checkpoint-1840/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.993215739484397,
  "eval_steps": 500,
  "global_step": 1840,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.3568521031207599,
      "grad_norm": 1.447092056274414,
      "learning_rate": 4.9800000000000004e-05,
      "loss": 0.0263,
      "step": 500
    },
    {
      "epoch": 2.7137042062415198,
      "grad_norm": 18.664369583129883,
      "learning_rate": 3.4810224965023004e-05,
      "loss": 0.0457,
      "step": 1000
    },
    {
      "epoch": 4.07055630936228,
      "grad_norm": 0.045205749571323395,
      "learning_rate": 7.614811590589446e-06,
      "loss": 0.0204,
      "step": 1500
    }
  ],
  "logging_steps": 500,
  "max_steps": 1840,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "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": 1.5591942691405824e+16,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}