trainer_state.json

{
  "best_metric": 0.9409933026979282,
  "best_model_checkpoint": "model_saves/deberta-large_spell_5k_4_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": 0.9409933026979282,
      "eval_loss": 0.42714807391166687,
      "eval_runtime": 4.9949,
      "eval_samples_per_second": 877.292,
      "eval_steps_per_second": 7.007,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9401156095215734,
      "eval_loss": 0.4541383385658264,
      "eval_runtime": 5.4718,
      "eval_samples_per_second": 800.836,
      "eval_steps_per_second": 6.396,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.939729871756169,
      "eval_loss": 0.49630552530288696,
      "eval_runtime": 4.8353,
      "eval_samples_per_second": 906.257,
      "eval_steps_per_second": 7.238,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.224,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9401603327407507,
      "eval_loss": 0.5561859607696533,
      "eval_runtime": 4.7211,
      "eval_samples_per_second": 928.164,
      "eval_steps_per_second": 7.413,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.6091023973285888e+16,
      "train_loss": 0.21816402940607782,
      "train_runtime": 612.0564,
      "train_samples_per_second": 840.707,
      "train_steps_per_second": 6.568
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.6091023973285888e+16,
  "trial_name": null,
  "trial_params": null
}