trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 12,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5925925925925926,
      "grad_norm": 8.73015022277832,
      "learning_rate": 0.0001666666666666667,
      "loss": 75.9863,
      "step": 2
    },
    {
      "epoch": 1.0,
      "grad_norm": 4.479891777038574,
      "learning_rate": 0.00013333333333333334,
      "loss": 50.6268,
      "step": 4
    },
    {
      "epoch": 1.5925925925925926,
      "grad_norm": 13.511935234069824,
      "learning_rate": 0.00011666666666666668,
      "loss": 70.6998,
      "step": 6
    },
    {
      "epoch": 2.0,
      "grad_norm": 7.3837199211120605,
      "learning_rate": 8.333333333333334e-05,
      "loss": 46.9192,
      "step": 8
    },
    {
      "epoch": 2.5925925925925926,
      "grad_norm": 17.328433990478516,
      "learning_rate": 5e-05,
      "loss": 64.6889,
      "step": 10
    },
    {
      "epoch": 3.0,
      "grad_norm": 5.950709819793701,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 43.3097,
      "step": 12
    },
    {
      "epoch": 3.0,
      "step": 12,
      "total_flos": 47431226450856.0,
      "train_loss": 58.705101013183594,
      "train_runtime": 281.4763,
      "train_samples_per_second": 0.767,
      "train_steps_per_second": 0.043
    }
  ],
  "logging_steps": 2,
  "max_steps": 12,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "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": 47431226450856.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}