trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 2826,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5307855626326964,
      "grad_norm": 7.908592224121094,
      "learning_rate": 8.333333333333334e-06,
      "loss": 2.288,
      "step": 500
    },
    {
      "epoch": 1.0615711252653928,
      "grad_norm": 9.661299705505371,
      "learning_rate": 8.203054806828391e-06,
      "loss": 1.1887,
      "step": 1000
    },
    {
      "epoch": 1.5923566878980893,
      "grad_norm": 6.738057613372803,
      "learning_rate": 5.9568733153638815e-06,
      "loss": 0.9571,
      "step": 1500
    },
    {
      "epoch": 2.1231422505307855,
      "grad_norm": 8.572662353515625,
      "learning_rate": 3.710691823899371e-06,
      "loss": 0.8923,
      "step": 2000
    },
    {
      "epoch": 2.653927813163482,
      "grad_norm": 7.6663737297058105,
      "learning_rate": 1.464510332434861e-06,
      "loss": 0.8164,
      "step": 2500
    },
    {
      "epoch": 3.0,
      "step": 2826,
      "total_flos": 1.1971550128324608e+16,
      "train_loss": 1.1780135147299904,
      "train_runtime": 24677.7647,
      "train_samples_per_second": 3.661,
      "train_steps_per_second": 0.115
    }
  ],
  "logging_steps": 500,
  "max_steps": 2826,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "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": 1.1971550128324608e+16,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}