trainer_state.json

{
  "best_metric": 0.9393888572099419,
  "best_model_checkpoint": "model_saves/electra-large_lemon_10k_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": 0.9393888572099419,
      "eval_loss": 0.44656530022621155,
      "eval_runtime": 3.4529,
      "eval_samples_per_second": 1269.087,
      "eval_steps_per_second": 10.136,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9385391160455729,
      "eval_loss": 0.4600903391838074,
      "eval_runtime": 3.4585,
      "eval_samples_per_second": 1267.016,
      "eval_steps_per_second": 10.12,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9384328984000269,
      "eval_loss": 0.4773552715778351,
      "eval_runtime": 3.4592,
      "eval_samples_per_second": 1266.763,
      "eval_steps_per_second": 10.118,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.3208,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9383769943760552,
      "eval_loss": 0.5144086480140686,
      "eval_runtime": 3.4653,
      "eval_samples_per_second": 1264.547,
      "eval_steps_per_second": 10.1,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.4147605630025728e+16,
      "train_loss": 0.31563160312709526,
      "train_runtime": 322.2044,
      "train_samples_per_second": 1596.999,
      "train_steps_per_second": 12.477
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.4147605630025728e+16,
  "trial_name": null,
  "trial_params": null
}