trainer_state.json

{
  "best_global_step": 2844,
  "best_metric": 0.3976808488368988,
  "best_model_checkpoint": "distilbert_rand_10_v2_qqp/checkpoint-2844",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 9954,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 2.127798557281494,
      "learning_rate": 4.9e-05,
      "loss": 0.478,
      "step": 1422
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7874845411822904,
      "eval_combined_score": 0.7256803036089059,
      "eval_f1": 0.6638760660355215,
      "eval_loss": 0.45048028230667114,
      "eval_runtime": 16.8437,
      "eval_samples_per_second": 2400.31,
      "eval_steps_per_second": 9.38,
      "step": 1422
    },
    {
      "epoch": 2.0,
      "grad_norm": 2.559284210205078,
      "learning_rate": 4.8e-05,
      "loss": 0.3705,
      "step": 2844
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8197872866683156,
      "eval_combined_score": 0.7861840535487734,
      "eval_f1": 0.7525808204292311,
      "eval_loss": 0.3976808488368988,
      "eval_runtime": 16.2844,
      "eval_samples_per_second": 2482.742,
      "eval_steps_per_second": 9.703,
      "step": 2844
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.4692444801330566,
      "learning_rate": 4.7e-05,
      "loss": 0.2963,
      "step": 4266
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8209250556517438,
      "eval_combined_score": 0.795585655523257,
      "eval_f1": 0.7702462553947702,
      "eval_loss": 0.40500912070274353,
      "eval_runtime": 16.287,
      "eval_samples_per_second": 2482.343,
      "eval_steps_per_second": 9.701,
      "step": 4266
    },
    {
      "epoch": 4.0,
      "grad_norm": 3.753375768661499,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.2384,
      "step": 5688
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8295819935691319,
      "eval_combined_score": 0.7914380893081981,
      "eval_f1": 0.7532941850472644,
      "eval_loss": 0.4475550353527069,
      "eval_runtime": 16.3502,
      "eval_samples_per_second": 2472.757,
      "eval_steps_per_second": 9.664,
      "step": 5688
    },
    {
      "epoch": 5.0,
      "grad_norm": 4.2039103507995605,
      "learning_rate": 4.5e-05,
      "loss": 0.1931,
      "step": 7110
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8325253524610438,
      "eval_combined_score": 0.7945477741992437,
      "eval_f1": 0.7565701959374438,
      "eval_loss": 0.48683643341064453,
      "eval_runtime": 16.2263,
      "eval_samples_per_second": 2491.633,
      "eval_steps_per_second": 9.737,
      "step": 7110
    },
    {
      "epoch": 6.0,
      "grad_norm": 4.1749186515808105,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.1571,
      "step": 8532
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8350482315112541,
      "eval_combined_score": 0.8055717907598238,
      "eval_f1": 0.7760953500083935,
      "eval_loss": 0.5057656764984131,
      "eval_runtime": 16.351,
      "eval_samples_per_second": 2472.635,
      "eval_steps_per_second": 9.663,
      "step": 8532
    },
    {
      "epoch": 7.0,
      "grad_norm": 4.48608922958374,
      "learning_rate": 4.3e-05,
      "loss": 0.1305,
      "step": 9954
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8364333415780361,
      "eval_combined_score": 0.8068529578793822,
      "eval_f1": 0.7772725741807281,
      "eval_loss": 0.5593815445899963,
      "eval_runtime": 16.2776,
      "eval_samples_per_second": 2483.776,
      "eval_steps_per_second": 9.707,
      "step": 9954
    },
    {
      "epoch": 7.0,
      "step": 9954,
      "total_flos": 1.686920659598684e+17,
      "train_loss": 0.26626481908331057,
      "train_runtime": 2423.2418,
      "train_samples_per_second": 7507.423,
      "train_steps_per_second": 29.341
    }
  ],
  "logging_steps": 1,
  "max_steps": 71100,
  "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": 1.686920659598684e+17,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}