run-1/checkpoint-660/trainer_state.json

{
  "best_metric": 0.6849886277482942,
  "best_model_checkpoint": "distilbert-base-uncased-finetuned-multiclass-classification/run-1/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": 0.44086429112964365,
      "eval_loss": 4.671164035797119,
      "eval_runtime": 42.4519,
      "eval_samples_per_second": 62.141,
      "eval_steps_per_second": 3.887,
      "step": 165
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6057619408642911,
      "eval_loss": 3.464550495147705,
      "eval_runtime": 39.1224,
      "eval_samples_per_second": 67.429,
      "eval_steps_per_second": 4.218,
      "step": 330
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6576952236542836,
      "eval_loss": 2.8911187648773193,
      "eval_runtime": 42.0047,
      "eval_samples_per_second": 62.802,
      "eval_steps_per_second": 3.928,
      "step": 495
    },
    {
      "epoch": 3.0303030303030303,
      "grad_norm": 3.1720504760742188,
      "learning_rate": 5.835366506489633e-06,
      "loss": 4.3235,
      "step": 500
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6849886277482942,
      "eval_loss": 2.7205638885498047,
      "eval_runtime": 38.9179,
      "eval_samples_per_second": 67.784,
      "eval_steps_per_second": 4.24,
      "step": 660
    }
  ],
  "logging_steps": 500,
  "max_steps": 660,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 355067764932608.0,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 2.4070886839269733e-05,
    "num_train_epochs": 4,
    "per_device_train_batch_size": 64,
    "seed": 20
  }
}