trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.5555555555555554,
  "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": 1.325890302658081,
      "learning_rate": 0.0001666666666666667,
      "loss": 7.803,
      "step": 2
    },
    {
      "epoch": 1.1851851851851851,
      "grad_norm": 0.9218242168426514,
      "learning_rate": 0.00013333333333333334,
      "loss": 7.5071,
      "step": 4
    },
    {
      "epoch": 1.7777777777777777,
      "grad_norm": 0.6321972608566284,
      "learning_rate": 0.0001,
      "loss": 7.303,
      "step": 6
    },
    {
      "epoch": 2.3703703703703702,
      "grad_norm": 1.051314115524292,
      "learning_rate": 6.666666666666667e-05,
      "loss": 7.3425,
      "step": 8
    },
    {
      "epoch": 2.962962962962963,
      "grad_norm": 0.732342541217804,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 7.2905,
      "step": 10
    },
    {
      "epoch": 3.5555555555555554,
      "grad_norm": 0.525382936000824,
      "learning_rate": 0.0,
      "loss": 7.2466,
      "step": 12
    },
    {
      "epoch": 3.5555555555555554,
      "step": 12,
      "total_flos": 56669875934856.0,
      "train_loss": 7.415418068567912,
      "train_runtime": 1065.7803,
      "train_samples_per_second": 0.203,
      "train_steps_per_second": 0.011
    }
  ],
  "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": false,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 56669875934856.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}