{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 19.946666666666665,
  "eval_steps": 500,
  "global_step": 3740,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5333333333333333,
      "grad_norm": 0.26794755458831787,
      "learning_rate": 3e-05,
      "loss": 1.8512,
      "step": 100
    },
    {
      "epoch": 0.9973333333333333,
      "eval_accuracy": 0.6031838565022422,
      "eval_loss": 1.7184041738510132,
      "eval_runtime": 6.0677,
      "eval_samples_per_second": 82.403,
      "eval_steps_per_second": 10.383,
      "step": 187
    },
    {
      "epoch": 1.0666666666666667,
      "grad_norm": 0.37397125363349915,
      "learning_rate": 3e-05,
      "loss": 1.7523,
      "step": 200
    },
    {
      "epoch": 1.6,
      "grad_norm": 0.35164713859558105,
      "learning_rate": 3e-05,
      "loss": 1.7174,
      "step": 300
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.604932735426009,
      "eval_loss": 1.70177161693573,
      "eval_runtime": 6.4641,
      "eval_samples_per_second": 77.351,
      "eval_steps_per_second": 9.746,
      "step": 375
    },
    {
      "epoch": 2.1333333333333333,
      "grad_norm": 0.3863197863101959,
      "learning_rate": 3e-05,
      "loss": 1.7212,
      "step": 400
    },
    {
      "epoch": 2.6666666666666665,
      "grad_norm": 0.4171089828014374,
      "learning_rate": 3e-05,
      "loss": 1.6805,
      "step": 500
    },
    {
      "epoch": 2.997333333333333,
      "eval_accuracy": 0.6061434977578475,
      "eval_loss": 1.693791151046753,
      "eval_runtime": 6.1109,
      "eval_samples_per_second": 81.821,
      "eval_steps_per_second": 10.309,
      "step": 562
    },
    {
      "epoch": 3.2,
      "grad_norm": 0.5697806477546692,
      "learning_rate": 3e-05,
      "loss": 1.6692,
      "step": 600
    },
    {
      "epoch": 3.7333333333333334,
      "grad_norm": 0.5351872444152832,
      "learning_rate": 3e-05,
      "loss": 1.6412,
      "step": 700
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6067354260089686,
      "eval_loss": 1.6924760341644287,
      "eval_runtime": 6.7092,
      "eval_samples_per_second": 74.525,
      "eval_steps_per_second": 9.39,
      "step": 750
    },
    {
      "epoch": 4.266666666666667,
      "grad_norm": 0.6729754209518433,
      "learning_rate": 3e-05,
      "loss": 1.6073,
      "step": 800
    },
    {
      "epoch": 4.8,
      "grad_norm": 0.7071828246116638,
      "learning_rate": 3e-05,
      "loss": 1.5834,
      "step": 900
    },
    {
      "epoch": 4.997333333333334,
      "eval_accuracy": 0.6062242152466367,
      "eval_loss": 1.7046513557434082,
      "eval_runtime": 6.7276,
      "eval_samples_per_second": 74.32,
      "eval_steps_per_second": 9.364,
      "step": 937
    },
    {
      "epoch": 5.333333333333333,
      "grad_norm": 0.8920167684555054,
      "learning_rate": 3e-05,
      "loss": 1.5161,
      "step": 1000
    },
    {
      "epoch": 5.866666666666667,
      "grad_norm": 0.8955753445625305,
      "learning_rate": 3e-05,
      "loss": 1.5304,
      "step": 1100
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6056143497757848,
      "eval_loss": 1.723886251449585,
      "eval_runtime": 6.7785,
      "eval_samples_per_second": 73.763,
      "eval_steps_per_second": 9.294,
      "step": 1125
    },
    {
      "epoch": 6.4,
      "grad_norm": 0.9878254532814026,
      "learning_rate": 3e-05,
      "loss": 1.4575,
      "step": 1200
    },
    {
      "epoch": 6.933333333333334,
      "grad_norm": 1.1449097394943237,
      "learning_rate": 3e-05,
      "loss": 1.452,
      "step": 1300
    },
    {
      "epoch": 6.997333333333334,
      "eval_accuracy": 0.6039192825112107,
      "eval_loss": 1.7508341073989868,
      "eval_runtime": 6.0652,
      "eval_samples_per_second": 82.437,
      "eval_steps_per_second": 10.387,
      "step": 1312
    },
    {
      "epoch": 7.466666666666667,
      "grad_norm": 1.1436129808425903,
      "learning_rate": 3e-05,
      "loss": 1.3916,
      "step": 1400
    },
    {
      "epoch": 8.0,
      "grad_norm": 1.1151554584503174,
      "learning_rate": 3e-05,
      "loss": 1.3847,
      "step": 1500
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.602762331838565,
      "eval_loss": 1.7711397409439087,
      "eval_runtime": 6.0614,
      "eval_samples_per_second": 82.489,
      "eval_steps_per_second": 10.394,
      "step": 1500
    },
    {
      "epoch": 8.533333333333333,
      "grad_norm": 1.3540856838226318,
      "learning_rate": 3e-05,
      "loss": 1.3177,
      "step": 1600
    },
    {
      "epoch": 8.997333333333334,
      "eval_accuracy": 0.6008699551569506,
      "eval_loss": 1.8049190044403076,
      "eval_runtime": 6.505,
      "eval_samples_per_second": 76.864,
      "eval_steps_per_second": 9.685,
      "step": 1687
    },
    {
      "epoch": 9.066666666666666,
      "grad_norm": 1.5734765529632568,
      "learning_rate": 3e-05,
      "loss": 1.326,
      "step": 1700
    },
    {
      "epoch": 9.6,
      "grad_norm": 1.6235650777816772,
      "learning_rate": 3e-05,
      "loss": 1.2747,
      "step": 1800
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.599847533632287,
      "eval_loss": 1.8297590017318726,
      "eval_runtime": 6.3451,
      "eval_samples_per_second": 78.801,
      "eval_steps_per_second": 9.929,
      "step": 1875
    },
    {
      "epoch": 10.133333333333333,
      "grad_norm": 1.6720285415649414,
      "learning_rate": 3e-05,
      "loss": 1.2459,
      "step": 1900
    },
    {
      "epoch": 10.666666666666666,
      "grad_norm": 1.6803735494613647,
      "learning_rate": 3e-05,
      "loss": 1.2202,
      "step": 2000
    },
    {
      "epoch": 10.997333333333334,
      "eval_accuracy": 0.5980717488789238,
      "eval_loss": 1.881359338760376,
      "eval_runtime": 5.6855,
      "eval_samples_per_second": 87.942,
      "eval_steps_per_second": 11.081,
      "step": 2062
    },
    {
      "epoch": 11.2,
      "grad_norm": 1.7732399702072144,
      "learning_rate": 3e-05,
      "loss": 1.1909,
      "step": 2100
    },
    {
      "epoch": 11.733333333333333,
      "grad_norm": 1.6892499923706055,
      "learning_rate": 3e-05,
      "loss": 1.1589,
      "step": 2200
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.5959372197309417,
      "eval_loss": 1.9310756921768188,
      "eval_runtime": 6.0847,
      "eval_samples_per_second": 82.173,
      "eval_steps_per_second": 10.354,
      "step": 2250
    },
    {
      "epoch": 12.266666666666667,
      "grad_norm": 1.7170026302337646,
      "learning_rate": 3e-05,
      "loss": 1.1364,
      "step": 2300
    },
    {
      "epoch": 12.8,
      "grad_norm": 1.7617255449295044,
      "learning_rate": 3e-05,
      "loss": 1.1231,
      "step": 2400
    },
    {
      "epoch": 12.997333333333334,
      "eval_accuracy": 0.5955336322869955,
      "eval_loss": 1.9429402351379395,
      "eval_runtime": 6.8196,
      "eval_samples_per_second": 73.318,
      "eval_steps_per_second": 9.238,
      "step": 2437
    },
    {
      "epoch": 13.333333333333334,
      "grad_norm": 2.0009663105010986,
      "learning_rate": 3e-05,
      "loss": 1.0823,
      "step": 2500
    },
    {
      "epoch": 13.866666666666667,
      "grad_norm": 2.044010877609253,
      "learning_rate": 3e-05,
      "loss": 1.0624,
      "step": 2600
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.593255605381166,
      "eval_loss": 1.996883511543274,
      "eval_runtime": 6.0345,
      "eval_samples_per_second": 82.857,
      "eval_steps_per_second": 10.44,
      "step": 2625
    },
    {
      "epoch": 14.4,
      "grad_norm": 2.0219309329986572,
      "learning_rate": 3e-05,
      "loss": 1.0217,
      "step": 2700
    },
    {
      "epoch": 14.933333333333334,
      "grad_norm": 2.0420074462890625,
      "learning_rate": 3e-05,
      "loss": 1.0185,
      "step": 2800
    },
    {
      "epoch": 14.997333333333334,
      "eval_accuracy": 0.592170403587444,
      "eval_loss": 2.031921625137329,
      "eval_runtime": 5.9337,
      "eval_samples_per_second": 84.265,
      "eval_steps_per_second": 10.617,
      "step": 2812
    },
    {
      "epoch": 15.466666666666667,
      "grad_norm": 2.2069709300994873,
      "learning_rate": 3e-05,
      "loss": 0.9645,
      "step": 2900
    },
    {
      "epoch": 16.0,
      "grad_norm": 1.863767147064209,
      "learning_rate": 3e-05,
      "loss": 0.9718,
      "step": 3000
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.5902959641255605,
      "eval_loss": 2.0798022747039795,
      "eval_runtime": 6.143,
      "eval_samples_per_second": 81.394,
      "eval_steps_per_second": 10.256,
      "step": 3000
    },
    {
      "epoch": 16.533333333333335,
      "grad_norm": 2.300877809524536,
      "learning_rate": 3e-05,
      "loss": 0.9101,
      "step": 3100
    },
    {
      "epoch": 16.997333333333334,
      "eval_accuracy": 0.5887354260089687,
      "eval_loss": 2.13960337638855,
      "eval_runtime": 6.0122,
      "eval_samples_per_second": 83.164,
      "eval_steps_per_second": 10.479,
      "step": 3187
    },
    {
      "epoch": 17.066666666666666,
      "grad_norm": 2.586768388748169,
      "learning_rate": 3e-05,
      "loss": 0.9136,
      "step": 3200
    },
    {
      "epoch": 17.6,
      "grad_norm": 2.3870368003845215,
      "learning_rate": 3e-05,
      "loss": 0.8606,
      "step": 3300
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.5870134529147982,
      "eval_loss": 2.1882169246673584,
      "eval_runtime": 6.0214,
      "eval_samples_per_second": 83.036,
      "eval_steps_per_second": 10.463,
      "step": 3375
    },
    {
      "epoch": 18.133333333333333,
      "grad_norm": 2.6867830753326416,
      "learning_rate": 3e-05,
      "loss": 0.8724,
      "step": 3400
    },
    {
      "epoch": 18.666666666666668,
      "grad_norm": 2.807380199432373,
      "learning_rate": 3e-05,
      "loss": 0.8168,
      "step": 3500
    },
    {
      "epoch": 18.997333333333334,
      "eval_accuracy": 0.5862600896860987,
      "eval_loss": 2.2291159629821777,
      "eval_runtime": 6.3423,
      "eval_samples_per_second": 78.836,
      "eval_steps_per_second": 9.933,
      "step": 3562
    },
    {
      "epoch": 19.2,
      "grad_norm": 2.6379778385162354,
      "learning_rate": 3e-05,
      "loss": 0.8061,
      "step": 3600
    },
    {
      "epoch": 19.733333333333334,
      "grad_norm": 3.161210775375366,
      "learning_rate": 3e-05,
      "loss": 0.777,
      "step": 3700
    },
    {
      "epoch": 19.946666666666665,
      "eval_accuracy": 0.5851838565022421,
      "eval_loss": 2.2508349418640137,
      "eval_runtime": 5.6415,
      "eval_samples_per_second": 88.63,
      "eval_steps_per_second": 11.167,
      "step": 3740
    },
    {
      "epoch": 19.946666666666665,
      "step": 3740,
      "total_flos": 3.767212755417825e+17,
      "train_loss": 1.2655855785716663,
      "train_runtime": 8167.3443,
      "train_samples_per_second": 14.693,
      "train_steps_per_second": 0.458
    }
  ],
  "logging_steps": 100,
  "max_steps": 3740,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "save_steps": 500,
  "total_flos": 3.767212755417825e+17,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}