trainer_state.json

{
  "best_metric": 0.9388298169702255,
  "best_model_checkpoint": "model_saves/roberta-large_spell_5k_5_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": 0.9388298169702255,
      "eval_loss": 0.44159987568855286,
      "eval_runtime": 3.3578,
      "eval_samples_per_second": 1305.032,
      "eval_steps_per_second": 10.424,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9383825847784524,
      "eval_loss": 0.4566632807254791,
      "eval_runtime": 3.3614,
      "eval_samples_per_second": 1303.619,
      "eval_steps_per_second": 10.412,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9385894296671474,
      "eval_loss": 0.5054126381874084,
      "eval_runtime": 3.365,
      "eval_samples_per_second": 1302.238,
      "eval_steps_per_second": 10.401,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2675,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9384943928263957,
      "eval_loss": 0.5354475378990173,
      "eval_runtime": 3.3723,
      "eval_samples_per_second": 1299.394,
      "eval_steps_per_second": 10.379,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.38598443646976e+16,
      "train_loss": 0.2625667621840292,
      "train_runtime": 318.2707,
      "train_samples_per_second": 1616.737,
      "train_steps_per_second": 12.631
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.38598443646976e+16,
  "trial_name": null,
  "trial_params": null
}