run-0/checkpoint-660/trainer_state.json

{
  "best_metric": 0.796057619408643,
  "best_model_checkpoint": "distilbert-base-uncased-finetuned-multiclass-classification/run-0/checkpoint-660",
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 660,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5113722517058378,
      "eval_loss": 4.086510181427002,
      "eval_runtime": 39.198,
      "eval_samples_per_second": 67.299,
      "eval_steps_per_second": 4.209,
      "step": 165
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6884003032600455,
      "eval_loss": 2.6795032024383545,
      "eval_runtime": 40.511,
      "eval_samples_per_second": 65.118,
      "eval_steps_per_second": 4.073,
      "step": 330
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7763457164518575,
      "eval_loss": 2.0228610038757324,
      "eval_runtime": 42.8401,
      "eval_samples_per_second": 61.578,
      "eval_steps_per_second": 3.852,
      "step": 495
    },
    {
      "epoch": 3.0303030303030303,
      "grad_norm": 3.224958896636963,
      "learning_rate": 1.255944459585537e-05,
      "loss": 3.7537,
      "step": 500
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.796057619408643,
      "eval_loss": 1.7170225381851196,
      "eval_runtime": 41.3607,
      "eval_samples_per_second": 63.78,
      "eval_steps_per_second": 3.989,
      "step": 660
    }
  ],
  "logging_steps": 500,
  "max_steps": 825,
  "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": 355824719364096.0,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 3.188166705101748e-05,
    "num_train_epochs": 5,
    "per_device_train_batch_size": 64,
    "seed": 29
  }
}