trainer_state.json

{
  "best_global_step": 10,
  "best_metric": 0.6905319690704346,
  "best_model_checkpoint": "tiny_bert_km_10_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": 0.641264796257019,
      "learning_rate": 4.9e-05,
      "loss": 0.6959,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5270758122743683,
      "eval_loss": 0.6905319690704346,
      "eval_runtime": 0.0937,
      "eval_samples_per_second": 2957.736,
      "eval_steps_per_second": 21.355,
      "step": 10
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.6615476608276367,
      "learning_rate": 4.8e-05,
      "loss": 0.6725,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.516245487364621,
      "eval_loss": 0.7031038403511047,
      "eval_runtime": 0.1007,
      "eval_samples_per_second": 2752.057,
      "eval_steps_per_second": 19.87,
      "step": 20
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.6798081398010254,
      "learning_rate": 4.7e-05,
      "loss": 0.6529,
      "step": 30
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.51985559566787,
      "eval_loss": 0.708575427532196,
      "eval_runtime": 0.0965,
      "eval_samples_per_second": 2869.937,
      "eval_steps_per_second": 20.722,
      "step": 30
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.6829408407211304,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6242,
      "step": 40
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5126353790613718,
      "eval_loss": 0.7211790680885315,
      "eval_runtime": 0.1144,
      "eval_samples_per_second": 2420.972,
      "eval_steps_per_second": 17.48,
      "step": 40
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.8751865029335022,
      "learning_rate": 4.5e-05,
      "loss": 0.5834,
      "step": 50
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.49097472924187724,
      "eval_loss": 0.7471743226051331,
      "eval_runtime": 0.0913,
      "eval_samples_per_second": 3033.756,
      "eval_steps_per_second": 21.904,
      "step": 50
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.2714381217956543,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.5316,
      "step": 60
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.5090252707581228,
      "eval_loss": 0.7853801250457764,
      "eval_runtime": 0.0918,
      "eval_samples_per_second": 3018.504,
      "eval_steps_per_second": 21.794,
      "step": 60
    },
    {
      "epoch": 6.0,
      "step": 60,
      "total_flos": 391780256993280.0,
      "train_loss": 0.6267409960428874,
      "train_runtime": 12.5605,
      "train_samples_per_second": 9912.024,
      "train_steps_per_second": 39.807
    }
  ],
  "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
}