trainer_state.json

{
  "best_metric": 0.9414405348897014,
  "best_model_checkpoint": "model_saves/deberta-large_lemon_5k_3_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": 0.9414405348897014,
      "eval_loss": 0.4162137806415558,
      "eval_runtime": 4.632,
      "eval_samples_per_second": 946.027,
      "eval_steps_per_second": 7.556,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9412448708058005,
      "eval_loss": 0.4352792203426361,
      "eval_runtime": 4.6349,
      "eval_samples_per_second": 945.428,
      "eval_steps_per_second": 7.551,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9402106463623252,
      "eval_loss": 0.47978702187538147,
      "eval_runtime": 4.645,
      "eval_samples_per_second": 943.388,
      "eval_steps_per_second": 7.535,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2573,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9398137277921265,
      "eval_loss": 0.5360333323478699,
      "eval_runtime": 4.6334,
      "eval_samples_per_second": 945.747,
      "eval_steps_per_second": 7.554,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.6078384287186944e+16,
      "train_loss": 0.2507487820155585,
      "train_runtime": 457.1298,
      "train_samples_per_second": 1125.632,
      "train_steps_per_second": 8.794
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.6078384287186944e+16,
  "trial_name": null,
  "trial_params": null
}