{
  "best_metric": 0.5939759612083435,
  "best_model_checkpoint": "satriadata/run-0/checkpoint-1000",
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 1000,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_f1": 0.520844900048434,
      "eval_loss": 1.3270556926727295,
      "eval_runtime": 15.2159,
      "eval_samples_per_second": 65.721,
      "eval_steps_per_second": 4.14,
      "step": 250
    },
    {
      "epoch": 2.0,
      "grad_norm": 8.648948669433594,
      "learning_rate": 5.086297130534548e-05,
      "loss": 1.4798,
      "step": 500
    },
    {
      "epoch": 2.0,
      "eval_f1": 0.7148278885110101,
      "eval_loss": 0.8877580165863037,
      "eval_runtime": 15.0754,
      "eval_samples_per_second": 66.333,
      "eval_steps_per_second": 4.179,
      "step": 500
    },
    {
      "epoch": 3.0,
      "eval_f1": 0.8147606395143462,
      "eval_loss": 0.6009902358055115,
      "eval_runtime": 15.0542,
      "eval_samples_per_second": 66.427,
      "eval_steps_per_second": 4.185,
      "step": 750
    },
    {
      "epoch": 4.0,
      "grad_norm": 12.858877182006836,
      "learning_rate": 1.6954323768448494e-05,
      "loss": 0.4927,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_f1": 0.8354816202915758,
      "eval_loss": 0.5939759612083435,
      "eval_runtime": 15.0559,
      "eval_samples_per_second": 66.419,
      "eval_steps_per_second": 4.184,
      "step": 1000
    }
  ],
  "logging_steps": 500,
  "max_steps": 1250,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 4210003673088000.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": {
    "gradient_accumulation_steps": 1,
    "learning_rate": 6.781729507379398e-05,
    "optimizer": "adamw",
    "per_device_train_batch_size": 16,
    "warmup_ratio": 0.2,
    "weight_decay": 0.01
  }
}