trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 500.0,
  "eval_steps": 500,
  "global_step": 4000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 62.5,
      "grad_norm": 9.630138397216797,
      "learning_rate": 4.375e-05,
      "loss": 0.8624,
      "step": 500
    },
    {
      "epoch": 125.0,
      "grad_norm": 1.9023737907409668,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.2825,
      "step": 1000
    },
    {
      "epoch": 187.5,
      "grad_norm": 4.60806941986084,
      "learning_rate": 3.125e-05,
      "loss": 0.1628,
      "step": 1500
    },
    {
      "epoch": 250.0,
      "grad_norm": 0.7057669758796692,
      "learning_rate": 2.5e-05,
      "loss": 0.0998,
      "step": 2000
    },
    {
      "epoch": 312.5,
      "grad_norm": 9.289773941040039,
      "learning_rate": 1.8750000000000002e-05,
      "loss": 0.0489,
      "step": 2500
    },
    {
      "epoch": 375.0,
      "grad_norm": 0.0008892796467989683,
      "learning_rate": 1.25e-05,
      "loss": 0.0117,
      "step": 3000
    },
    {
      "epoch": 437.5,
      "grad_norm": 0.008542646653950214,
      "learning_rate": 6.25e-06,
      "loss": 0.001,
      "step": 3500
    },
    {
      "epoch": 500.0,
      "grad_norm": 3.78551339963451e-05,
      "learning_rate": 0.0,
      "loss": 0.0001,
      "step": 4000
    },
    {
      "epoch": 500.0,
      "step": 4000,
      "total_flos": 2.176538025123e+16,
      "train_loss": 0.18364764293190092,
      "train_runtime": 2711.45,
      "train_samples_per_second": 44.626,
      "train_steps_per_second": 1.475
    }
  ],
  "logging_steps": 500,
  "max_steps": 4000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 500,
  "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": 2.176538025123e+16,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}