trainer_state.json

{
  "best_global_step": 10,
  "best_metric": 0.6871192455291748,
  "best_model_checkpoint": "tiny_bert_km_10_v1_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": 0.531049370765686,
      "learning_rate": 4.9e-05,
      "loss": 0.7014,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5306859205776173,
      "eval_loss": 0.6871192455291748,
      "eval_runtime": 0.1032,
      "eval_samples_per_second": 2684.035,
      "eval_steps_per_second": 19.379,
      "step": 10
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.696516752243042,
      "learning_rate": 4.8e-05,
      "loss": 0.6871,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.48375451263537905,
      "eval_loss": 0.6926754117012024,
      "eval_runtime": 0.0913,
      "eval_samples_per_second": 3034.881,
      "eval_steps_per_second": 21.912,
      "step": 20
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.6348766684532166,
      "learning_rate": 4.7e-05,
      "loss": 0.6715,
      "step": 30
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5487364620938628,
      "eval_loss": 0.687958300113678,
      "eval_runtime": 0.1003,
      "eval_samples_per_second": 2761.444,
      "eval_steps_per_second": 19.938,
      "step": 30
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.6822134256362915,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.654,
      "step": 40
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5306859205776173,
      "eval_loss": 0.6973008513450623,
      "eval_runtime": 0.104,
      "eval_samples_per_second": 2664.241,
      "eval_steps_per_second": 19.236,
      "step": 40
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.0210248231887817,
      "learning_rate": 4.5e-05,
      "loss": 0.6228,
      "step": 50
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.5487364620938628,
      "eval_loss": 0.7094991207122803,
      "eval_runtime": 0.0968,
      "eval_samples_per_second": 2862.28,
      "eval_steps_per_second": 20.666,
      "step": 50
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.0849286317825317,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.5753,
      "step": 60
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.5415162454873647,
      "eval_loss": 0.7562154531478882,
      "eval_runtime": 0.1017,
      "eval_samples_per_second": 2724.475,
      "eval_steps_per_second": 19.671,
      "step": 60
    },
    {
      "epoch": 6.0,
      "step": 60,
      "total_flos": 391780256993280.0,
      "train_loss": 0.6520156860351562,
      "train_runtime": 12.9147,
      "train_samples_per_second": 9640.154,
      "train_steps_per_second": 38.715
    }
  ],
  "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
}