trainer_state.json

{
  "best_metric": 0.9400205726808216,
  "best_model_checkpoint": "model_saves/roberta-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": 0.9400205726808216,
      "eval_loss": 0.43053317070007324,
      "eval_runtime": 3.3658,
      "eval_samples_per_second": 1301.938,
      "eval_steps_per_second": 10.399,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9392490971500128,
      "eval_loss": 0.4418651759624481,
      "eval_runtime": 3.3543,
      "eval_samples_per_second": 1306.366,
      "eval_steps_per_second": 10.434,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9386341528863248,
      "eval_loss": 0.4717348515987396,
      "eval_runtime": 3.3519,
      "eval_samples_per_second": 1307.332,
      "eval_steps_per_second": 10.442,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2908,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9389472154205659,
      "eval_loss": 0.512559175491333,
      "eval_runtime": 3.3494,
      "eval_samples_per_second": 1308.284,
      "eval_steps_per_second": 10.45,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.3886616305139712e+16,
      "train_loss": 0.2856182295884659,
      "train_runtime": 329.4961,
      "train_samples_per_second": 1561.657,
      "train_steps_per_second": 12.2
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.3886616305139712e+16,
  "trial_name": null,
  "trial_params": null
}