trainer_state.json

{
  "best_global_step": 820,
  "best_metric": 0.6271847486495972,
  "best_model_checkpoint": "bert_base_km_10_v2_qnli/checkpoint-820",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 2870,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 2.019942283630371,
      "learning_rate": 4.9e-05,
      "loss": 0.6637,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6196229178107268,
      "eval_loss": 0.6429257988929749,
      "eval_runtime": 3.4661,
      "eval_samples_per_second": 1576.116,
      "eval_steps_per_second": 6.347,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.9370707273483276,
      "learning_rate": 4.8e-05,
      "loss": 0.6245,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.643053267435475,
      "eval_loss": 0.6271847486495972,
      "eval_runtime": 3.426,
      "eval_samples_per_second": 1594.579,
      "eval_steps_per_second": 6.422,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.826284646987915,
      "learning_rate": 4.7e-05,
      "loss": 0.565,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6313380926231008,
      "eval_loss": 0.6558482646942139,
      "eval_runtime": 3.4266,
      "eval_samples_per_second": 1594.283,
      "eval_steps_per_second": 6.42,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 3.313979148864746,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4615,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6454329123192385,
      "eval_loss": 0.6744264364242554,
      "eval_runtime": 3.4493,
      "eval_samples_per_second": 1583.79,
      "eval_steps_per_second": 6.378,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 6.450065612792969,
      "learning_rate": 4.5e-05,
      "loss": 0.3397,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.637195680029288,
      "eval_loss": 0.7911816835403442,
      "eval_runtime": 3.4226,
      "eval_samples_per_second": 1596.172,
      "eval_steps_per_second": 6.428,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 6.9949235916137695,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.2395,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6322533406553176,
      "eval_loss": 0.9379204511642456,
      "eval_runtime": 3.428,
      "eval_samples_per_second": 1593.659,
      "eval_steps_per_second": 6.418,
      "step": 2460
    },
    {
      "epoch": 7.0,
      "grad_norm": 7.427356719970703,
      "learning_rate": 4.3e-05,
      "loss": 0.1681,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6366465312099578,
      "eval_loss": 1.1371195316314697,
      "eval_runtime": 3.4407,
      "eval_samples_per_second": 1587.742,
      "eval_steps_per_second": 6.394,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "step": 2870,
      "total_flos": 9.645664445050368e+16,
      "train_loss": 0.43743079315079214,
      "train_runtime": 1216.8802,
      "train_samples_per_second": 4303.752,
      "train_steps_per_second": 16.846
    }
  ],
  "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": 9.645664445050368e+16,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}