trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.764705882352941,
  "eval_steps": 25,
  "global_step": 8,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.47058823529411764,
      "grad_norm": 1.2098472118377686,
      "learning_rate": 0.000175,
      "loss": 9.5796,
      "step": 1
    },
    {
      "epoch": 0.9411764705882353,
      "grad_norm": 1.1542775630950928,
      "learning_rate": 0.00015000000000000001,
      "loss": 9.4541,
      "step": 2
    },
    {
      "epoch": 1.4117647058823528,
      "grad_norm": 1.1577507257461548,
      "learning_rate": 0.000125,
      "loss": 9.2188,
      "step": 3
    },
    {
      "epoch": 1.8823529411764706,
      "grad_norm": 1.3564627170562744,
      "learning_rate": 0.0001,
      "loss": 9.02,
      "step": 4
    },
    {
      "epoch": 2.3529411764705883,
      "grad_norm": 1.611424446105957,
      "learning_rate": 7.500000000000001e-05,
      "loss": 8.8105,
      "step": 5
    },
    {
      "epoch": 2.8235294117647056,
      "grad_norm": 1.7456287145614624,
      "learning_rate": 5e-05,
      "loss": 8.7183,
      "step": 6
    },
    {
      "epoch": 3.2941176470588234,
      "grad_norm": 2.1504805088043213,
      "learning_rate": 2.5e-05,
      "loss": 8.4925,
      "step": 7
    },
    {
      "epoch": 3.764705882352941,
      "grad_norm": 2.3203675746917725,
      "learning_rate": 0.0,
      "loss": 8.4446,
      "step": 8
    },
    {
      "epoch": 3.764705882352941,
      "step": 8,
      "total_flos": 3464493556500.0,
      "train_loss": 8.967305541038513,
      "train_runtime": 43.8097,
      "train_samples_per_second": 3.013,
      "train_steps_per_second": 0.183
    }
  ],
  "logging_steps": 1.0,
  "max_steps": 8,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": false,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 3464493556500.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}