trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 2508,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.39872408293460926,
      "grad_norm": 9.840682983398438,
      "learning_rate": 4.63429565483241e-05,
      "loss": 0.5605,
      "step": 500
    },
    {
      "epoch": 0.7974481658692185,
      "grad_norm": 4.425215721130371,
      "learning_rate": 3.4207917862252084e-05,
      "loss": 0.2413,
      "step": 1000
    },
    {
      "epoch": 1.1961722488038278,
      "grad_norm": 0.04050261154770851,
      "learning_rate": 1.83633111222982e-05,
      "loss": 0.1967,
      "step": 1500
    },
    {
      "epoch": 1.594896331738437,
      "grad_norm": 0.4518131911754608,
      "learning_rate": 5.1924086239082456e-06,
      "loss": 0.1003,
      "step": 2000
    },
    {
      "epoch": 1.9936204146730463,
      "grad_norm": 0.030251897871494293,
      "learning_rate": 1.334930716231919e-09,
      "loss": 0.1014,
      "step": 2500
    },
    {
      "epoch": 2.0,
      "step": 2508,
      "total_flos": 9345407711453184.0,
      "train_loss": 0.239293363770066,
      "train_runtime": 1172.3192,
      "train_samples_per_second": 8.554,
      "train_steps_per_second": 2.139
    }
  ],
  "logging_steps": 500,
  "max_steps": 2508,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 1254,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 9345407711453184.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}