trainer_state.json

{
  "best_global_step": 30,
  "best_metric": 0.5933095812797546,
  "best_model_checkpoint": "bert_base_rand_10_v2_mrpc/checkpoint-30",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 105,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.0715363025665283,
      "learning_rate": 4.9e-05,
      "loss": 0.635,
      "step": 15
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7009803921568627,
      "eval_combined_score": 0.7530461066215624,
      "eval_f1": 0.805111821086262,
      "eval_loss": 0.5951478481292725,
      "eval_runtime": 0.2729,
      "eval_samples_per_second": 1495.251,
      "eval_steps_per_second": 7.33,
      "step": 15
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.2956693172454834,
      "learning_rate": 4.8e-05,
      "loss": 0.5764,
      "step": 30
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6985294117647058,
      "eval_combined_score": 0.743047089302042,
      "eval_f1": 0.7875647668393783,
      "eval_loss": 0.5933095812797546,
      "eval_runtime": 0.2869,
      "eval_samples_per_second": 1421.968,
      "eval_steps_per_second": 6.97,
      "step": 30
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.7342514991760254,
      "learning_rate": 4.7e-05,
      "loss": 0.4976,
      "step": 45
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6740196078431373,
      "eval_combined_score": 0.7136331805449453,
      "eval_f1": 0.7532467532467533,
      "eval_loss": 0.6354292035102844,
      "eval_runtime": 0.2922,
      "eval_samples_per_second": 1396.108,
      "eval_steps_per_second": 6.844,
      "step": 45
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.9168171882629395,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.3711,
      "step": 60
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.678921568627451,
      "eval_combined_score": 0.7231197541183436,
      "eval_f1": 0.7673179396092362,
      "eval_loss": 0.6944642663002014,
      "eval_runtime": 0.276,
      "eval_samples_per_second": 1478.386,
      "eval_steps_per_second": 7.247,
      "step": 60
    },
    {
      "epoch": 5.0,
      "grad_norm": 3.2907121181488037,
      "learning_rate": 4.5e-05,
      "loss": 0.24,
      "step": 75
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6568627450980392,
      "eval_combined_score": 0.6953515246402744,
      "eval_f1": 0.7338403041825095,
      "eval_loss": 1.000902771949768,
      "eval_runtime": 0.2956,
      "eval_samples_per_second": 1380.18,
      "eval_steps_per_second": 6.766,
      "step": 75
    },
    {
      "epoch": 6.0,
      "grad_norm": 2.1583664417266846,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.1426,
      "step": 90
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6299019607843137,
      "eval_combined_score": 0.6705915157650058,
      "eval_f1": 0.7112810707456979,
      "eval_loss": 1.191858172416687,
      "eval_runtime": 0.2778,
      "eval_samples_per_second": 1468.479,
      "eval_steps_per_second": 7.198,
      "step": 90
    },
    {
      "epoch": 7.0,
      "grad_norm": 5.813288688659668,
      "learning_rate": 4.3e-05,
      "loss": 0.0893,
      "step": 105
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6617647058823529,
      "eval_combined_score": 0.7067816335167161,
      "eval_f1": 0.7517985611510791,
      "eval_loss": 1.2962353229522705,
      "eval_runtime": 0.2718,
      "eval_samples_per_second": 1501.172,
      "eval_steps_per_second": 7.359,
      "step": 105
    },
    {
      "epoch": 7.0,
      "step": 105,
      "total_flos": 3377819728711680.0,
      "train_loss": 0.36457218329111735,
      "train_runtime": 58.779,
      "train_samples_per_second": 3120.16,
      "train_steps_per_second": 12.76
    }
  ],
  "logging_steps": 1,
  "max_steps": 750,
  "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": 3377819728711680.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}