trainer_state.json

{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "eval_steps": 500,
  "global_step": 450,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.022222222222222223,
      "grad_norm": 33.61671447753906,
      "learning_rate": 2e-05,
      "loss": 5.4203,
      "mean_token_accuracy": 0.21561338007450104,
      "num_tokens": 1071.0,
      "step": 1
    },
    {
      "epoch": 1.0,
      "grad_norm": 39.851768493652344,
      "learning_rate": 1.8044444444444445e-05,
      "loss": 2.8638,
      "mean_token_accuracy": 0.4754302335733717,
      "num_tokens": 72222.0,
      "step": 45
    },
    {
      "epoch": 2.0,
      "grad_norm": 23.199281692504883,
      "learning_rate": 1.6044444444444444e-05,
      "loss": 1.3918,
      "mean_token_accuracy": 0.6957559704780578,
      "num_tokens": 144444.0,
      "step": 90
    },
    {
      "epoch": 3.0,
      "grad_norm": 11.660584449768066,
      "learning_rate": 1.4044444444444445e-05,
      "loss": 0.6698,
      "mean_token_accuracy": 0.8350067456563314,
      "num_tokens": 216666.0,
      "step": 135
    },
    {
      "epoch": 4.0,
      "grad_norm": 4.399273872375488,
      "learning_rate": 1.2044444444444445e-05,
      "loss": 0.3166,
      "mean_token_accuracy": 0.9196972608566284,
      "num_tokens": 288888.0,
      "step": 180
    },
    {
      "epoch": 5.0,
      "grad_norm": 13.878437995910645,
      "learning_rate": 1.0044444444444446e-05,
      "loss": 0.1585,
      "mean_token_accuracy": 0.9616732305950588,
      "num_tokens": 361110.0,
      "step": 225
    },
    {
      "epoch": 6.0,
      "grad_norm": 7.022050380706787,
      "learning_rate": 8.044444444444444e-06,
      "loss": 0.0867,
      "mean_token_accuracy": 0.9774630414115058,
      "num_tokens": 433332.0,
      "step": 270
    },
    {
      "epoch": 7.0,
      "grad_norm": 2.5981574058532715,
      "learning_rate": 6.044444444444445e-06,
      "loss": 0.0484,
      "mean_token_accuracy": 0.9874586674902174,
      "num_tokens": 505554.0,
      "step": 315
    },
    {
      "epoch": 8.0,
      "grad_norm": 9.60135555267334,
      "learning_rate": 4.044444444444445e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9892550269762675,
      "num_tokens": 577776.0,
      "step": 360
    },
    {
      "epoch": 9.0,
      "grad_norm": 4.281480312347412,
      "learning_rate": 2.0444444444444447e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9917015499538846,
      "num_tokens": 649998.0,
      "step": 405
    },
    {
      "epoch": 10.0,
      "grad_norm": 0.3038378059864044,
      "learning_rate": 4.444444444444445e-08,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9933217287063598,
      "num_tokens": 722220.0,
      "step": 450
    },
    {
      "epoch": 10.0,
      "step": 450,
      "total_flos": 3625341809836032.0,
      "train_loss": 0.5666879651281569,
      "train_runtime": 426.1414,
      "train_samples_per_second": 8.354,
      "train_steps_per_second": 1.056
    }
  ],
  "logging_steps": 10,
  "max_steps": 450,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 1000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": false,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 3625341809836032.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}