trainer_state.json

{
  "best_metric": 0.9394168092219278,
  "best_model_checkpoint": "model_saves/electra-large_lemon_5k_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": 0.9394168092219278,
      "eval_loss": 0.4363527297973633,
      "eval_runtime": 3.4205,
      "eval_samples_per_second": 1281.112,
      "eval_steps_per_second": 10.233,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9385838392647503,
      "eval_loss": 0.451455682516098,
      "eval_runtime": 3.416,
      "eval_samples_per_second": 1282.777,
      "eval_steps_per_second": 10.246,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9385111640335871,
      "eval_loss": 0.4689089059829712,
      "eval_runtime": 3.4314,
      "eval_samples_per_second": 1277.027,
      "eval_steps_per_second": 10.2,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.3218,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.937996847013048,
      "eval_loss": 0.4999893009662628,
      "eval_runtime": 3.4268,
      "eval_samples_per_second": 1278.747,
      "eval_steps_per_second": 10.214,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.3915633573953536e+16,
      "train_loss": 0.31665303102180137,
      "train_runtime": 314.2905,
      "train_samples_per_second": 1637.212,
      "train_steps_per_second": 12.791
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.3915633573953536e+16,
  "trial_name": null,
  "trial_params": null
}