trainer_state.json

{
  "best_metric": 0.941295184427375,
  "best_model_checkpoint": "model_saves/deberta-large_lemon_10k_1_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": 0.941295184427375,
      "eval_loss": 0.42443111538887024,
      "eval_runtime": 4.6909,
      "eval_samples_per_second": 934.156,
      "eval_steps_per_second": 7.461,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9408200002236161,
      "eval_loss": 0.44897493720054626,
      "eval_runtime": 4.6767,
      "eval_samples_per_second": 936.98,
      "eval_steps_per_second": 7.484,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9409318082715594,
      "eval_loss": 0.5006834268569946,
      "eval_runtime": 4.6798,
      "eval_samples_per_second": 936.372,
      "eval_steps_per_second": 7.479,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.249,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9406411073469069,
      "eval_loss": 0.5360766053199768,
      "eval_runtime": 4.6933,
      "eval_samples_per_second": 933.664,
      "eval_steps_per_second": 7.457,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.6315392326631424e+16,
      "train_loss": 0.24265052044569557,
      "train_runtime": 470.3957,
      "train_samples_per_second": 1093.888,
      "train_steps_per_second": 8.546
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.6315392326631424e+16,
  "trial_name": null,
  "trial_params": null
}