trainer_state.json

{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 399.72727272727275,
  "eval_steps": 500,
  "global_step": 800,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 49.72727272727273,
      "grad_norm": 8.328217506408691,
      "learning_rate": 8.775e-06,
      "loss": 1.8952,
      "step": 100
    },
    {
      "epoch": 99.72727272727273,
      "grad_norm": 4.615832805633545,
      "learning_rate": 7.525e-06,
      "loss": 0.4619,
      "step": 200
    },
    {
      "epoch": 149.72727272727272,
      "grad_norm": 2.132920503616333,
      "learning_rate": 6.275e-06,
      "loss": 0.0905,
      "step": 300
    },
    {
      "epoch": 199.72727272727272,
      "grad_norm": 0.5747965574264526,
      "learning_rate": 5.025e-06,
      "loss": 0.0256,
      "step": 400
    },
    {
      "epoch": 249.72727272727272,
      "grad_norm": 0.17656728625297546,
      "learning_rate": 3.7750000000000003e-06,
      "loss": 0.013,
      "step": 500
    },
    {
      "epoch": 299.72727272727275,
      "grad_norm": 0.22439385950565338,
      "learning_rate": 2.5250000000000004e-06,
      "loss": 0.0083,
      "step": 600
    },
    {
      "epoch": 349.72727272727275,
      "grad_norm": 0.07200386375188828,
      "learning_rate": 1.275e-06,
      "loss": 0.0059,
      "step": 700
    },
    {
      "epoch": 399.72727272727275,
      "grad_norm": 1.7082760334014893,
      "learning_rate": 2.5000000000000002e-08,
      "loss": 0.0035,
      "step": 800
    }
  ],
  "logging_steps": 100,
  "max_steps": 800,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 400,
  "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": 8774662198149120.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}