trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 12,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5925925925925926,
      "grad_norm": 8.690666198730469,
      "learning_rate": 0.0001666666666666667,
      "loss": 75.8904,
      "step": 2
    },
    {
      "epoch": 1.0,
      "grad_norm": 4.476772308349609,
      "learning_rate": 0.00013333333333333334,
      "loss": 50.6115,
      "step": 4
    },
    {
      "epoch": 1.5925925925925926,
      "grad_norm": 17.084613800048828,
      "learning_rate": 0.0001,
      "loss": 69.5651,
      "step": 6
    },
    {
      "epoch": 2.0,
      "grad_norm": 6.6332879066467285,
      "learning_rate": 6.666666666666667e-05,
      "loss": 45.2883,
      "step": 8
    },
    {
      "epoch": 2.5925925925925926,
      "grad_norm": 12.756552696228027,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 63.1417,
      "step": 10
    },
    {
      "epoch": 3.0,
      "grad_norm": 5.396965026855469,
      "learning_rate": 0.0,
      "loss": 42.7844,
      "step": 12
    },
    {
      "epoch": 3.0,
      "step": 12,
      "total_flos": 46581270698328.0,
      "train_loss": 57.88021469116211,
      "train_runtime": 51.0883,
      "train_samples_per_second": 4.228,
      "train_steps_per_second": 0.235
    }
  ],
  "logging_steps": 2,
  "max_steps": 12,
  "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": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 46581270698328.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}