{
  "best_metric": 0.7850534531405503,
  "best_model_checkpoint": "test-klue/ynat/run-3/checkpoint-500",
  "epoch": 0.8756567425569177,
  "eval_steps": 50,
  "global_step": 500,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.09,
      "learning_rate": 3.851823303957499e-06,
      "loss": 1.934,
      "step": 50
    },
    {
      "epoch": 0.09,
      "eval_f1": 0.01705140497086806,
      "eval_loss": 2.002474546432495,
      "eval_runtime": 12.8723,
      "eval_samples_per_second": 707.486,
      "eval_steps_per_second": 1.398,
      "step": 50
    },
    {
      "epoch": 0.18,
      "learning_rate": 3.783163352015298e-06,
      "loss": 1.9174,
      "step": 100
    },
    {
      "epoch": 0.18,
      "eval_f1": 0.16773144497786774,
      "eval_loss": 1.9946870803833008,
      "eval_runtime": 13.4529,
      "eval_samples_per_second": 676.953,
      "eval_steps_per_second": 1.338,
      "step": 100
    },
    {
      "epoch": 0.26,
      "learning_rate": 3.7145034000730966e-06,
      "loss": 1.6984,
      "step": 150
    },
    {
      "epoch": 0.26,
      "eval_f1": 0.579749057690188,
      "eval_loss": 1.6373496055603027,
      "eval_runtime": 13.0114,
      "eval_samples_per_second": 699.924,
      "eval_steps_per_second": 1.383,
      "step": 150
    },
    {
      "epoch": 0.35,
      "learning_rate": 3.645843448130895e-06,
      "loss": 1.2585,
      "step": 200
    },
    {
      "epoch": 0.35,
      "eval_f1": 0.6734776928120875,
      "eval_loss": 1.2733124494552612,
      "eval_runtime": 12.6224,
      "eval_samples_per_second": 721.493,
      "eval_steps_per_second": 1.426,
      "step": 200
    },
    {
      "epoch": 0.44,
      "learning_rate": 3.5771834961886936e-06,
      "loss": 0.9487,
      "step": 250
    },
    {
      "epoch": 0.44,
      "eval_f1": 0.6814953813845488,
      "eval_loss": 1.1104496717453003,
      "eval_runtime": 12.492,
      "eval_samples_per_second": 729.025,
      "eval_steps_per_second": 1.441,
      "step": 250
    },
    {
      "epoch": 0.53,
      "learning_rate": 3.5085235442464923e-06,
      "loss": 0.8072,
      "step": 300
    },
    {
      "epoch": 0.53,
      "eval_f1": 0.6717305786202503,
      "eval_loss": 1.0609983205795288,
      "eval_runtime": 12.6873,
      "eval_samples_per_second": 717.806,
      "eval_steps_per_second": 1.419,
      "step": 300
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.439863592304291e-06,
      "loss": 0.7549,
      "step": 350
    },
    {
      "epoch": 0.61,
      "eval_f1": 0.7228793656517608,
      "eval_loss": 0.9494264721870422,
      "eval_runtime": 12.8373,
      "eval_samples_per_second": 709.416,
      "eval_steps_per_second": 1.402,
      "step": 350
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.3712036403620893e-06,
      "loss": 0.6644,
      "step": 400
    },
    {
      "epoch": 0.7,
      "eval_f1": 0.7719549711033966,
      "eval_loss": 0.8139678239822388,
      "eval_runtime": 12.9124,
      "eval_samples_per_second": 705.29,
      "eval_steps_per_second": 1.394,
      "step": 400
    },
    {
      "epoch": 0.79,
      "learning_rate": 3.302543688419888e-06,
      "loss": 0.6306,
      "step": 450
    },
    {
      "epoch": 0.79,
      "eval_f1": 0.7738311952589042,
      "eval_loss": 0.7844408750534058,
      "eval_runtime": 12.8236,
      "eval_samples_per_second": 710.174,
      "eval_steps_per_second": 1.404,
      "step": 450
    },
    {
      "epoch": 0.88,
      "learning_rate": 3.2338837364776867e-06,
      "loss": 0.6093,
      "step": 500
    },
    {
      "epoch": 0.88,
      "eval_f1": 0.7850534531405503,
      "eval_loss": 0.7450836300849915,
      "eval_runtime": 12.7168,
      "eval_samples_per_second": 716.14,
      "eval_steps_per_second": 1.415,
      "step": 500
    }
  ],
  "logging_steps": 50,
  "max_steps": 2855,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 40829268658800.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 3.851823303957499e-06,
    "num_train_epochs": 5,
    "per_device_train_batch_size": 8,
    "seed": 30
  }
}