{
  "best_global_step": 1230,
  "best_metric": 0.5545818209648132,
  "best_model_checkpoint": "bert_base_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": 3.1364433765411377,
      "learning_rate": 4.9e-05,
      "loss": 0.6587,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.653120995789859,
      "eval_loss": 0.6231846809387207,
      "eval_runtime": 3.4944,
      "eval_samples_per_second": 1563.373,
      "eval_steps_per_second": 6.296,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 3.362776279449463,
      "learning_rate": 4.8e-05,
      "loss": 0.5937,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7095002745744097,
      "eval_loss": 0.5632237792015076,
      "eval_runtime": 3.4737,
      "eval_samples_per_second": 1572.683,
      "eval_steps_per_second": 6.333,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 4.1579790115356445,
      "learning_rate": 4.7e-05,
      "loss": 0.4625,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7230459454512173,
      "eval_loss": 0.5545818209648132,
      "eval_runtime": 3.4852,
      "eval_samples_per_second": 1567.485,
      "eval_steps_per_second": 6.312,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 5.969952583312988,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.3374,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7349441698700347,
      "eval_loss": 0.6059415936470032,
      "eval_runtime": 3.4663,
      "eval_samples_per_second": 1576.053,
      "eval_steps_per_second": 6.347,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 5.252902507781982,
      "learning_rate": 4.5e-05,
      "loss": 0.2281,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.7358594179022515,
      "eval_loss": 0.6952667832374573,
      "eval_runtime": 3.4644,
      "eval_samples_per_second": 1576.908,
      "eval_steps_per_second": 6.35,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 9.05676555633545,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.1522,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.7316492769540546,
      "eval_loss": 0.8388772010803223,
      "eval_runtime": 3.4941,
      "eval_samples_per_second": 1563.513,
      "eval_steps_per_second": 6.296,
      "step": 2460
    },
    {
      "epoch": 7.0,
      "grad_norm": 11.310493469238281,
      "learning_rate": 4.3e-05,
      "loss": 0.113,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.7254255903349808,
      "eval_loss": 0.987945556640625,
      "eval_runtime": 3.4671,
      "eval_samples_per_second": 1575.652,
      "eval_steps_per_second": 6.345,
      "step": 2870
    },
    {
      "epoch": 8.0,
      "grad_norm": 6.990315914154053,
      "learning_rate": 4.2e-05,
      "loss": 0.0849,
      "step": 3280
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.7204832509610104,
      "eval_loss": 1.0977376699447632,
      "eval_runtime": 3.4638,
      "eval_samples_per_second": 1577.179,
      "eval_steps_per_second": 6.351,
      "step": 3280
    },
    {
      "epoch": 8.0,
      "step": 3280,
      "total_flos": 1.1023616508628992e+17,
      "train_loss": 0.3288262250946789,
      "train_runtime": 1406.5934,
      "train_samples_per_second": 3723.286,
      "train_steps_per_second": 14.574
    }
  ],
  "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": 1.1023616508628992e+17,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}