checkpoint-189/trainer_state.json

{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 189,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 0.8975155434757471,
      "epoch": 0.8,
      "grad_norm": 2.0314748287200928,
      "learning_rate": 0.0001739688860864967,
      "loss": 0.8406,
      "mean_token_accuracy": 0.8197277544438839,
      "num_tokens": 397069.0,
      "step": 50
    },
    {
      "entropy": 0.5006740084952779,
      "epoch": 1.592,
      "grad_norm": 0.6300589442253113,
      "learning_rate": 9.742520863450115e-05,
      "loss": 0.4951,
      "mean_token_accuracy": 0.8895806454949908,
      "num_tokens": 798321.0,
      "step": 100
    },
    {
      "entropy": 0.4396799075031521,
      "epoch": 2.384,
      "grad_norm": 0.615988552570343,
      "learning_rate": 2.2664932229027024e-05,
      "loss": 0.4459,
      "mean_token_accuracy": 0.8991429293998564,
      "num_tokens": 1198472.0,
      "step": 150
    }
  ],
  "logging_steps": 50,
  "max_steps": 189,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 1000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 3.07881924006912e+16,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}