trainer_state.json

{
  "best_metric": 0.9391876027236441,
  "best_model_checkpoint": "model_saves/electra-large_spell_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": 0.9391876027236441,
      "eval_loss": 0.44028031826019287,
      "eval_runtime": 3.4892,
      "eval_samples_per_second": 1255.875,
      "eval_steps_per_second": 10.031,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.938371403973658,
      "eval_loss": 0.453914076089859,
      "eval_runtime": 3.4806,
      "eval_samples_per_second": 1258.993,
      "eval_steps_per_second": 10.056,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9380695222442111,
      "eval_loss": 0.4717390835285187,
      "eval_runtime": 3.507,
      "eval_samples_per_second": 1249.519,
      "eval_steps_per_second": 9.98,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.3207,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9378347253435302,
      "eval_loss": 0.5070739388465881,
      "eval_runtime": 3.4908,
      "eval_samples_per_second": 1255.31,
      "eval_steps_per_second": 10.026,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.391988398096384e+16,
      "train_loss": 0.3156101401172467,
      "train_runtime": 427.104,
      "train_samples_per_second": 1204.765,
      "train_steps_per_second": 9.412
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.391988398096384e+16,
  "trial_name": null,
  "trial_params": null
}