trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 3360,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.44642857142857145,
      "grad_norm": 0.4475659728050232,
      "learning_rate": 4.255952380952381e-05,
      "loss": 0.4396,
      "step": 500
    },
    {
      "epoch": 0.8928571428571429,
      "grad_norm": 0.3704160749912262,
      "learning_rate": 3.511904761904762e-05,
      "loss": 0.4339,
      "step": 1000
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.2521186440677966,
      "eval_loss": 0.4316615164279938,
      "eval_runtime": 8.5382,
      "eval_samples_per_second": 55.281,
      "eval_steps_per_second": 6.91,
      "step": 1120
    },
    {
      "epoch": 1.3392857142857144,
      "grad_norm": 0.3202057182788849,
      "learning_rate": 2.767857142857143e-05,
      "loss": 0.4277,
      "step": 1500
    },
    {
      "epoch": 1.7857142857142856,
      "grad_norm": 0.6411174535751343,
      "learning_rate": 2.023809523809524e-05,
      "loss": 0.4246,
      "step": 2000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.23940677966101695,
      "eval_loss": 0.42827531695365906,
      "eval_runtime": 8.5377,
      "eval_samples_per_second": 55.284,
      "eval_steps_per_second": 6.911,
      "step": 2240
    },
    {
      "epoch": 2.232142857142857,
      "grad_norm": 0.7362692952156067,
      "learning_rate": 1.2797619047619047e-05,
      "loss": 0.4236,
      "step": 2500
    },
    {
      "epoch": 2.678571428571429,
      "grad_norm": 1.011610984802246,
      "learning_rate": 5.357142857142857e-06,
      "loss": 0.418,
      "step": 3000
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.2796610169491525,
      "eval_loss": 0.4260082542896271,
      "eval_runtime": 8.5784,
      "eval_samples_per_second": 55.022,
      "eval_steps_per_second": 6.878,
      "step": 3360
    }
  ],
  "logging_steps": 500,
  "max_steps": 3360,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "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": 7071975903283200.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}