trainer_state.json

{
  "best_global_step": 6,
  "best_metric": 0.6948724985122681,
  "best_model_checkpoint": "tiny_bert_rand_10_v2_wnli/checkpoint-6",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 21,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.6399606466293335,
      "learning_rate": 4.9e-05,
      "loss": 0.7086,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.43661971830985913,
      "eval_loss": 0.710277259349823,
      "eval_runtime": 0.0333,
      "eval_samples_per_second": 2129.576,
      "eval_steps_per_second": 29.994,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.49827128648757935,
      "learning_rate": 4.8e-05,
      "loss": 0.6975,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5633802816901409,
      "eval_loss": 0.6948724985122681,
      "eval_runtime": 0.0315,
      "eval_samples_per_second": 2252.819,
      "eval_steps_per_second": 31.73,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.573650598526001,
      "learning_rate": 4.7e-05,
      "loss": 0.6997,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.4788732394366197,
      "eval_loss": 0.7018594145774841,
      "eval_runtime": 0.0336,
      "eval_samples_per_second": 2111.097,
      "eval_steps_per_second": 29.734,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.6902194023132324,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.696,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.323943661971831,
      "eval_loss": 0.7099470496177673,
      "eval_runtime": 0.0321,
      "eval_samples_per_second": 2213.221,
      "eval_steps_per_second": 31.172,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.6926400065422058,
      "learning_rate": 4.5e-05,
      "loss": 0.6961,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.352112676056338,
      "eval_loss": 0.7143486142158508,
      "eval_runtime": 0.0319,
      "eval_samples_per_second": 2225.112,
      "eval_steps_per_second": 31.34,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.8389093279838562,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6925,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.352112676056338,
      "eval_loss": 0.7172093987464905,
      "eval_runtime": 0.0311,
      "eval_samples_per_second": 2282.326,
      "eval_steps_per_second": 32.145,
      "step": 18
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.4832870364189148,
      "learning_rate": 4.3e-05,
      "loss": 0.695,
      "step": 21
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.4507042253521127,
      "eval_loss": 0.7171546220779419,
      "eval_runtime": 0.0309,
      "eval_samples_per_second": 2301.072,
      "eval_steps_per_second": 32.409,
      "step": 21
    },
    {
      "epoch": 7.0,
      "step": 21,
      "total_flos": 116563804707840.0,
      "train_loss": 0.6979138510567802,
      "train_runtime": 7.8489,
      "train_samples_per_second": 4045.176,
      "train_steps_per_second": 19.111
    }
  ],
  "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": 116563804707840.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}