trainer_state.json

{
  "best_global_step": 20,
  "best_metric": 0.7182247042655945,
  "best_model_checkpoint": "bert_base_km_10_v2_rte/checkpoint-20",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 70,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.4369864463806152,
      "learning_rate": 4.9e-05,
      "loss": 0.7122,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5054151624548736,
      "eval_loss": 0.7219687700271606,
      "eval_runtime": 0.2097,
      "eval_samples_per_second": 1320.698,
      "eval_steps_per_second": 9.536,
      "step": 10
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.8543198108673096,
      "learning_rate": 4.8e-05,
      "loss": 0.6651,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5234657039711191,
      "eval_loss": 0.7182247042655945,
      "eval_runtime": 0.2176,
      "eval_samples_per_second": 1273.124,
      "eval_steps_per_second": 9.192,
      "step": 20
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.227674126625061,
      "learning_rate": 4.7e-05,
      "loss": 0.6283,
      "step": 30
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5523465703971119,
      "eval_loss": 0.7291184663772583,
      "eval_runtime": 0.2093,
      "eval_samples_per_second": 1323.276,
      "eval_steps_per_second": 9.554,
      "step": 30
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.091334342956543,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.558,
      "step": 40
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5126353790613718,
      "eval_loss": 0.7982825636863708,
      "eval_runtime": 0.1961,
      "eval_samples_per_second": 1412.327,
      "eval_steps_per_second": 10.197,
      "step": 40
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.4021227359771729,
      "learning_rate": 4.5e-05,
      "loss": 0.4573,
      "step": 50
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.5342960288808665,
      "eval_loss": 0.8778300881385803,
      "eval_runtime": 0.2255,
      "eval_samples_per_second": 1228.133,
      "eval_steps_per_second": 8.867,
      "step": 50
    },
    {
      "epoch": 6.0,
      "grad_norm": 3.7581570148468018,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.3427,
      "step": 60
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.5054151624548736,
      "eval_loss": 1.0947409868240356,
      "eval_runtime": 0.1945,
      "eval_samples_per_second": 1424.095,
      "eval_steps_per_second": 10.282,
      "step": 60
    },
    {
      "epoch": 7.0,
      "grad_norm": 2.102663993835449,
      "learning_rate": 4.3e-05,
      "loss": 0.2296,
      "step": 70
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.516245487364621,
      "eval_loss": 1.2676050662994385,
      "eval_runtime": 0.1976,
      "eval_samples_per_second": 1401.474,
      "eval_steps_per_second": 10.119,
      "step": 70
    },
    {
      "epoch": 7.0,
      "step": 70,
      "total_flos": 2293012847462400.0,
      "train_loss": 0.5133278846740723,
      "train_runtime": 44.121,
      "train_samples_per_second": 2821.784,
      "train_steps_per_second": 11.332
    }
  ],
  "logging_steps": 1,
  "max_steps": 500,
  "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": 2293012847462400.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}