{
  "best_global_step": 4602,
  "best_metric": 0.8094508051872253,
  "best_model_checkpoint": "bert_base_km_100_v1_mnli/checkpoint-4602",
  "epoch": 8.0,
  "eval_steps": 500,
  "global_step": 12272,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.1032648086547852,
      "learning_rate": 4.9e-05,
      "loss": 0.9905,
      "step": 1534
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.567906265919511,
      "eval_loss": 0.9129114151000977,
      "eval_runtime": 6.2045,
      "eval_samples_per_second": 1581.905,
      "eval_steps_per_second": 6.286,
      "step": 1534
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.5337332487106323,
      "learning_rate": 4.8e-05,
      "loss": 0.8746,
      "step": 3068
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6136525725929699,
      "eval_loss": 0.8563531041145325,
      "eval_runtime": 6.1268,
      "eval_samples_per_second": 1601.989,
      "eval_steps_per_second": 6.366,
      "step": 3068
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.8543452024459839,
      "learning_rate": 4.7e-05,
      "loss": 0.788,
      "step": 4602
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6407539480387162,
      "eval_loss": 0.8094508051872253,
      "eval_runtime": 6.1538,
      "eval_samples_per_second": 1594.962,
      "eval_steps_per_second": 6.338,
      "step": 4602
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.3306071758270264,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.7016,
      "step": 6136
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6445236882322976,
      "eval_loss": 0.8233607411384583,
      "eval_runtime": 6.1693,
      "eval_samples_per_second": 1590.93,
      "eval_steps_per_second": 6.322,
      "step": 6136
    },
    {
      "epoch": 5.0,
      "grad_norm": 2.266649007797241,
      "learning_rate": 4.5e-05,
      "loss": 0.6148,
      "step": 7670
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6544065206316861,
      "eval_loss": 0.8329027891159058,
      "eval_runtime": 6.1542,
      "eval_samples_per_second": 1594.848,
      "eval_steps_per_second": 6.337,
      "step": 7670
    },
    {
      "epoch": 6.0,
      "grad_norm": 3.5796010494232178,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.5236,
      "step": 9204
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6468670402445237,
      "eval_loss": 0.9532320499420166,
      "eval_runtime": 6.1401,
      "eval_samples_per_second": 1598.515,
      "eval_steps_per_second": 6.352,
      "step": 9204
    },
    {
      "epoch": 7.0,
      "grad_norm": 3.174539566040039,
      "learning_rate": 4.3e-05,
      "loss": 0.4384,
      "step": 10738
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6430973000509425,
      "eval_loss": 0.9941398501396179,
      "eval_runtime": 6.134,
      "eval_samples_per_second": 1600.094,
      "eval_steps_per_second": 6.358,
      "step": 10738
    },
    {
      "epoch": 8.0,
      "grad_norm": 3.524613618850708,
      "learning_rate": 4.2e-05,
      "loss": 0.3583,
      "step": 12272
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.6404482934284259,
      "eval_loss": 1.1222126483917236,
      "eval_runtime": 6.1541,
      "eval_samples_per_second": 1594.877,
      "eval_steps_per_second": 6.337,
      "step": 12272
    },
    {
      "epoch": 8.0,
      "step": 12272,
      "total_flos": 4.133006614744842e+17,
      "train_loss": 0.6612219829136576,
      "train_runtime": 5094.6925,
      "train_samples_per_second": 3854.03,
      "train_steps_per_second": 15.055
    }
  ],
  "logging_steps": 1,
  "max_steps": 76700,
  "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": 4.133006614744842e+17,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}