trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.769230769230769,
  "eval_steps": 500,
  "global_step": 18,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.15,
      "grad_norm": 10.606050006078899,
      "learning_rate": 1e-05,
      "loss": 1.6582,
      "step": 1
    },
    {
      "epoch": 0.77,
      "grad_norm": 7.392104831059971,
      "learning_rate": 1.8314696123025456e-05,
      "loss": 1.6842,
      "step": 5
    },
    {
      "epoch": 0.92,
      "eval_loss": 2.2982828617095947,
      "eval_runtime": 5.7007,
      "eval_samples_per_second": 8.069,
      "eval_steps_per_second": 0.526,
      "step": 6
    },
    {
      "epoch": 1.54,
      "grad_norm": 5.487530145939054,
      "learning_rate": 1e-05,
      "loss": 1.312,
      "step": 10
    },
    {
      "epoch": 2.0,
      "eval_loss": 2.358544111251831,
      "eval_runtime": 5.6741,
      "eval_samples_per_second": 8.107,
      "eval_steps_per_second": 0.529,
      "step": 13
    },
    {
      "epoch": 2.31,
      "grad_norm": 5.752888566223131,
      "learning_rate": 1.6853038769745466e-06,
      "loss": 0.8758,
      "step": 15
    },
    {
      "epoch": 2.77,
      "eval_loss": 2.4065513610839844,
      "eval_runtime": 5.6783,
      "eval_samples_per_second": 8.101,
      "eval_steps_per_second": 0.528,
      "step": 18
    },
    {
      "epoch": 2.77,
      "step": 18,
      "total_flos": 3716488888320.0,
      "train_loss": 1.1936739020877414,
      "train_runtime": 399.8921,
      "train_samples_per_second": 3.031,
      "train_steps_per_second": 0.045
    }
  ],
  "logging_steps": 5,
  "max_steps": 18,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 3716488888320.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}