trainer_state.json

{
  "best_metric": 0.9413678596585382,
  "best_model_checkpoint": "model_saves/deberta-large_lemon-spell_10k_2_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": 0.9413678596585382,
      "eval_loss": 0.4281282126903534,
      "eval_runtime": 4.7025,
      "eval_samples_per_second": 931.836,
      "eval_steps_per_second": 7.443,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9402274175695167,
      "eval_loss": 0.45566827058792114,
      "eval_runtime": 4.8079,
      "eval_samples_per_second": 911.423,
      "eval_steps_per_second": 7.28,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9399478974496585,
      "eval_loss": 0.4906838834285736,
      "eval_runtime": 4.7024,
      "eval_samples_per_second": 931.861,
      "eval_steps_per_second": 7.443,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.249,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9403224544102684,
      "eval_loss": 0.5484734177589417,
      "eval_runtime": 4.694,
      "eval_samples_per_second": 933.536,
      "eval_steps_per_second": 7.456,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.6335737586712576e+16,
      "train_loss": 0.24266756203637194,
      "train_runtime": 495.9831,
      "train_samples_per_second": 1037.455,
      "train_steps_per_second": 8.105
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.6335737586712576e+16,
  "trial_name": null,
  "trial_params": null
}