trainer_state.json

{
  "best_metric": 0.8629648497198166,
  "best_model_checkpoint": "./outputs/xlnet-base-cased/fine_tuned_models/checkpoint-49088",
  "epoch": 5.0,
  "global_step": 61360,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "learning_rate": 4.34714842366277e-05,
      "loss": 0.545,
      "step": 12272
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8356597045338767,
      "eval_loss": 0.4324648678302765,
      "eval_runtime": 36.38,
      "eval_samples_per_second": 269.791,
      "eval_steps_per_second": 33.727,
      "step": 12272
    },
    {
      "epoch": 2.0,
      "learning_rate": 3.260361317747078e-05,
      "loss": 0.3777,
      "step": 24544
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8493122771268466,
      "eval_loss": 0.4276026785373688,
      "eval_runtime": 36.3876,
      "eval_samples_per_second": 269.735,
      "eval_steps_per_second": 33.72,
      "step": 24544
    },
    {
      "epoch": 3.0,
      "learning_rate": 2.173574211831385e-05,
      "loss": 0.2632,
      "step": 36816
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8523688232297504,
      "eval_loss": 0.4237123131752014,
      "eval_runtime": 36.3746,
      "eval_samples_per_second": 269.831,
      "eval_steps_per_second": 33.732,
      "step": 36816
    },
    {
      "epoch": 4.0,
      "learning_rate": 1.0867871059156925e-05,
      "loss": 0.1704,
      "step": 49088
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8629648497198166,
      "eval_loss": 0.5061570405960083,
      "eval_runtime": 36.3729,
      "eval_samples_per_second": 269.844,
      "eval_steps_per_second": 33.734,
      "step": 49088
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0,
      "loss": 0.1082,
      "step": 61360
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8607233825776872,
      "eval_loss": 0.6351755857467651,
      "eval_runtime": 36.4174,
      "eval_samples_per_second": 269.514,
      "eval_steps_per_second": 33.693,
      "step": 61360
    },
    {
      "epoch": 5.0,
      "step": 61360,
      "total_flos": 1.769024893707648e+17,
      "train_loss": 0.29289959804330995,
      "train_runtime": 20293.0578,
      "train_samples_per_second": 96.758,
      "train_steps_per_second": 3.024
    }
  ],
  "max_steps": 61360,
  "num_train_epochs": 5,
  "total_flos": 1.769024893707648e+17,
  "trial_name": null,
  "trial_params": null
}