{
  "best_global_step": 1230,
  "best_metric": 0.5834688544273376,
  "best_model_checkpoint": "tiny_bert_km_5_v1_qnli/checkpoint-1230",
  "epoch": 8.0,
  "eval_steps": 500,
  "global_step": 3280,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.540867567062378,
      "learning_rate": 4.9e-05,
      "loss": 0.6651,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.61870766977851,
      "eval_loss": 0.6433318853378296,
      "eval_runtime": 1.5735,
      "eval_samples_per_second": 3471.792,
      "eval_steps_per_second": 13.981,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.6758782863616943,
      "learning_rate": 4.8e-05,
      "loss": 0.63,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6426871682225883,
      "eval_loss": 0.6245182156562805,
      "eval_runtime": 1.5486,
      "eval_samples_per_second": 3527.65,
      "eval_steps_per_second": 14.206,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.1263701915740967,
      "learning_rate": 4.7e-05,
      "loss": 0.5711,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6994325462200256,
      "eval_loss": 0.5834688544273376,
      "eval_runtime": 1.5561,
      "eval_samples_per_second": 3510.739,
      "eval_steps_per_second": 14.138,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 5.254637241363525,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4752,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7082189273293062,
      "eval_loss": 0.5910441279411316,
      "eval_runtime": 1.5491,
      "eval_samples_per_second": 3526.58,
      "eval_steps_per_second": 14.202,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 5.907124042510986,
      "learning_rate": 4.5e-05,
      "loss": 0.3883,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.7129782170968333,
      "eval_loss": 0.6258730888366699,
      "eval_runtime": 1.5681,
      "eval_samples_per_second": 3483.82,
      "eval_steps_per_second": 14.03,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 6.061448574066162,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.3147,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.7067545304777595,
      "eval_loss": 0.688542902469635,
      "eval_runtime": 1.554,
      "eval_samples_per_second": 3515.533,
      "eval_steps_per_second": 14.157,
      "step": 2460
    },
    {
      "epoch": 7.0,
      "grad_norm": 7.78029727935791,
      "learning_rate": 4.3e-05,
      "loss": 0.2538,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6972359509427055,
      "eval_loss": 0.8067191243171692,
      "eval_runtime": 1.5468,
      "eval_samples_per_second": 3531.879,
      "eval_steps_per_second": 14.223,
      "step": 2870
    },
    {
      "epoch": 8.0,
      "grad_norm": 4.341258525848389,
      "learning_rate": 4.2e-05,
      "loss": 0.2075,
      "step": 3280
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.6996155958264689,
      "eval_loss": 0.8747366070747375,
      "eval_runtime": 1.5509,
      "eval_samples_per_second": 3522.396,
      "eval_steps_per_second": 14.185,
      "step": 3280
    },
    {
      "epoch": 8.0,
      "step": 3280,
      "total_flos": 2.197388993748173e+16,
      "train_loss": 0.43822247574969037,
      "train_runtime": 458.6842,
      "train_samples_per_second": 11417.769,
      "train_steps_per_second": 44.693
    }
  ],
  "logging_steps": 1,
  "max_steps": 20500,
  "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": 2.197388993748173e+16,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}