{
  "best_global_step": 1230,
  "best_metric": 0.636319100856781,
  "best_model_checkpoint": "tiny_bert_km_10_v2_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": 1.503872036933899,
      "learning_rate": 4.9e-05,
      "loss": 0.6657,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6165110745011898,
      "eval_loss": 0.6487873196601868,
      "eval_runtime": 1.5967,
      "eval_samples_per_second": 3421.435,
      "eval_steps_per_second": 13.778,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.6381971836090088,
      "learning_rate": 4.8e-05,
      "loss": 0.6357,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6258466044298004,
      "eval_loss": 0.6424305438995361,
      "eval_runtime": 1.5551,
      "eval_samples_per_second": 3512.877,
      "eval_steps_per_second": 14.147,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.8252344131469727,
      "learning_rate": 4.7e-05,
      "loss": 0.5908,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6320702910488742,
      "eval_loss": 0.636319100856781,
      "eval_runtime": 1.5823,
      "eval_samples_per_second": 3452.462,
      "eval_steps_per_second": 13.903,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 3.978731393814087,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.5165,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6267618524620172,
      "eval_loss": 0.6967955827713013,
      "eval_runtime": 1.5604,
      "eval_samples_per_second": 3500.944,
      "eval_steps_per_second": 14.099,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 3.8373589515686035,
      "learning_rate": 4.5e-05,
      "loss": 0.4376,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.624016108365367,
      "eval_loss": 0.7547832727432251,
      "eval_runtime": 1.5541,
      "eval_samples_per_second": 3515.259,
      "eval_steps_per_second": 14.156,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 7.083200931549072,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.3658,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6198059674171701,
      "eval_loss": 0.8519981503486633,
      "eval_runtime": 1.5897,
      "eval_samples_per_second": 3436.445,
      "eval_steps_per_second": 13.839,
      "step": 2460
    },
    {
      "epoch": 7.0,
      "grad_norm": 7.9356231689453125,
      "learning_rate": 4.3e-05,
      "loss": 0.3003,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6168771737140765,
      "eval_loss": 1.0204379558563232,
      "eval_runtime": 1.5859,
      "eval_samples_per_second": 3444.814,
      "eval_steps_per_second": 13.873,
      "step": 2870
    },
    {
      "epoch": 8.0,
      "grad_norm": 7.396778106689453,
      "learning_rate": 4.2e-05,
      "loss": 0.2448,
      "step": 3280
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.6241991579718104,
      "eval_loss": 1.104865312576294,
      "eval_runtime": 1.5791,
      "eval_samples_per_second": 3459.575,
      "eval_steps_per_second": 13.932,
      "step": 3280
    },
    {
      "epoch": 8.0,
      "step": 3280,
      "total_flos": 2.197388993748173e+16,
      "train_loss": 0.46965792120956795,
      "train_runtime": 454.0814,
      "train_samples_per_second": 11533.504,
      "train_steps_per_second": 45.146
    }
  ],
  "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
}