trainer_state.json

{
  "best_global_step": 20,
  "best_metric": 0.6904613971710205,
  "best_model_checkpoint": "bert_base_km_5_v1_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": 1.099122405052185,
      "learning_rate": 4.9e-05,
      "loss": 0.71,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.4981949458483754,
      "eval_loss": 0.7088927030563354,
      "eval_runtime": 0.1983,
      "eval_samples_per_second": 1396.898,
      "eval_steps_per_second": 10.086,
      "step": 10
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.0518513917922974,
      "learning_rate": 4.8e-05,
      "loss": 0.6574,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5740072202166066,
      "eval_loss": 0.6904613971710205,
      "eval_runtime": 0.1904,
      "eval_samples_per_second": 1454.749,
      "eval_steps_per_second": 10.504,
      "step": 20
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.526757001876831,
      "learning_rate": 4.7e-05,
      "loss": 0.5802,
      "step": 30
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5415162454873647,
      "eval_loss": 0.7469769716262817,
      "eval_runtime": 0.192,
      "eval_samples_per_second": 1442.709,
      "eval_steps_per_second": 10.417,
      "step": 30
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.3605384826660156,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4492,
      "step": 40
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5451263537906137,
      "eval_loss": 0.8843734264373779,
      "eval_runtime": 0.2034,
      "eval_samples_per_second": 1362.047,
      "eval_steps_per_second": 9.834,
      "step": 40
    },
    {
      "epoch": 5.0,
      "grad_norm": 2.213836908340454,
      "learning_rate": 4.5e-05,
      "loss": 0.279,
      "step": 50
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.5776173285198556,
      "eval_loss": 1.0972446203231812,
      "eval_runtime": 0.2037,
      "eval_samples_per_second": 1360.024,
      "eval_steps_per_second": 9.82,
      "step": 50
    },
    {
      "epoch": 6.0,
      "grad_norm": 3.272091865539551,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.1748,
      "step": 60
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.5379061371841155,
      "eval_loss": 1.4186111688613892,
      "eval_runtime": 0.2128,
      "eval_samples_per_second": 1301.641,
      "eval_steps_per_second": 9.398,
      "step": 60
    },
    {
      "epoch": 7.0,
      "grad_norm": 2.371623992919922,
      "learning_rate": 4.3e-05,
      "loss": 0.1045,
      "step": 70
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.48736462093862815,
      "eval_loss": 1.7102928161621094,
      "eval_runtime": 0.2074,
      "eval_samples_per_second": 1335.75,
      "eval_steps_per_second": 9.644,
      "step": 70
    },
    {
      "epoch": 7.0,
      "step": 70,
      "total_flos": 2293012847462400.0,
      "train_loss": 0.4221469061715262,
      "train_runtime": 43.7339,
      "train_samples_per_second": 2846.759,
      "train_steps_per_second": 11.433
    }
  ],
  "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
}