trainer_state.json

{
  "best_global_step": 10,
  "best_metric": 0.6880781650543213,
  "best_model_checkpoint": "tiny_bert_km_100_v2_rte/checkpoint-10",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 60,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.6009523868560791,
      "learning_rate": 4.9e-05,
      "loss": 0.6985,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5379061371841155,
      "eval_loss": 0.6880781650543213,
      "eval_runtime": 0.0952,
      "eval_samples_per_second": 2909.946,
      "eval_steps_per_second": 21.01,
      "step": 10
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.6875553131103516,
      "learning_rate": 4.8e-05,
      "loss": 0.6784,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5487364620938628,
      "eval_loss": 0.6939728260040283,
      "eval_runtime": 0.099,
      "eval_samples_per_second": 2796.963,
      "eval_steps_per_second": 20.195,
      "step": 20
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.5809323191642761,
      "learning_rate": 4.7e-05,
      "loss": 0.6618,
      "step": 30
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5776173285198556,
      "eval_loss": 0.6917306184768677,
      "eval_runtime": 0.1437,
      "eval_samples_per_second": 1927.961,
      "eval_steps_per_second": 13.92,
      "step": 30
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.6937026381492615,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6379,
      "step": 40
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5451263537906137,
      "eval_loss": 0.6998885869979858,
      "eval_runtime": 0.1194,
      "eval_samples_per_second": 2320.382,
      "eval_steps_per_second": 16.754,
      "step": 40
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.9628274440765381,
      "learning_rate": 4.5e-05,
      "loss": 0.6094,
      "step": 50
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.5487364620938628,
      "eval_loss": 0.7120586037635803,
      "eval_runtime": 0.1001,
      "eval_samples_per_second": 2766.764,
      "eval_steps_per_second": 19.977,
      "step": 50
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.1521601676940918,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.5673,
      "step": 60
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.5306859205776173,
      "eval_loss": 0.7469751238822937,
      "eval_runtime": 0.1,
      "eval_samples_per_second": 2770.868,
      "eval_steps_per_second": 20.006,
      "step": 60
    },
    {
      "epoch": 6.0,
      "step": 60,
      "total_flos": 391780256993280.0,
      "train_loss": 0.6422052065531413,
      "train_runtime": 12.7203,
      "train_samples_per_second": 9787.505,
      "train_steps_per_second": 39.307
    }
  ],
  "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": 391780256993280.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}