trainer_state.json

{
  "best_global_step": 20,
  "best_metric": 0.6972938179969788,
  "best_model_checkpoint": "tiny_bert_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": 0.6201291084289551,
      "learning_rate": 4.9e-05,
      "loss": 0.6988,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.48736462093862815,
      "eval_loss": 0.7008051872253418,
      "eval_runtime": 0.098,
      "eval_samples_per_second": 2825.189,
      "eval_steps_per_second": 20.398,
      "step": 10
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.6418693661689758,
      "learning_rate": 4.8e-05,
      "loss": 0.6769,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.4548736462093863,
      "eval_loss": 0.6972938179969788,
      "eval_runtime": 0.0969,
      "eval_samples_per_second": 2858.139,
      "eval_steps_per_second": 20.636,
      "step": 20
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.6856212615966797,
      "learning_rate": 4.7e-05,
      "loss": 0.6577,
      "step": 30
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.49097472924187724,
      "eval_loss": 0.7094991207122803,
      "eval_runtime": 0.098,
      "eval_samples_per_second": 2826.797,
      "eval_steps_per_second": 20.41,
      "step": 30
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.6920773983001709,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6243,
      "step": 40
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.48375451263537905,
      "eval_loss": 0.7321539521217346,
      "eval_runtime": 0.108,
      "eval_samples_per_second": 2565.607,
      "eval_steps_per_second": 18.524,
      "step": 40
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.8982360363006592,
      "learning_rate": 4.5e-05,
      "loss": 0.5811,
      "step": 50
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.48014440433212996,
      "eval_loss": 0.7696899175643921,
      "eval_runtime": 0.1156,
      "eval_samples_per_second": 2395.544,
      "eval_steps_per_second": 17.296,
      "step": 50
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.2707350254058838,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.5124,
      "step": 60
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.4657039711191336,
      "eval_loss": 0.8588778972625732,
      "eval_runtime": 0.0964,
      "eval_samples_per_second": 2873.095,
      "eval_steps_per_second": 20.744,
      "step": 60
    },
    {
      "epoch": 7.0,
      "grad_norm": 1.3460285663604736,
      "learning_rate": 4.3e-05,
      "loss": 0.4223,
      "step": 70
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.4981949458483754,
      "eval_loss": 0.974549412727356,
      "eval_runtime": 0.0906,
      "eval_samples_per_second": 3056.558,
      "eval_steps_per_second": 22.069,
      "step": 70
    },
    {
      "epoch": 7.0,
      "step": 70,
      "total_flos": 457076966492160.0,
      "train_loss": 0.5962256976536342,
      "train_runtime": 15.1911,
      "train_samples_per_second": 8195.565,
      "train_steps_per_second": 32.914
    }
  ],
  "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": 457076966492160.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}