trainer_state.json

{
  "best_metric": 0.9402162367647223,
  "best_model_checkpoint": "model_saves/roberta-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": 0.9402162367647223,
      "eval_loss": 0.42938709259033203,
      "eval_runtime": 3.3638,
      "eval_samples_per_second": 1302.688,
      "eval_steps_per_second": 10.405,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9395509788794597,
      "eval_loss": 0.4404939115047455,
      "eval_runtime": 3.3672,
      "eval_samples_per_second": 1301.368,
      "eval_steps_per_second": 10.394,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9392435067476157,
      "eval_loss": 0.4706648588180542,
      "eval_runtime": 3.366,
      "eval_samples_per_second": 1301.851,
      "eval_steps_per_second": 10.398,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.29,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9388074553606368,
      "eval_loss": 0.5094670653343201,
      "eval_runtime": 3.3584,
      "eval_samples_per_second": 1304.803,
      "eval_steps_per_second": 10.422,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.3886616305139712e+16,
      "train_loss": 0.284751944577516,
      "train_runtime": 319.1108,
      "train_samples_per_second": 1612.48,
      "train_steps_per_second": 12.598
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.3886616305139712e+16,
  "trial_name": null,
  "trial_params": null
}