trainer_state.json

{
  "best_metric": 0.9406690593588927,
  "best_model_checkpoint": "model_saves/roberta-large_lemon_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": 0.9406690593588927,
      "eval_loss": 0.4224521517753601,
      "eval_runtime": 3.3664,
      "eval_samples_per_second": 1301.703,
      "eval_steps_per_second": 10.397,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9404119008486231,
      "eval_loss": 0.4325179159641266,
      "eval_runtime": 3.3738,
      "eval_samples_per_second": 1298.829,
      "eval_steps_per_second": 10.374,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9399478974496585,
      "eval_loss": 0.4515843391418457,
      "eval_runtime": 3.3677,
      "eval_samples_per_second": 1301.201,
      "eval_steps_per_second": 10.393,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.3173,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9387571417390623,
      "eval_loss": 0.4898863732814789,
      "eval_runtime": 3.3685,
      "eval_samples_per_second": 1300.871,
      "eval_steps_per_second": 10.39,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.3860081393205248e+16,
      "train_loss": 0.31179518308212506,
      "train_runtime": 325.3041,
      "train_samples_per_second": 1581.782,
      "train_steps_per_second": 12.358
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.3860081393205248e+16,
  "trial_name": null,
  "trial_params": null
}