trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.9999175076305442,
  "global_step": 18182,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "learning_rate": 5.000549994500056e-06,
      "loss": 0.9118,
      "step": 9091
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.8586767315864563,
      "eval_runtime": 1119.8237,
      "eval_samples_per_second": 7.183,
      "eval_steps_per_second": 7.183,
      "step": 9091
    },
    {
      "epoch": 2.0,
      "learning_rate": 3.2999670003299967e-09,
      "loss": 0.8033,
      "step": 18182
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.8357626795768738,
      "eval_runtime": 1139.4211,
      "eval_samples_per_second": 7.06,
      "eval_steps_per_second": 7.06,
      "step": 18182
    },
    {
      "epoch": 2.0,
      "step": 18182,
      "total_flos": 1.0800187738989527e+18,
      "train_loss": 0.8575281141719833,
      "train_runtime": 49942.9507,
      "train_samples_per_second": 2.913,
      "train_steps_per_second": 0.364
    }
  ],
  "max_steps": 18182,
  "num_train_epochs": 2,
  "total_flos": 1.0800187738989527e+18,
  "trial_name": null,
  "trial_params": null
}