trainer_state.json

{
  "best_metric": 0.9416361989736021,
  "best_model_checkpoint": "model_saves/deberta-large_lemon-spell_5k_6_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": 0.9416361989736021,
      "eval_loss": 0.4194476902484894,
      "eval_runtime": 4.6377,
      "eval_samples_per_second": 944.856,
      "eval_steps_per_second": 7.547,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9409933026979282,
      "eval_loss": 0.4475340247154236,
      "eval_runtime": 4.6497,
      "eval_samples_per_second": 942.428,
      "eval_steps_per_second": 7.527,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9404063104462259,
      "eval_loss": 0.48383891582489014,
      "eval_runtime": 4.6485,
      "eval_samples_per_second": 942.67,
      "eval_steps_per_second": 7.529,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2547,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9390087098469347,
      "eval_loss": 0.5387859344482422,
      "eval_runtime": 4.6443,
      "eval_samples_per_second": 943.514,
      "eval_steps_per_second": 7.536,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.6074262796304384e+16,
      "train_loss": 0.24829398251291532,
      "train_runtime": 456.1591,
      "train_samples_per_second": 1128.027,
      "train_steps_per_second": 8.813
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.6074262796304384e+16,
  "trial_name": null,
  "trial_params": null
}