trainer_state.json

{
  "best_metric": 0.940171513545545,
  "best_model_checkpoint": "model_saves/roberta-large_lemon_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": 0.940171513545545,
      "eval_loss": 0.41954633593559265,
      "eval_runtime": 3.3959,
      "eval_samples_per_second": 1290.368,
      "eval_steps_per_second": 10.306,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9393050011739845,
      "eval_loss": 0.43967413902282715,
      "eval_runtime": 3.4038,
      "eval_samples_per_second": 1287.375,
      "eval_steps_per_second": 10.283,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9397186909513747,
      "eval_loss": 0.46832728385925293,
      "eval_runtime": 3.4056,
      "eval_samples_per_second": 1286.689,
      "eval_steps_per_second": 10.277,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.29,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9380751126466083,
      "eval_loss": 0.5288189649581909,
      "eval_runtime": 3.4067,
      "eval_samples_per_second": 1286.273,
      "eval_steps_per_second": 10.274,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.3870742575775744e+16,
      "train_loss": 0.28459733012896865,
      "train_runtime": 320.8579,
      "train_samples_per_second": 1603.701,
      "train_steps_per_second": 12.529
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.3870742575775744e+16,
  "trial_name": null,
  "trial_params": null
}