trainer_state.json

{
  "best_metric": 0.9394783036482967,
  "best_model_checkpoint": "model_saves/electra-large_spell_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": 0.9394783036482967,
      "eval_loss": 0.4361410140991211,
      "eval_runtime": 3.4839,
      "eval_samples_per_second": 1257.772,
      "eval_steps_per_second": 10.046,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.93850557363119,
      "eval_loss": 0.4486663341522217,
      "eval_runtime": 3.4696,
      "eval_samples_per_second": 1262.962,
      "eval_steps_per_second": 10.088,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9387571417390623,
      "eval_loss": 0.4750184416770935,
      "eval_runtime": 3.4594,
      "eval_samples_per_second": 1266.704,
      "eval_steps_per_second": 10.117,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.3204,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9371247442390903,
      "eval_loss": 0.4949221909046173,
      "eval_runtime": 3.4777,
      "eval_samples_per_second": 1260.039,
      "eval_steps_per_second": 10.064,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.3926494673829888e+16,
      "train_loss": 0.31591533902865737,
      "train_runtime": 329.3261,
      "train_samples_per_second": 1562.463,
      "train_steps_per_second": 12.207
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.3926494673829888e+16,
  "trial_name": null,
  "trial_params": null
}