trainer_state.json

{
  "best_metric": 0.9413399076465524,
  "best_model_checkpoint": "model_saves/deberta-large_lemon_5k_1_p3/checkpoint-268",
  "epoch": 4.0,
  "global_step": 1072,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9413399076465524,
      "eval_loss": 0.41461190581321716,
      "eval_runtime": 4.6477,
      "eval_samples_per_second": 942.83,
      "eval_steps_per_second": 7.531,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9409653506859423,
      "eval_loss": 0.4394238293170929,
      "eval_runtime": 4.6401,
      "eval_samples_per_second": 944.381,
      "eval_steps_per_second": 7.543,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.94033922561746,
      "eval_loss": 0.49036818742752075,
      "eval_runtime": 4.6422,
      "eval_samples_per_second": 943.959,
      "eval_steps_per_second": 7.54,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2551,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9403112736054742,
      "eval_loss": 0.5281780362129211,
      "eval_runtime": 4.6456,
      "eval_samples_per_second": 943.261,
      "eval_steps_per_second": 7.534,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.6084978873925632e+16,
      "train_loss": 0.2486675666339362,
      "train_runtime": 461.7286,
      "train_samples_per_second": 1114.421,
      "train_steps_per_second": 8.706
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.6084978873925632e+16,
  "trial_name": null,
  "trial_params": null
}