{
  "best_global_step": 3000,
  "best_metric": 0.8006341340755991,
  "best_model_checkpoint": "./results/run-2/checkpoint-3000",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 3000,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.1,
      "grad_norm": 3.6001648902893066,
      "learning_rate": 3.1887409174919234e-06,
      "loss": 0.0412,
      "step": 100
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.6239789128303528,
      "learning_rate": 6.409691339200936e-06,
      "loss": 0.1091,
      "step": 200
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.6479267477989197,
      "learning_rate": 9.63064176090995e-06,
      "loss": 0.0,
      "step": 300
    },
    {
      "epoch": 0.4,
      "grad_norm": 20.527082443237305,
      "learning_rate": 1.054959200884952e-05,
      "loss": 0.0,
      "step": 400
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.8853731751441956,
      "learning_rate": 1.0481970753745042e-05,
      "loss": 0.0,
      "step": 500
    },
    {
      "epoch": 0.6,
      "grad_norm": 1.839735507965088,
      "learning_rate": 1.0414349498640566e-05,
      "loss": 0.0,
      "step": 600
    },
    {
      "epoch": 0.7,
      "grad_norm": 2.7137391567230225,
      "learning_rate": 1.0346728243536088e-05,
      "loss": 0.0284,
      "step": 700
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.6026535630226135,
      "learning_rate": 1.027910698843161e-05,
      "loss": 0.1612,
      "step": 800
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.5444966554641724,
      "learning_rate": 1.0211485733327134e-05,
      "loss": 0.0,
      "step": 900
    },
    {
      "epoch": 1.0,
      "grad_norm": 1.2413469552993774,
      "learning_rate": 1.0143864478222655e-05,
      "loss": 0.0986,
      "step": 1000
    },
    {
      "epoch": 1.0,
      "eval_economic_inequality_accuracy": 0.8047858942065491,
      "eval_economic_inequality_f1": 0.7538941327354425,
      "eval_economic_policy_benefits_accuracy": 0.8282828282828283,
      "eval_economic_policy_benefits_f1": 0.764129475069552,
      "eval_ethnic_boundaries_accuracy": 0.9493464052287581,
      "eval_ethnic_boundaries_f1": 0.9281293325410972,
      "eval_language_policy_accuracy": 0.7058823529411765,
      "eval_language_policy_f1": 0.6071428571428571,
      "eval_loss": NaN,
      "eval_mother_tongue_education_accuracy": 0.8888888888888888,
      "eval_mother_tongue_education_f1": 0.837037037037037,
      "eval_overall_accuracy": 0.8390088178776338,
      "eval_overall_f1": 0.7839408748610676,
      "eval_religion_ethnic_identity_accuracy": 0.8568665377176016,
      "eval_religion_ethnic_identity_f1": 0.8133124146404199,
      "eval_runtime": 4.4593,
      "eval_samples_per_second": 897.004,
      "eval_steps_per_second": 56.063,
      "step": 1000
    },
    {
      "epoch": 1.1,
      "grad_norm": 0.5792437195777893,
      "learning_rate": 1.0076243223118178e-05,
      "loss": 0.0,
      "step": 1100
    },
    {
      "epoch": 1.2,
      "grad_norm": 8.802961349487305,
      "learning_rate": 1.0008621968013701e-05,
      "loss": 0.0297,
      "step": 1200
    },
    {
      "epoch": 1.3,
      "grad_norm": 3.981771469116211,
      "learning_rate": 9.941000712909224e-06,
      "loss": 0.0,
      "step": 1300
    },
    {
      "epoch": 1.4,
      "grad_norm": 0.9215607047080994,
      "learning_rate": 9.873379457804745e-06,
      "loss": 0.263,
      "step": 1400
    },
    {
      "epoch": 1.5,
      "grad_norm": 1.5474179983139038,
      "learning_rate": 9.805758202700268e-06,
      "loss": 0.0,
      "step": 1500
    },
    {
      "epoch": 1.6,
      "grad_norm": 1.1813520193099976,
      "learning_rate": 9.738136947595791e-06,
      "loss": 0.0,
      "step": 1600
    },
    {
      "epoch": 1.7,
      "grad_norm": 5.511495113372803,
      "learning_rate": 9.670515692491313e-06,
      "loss": 0.0037,
      "step": 1700
    },
    {
      "epoch": 1.8,
      "grad_norm": 3.1629152297973633,
      "learning_rate": 9.602894437386836e-06,
      "loss": 0.0027,
      "step": 1800
    },
    {
      "epoch": 1.9,
      "grad_norm": 0.6996293663978577,
      "learning_rate": 9.535273182282359e-06,
      "loss": 19.1531,
      "step": 1900
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.3346213102340698,
      "learning_rate": 9.467651927177882e-06,
      "loss": 2.9418,
      "step": 2000
    },
    {
      "epoch": 2.0,
      "eval_economic_inequality_accuracy": 0.7896725440806045,
      "eval_economic_inequality_f1": 0.7951291586358689,
      "eval_economic_policy_benefits_accuracy": 0.8148148148148148,
      "eval_economic_policy_benefits_f1": 0.8131577832835695,
      "eval_ethnic_boundaries_accuracy": 0.9264705882352942,
      "eval_ethnic_boundaries_f1": 0.9324831159457648,
      "eval_language_policy_accuracy": 0.7058823529411765,
      "eval_language_policy_f1": 0.6071428571428571,
      "eval_loss": NaN,
      "eval_mother_tongue_education_accuracy": 0.7777777777777778,
      "eval_mother_tongue_education_f1": 0.7407407407407408,
      "eval_overall_accuracy": 0.8103022392805538,
      "eval_overall_f1": 0.7884819842273197,
      "eval_religion_ethnic_identity_accuracy": 0.8471953578336557,
      "eval_religion_ethnic_identity_f1": 0.8422382496151178,
      "eval_runtime": 4.499,
      "eval_samples_per_second": 889.087,
      "eval_steps_per_second": 55.568,
      "step": 2000
    },
    {
      "epoch": 2.1,
      "grad_norm": 133.0632781982422,
      "learning_rate": 9.400030672073403e-06,
      "loss": 0.0,
      "step": 2100
    },
    {
      "epoch": 2.2,
      "grad_norm": 1.1462996006011963,
      "learning_rate": 9.332409416968926e-06,
      "loss": 0.0,
      "step": 2200
    },
    {
      "epoch": 2.3,
      "grad_norm": 0.11461023986339569,
      "learning_rate": 9.26478816186445e-06,
      "loss": 0.0,
      "step": 2300
    },
    {
      "epoch": 2.4,
      "grad_norm": 0.37900322675704956,
      "learning_rate": 9.197166906759972e-06,
      "loss": 0.0046,
      "step": 2400
    },
    {
      "epoch": 2.5,
      "grad_norm": 1.0005897283554077,
      "learning_rate": 9.129545651655495e-06,
      "loss": 0.0,
      "step": 2500
    },
    {
      "epoch": 2.6,
      "grad_norm": 0.15003152191638947,
      "learning_rate": 9.061924396551017e-06,
      "loss": 0.0,
      "step": 2600
    },
    {
      "epoch": 2.7,
      "grad_norm": 1.9842479228973389,
      "learning_rate": 8.99430314144654e-06,
      "loss": 0.0017,
      "step": 2700
    },
    {
      "epoch": 2.8,
      "grad_norm": 4.747731685638428,
      "learning_rate": 8.926681886342063e-06,
      "loss": 0.0,
      "step": 2800
    },
    {
      "epoch": 2.9,
      "grad_norm": 0.4671611189842224,
      "learning_rate": 8.859060631237586e-06,
      "loss": 0.0,
      "step": 2900
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.9774578213691711,
      "learning_rate": 8.791439376133107e-06,
      "loss": 0.0,
      "step": 3000
    },
    {
      "epoch": 3.0,
      "eval_economic_inequality_accuracy": 0.8173803526448362,
      "eval_economic_inequality_f1": 0.8105151543481738,
      "eval_economic_policy_benefits_accuracy": 0.8686868686868687,
      "eval_economic_policy_benefits_f1": 0.8544859369266805,
      "eval_ethnic_boundaries_accuracy": 0.9330065359477124,
      "eval_ethnic_boundaries_f1": 0.9342620574580529,
      "eval_language_policy_accuracy": 0.7058823529411765,
      "eval_language_policy_f1": 0.6071428571428571,
      "eval_loss": NaN,
      "eval_mother_tongue_education_accuracy": 0.7777777777777778,
      "eval_mother_tongue_education_f1": 0.7407407407407408,
      "eval_overall_accuracy": 0.8266000709526621,
      "eval_overall_f1": 0.8006341340755991,
      "eval_religion_ethnic_identity_accuracy": 0.8568665377176016,
      "eval_religion_ethnic_identity_f1": 0.85665805783709,
      "eval_runtime": 4.4861,
      "eval_samples_per_second": 891.646,
      "eval_steps_per_second": 55.728,
      "step": 3000
    }
  ],
  "logging_steps": 100,
  "max_steps": 16000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 16,
  "save_steps": 500,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 2,
        "early_stopping_threshold": 0.0
      },
      "attributes": {
        "early_stopping_patience_counter": 0
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.2631144955904e+16,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": {
    "gradient_accumulation_steps": 2,
    "learning_rate": 1.0596926887422654e-05,
    "num_train_epochs": 16,
    "per_device_train_batch_size": 8,
    "warmup_steps": 329,
    "weight_decay": 0.059185379025811405
  }
}