trainer_state.json

{
  "best_metric": 0.9400093918760273,
  "best_model_checkpoint": "model_saves/electra-large_lemon_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": 0.9400093918760273,
      "eval_loss": 0.4350626468658447,
      "eval_runtime": 3.4299,
      "eval_samples_per_second": 1277.585,
      "eval_steps_per_second": 10.204,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9391372891020696,
      "eval_loss": 0.4442809820175171,
      "eval_runtime": 3.427,
      "eval_samples_per_second": 1278.686,
      "eval_steps_per_second": 10.213,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9394838940506938,
      "eval_loss": 0.4578820466995239,
      "eval_runtime": 3.4195,
      "eval_samples_per_second": 1281.475,
      "eval_steps_per_second": 10.235,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.3493,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9384608504120127,
      "eval_loss": 0.48522841930389404,
      "eval_runtime": 3.4121,
      "eval_samples_per_second": 1284.239,
      "eval_steps_per_second": 10.257,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.3892022125461504e+16,
      "train_loss": 0.34391611725536747,
      "train_runtime": 316.6274,
      "train_samples_per_second": 1625.128,
      "train_steps_per_second": 12.696
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.3892022125461504e+16,
  "trial_name": null,
  "trial_params": null
}