trainer_state.json

{
  "best_global_step": 3,
  "best_metric": 0.6978431344032288,
  "best_model_checkpoint": "bert_base_rand_100_v2_wnli/checkpoint-3",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 18,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.7214457392692566,
      "learning_rate": 4.9e-05,
      "loss": 0.7033,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5633802816901409,
      "eval_loss": 0.6978431344032288,
      "eval_runtime": 0.0556,
      "eval_samples_per_second": 1277.796,
      "eval_steps_per_second": 17.997,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.7904929518699646,
      "learning_rate": 4.8e-05,
      "loss": 0.6943,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.39436619718309857,
      "eval_loss": 0.7144037485122681,
      "eval_runtime": 0.0622,
      "eval_samples_per_second": 1140.957,
      "eval_steps_per_second": 16.07,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.7545483708381653,
      "learning_rate": 4.7e-05,
      "loss": 0.6942,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.28169014084507044,
      "eval_loss": 0.7125880718231201,
      "eval_runtime": 0.0553,
      "eval_samples_per_second": 1284.615,
      "eval_steps_per_second": 18.093,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.7479008436203003,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6971,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5352112676056338,
      "eval_loss": 0.713138222694397,
      "eval_runtime": 0.057,
      "eval_samples_per_second": 1245.272,
      "eval_steps_per_second": 17.539,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.3930493593215942,
      "learning_rate": 4.5e-05,
      "loss": 0.693,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.19718309859154928,
      "eval_loss": 0.7353102564811707,
      "eval_runtime": 0.0594,
      "eval_samples_per_second": 1195.62,
      "eval_steps_per_second": 16.84,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.8294757604599,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.693,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.2535211267605634,
      "eval_loss": 0.764964759349823,
      "eval_runtime": 0.058,
      "eval_samples_per_second": 1223.1,
      "eval_steps_per_second": 17.227,
      "step": 18
    },
    {
      "epoch": 6.0,
      "step": 18,
      "total_flos": 501226560460800.0,
      "train_loss": 0.6958277755313449,
      "train_runtime": 18.7722,
      "train_samples_per_second": 1691.332,
      "train_steps_per_second": 7.991
    }
  ],
  "logging_steps": 1,
  "max_steps": 150,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 50,
  "save_steps": 500,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 5,
        "early_stopping_threshold": 0.0
      },
      "attributes": {
        "early_stopping_patience_counter": 5
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 501226560460800.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}