training_checkpoints/checkpoint-2572/trainer_state.json

{
  "best_metric": 0.7610168782310045,
  "best_model_checkpoint": "tiny-llama-lora-no-grad/checkpoint-1286",
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 2572,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.78,
      "learning_rate": 4.740798341109383e-05,
      "loss": 1.1276,
      "step": 500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8086754453911696,
      "eval_f1_macro": 0.6806109083601769,
      "eval_f1_micro": 0.8086754453911695,
      "eval_f1_weighted": 0.8034368965438116,
      "eval_loss": 0.6704964637756348,
      "eval_macro_fpr": 0.017184987145855975,
      "eval_macro_sensitivity": 0.6852607939619662,
      "eval_macro_specificity": 0.985525532858602,
      "eval_precision": 0.8055191978891799,
      "eval_precision_macro": 0.7052877886582329,
      "eval_recall": 0.8086754453911696,
      "eval_recall_macro": 0.6852607939619662,
      "eval_runtime": 225.0756,
      "eval_samples_per_second": 5.736,
      "eval_steps_per_second": 0.72,
      "eval_weighted_fpr": 0.016618448496265894,
      "eval_weighted_sensitivity": 0.8086754453911696,
      "eval_weighted_specificity": 0.9742075474878604,
      "step": 643
    },
    {
      "epoch": 1.56,
      "learning_rate": 4.481596682218767e-05,
      "loss": 0.503,
      "step": 1000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8164213787761425,
      "eval_f1_macro": 0.7610168782310045,
      "eval_f1_micro": 0.8164213787761425,
      "eval_f1_weighted": 0.8154049040588042,
      "eval_loss": 0.7205804586410522,
      "eval_macro_fpr": 0.01631511703260385,
      "eval_macro_sensitivity": 0.7641451876525827,
      "eval_macro_specificity": 0.9862465922523334,
      "eval_precision": 0.8231306684180838,
      "eval_precision_macro": 0.7746203157296797,
      "eval_recall": 0.8164213787761425,
      "eval_recall_macro": 0.7641451876525827,
      "eval_runtime": 230.0342,
      "eval_samples_per_second": 5.612,
      "eval_steps_per_second": 0.704,
      "eval_weighted_fpr": 0.01580737677582872,
      "eval_weighted_sensitivity": 0.8164213787761425,
      "eval_weighted_specificity": 0.977277505008855,
      "step": 1286
    },
    {
      "epoch": 2.33,
      "learning_rate": 4.22239502332815e-05,
      "loss": 0.3617,
      "step": 1500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8164213787761425,
      "eval_f1_macro": 0.7242203392667825,
      "eval_f1_micro": 0.8164213787761425,
      "eval_f1_weighted": 0.8123838748630571,
      "eval_loss": 0.8818831443786621,
      "eval_macro_fpr": 0.01641791191580845,
      "eval_macro_sensitivity": 0.7170126291644698,
      "eval_macro_specificity": 0.9861058068231509,
      "eval_precision": 0.8137334709054306,
      "eval_precision_macro": 0.7499445969412153,
      "eval_recall": 0.8164213787761425,
      "eval_recall_macro": 0.7170126291644698,
      "eval_runtime": 231.2189,
      "eval_samples_per_second": 5.583,
      "eval_steps_per_second": 0.701,
      "eval_weighted_fpr": 0.01580737677582872,
      "eval_weighted_sensitivity": 0.8164213787761425,
      "eval_weighted_specificity": 0.975165723571122,
      "step": 1929
    },
    {
      "epoch": 3.11,
      "learning_rate": 3.963193364437533e-05,
      "loss": 0.2293,
      "step": 2000
    },
    {
      "epoch": 3.89,
      "learning_rate": 3.7039917055469156e-05,
      "loss": 0.0618,
      "step": 2500
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8086754453911696,
      "eval_f1_macro": 0.740118070206755,
      "eval_f1_micro": 0.8086754453911695,
      "eval_f1_weighted": 0.8074036218868607,
      "eval_loss": 1.1433583498001099,
      "eval_macro_fpr": 0.017252929817186225,
      "eval_macro_sensitivity": 0.729298059210823,
      "eval_macro_specificity": 0.9854230690678942,
      "eval_precision": 0.8106816369031802,
      "eval_precision_macro": 0.7673360060599891,
      "eval_recall": 0.8086754453911696,
      "eval_recall_macro": 0.729298059210823,
      "eval_runtime": 230.3503,
      "eval_samples_per_second": 5.605,
      "eval_steps_per_second": 0.703,
      "eval_weighted_fpr": 0.016618448496265894,
      "eval_weighted_sensitivity": 0.8086754453911696,
      "eval_weighted_specificity": 0.972670590627246,
      "step": 2572
    }
  ],
  "logging_steps": 500,
  "max_steps": 9645,
  "num_train_epochs": 15,
  "save_steps": 500,
  "total_flos": 6.197742432524698e+16,
  "trial_name": null,
  "trial_params": null
}