trainer_state.json

{
  "best_metric": 0.9415858853520276,
  "best_model_checkpoint": "model_saves/deberta-large_lemon_5k_4_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": 0.9415858853520276,
      "eval_loss": 0.4267314672470093,
      "eval_runtime": 4.6709,
      "eval_samples_per_second": 938.154,
      "eval_steps_per_second": 7.493,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9402553695815025,
      "eval_loss": 0.4596344828605652,
      "eval_runtime": 4.6631,
      "eval_samples_per_second": 939.71,
      "eval_steps_per_second": 7.506,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9400652958999989,
      "eval_loss": 0.5083373188972473,
      "eval_runtime": 4.6625,
      "eval_samples_per_second": 939.839,
      "eval_steps_per_second": 7.507,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2208,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9393720860027505,
      "eval_loss": 0.5561937093734741,
      "eval_runtime": 4.6609,
      "eval_samples_per_second": 940.166,
      "eval_steps_per_second": 7.509,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.60866272018432e+16,
      "train_loss": 0.21498672375038488,
      "train_runtime": 468.6637,
      "train_samples_per_second": 1097.93,
      "train_steps_per_second": 8.578
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.60866272018432e+16,
  "trial_name": null,
  "trial_params": null
}