trainer_state.json

{
  "best_metric": 0.9403559968246514,
  "best_model_checkpoint": "model_saves/roberta-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": 0.9403559968246514,
      "eval_loss": 0.42721834778785706,
      "eval_runtime": 3.3595,
      "eval_samples_per_second": 1304.35,
      "eval_steps_per_second": 10.418,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9393329531859703,
      "eval_loss": 0.43651333451271057,
      "eval_runtime": 3.3601,
      "eval_samples_per_second": 1304.114,
      "eval_steps_per_second": 10.416,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9397019197441832,
      "eval_loss": 0.47140324115753174,
      "eval_runtime": 3.3604,
      "eval_samples_per_second": 1304.026,
      "eval_steps_per_second": 10.416,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2903,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9390087098469347,
      "eval_loss": 0.507944643497467,
      "eval_runtime": 3.3695,
      "eval_samples_per_second": 1300.49,
      "eval_steps_per_second": 10.387,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.3874533421285376e+16,
      "train_loss": 0.2849151705628011,
      "train_runtime": 318.1632,
      "train_samples_per_second": 1617.283,
      "train_steps_per_second": 12.635
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.3874533421285376e+16,
  "trial_name": null,
  "trial_params": null
}