trainer_state.json

{
  "best_global_step": 30,
  "best_metric": 0.5967610478401184,
  "best_model_checkpoint": "tiny_bert_km_5_v1_mrpc/checkpoint-30",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 105,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.9092402458190918,
      "learning_rate": 4.9e-05,
      "loss": 0.6303,
      "step": 15
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.696078431372549,
      "eval_combined_score": 0.7529471911463972,
      "eval_f1": 0.8098159509202454,
      "eval_loss": 0.6028932929039001,
      "eval_runtime": 0.1364,
      "eval_samples_per_second": 2990.952,
      "eval_steps_per_second": 14.662,
      "step": 15
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.167948603630066,
      "learning_rate": 4.8e-05,
      "loss": 0.5897,
      "step": 30
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6936274509803921,
      "eval_combined_score": 0.7508075810969549,
      "eval_f1": 0.8079877112135176,
      "eval_loss": 0.5967610478401184,
      "eval_runtime": 0.1412,
      "eval_samples_per_second": 2889.795,
      "eval_steps_per_second": 14.166,
      "step": 30
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.9850865602493286,
      "learning_rate": 4.7e-05,
      "loss": 0.5512,
      "step": 45
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6936274509803921,
      "eval_combined_score": 0.7522600189849011,
      "eval_f1": 0.81089258698941,
      "eval_loss": 0.6064477562904358,
      "eval_runtime": 0.1457,
      "eval_samples_per_second": 2799.386,
      "eval_steps_per_second": 13.722,
      "step": 45
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.3615449666976929,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.5077,
      "step": 60
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6911764705882353,
      "eval_combined_score": 0.7380797028709094,
      "eval_f1": 0.7849829351535836,
      "eval_loss": 0.625641405582428,
      "eval_runtime": 0.133,
      "eval_samples_per_second": 3068.15,
      "eval_steps_per_second": 15.04,
      "step": 60
    },
    {
      "epoch": 5.0,
      "grad_norm": 2.8198459148406982,
      "learning_rate": 4.5e-05,
      "loss": 0.4312,
      "step": 75
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.678921568627451,
      "eval_combined_score": 0.7282553513765438,
      "eval_f1": 0.7775891341256367,
      "eval_loss": 0.6735367178916931,
      "eval_runtime": 0.1322,
      "eval_samples_per_second": 3085.192,
      "eval_steps_per_second": 15.123,
      "step": 75
    },
    {
      "epoch": 6.0,
      "grad_norm": 2.6575522422790527,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.3364,
      "step": 90
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6813725490196079,
      "eval_combined_score": 0.7282295271049597,
      "eval_f1": 0.7750865051903114,
      "eval_loss": 0.7616596817970276,
      "eval_runtime": 0.1312,
      "eval_samples_per_second": 3109.02,
      "eval_steps_per_second": 15.24,
      "step": 90
    },
    {
      "epoch": 7.0,
      "grad_norm": 4.018059253692627,
      "learning_rate": 4.3e-05,
      "loss": 0.2516,
      "step": 105
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6372549019607843,
      "eval_combined_score": 0.6888028895768834,
      "eval_f1": 0.7403508771929824,
      "eval_loss": 0.8654133677482605,
      "eval_runtime": 0.1519,
      "eval_samples_per_second": 2685.121,
      "eval_steps_per_second": 13.162,
      "step": 105
    },
    {
      "epoch": 7.0,
      "step": 105,
      "total_flos": 673316591603712.0,
      "train_loss": 0.4711543469201951,
      "train_runtime": 19.9202,
      "train_samples_per_second": 9206.727,
      "train_steps_per_second": 37.65
    }
  ],
  "logging_steps": 1,
  "max_steps": 750,
  "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": 673316591603712.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}