trainer_state.json

{
  "best_global_step": 30,
  "best_metric": 0.607943594455719,
  "best_model_checkpoint": "distilbert_km_20_v2_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": 1.3289546966552734,
      "learning_rate": 4.9e-05,
      "loss": 0.6244,
      "step": 15
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7034313725490197,
      "eval_combined_score": 0.7584953472914591,
      "eval_f1": 0.8135593220338984,
      "eval_loss": 0.6129724979400635,
      "eval_runtime": 0.1803,
      "eval_samples_per_second": 2262.497,
      "eval_steps_per_second": 11.091,
      "step": 15
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.6791658401489258,
      "learning_rate": 4.8e-05,
      "loss": 0.5689,
      "step": 30
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6813725490196079,
      "eval_combined_score": 0.740686274509804,
      "eval_f1": 0.8,
      "eval_loss": 0.607943594455719,
      "eval_runtime": 0.1831,
      "eval_samples_per_second": 2227.87,
      "eval_steps_per_second": 10.921,
      "step": 30
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.4310699701309204,
      "learning_rate": 4.7e-05,
      "loss": 0.5055,
      "step": 45
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7107843137254902,
      "eval_combined_score": 0.7637772500304469,
      "eval_f1": 0.8167701863354038,
      "eval_loss": 0.6476616859436035,
      "eval_runtime": 0.1835,
      "eval_samples_per_second": 2223.251,
      "eval_steps_per_second": 10.898,
      "step": 45
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.9108912944793701,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4324,
      "step": 60
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6691176470588235,
      "eval_combined_score": 0.7175744214496891,
      "eval_f1": 0.7660311958405546,
      "eval_loss": 0.6749581694602966,
      "eval_runtime": 0.1871,
      "eval_samples_per_second": 2181.206,
      "eval_steps_per_second": 10.692,
      "step": 60
    },
    {
      "epoch": 5.0,
      "grad_norm": 2.127835512161255,
      "learning_rate": 4.5e-05,
      "loss": 0.3194,
      "step": 75
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6666666666666666,
      "eval_combined_score": 0.7148664343786295,
      "eval_f1": 0.7630662020905923,
      "eval_loss": 0.7786697149276733,
      "eval_runtime": 0.1998,
      "eval_samples_per_second": 2041.991,
      "eval_steps_per_second": 10.01,
      "step": 75
    },
    {
      "epoch": 6.0,
      "grad_norm": 2.1237330436706543,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.1929,
      "step": 90
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6691176470588235,
      "eval_combined_score": 0.7133739043193579,
      "eval_f1": 0.7576301615798923,
      "eval_loss": 1.000406265258789,
      "eval_runtime": 0.18,
      "eval_samples_per_second": 2266.474,
      "eval_steps_per_second": 11.11,
      "step": 90
    },
    {
      "epoch": 7.0,
      "grad_norm": 2.40433931350708,
      "learning_rate": 4.3e-05,
      "loss": 0.1087,
      "step": 105
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6274509803921569,
      "eval_combined_score": 0.6740196078431373,
      "eval_f1": 0.7205882352941176,
      "eval_loss": 1.2845765352249146,
      "eval_runtime": 0.1901,
      "eval_samples_per_second": 2146.134,
      "eval_steps_per_second": 10.52,
      "step": 105
    },
    {
      "epoch": 7.0,
      "step": 105,
      "total_flos": 1700616463945728.0,
      "train_loss": 0.39318862301962715,
      "train_runtime": 34.179,
      "train_samples_per_second": 5365.876,
      "train_steps_per_second": 21.943
    }
  ],
  "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": 1700616463945728.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}