trainer_state.json

{
  "best_metric": 0.9405181184941692,
  "best_model_checkpoint": "model_saves/deberta-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": 0.9405181184941692,
      "eval_loss": 0.42925789952278137,
      "eval_runtime": 4.6465,
      "eval_samples_per_second": 943.085,
      "eval_steps_per_second": 7.533,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9391428795044667,
      "eval_loss": 0.4665437638759613,
      "eval_runtime": 4.6371,
      "eval_samples_per_second": 944.99,
      "eval_steps_per_second": 7.548,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9388242265678284,
      "eval_loss": 0.5086150169372559,
      "eval_runtime": 4.646,
      "eval_samples_per_second": 943.168,
      "eval_steps_per_second": 7.533,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2208,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9393888572099419,
      "eval_loss": 0.5620639324188232,
      "eval_runtime": 4.6383,
      "eval_samples_per_second": 944.734,
      "eval_steps_per_second": 7.546,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.6105037109395456e+16,
      "train_loss": 0.2150960520132264,
      "train_runtime": 460.8647,
      "train_samples_per_second": 1116.51,
      "train_steps_per_second": 8.723
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.6105037109395456e+16,
  "trial_name": null,
  "trial_params": null
}