trainer_state.json

{
  "best_global_step": 2844,
  "best_metric": 0.39146092534065247,
  "best_model_checkpoint": "distilbert_rand_5_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": 1.8924601078033447,
      "learning_rate": 4.9e-05,
      "loss": 0.477,
      "step": 1422
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7876082117239673,
      "eval_combined_score": 0.7252266195060215,
      "eval_f1": 0.6628450272880757,
      "eval_loss": 0.44889965653419495,
      "eval_runtime": 16.4962,
      "eval_samples_per_second": 2450.863,
      "eval_steps_per_second": 9.578,
      "step": 1422
    },
    {
      "epoch": 2.0,
      "grad_norm": 2.7057721614837646,
      "learning_rate": 4.8e-05,
      "loss": 0.3698,
      "step": 2844
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.820751916893396,
      "eval_combined_score": 0.7880224213877751,
      "eval_f1": 0.7552929258821544,
      "eval_loss": 0.39146092534065247,
      "eval_runtime": 16.416,
      "eval_samples_per_second": 2462.836,
      "eval_steps_per_second": 9.625,
      "step": 2844
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.0676562786102295,
      "learning_rate": 4.7e-05,
      "loss": 0.2968,
      "step": 4266
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8283700222606974,
      "eval_combined_score": 0.7970948015525416,
      "eval_f1": 0.765819580844386,
      "eval_loss": 0.3985094130039215,
      "eval_runtime": 16.2325,
      "eval_samples_per_second": 2490.678,
      "eval_steps_per_second": 9.734,
      "step": 4266
    },
    {
      "epoch": 4.0,
      "grad_norm": 3.1476473808288574,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.2383,
      "step": 5688
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8310907741775909,
      "eval_combined_score": 0.7958349779452972,
      "eval_f1": 0.7605791817130035,
      "eval_loss": 0.42779186367988586,
      "eval_runtime": 16.3555,
      "eval_samples_per_second": 2471.947,
      "eval_steps_per_second": 9.66,
      "step": 5688
    },
    {
      "epoch": 5.0,
      "grad_norm": 4.090065956115723,
      "learning_rate": 4.5e-05,
      "loss": 0.1921,
      "step": 7110
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8350976997279248,
      "eval_combined_score": 0.8002092743773208,
      "eval_f1": 0.7653208490267169,
      "eval_loss": 0.4705563485622406,
      "eval_runtime": 16.3723,
      "eval_samples_per_second": 2469.42,
      "eval_steps_per_second": 9.65,
      "step": 7110
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.9940907955169678,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.1577,
      "step": 8532
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8349492950779125,
      "eval_combined_score": 0.8000460057200705,
      "eval_f1": 0.7651427163622285,
      "eval_loss": 0.5084697604179382,
      "eval_runtime": 16.3199,
      "eval_samples_per_second": 2477.342,
      "eval_steps_per_second": 9.681,
      "step": 8532
    },
    {
      "epoch": 7.0,
      "grad_norm": 7.353980541229248,
      "learning_rate": 4.3e-05,
      "loss": 0.1306,
      "step": 9954
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8374969082364581,
      "eval_combined_score": 0.8091095782087436,
      "eval_f1": 0.7807222481810293,
      "eval_loss": 0.5504053235054016,
      "eval_runtime": 16.2951,
      "eval_samples_per_second": 2481.107,
      "eval_steps_per_second": 9.696,
      "step": 9954
    },
    {
      "epoch": 7.0,
      "step": 9954,
      "total_flos": 1.686920659598684e+17,
      "train_loss": 0.26605301102601264,
      "train_runtime": 2423.0976,
      "train_samples_per_second": 7507.869,
      "train_steps_per_second": 29.343
    }
  ],
  "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
}