.3160493827160494,
      "grad_norm": 0.6297398805618286,
      "learning_rate": 0.0009993089770195807,
      "loss": 1.1104586124420166,
      "step": 640
    },
    {
      "epoch": 0.37925925925925924,
      "grad_norm": 0.706788957118988,
      "learning_rate": 0.0009972160460972733,
      "loss": 1.1022820472717285,
      "step": 768
    },
    {
      "epoch": 0.44246913580246916,
      "grad_norm": 0.5103466510772705,
      "learning_rate": 0.0009937270408736224,
      "loss": 1.0948997735977173,
      "step": 896
    },
    {
      "epoch": 0.505679012345679,
      "grad_norm": 0.47909650206565857,
      "learning_rate": 0.0009888517577149526,
      "loss": 1.0888899564743042,
      "step": 1024
    },
    {
      "epoch": 0.505679012345679,
      "eval_bleu": 0.0,
      "eval_cos_loss": 0.3223608470544582,
      "eval_loss": 1.0882294381537088,
      "eval_mse_loss": 1.0882294381537088,
      "step": 1024
    },
    {
      "epoch": 0.505679012345679,
      "eval_bleu": 0.0,
      "eval_cos_loss": 0.3223608470544582,
      "eval_loss": 1.0882294381537088,
      "eval_mse_loss": 1.0882294381537088,
      "eval_runtime": 17.5658,
      "eval_samples_per_second": 149.04,
      "eval_steps_per_second": 2.334,
      "step": 1024
    },
    {
      "epoch": 0.5688888888888889,
      "grad_norm": 0.5776141881942749,
      "learning_rate": 0.0009826038853539248,
      "loss": 1.0844441652297974,
      "step": 1152
    },
    {
      "epoch": 0.6320987654320988,
      "grad_norm": 0.39693504571914673,
      "learning_rate": 0.0009750009664545572,
      "loss": 1.0849089622497559,
      "step": 1280
    },
    {
      "epoch": 0.6953086419753086,
      "grad_norm": 0.5093810558319092,
      "learning_rate": 0.0009660643483562486,
      "loss": 1.0845346450805664,
      "step": 1408
    },
    {
      "epoch": 0.7585185185185185,
      "grad_norm": 0.6632571816444397,
      "learning_rate": 0.0009558191231351013,
      "loss": 1.079752802848816,
      "step": 1536
    },
    {
      "epoch": 0.8217283950617283,
      "grad_norm": 0.5488501787185669,
      "learning_rate": 0.0009442940571508399,
      "loss": 1.0838395357131958,
      "step": 1664
    },
    {
      "epoch": 0.8849382716049383,
      "grad_norm": 0.7813119888305664,
      "learning_rate": 0.0009315215102771411,
      "loss": 1.0782536268234253,
      "step": 1792
    },
    {
      "epoch": 0.9481481481481482,
      "grad_norm": 0.6261882185935974,
      "learning_rate": 0.0009175373450421618,
      "loss": 1.0775682926177979,
      "step": 1920
    },
    {
      "epoch": 1.011358024691358,
      "grad_norm": 0.613799512386322,
      "learning_rate": 0.0009023808259343743,
      "loss": 1.0777174234390259,
      "step": 2048
    },
    {
      "epoch": 1.011358024691358,
      "eval_bleu": 0.0,
      "eval_cos_loss": 0.3193167927788525,
      "eval_loss": 1.0737163497180473,
      "eval_mse_loss": 1.0737163497180473,
      "step": 2048
    },
    {
      "epoch": 1.011358024691358,
      "eval_bleu": 0.0,
      "eval_cos_loss": 0.3193167927788525,
      "eval_loss": 1.0737163497180473,
      "eval_mse_loss": 1.0737163497180473,
      "eval_runtime": 17.0019,
      "eval_samples_per_second": 153.982,
      "eval_steps_per_second": 2.411,
      "step": 2048
    },
    {
      "epoch": 1.074567901234568,
      "grad_norm": 0.4670981764793396,
      "learning_rate": 0.0008860945091564399,
      "loss": 1.0761820077896118,
      "step": 2176
    },
    {
      "epoch": 1.1377777777777778,
      "grad_norm": 0.6427945494651794,
      "learning_rate": 0.0008687241231366662,
      "loss": 1.0726263523101807,
      "step": 2304
    },
    {
      "epoch": 1.2009876543209876,
      "grad_norm": 0.46963557600975037,
      "learning_rate": 0.0008503184401335448,
      "loss": 1.0703215599060059,
      "step": 2432
    },
    {
      "epoch": 1.2641975308641975,
      "grad_norm": 0.785935640335083,
      "learning_rate": 0.0008309291392938795,
      "loss": 1.070408821105957,
      "step": 2560
    },
    {
      "epoch": 1.3274074074074074,
      "grad_norm": 0.4883745014667511,
      "learning_rate": 0.0008106106615490032,
      "loss": 1.069784164428711,
      "step": 2688
    },
    {
      "epoch": 1.3906172839506172,
      "grad_norm": 0.5132769346237183,
      "learning_rate": 0.0007894200567565075,
      "loss": 1.0658392906188965,
      "step": 2816
    },
    {
      "epoch": 1.453827160493827,
      "grad_norm": 0.7683017253875732,
      "learning_rate": 0.0007674168235166747,
      "loss": 1.0659981966018677,
      "step": 2944
    },
    {
      "epoch": 1.5170370370370372,
      "grad_norm": 0.49547040462493896,
      "learning_rate": 0.000744662742113375,
      "loss": 1.0679717063903809,
      "step": 3072
    },
    {
      "epoch": 1.5170370370370372,
      "eval_bleu": 0.0009612199755714339,
      "eval_cos_loss": 0.31748645988906304,
      "eval_loss": 1.066777130452598,
      "eval_mse_loss": 1.066777130452598,
      "step": 3072
    },
    {
      "epoch": 1.5170370370370372,
      "eval_bleu": 0.0009612199755714339,
      "eval_cos_loss": 0.31748645988906304,
      "eval_loss": 1.066777130452598,
      "eval_mse_loss": 1.066777130452598,
      "eval_runtime": 17.323,
      "eval_samples_per_second": 151.128,
      "eval_steps_per_second": 2.367,
      "step": 3072
    },
    {
      "epoch": 1.5802469135802468,
      "grad_norm": 0.33905908465385437,
      "learning_rate": 0.0007212217010484917,
      "loss": 1.062569260597229,
      "step": 3200
    },
    {
      "epoch": 1.643456790123457,
      "grad_norm": 0.5646777749061584,
      "learning_rate": 0.0006971595176569332,
      "loss": 1.0659024715423584,
      "step": 3328
    },
    {
      "epoch": 1.7066666666666666,
      "grad_norm": 0.4753640294075012,
      "learning_rate": 0.0006725437533058971,
      "loss": 1.0629714727401733,
      "step": 3456
    },
    {
      "epoch": 1.7698765432098766,
      "grad_norm": 0.4937284290790558,
      "learning_rate": 0.0006474435236972767,
      "loss": 1.0599957704544067,
      "step": 3584
    },
    {
      "epoch": 1.8330864197530863,
      "grad_norm": 0.6321665644645691,
      "learning_rate": 0.0006219293048058301,
      "loss": 1.0651935338974,
      "step": 3712
    },
    {
      "epoch": 1.8962962962962964,
      "grad_norm": 0.4198386073112488,
      "learning_rate": 0.0005960727349980042,
      "loss": 1.0622215270996094,
      "step": 3840
    },
    {
      "epoch": 1.959506172839506,
      "grad_norm": 0.44022831320762634,
      "learning_rate": 0.0005699464138870106,
      "loss": 1.0583852529525757,
      "step": 3968
    },
    {
      "epoch": 2.022716049382716,
      "grad_norm": 0.47051143646240234,
      "learning_rate": 0.0005436236984889356,
      "loss": 1.0613298416137695,
      "step": 4096
    },
    {
      "epoch": 2.022716049382716,
      "eval_bleu": 0.0011512758175346559,
      "eval_cos_loss": 0.31534567766073274,
      "eval_loss": 1.0575819248106422,
      "eval_mse_loss": 1.0575819248106422,
      "step": 4096
    },
    {
      "epoch": 2.022716049382716,
      "eval_bleu": 0.0011512758175346559,
      "eval_cos_loss": 0.31534567766073274,
      "eval_loss": 1.0575819248106422,
      "eval_mse_loss": 1.0575819248106422,
      "eval_runtime": 17.0064,
      "eval_samples_per_second": 153.942,
      "eval_steps_per_second": 2.411,
      "step": 4096
    },
    {
      "epoch": 2.0859259259259257,
      "grad_norm": 0.5936776995658875,
      "learning_rate": 0.0005171784972522235,
      "loss": 1.059881567955017,
      "step": 4224
    },
    {
      "epoch": 2.149135802469136,
      "grad_norm": 0.50244140625,
      "learning_rate": 0.0004906850625388614,
      "loss": 1.0667545795440674,
      "step": 4352
    },
    {
      "epoch": 2.212345679012346,
      "grad_norm": 0.3169984817504883,
      "learning_rate": 0.0004642177821399269,
      "loss": 1.058973789215088,
      "step": 4480
    },
    {
      "epoch": 2.2755555555555556,
      "grad_norm": 0.6188138127326965,
      "learning_rate": 0.00043785097041088293,
      "loss": 1.0577284097671509,
      "step": 4608
    },
    {
      "epoch": 2.338765432098765,
      "grad_norm": 0.6315109729766846,
      "learning_rate": 0.00041165865961306135,
      "loss": 1.0569205284118652,
      "step": 4736
    },
    {
      "epoch": 2.4019753086419753,
      "grad_norm": 0.5078902840614319,
      "learning_rate": 0.000385714392047208,
      "loss": 1.0600513219833374,
      "step": 4864
    },
    {
      "epoch": 2.4651851851851854,
      "grad_norm": 0.4911118447780609,
      "learning_rate": 0.00036009101356272814,
      "loss": 1.0586142539978027,
      "step": 4992
    },
    {
      "epoch": 2.528395061728395,
      "grad_norm": 0.3357234299182892,
      "learning_rate": 0.00033486046902241664,
      "loss": 1.0536247491836548,
      "step": 5120
    },
    {
      "epoch": 2.528395061728395,
      "eval_bleu": 0.0025068600746499426,
      "eval_cos_loss": 0.30963056407323697,
      "eval_loss": 1.0551867717649879,
      "eval_mse_loss": 1.0551867717649879,
      "step": 5120
    },
    {
      "epoch": 2.528395061728395,
      "eval_bleu": 0.0025068600746499426,
      "eval_cos_loss": 0.30963056407323697,
      "eval_loss": 1.0551867717649879,
      "eval_mse_loss": 1.0551867717649879,
      "eval_runtime": 17.2497,
      "eval_samples_per_second": 151.771,
      "eval_steps_per_second": 2.377,
      "step": 5120
    },
    {
      "epoch": 2.591604938271605,
      "grad_norm": 0.7107070088386536,
      "learning_rate": 0.00031009360029696107,
      "loss": 1.054158329963684,
      "step": 5248
    },
    {
      "epoch": 2.6548148148148147,
      "grad_norm": 0.38434842228889465,
      "learning_rate": 0.00028585994735640823,
      "loss": 1.0530760288238525,
      "step": 5376
    },
    {
      "epoch": 2.718024691358025,
      "grad_norm": 0.5316303372383118,
      "learning_rate": 0.0002622275530170825,
      "loss": 1.0520426034927368,
      "step": 5504
    },
    {
      "epoch": 2.7812345679012345,
      "grad_norm": 0.47037559747695923,
      "learning_rate": 0.0002392627718921852,
      "loss": 1.0520830154418945,
      "step": 5632
    },
    {
      "epoch": 2.8444444444444446,
      "grad_norm": 0.27498167753219604,
      "learning_rate": 0.00021703008408250187,
      "loss": 1.0527888536453247,
      "step": 5760
    },
    {
      "epoch": 2.907654320987654,
      "grad_norm": 0.35426315665245056,
      "learning_rate": 0.00019559191413033017,
      "loss": 1.0490405559539795,
      "step": 5888
    },
    {
      "epoch": 2.9708641975308643,
      "grad_norm": 0.40470120310783386,
      "learning_rate": 0.00017500845574497032,
      "loss": 1.0491385459899902,
      "step": 6016
    },
    {
      "epoch": 3.034074074074074,
      "grad_norm": 0.3126714825630188,
      "learning_rate": 0.00015533750279190617,
      "loss": 1.0497562885284424,
      "step": 6144
    },
    {
      "epoch": 3.034074074074074,
      "eval_bleu": 0.0032385289929989175,
      "eval_cos_loss": 0.3115836839850356,
      "eval_loss": 1.0473006324070255,
      "eval_mse_loss": 1.0473006324070255,
      "step": 6144
    },
    {
      "epoch": 3.034074074074074,
      "eval_bleu": 0.0032385289929989175,
      "eval_cos_loss": 0.3115836839850356,
      "eval_loss": 1.0473006324070255,
      "eval_mse_loss": 1.0473006324070255,
      "eval_runtime": 17.416,
      "eval_samples_per_second": 150.322,
      "eval_steps_per_second": 2.354,
      "step": 6144
    },
    {
      "epoch": 3.097283950617284,
      "grad_norm": 0.4265057146549225,
      "learning_rate": 0.000136634287020226,
      "loss": 1.049530029296875,
      "step": 6272
    },
    {
      "epoch": 3.1604938271604937,
      "grad_norm": 0.8321629762649536,
      "learning_rate": 0.00011895132298390743,
      "loss": 1.0472538471221924,
      "step": 6400
    },
    {
      "epoch": 3.2237037037037037,
      "grad_norm": 0.3349120318889618,
      "learning_rate": 0.00010233826059239426,
      "loss": 1.047644853591919,
      "step": 6528
    },
    {
      "epoch": 3.286913580246914,
      "grad_norm": 0.39967530965805054,
      "learning_rate": 8.68417457044704e-05,
      "loss": 1.0478882789611816,
      "step": 6656
    },
    {
      "epoch": 3.3501234567901235,
      "grad_norm": 0.25927555561065674,
      "learning_rate": 7.250528915685422e-05,
      "loss": 1.0449655055999756,
      "step": 6784
    },
    {
      "epoch": 3.413333333333333,
      "grad_norm": 1.115824580192566,
      "learning_rate": 5.93691445952505e-05,
      "loss": 1.0455702543258667,
      "step": 6912
    },
    {
      "epoch": 3.476543209876543,
      "grad_norm": 0.306415319442749,
      "learning_rate": 4.7470195450886376e-05,
      "loss": 1.0458422899246216,
      "step": 7040
    },
    {
      "epoch": 3.5397530864197533,
      "grad_norm": 0.2206840217113495,
      "learning_rate": 3.684185137987378e-05,
      "loss": 1.0450292825698853,
      "step": 7168
    },
    {
      "epoch": 3.5397530864197533,
      "eval_bleu": 0.0034683603537757966,
      "eval_cos_loss": 0.3089482319064257,
      "eval_loss": 1.0452435103858388,
      "eval_mse_loss": 1.0452435103858388,
      "step": 7168
    },
    {
      "epoch": 3.5397530864197533,
      "eval_bleu": 0.0034683603537757966,
      "eval_cos_loss": 0.3089482319064257,
      "eval_loss": 1.0452435103858388,
      "eval_mse_loss": 1.0452435103858388,
      "eval_runtime": 17.0266,
      "eval_samples_per_second": 153.759,
      "eval_steps_per_second": 2.408,
      "step": 7168
    },
    {
      "epoch": 3.602962962962963,
      "grad_norm": 0.24500082433223724,
      "learning_rate": 2.751395445617594e-05,
      "loss": 1.0446120500564575,
      "step": 7296
    },
    {
      "epoch": 3.6661728395061726,
      "grad_norm": 0.2302686870098114,
      "learning_rate": 1.9512695381567302e-05,
      "loss": 1.0452312231063843,
      "step": 7424
    },
    {
      "epoch": 3.7293827160493827,
      "grad_norm": 0.28751441836357117,
      "learning_rate": 1.2860539947850303e-05,
      "loss": 1.044912576675415,
      "step": 7552
    },
    {
      "epoch": 3.7925925925925927,
      "grad_norm": 0.1986972540616989,
      "learning_rate": 7.5761659578078565e-06,
      "loss": 1.0442619323730469,
      "step": 7680
    },
    {
      "epoch": 3.8558024691358024,
      "grad_norm": 0.24690282344818115,
      "learning_rate": 3.674410782003812e-06,
      "loss": 1.0443907976150513,
      "step": 7808
    },
    {
      "epoch": 3.9190123456790125,
      "grad_norm": 0.19394594430923462,
      "learning_rate": 1.1662296986795217e-06,
      "loss": 1.044499397277832,
      "step": 7936
    },
    {
      "epoch": 3.982222222222222,
      "grad_norm": 0.28466951847076416,
      "learning_rate": 5.866513372004834e-08,
      "loss": 1.0441296100616455,
      "step": 8064
    }
  ],
  "logging_steps": 128,
  "max_steps": 8100,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 1024,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}