trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.6984126984126986,
  "eval_steps": 500,
  "global_step": 21,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.8888888888888888,
      "eval_loss": 0.2620887756347656,
      "eval_runtime": 10.139,
      "eval_samples_per_second": 21.008,
      "eval_steps_per_second": 0.395,
      "step": 7
    },
    {
      "epoch": 1.2857142857142856,
      "grad_norm": 137.20972468665934,
      "learning_rate": 5e-06,
      "loss": 0.4273,
      "step": 10
    },
    {
      "epoch": 1.9206349206349205,
      "eval_loss": 0.2767665386199951,
      "eval_runtime": 10.1155,
      "eval_samples_per_second": 21.057,
      "eval_steps_per_second": 0.395,
      "step": 15
    },
    {
      "epoch": 2.571428571428571,
      "grad_norm": 14.872716775616082,
      "learning_rate": 5e-06,
      "loss": 0.2692,
      "step": 20
    },
    {
      "epoch": 2.6984126984126986,
      "eval_loss": 0.21467676758766174,
      "eval_runtime": 9.4027,
      "eval_samples_per_second": 22.653,
      "eval_steps_per_second": 0.425,
      "step": 21
    },
    {
      "epoch": 2.6984126984126986,
      "step": 21,
      "total_flos": 34966402498560.0,
      "train_loss": 0.34268649844896226,
      "train_runtime": 1534.2684,
      "train_samples_per_second": 7.88,
      "train_steps_per_second": 0.014
    }
  ],
  "logging_steps": 10,
  "max_steps": 21,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 34966402498560.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}