trainer_state.json

{
  "best_global_step": 30,
  "best_metric": 0.5979148149490356,
  "best_model_checkpoint": "distilbert_km_20_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": 1.1862431764602661,
      "learning_rate": 4.9e-05,
      "loss": 0.6209,
      "step": 15
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7083333333333334,
      "eval_combined_score": 0.7619186046511628,
      "eval_f1": 0.8155038759689922,
      "eval_loss": 0.6047530174255371,
      "eval_runtime": 0.1882,
      "eval_samples_per_second": 2167.695,
      "eval_steps_per_second": 10.626,
      "step": 15
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.7641983032226562,
      "learning_rate": 4.8e-05,
      "loss": 0.5735,
      "step": 30
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7009803921568627,
      "eval_combined_score": 0.7563543936092956,
      "eval_f1": 0.8117283950617284,
      "eval_loss": 0.5979148149490356,
      "eval_runtime": 0.1848,
      "eval_samples_per_second": 2207.981,
      "eval_steps_per_second": 10.823,
      "step": 30
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.3815338611602783,
      "learning_rate": 4.7e-05,
      "loss": 0.5215,
      "step": 45
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7107843137254902,
      "eval_combined_score": 0.7649013593167329,
      "eval_f1": 0.8190184049079755,
      "eval_loss": 0.6131736040115356,
      "eval_runtime": 0.1931,
      "eval_samples_per_second": 2113.096,
      "eval_steps_per_second": 10.358,
      "step": 45
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.9419413805007935,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4693,
      "step": 60
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6691176470588235,
      "eval_combined_score": 0.7155112044817926,
      "eval_f1": 0.7619047619047619,
      "eval_loss": 0.6505874991416931,
      "eval_runtime": 0.1845,
      "eval_samples_per_second": 2211.648,
      "eval_steps_per_second": 10.841,
      "step": 60
    },
    {
      "epoch": 5.0,
      "grad_norm": 2.461724281311035,
      "learning_rate": 4.5e-05,
      "loss": 0.3572,
      "step": 75
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6593137254901961,
      "eval_combined_score": 0.7092062561593095,
      "eval_f1": 0.7590987868284229,
      "eval_loss": 0.733345627784729,
      "eval_runtime": 0.1826,
      "eval_samples_per_second": 2233.966,
      "eval_steps_per_second": 10.951,
      "step": 75
    },
    {
      "epoch": 6.0,
      "grad_norm": 2.122968912124634,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.2262,
      "step": 90
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6029411764705882,
      "eval_combined_score": 0.6492149491375497,
      "eval_f1": 0.6954887218045113,
      "eval_loss": 0.9685078263282776,
      "eval_runtime": 0.1825,
      "eval_samples_per_second": 2235.083,
      "eval_steps_per_second": 10.956,
      "step": 90
    },
    {
      "epoch": 7.0,
      "grad_norm": 3.679675579071045,
      "learning_rate": 4.3e-05,
      "loss": 0.1354,
      "step": 105
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6593137254901961,
      "eval_combined_score": 0.7155353520718633,
      "eval_f1": 0.7717569786535303,
      "eval_loss": 1.1398566961288452,
      "eval_runtime": 0.1861,
      "eval_samples_per_second": 2192.381,
      "eval_steps_per_second": 10.747,
      "step": 105
    },
    {
      "epoch": 7.0,
      "step": 105,
      "total_flos": 1700616463945728.0,
      "train_loss": 0.41487605004083544,
      "train_runtime": 34.357,
      "train_samples_per_second": 5338.07,
      "train_steps_per_second": 21.83
    }
  ],
  "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
}