{
  "best_metric": 0.8734693877551021,
  "best_model_checkpoint": "rubert-tiny2-finetuned-classification\\run-2\\checkpoint-1650",
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 2200,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.91,
      "learning_rate": 1.5039435145104783e-05,
      "loss": 1.8268,
      "step": 500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7224489795918367,
      "eval_loss": 1.1639864444732666,
      "eval_runtime": 1.3165,
      "eval_samples_per_second": 186.096,
      "eval_steps_per_second": 12.153,
      "step": 550
    },
    {
      "epoch": 1.82,
      "learning_rate": 1.0616071867132787e-05,
      "loss": 1.0445,
      "step": 1000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8571428571428571,
      "eval_loss": 0.730604887008667,
      "eval_runtime": 1.103,
      "eval_samples_per_second": 222.131,
      "eval_steps_per_second": 14.507,
      "step": 1100
    },
    {
      "epoch": 2.73,
      "learning_rate": 6.192708589160794e-06,
      "loss": 0.7493,
      "step": 1500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8734693877551021,
      "eval_loss": 0.5868724584579468,
      "eval_runtime": 1.1252,
      "eval_samples_per_second": 217.733,
      "eval_steps_per_second": 14.219,
      "step": 1650
    },
    {
      "epoch": 3.64,
      "learning_rate": 1.7693453111887982e-06,
      "loss": 0.5909,
      "step": 2000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8693877551020408,
      "eval_loss": 0.5535635352134705,
      "eval_runtime": 1.079,
      "eval_samples_per_second": 227.063,
      "eval_steps_per_second": 14.829,
      "step": 2200
    }
  ],
  "logging_steps": 500,
  "max_steps": 2200,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 500,
  "total_flos": 3892671035400.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 1.946279842307678e-05,
    "num_train_epochs": 4,
    "per_device_train_batch_size": 4,
    "seed": 23
  }
}