{
  "best_global_step": 3000,
  "best_metric": 0.7988061056554884,
  "best_model_checkpoint": "./results/run-4/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": 3.002380609512329,
      "learning_rate": 4.572943108695439e-06,
      "loss": 0.296,
      "step": 100
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.9367781281471252,
      "learning_rate": 9.192077561923155e-06,
      "loss": 1.0274,
      "step": 200
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.7274576425552368,
      "learning_rate": 1.3811212015150872e-05,
      "loss": 0.0,
      "step": 300
    },
    {
      "epoch": 0.4,
      "grad_norm": 44.297061920166016,
      "learning_rate": 1.8430346468378586e-05,
      "loss": 0.0,
      "step": 400
    },
    {
      "epoch": 0.5,
      "grad_norm": 1.0151411294937134,
      "learning_rate": 2.3049480921606306e-05,
      "loss": 0.0,
      "step": 500
    },
    {
      "epoch": 0.6,
      "grad_norm": 1.208032488822937,
      "learning_rate": 2.766861537483402e-05,
      "loss": 0.0,
      "step": 600
    },
    {
      "epoch": 0.7,
      "grad_norm": 6.984821319580078,
      "learning_rate": 3.2287749828061733e-05,
      "loss": 0.0342,
      "step": 700
    },
    {
      "epoch": 0.8,
      "grad_norm": 1.5860588550567627,
      "learning_rate": 3.690688428128945e-05,
      "loss": 0.0488,
      "step": 800
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.7841507196426392,
      "learning_rate": 4.152601873451717e-05,
      "loss": 0.0314,
      "step": 900
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.9344207644462585,
      "learning_rate": 4.414230747586468e-05,
      "loss": 0.0356,
      "step": 1000
    },
    {
      "epoch": 1.0,
      "eval_economic_inequality_accuracy": 0.7783375314861462,
      "eval_economic_inequality_f1": 0.6829252350399464,
      "eval_economic_policy_benefits_accuracy": 0.8215488215488216,
      "eval_economic_policy_benefits_f1": 0.752833443442648,
      "eval_ethnic_boundaries_accuracy": 0.9493464052287581,
      "eval_ethnic_boundaries_f1": 0.9305506350938352,
      "eval_language_policy_accuracy": 0.6764705882352942,
      "eval_language_policy_f1": 0.555672268907563,
      "eval_loss": NaN,
      "eval_mother_tongue_education_accuracy": 0.7777777777777778,
      "eval_mother_tongue_education_f1": 0.6805555555555556,
      "eval_overall_accuracy": 0.8090908256773385,
      "eval_overall_f1": 0.7331392486819855,
      "eval_religion_ethnic_identity_accuracy": 0.851063829787234,
      "eval_religion_ethnic_identity_f1": 0.7962983540523647,
      "eval_runtime": 4.4716,
      "eval_samples_per_second": 894.535,
      "eval_steps_per_second": 55.908,
      "step": 1000
    },
    {
      "epoch": 1.1,
      "grad_norm": 0.8767189383506775,
      "learning_rate": 4.303902560943467e-05,
      "loss": 0.3117,
      "step": 1100
    },
    {
      "epoch": 1.2,
      "grad_norm": 32.66496658325195,
      "learning_rate": 4.193574374300466e-05,
      "loss": 0.0098,
      "step": 1200
    },
    {
      "epoch": 1.3,
      "grad_norm": 6.170604705810547,
      "learning_rate": 4.083246187657465e-05,
      "loss": 0.0,
      "step": 1300
    },
    {
      "epoch": 1.4,
      "grad_norm": 0.40170523524284363,
      "learning_rate": 3.972918001014464e-05,
      "loss": 0.2259,
      "step": 1400
    },
    {
      "epoch": 1.5,
      "grad_norm": 3.7784886360168457,
      "learning_rate": 3.862589814371463e-05,
      "loss": 0.0,
      "step": 1500
    },
    {
      "epoch": 1.6,
      "grad_norm": 1.0622837543487549,
      "learning_rate": 3.752261627728462e-05,
      "loss": 0.0,
      "step": 1600
    },
    {
      "epoch": 1.7,
      "grad_norm": 0.3284582793712616,
      "learning_rate": 3.641933441085462e-05,
      "loss": 0.0012,
      "step": 1700
    },
    {
      "epoch": 1.8,
      "grad_norm": 3.2630698680877686,
      "learning_rate": 3.53160525444246e-05,
      "loss": 0.0045,
      "step": 1800
    },
    {
      "epoch": 1.9,
      "grad_norm": 0.4251757860183716,
      "learning_rate": 3.421277067799459e-05,
      "loss": 0.1053,
      "step": 1900
    },
    {
      "epoch": 2.0,
      "grad_norm": 2.5576839447021484,
      "learning_rate": 3.310948881156459e-05,
      "loss": 0.0826,
      "step": 2000
    },
    {
      "epoch": 2.0,
      "eval_economic_inequality_accuracy": 0.7984886649874056,
      "eval_economic_inequality_f1": 0.792653055614301,
      "eval_economic_policy_benefits_accuracy": 0.8215488215488216,
      "eval_economic_policy_benefits_f1": 0.8093298577169545,
      "eval_ethnic_boundaries_accuracy": 0.9444444444444444,
      "eval_ethnic_boundaries_f1": 0.9370640143333165,
      "eval_language_policy_accuracy": 0.5882352941176471,
      "eval_language_policy_f1": 0.5458823529411765,
      "eval_loss": NaN,
      "eval_mother_tongue_education_accuracy": 0.8888888888888888,
      "eval_mother_tongue_education_f1": 0.837037037037037,
      "eval_overall_accuracy": 0.8148002453034774,
      "eval_overall_f1": 0.7921517057928192,
      "eval_religion_ethnic_identity_accuracy": 0.8471953578336557,
      "eval_religion_ethnic_identity_f1": 0.8309439171141297,
      "eval_runtime": 4.6859,
      "eval_samples_per_second": 853.616,
      "eval_steps_per_second": 53.351,
      "step": 2000
    },
    {
      "epoch": 2.1,
      "grad_norm": 210.1742706298828,
      "learning_rate": 3.200620694513457e-05,
      "loss": 1.2274,
      "step": 2100
    },
    {
      "epoch": 2.2,
      "grad_norm": 3.0127012729644775,
      "learning_rate": 3.090292507870457e-05,
      "loss": 0.0,
      "step": 2200
    },
    {
      "epoch": 2.3,
      "grad_norm": 0.38292792439460754,
      "learning_rate": 2.9799643212274557e-05,
      "loss": 0.0351,
      "step": 2300
    },
    {
      "epoch": 2.4,
      "grad_norm": 2.8801395893096924,
      "learning_rate": 2.8696361345844544e-05,
      "loss": 0.0026,
      "step": 2400
    },
    {
      "epoch": 2.5,
      "grad_norm": 1.5863370895385742,
      "learning_rate": 2.7593079479414537e-05,
      "loss": 0.0,
      "step": 2500
    },
    {
      "epoch": 2.6,
      "grad_norm": 0.23432491719722748,
      "learning_rate": 2.6489797612984527e-05,
      "loss": 0.1192,
      "step": 2600
    },
    {
      "epoch": 2.7,
      "grad_norm": 1.943647027015686,
      "learning_rate": 2.5386515746554517e-05,
      "loss": 0.0195,
      "step": 2700
    },
    {
      "epoch": 2.8,
      "grad_norm": 0.5006923079490662,
      "learning_rate": 2.4283233880124507e-05,
      "loss": 0.0,
      "step": 2800
    },
    {
      "epoch": 2.9,
      "grad_norm": 0.26087385416030884,
      "learning_rate": 2.31799520136945e-05,
      "loss": 0.0,
      "step": 2900
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.207767367362976,
      "learning_rate": 2.207667014726449e-05,
      "loss": 0.1565,
      "step": 3000
    },
    {
      "epoch": 3.0,
      "eval_economic_inequality_accuracy": 0.8148614609571788,
      "eval_economic_inequality_f1": 0.7997421102149133,
      "eval_economic_policy_benefits_accuracy": 0.8518518518518519,
      "eval_economic_policy_benefits_f1": 0.8175032474456659,
      "eval_ethnic_boundaries_accuracy": 0.9477124183006536,
      "eval_ethnic_boundaries_f1": 0.9334997758629278,
      "eval_language_policy_accuracy": 0.6764705882352942,
      "eval_language_policy_f1": 0.5869281045751634,
      "eval_loss": NaN,
      "eval_mother_tongue_education_accuracy": 0.8888888888888888,
      "eval_mother_tongue_education_f1": 0.837037037037037,
      "eval_overall_accuracy": 0.8371853490189908,
      "eval_overall_f1": 0.7988061056554884,
      "eval_religion_ethnic_identity_accuracy": 0.8433268858800773,
      "eval_religion_ethnic_identity_f1": 0.8181263587972226,
      "eval_runtime": 4.4894,
      "eval_samples_per_second": 890.983,
      "eval_steps_per_second": 55.686,
      "step": 3000
    }
  ],
  "logging_steps": 100,
  "max_steps": 5000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "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": 16,
  "trial_name": null,
  "trial_params": {
    "gradient_accumulation_steps": 1,
    "learning_rate": 4.452845612911518e-05,
    "num_train_epochs": 5,
    "per_device_train_batch_size": 16,
    "warmup_steps": 964,
    "weight_decay": 0.08427803271887896
  }
}