trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.3220125786163522,
  "eval_steps": 500,
  "global_step": 10,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.03,
      "learning_rate": 1.994869323391895e-05,
      "loss": 2.5826,
      "step": 1
    },
    {
      "epoch": 0.16,
      "learning_rate": 1.8743466161445823e-05,
      "loss": 2.4436,
      "step": 5
    },
    {
      "epoch": 0.32,
      "learning_rate": 1.5289640103269626e-05,
      "loss": 2.0318,
      "step": 10
    },
    {
      "epoch": 0.32,
      "eval_loss": 1.780266523361206,
      "eval_runtime": 154.3285,
      "eval_samples_per_second": 11.463,
      "eval_steps_per_second": 1.438,
      "step": 10
    },
    {
      "epoch": 0.32,
      "step": 10,
      "total_flos": 4.903271030325248e+17,
      "train_loss": 2.409137284755707,
      "train_runtime": 4600.7142,
      "train_samples_per_second": 3.456,
      "train_steps_per_second": 0.007
    }
  ],
  "logging_steps": 5,
  "max_steps": 31,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 4.903271030325248e+17,
  "trial_name": null,
  "trial_params": null
}