trainer_state.json

{
  "best_metric": 0.9397690045729492,
  "best_model_checkpoint": "model_saves/electra-large_lemon-spell_5k_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": 0.9397690045729492,
      "eval_loss": 0.4357127845287323,
      "eval_runtime": 3.4067,
      "eval_samples_per_second": 1286.303,
      "eval_steps_per_second": 10.274,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9391428795044667,
      "eval_loss": 0.4499569833278656,
      "eval_runtime": 3.4041,
      "eval_samples_per_second": 1287.272,
      "eval_steps_per_second": 10.282,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9388018649582397,
      "eval_loss": 0.4677518308162689,
      "eval_runtime": 3.4064,
      "eval_samples_per_second": 1286.42,
      "eval_steps_per_second": 10.275,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.3213,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9383658135712608,
      "eval_loss": 0.5005788207054138,
      "eval_runtime": 3.4088,
      "eval_samples_per_second": 1285.513,
      "eval_steps_per_second": 10.268,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.3915633573953536e+16,
      "train_loss": 0.31619261271918,
      "train_runtime": 315.7506,
      "train_samples_per_second": 1629.641,
      "train_steps_per_second": 12.732
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.3915633573953536e+16,
  "trial_name": null,
  "trial_params": null
}