trainer_state.json

{
  "best_global_step": 30,
  "best_metric": 0.5865323543548584,
  "best_model_checkpoint": "bert_base_rand_5_v1_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": 1.123695969581604,
      "learning_rate": 4.9e-05,
      "loss": 0.6331,
      "step": 15
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7083333333333334,
      "eval_combined_score": 0.7616316744427165,
      "eval_f1": 0.8149300155520995,
      "eval_loss": 0.5937906503677368,
      "eval_runtime": 0.2817,
      "eval_samples_per_second": 1448.604,
      "eval_steps_per_second": 7.101,
      "step": 15
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.7262179851531982,
      "learning_rate": 4.8e-05,
      "loss": 0.5749,
      "step": 30
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7058823529411765,
      "eval_combined_score": 0.7505520979722947,
      "eval_f1": 0.7952218430034129,
      "eval_loss": 0.5865323543548584,
      "eval_runtime": 0.3021,
      "eval_samples_per_second": 1350.431,
      "eval_steps_per_second": 6.62,
      "step": 30
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.8776628971099854,
      "learning_rate": 4.7e-05,
      "loss": 0.4908,
      "step": 45
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6936274509803921,
      "eval_combined_score": 0.7417717086834734,
      "eval_f1": 0.7899159663865546,
      "eval_loss": 0.599459171295166,
      "eval_runtime": 0.2815,
      "eval_samples_per_second": 1449.589,
      "eval_steps_per_second": 7.106,
      "step": 45
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.8931710720062256,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.3591,
      "step": 60
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6666666666666666,
      "eval_combined_score": 0.7114695340501792,
      "eval_f1": 0.7562724014336918,
      "eval_loss": 0.7469572424888611,
      "eval_runtime": 0.2848,
      "eval_samples_per_second": 1432.49,
      "eval_steps_per_second": 7.022,
      "step": 60
    },
    {
      "epoch": 5.0,
      "grad_norm": 4.633615016937256,
      "learning_rate": 4.5e-05,
      "loss": 0.2263,
      "step": 75
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6691176470588235,
      "eval_combined_score": 0.7124973407084352,
      "eval_f1": 0.755877034358047,
      "eval_loss": 0.9707262516021729,
      "eval_runtime": 0.2732,
      "eval_samples_per_second": 1493.151,
      "eval_steps_per_second": 7.319,
      "step": 75
    },
    {
      "epoch": 6.0,
      "grad_norm": 6.974949359893799,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.1315,
      "step": 90
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6519607843137255,
      "eval_combined_score": 0.6954656862745099,
      "eval_f1": 0.7389705882352942,
      "eval_loss": 1.2854925394058228,
      "eval_runtime": 0.2837,
      "eval_samples_per_second": 1438.04,
      "eval_steps_per_second": 7.049,
      "step": 90
    },
    {
      "epoch": 7.0,
      "grad_norm": 9.64598560333252,
      "learning_rate": 4.3e-05,
      "loss": 0.0994,
      "step": 105
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6151960784313726,
      "eval_combined_score": 0.653977686965197,
      "eval_f1": 0.6927592954990215,
      "eval_loss": 1.3981316089630127,
      "eval_runtime": 0.2905,
      "eval_samples_per_second": 1404.535,
      "eval_steps_per_second": 6.885,
      "step": 105
    },
    {
      "epoch": 7.0,
      "step": 105,
      "total_flos": 3377819728711680.0,
      "train_loss": 0.3592911935987927,
      "train_runtime": 58.5352,
      "train_samples_per_second": 3133.16,
      "train_steps_per_second": 12.813
    }
  ],
  "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
}