{
  "best_metric": 0.8981744694017328,
  "best_model_checkpoint": "en_output/run-1/checkpoint-4015",
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 4015,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.12,
      "grad_norm": 9.016681671142578,
      "learning_rate": 3.137975621127507e-05,
      "loss": 1.3969,
      "step": 100
    },
    {
      "epoch": 0.25,
      "grad_norm": 11.532567024230957,
      "learning_rate": 3.057822987126805e-05,
      "loss": 1.12,
      "step": 200
    },
    {
      "epoch": 0.37,
      "grad_norm": 13.854158401489258,
      "learning_rate": 2.977670353126102e-05,
      "loss": 1.0118,
      "step": 300
    },
    {
      "epoch": 0.5,
      "grad_norm": 15.174864768981934,
      "learning_rate": 2.8975177191253995e-05,
      "loss": 0.8854,
      "step": 400
    },
    {
      "epoch": 0.62,
      "grad_norm": 10.61332893371582,
      "learning_rate": 2.8173650851246964e-05,
      "loss": 0.8194,
      "step": 500
    },
    {
      "epoch": 0.75,
      "grad_norm": 8.371576309204102,
      "learning_rate": 2.737212451123994e-05,
      "loss": 0.8095,
      "step": 600
    },
    {
      "epoch": 0.87,
      "grad_norm": 7.02115535736084,
      "learning_rate": 2.6570598171232915e-05,
      "loss": 0.7585,
      "step": 700
    },
    {
      "epoch": 1.0,
      "grad_norm": 9.261237144470215,
      "learning_rate": 2.5769071831225887e-05,
      "loss": 0.6611,
      "step": 800
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7717528373266078,
      "eval_f1": 0.8197696469069372,
      "eval_loss": 0.631992518901825,
      "eval_runtime": 6.3017,
      "eval_samples_per_second": 503.354,
      "eval_steps_per_second": 15.869,
      "step": 803
    },
    {
      "epoch": 1.12,
      "grad_norm": 11.751930236816406,
      "learning_rate": 2.4967545491218862e-05,
      "loss": 0.5042,
      "step": 900
    },
    {
      "epoch": 1.25,
      "grad_norm": 9.949508666992188,
      "learning_rate": 2.4166019151211838e-05,
      "loss": 0.4727,
      "step": 1000
    },
    {
      "epoch": 1.37,
      "grad_norm": 9.066802024841309,
      "learning_rate": 2.336449281120481e-05,
      "loss": 0.4475,
      "step": 1100
    },
    {
      "epoch": 1.49,
      "grad_norm": 13.22970962524414,
      "learning_rate": 2.2562966471197785e-05,
      "loss": 0.4549,
      "step": 1200
    },
    {
      "epoch": 1.62,
      "grad_norm": 5.078187942504883,
      "learning_rate": 2.176144013119076e-05,
      "loss": 0.4522,
      "step": 1300
    },
    {
      "epoch": 1.74,
      "grad_norm": 10.427614212036133,
      "learning_rate": 2.0959913791183733e-05,
      "loss": 0.4271,
      "step": 1400
    },
    {
      "epoch": 1.87,
      "grad_norm": 4.8712615966796875,
      "learning_rate": 2.0158387451176705e-05,
      "loss": 0.4391,
      "step": 1500
    },
    {
      "epoch": 1.99,
      "grad_norm": 6.465252876281738,
      "learning_rate": 1.9356861111169677e-05,
      "loss": 0.4389,
      "step": 1600
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8193568726355611,
      "eval_f1": 0.8738583019819207,
      "eval_loss": 0.486948698759079,
      "eval_runtime": 6.3232,
      "eval_samples_per_second": 501.641,
      "eval_steps_per_second": 15.815,
      "step": 1606
    },
    {
      "epoch": 2.12,
      "grad_norm": 4.6454176902771,
      "learning_rate": 1.8555334771162652e-05,
      "loss": 0.284,
      "step": 1700
    },
    {
      "epoch": 2.24,
      "grad_norm": 8.212606430053711,
      "learning_rate": 1.7753808431155628e-05,
      "loss": 0.276,
      "step": 1800
    },
    {
      "epoch": 2.37,
      "grad_norm": 5.965590000152588,
      "learning_rate": 1.69522820911486e-05,
      "loss": 0.2434,
      "step": 1900
    },
    {
      "epoch": 2.49,
      "grad_norm": 9.97062873840332,
      "learning_rate": 1.6150755751141575e-05,
      "loss": 0.249,
      "step": 2000
    },
    {
      "epoch": 2.62,
      "grad_norm": 7.969445705413818,
      "learning_rate": 1.534922941113455e-05,
      "loss": 0.244,
      "step": 2100
    },
    {
      "epoch": 2.74,
      "grad_norm": 8.639568328857422,
      "learning_rate": 1.4547703071127523e-05,
      "loss": 0.244,
      "step": 2200
    },
    {
      "epoch": 2.86,
      "grad_norm": 9.686100006103516,
      "learning_rate": 1.3746176731120497e-05,
      "loss": 0.2373,
      "step": 2300
    },
    {
      "epoch": 2.99,
      "grad_norm": 9.704062461853027,
      "learning_rate": 1.294465039111347e-05,
      "loss": 0.2389,
      "step": 2400
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8199873896595208,
      "eval_f1": 0.8793961295450606,
      "eval_loss": 0.5388198494911194,
      "eval_runtime": 6.3147,
      "eval_samples_per_second": 502.323,
      "eval_steps_per_second": 15.836,
      "step": 2409
    },
    {
      "epoch": 3.11,
      "grad_norm": 7.037457466125488,
      "learning_rate": 1.2143124051106446e-05,
      "loss": 0.1588,
      "step": 2500
    },
    {
      "epoch": 3.24,
      "grad_norm": 6.776266574859619,
      "learning_rate": 1.134159771109942e-05,
      "loss": 0.128,
      "step": 2600
    },
    {
      "epoch": 3.36,
      "grad_norm": 6.907045841217041,
      "learning_rate": 1.0540071371092392e-05,
      "loss": 0.1408,
      "step": 2700
    },
    {
      "epoch": 3.49,
      "grad_norm": 7.6478800773620605,
      "learning_rate": 9.738545031085367e-06,
      "loss": 0.1214,
      "step": 2800
    },
    {
      "epoch": 3.61,
      "grad_norm": 13.469510078430176,
      "learning_rate": 8.93701869107834e-06,
      "loss": 0.1395,
      "step": 2900
    },
    {
      "epoch": 3.74,
      "grad_norm": 9.527716636657715,
      "learning_rate": 8.135492351071314e-06,
      "loss": 0.1342,
      "step": 3000
    },
    {
      "epoch": 3.86,
      "grad_norm": 4.901115417480469,
      "learning_rate": 7.333966011064288e-06,
      "loss": 0.1263,
      "step": 3100
    },
    {
      "epoch": 3.99,
      "grad_norm": 7.46558141708374,
      "learning_rate": 6.532439671057263e-06,
      "loss": 0.1405,
      "step": 3200
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8354350567465322,
      "eval_f1": 0.8956331776920504,
      "eval_loss": 0.5953832268714905,
      "eval_runtime": 6.2939,
      "eval_samples_per_second": 503.981,
      "eval_steps_per_second": 15.888,
      "step": 3212
    },
    {
      "epoch": 4.11,
      "grad_norm": 0.39140334725379944,
      "learning_rate": 5.7309133310502365e-06,
      "loss": 0.0853,
      "step": 3300
    },
    {
      "epoch": 4.23,
      "grad_norm": 7.5608391761779785,
      "learning_rate": 4.92938699104321e-06,
      "loss": 0.0715,
      "step": 3400
    },
    {
      "epoch": 4.36,
      "grad_norm": 15.901948928833008,
      "learning_rate": 4.127860651036185e-06,
      "loss": 0.0504,
      "step": 3500
    },
    {
      "epoch": 4.48,
      "grad_norm": 1.3447962999343872,
      "learning_rate": 3.326334311029158e-06,
      "loss": 0.0616,
      "step": 3600
    },
    {
      "epoch": 4.61,
      "grad_norm": 0.6756995916366577,
      "learning_rate": 2.524807971022132e-06,
      "loss": 0.0783,
      "step": 3700
    },
    {
      "epoch": 4.73,
      "grad_norm": 0.16733744740486145,
      "learning_rate": 1.723281631015106e-06,
      "loss": 0.0812,
      "step": 3800
    },
    {
      "epoch": 4.86,
      "grad_norm": 0.649143636226654,
      "learning_rate": 9.2175529100808e-07,
      "loss": 0.0674,
      "step": 3900
    },
    {
      "epoch": 4.98,
      "grad_norm": 0.20296114683151245,
      "learning_rate": 1.202289510010539e-07,
      "loss": 0.0685,
      "step": 4000
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8385876418663304,
      "eval_f1": 0.8981744694017328,
      "eval_loss": 0.7969151735305786,
      "eval_runtime": 6.2809,
      "eval_samples_per_second": 505.024,
      "eval_steps_per_second": 15.921,
      "step": 4015
    }
  ],
  "logging_steps": 100,
  "max_steps": 4015,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 3783455887092102.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 3.21812825512821e-05,
    "lr_scheduler_type": "linear",
    "num_train_epochs": 5,
    "per_device_train_batch_size": 32
  }
}