trainer_state.json

{
  "best_global_step": 3420,
  "best_metric": 0.7986006996501749,
  "best_model_checkpoint": "evaluation_results/run_20250604_000220/checkpoint-3420",
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 3420,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.7309941520467836,
      "grad_norm": 2.736999988555908,
      "learning_rate": 8.540935672514621e-06,
      "loss": 0.4437,
      "step": 500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8245955355314356,
      "eval_f1": 0.7761662093296746,
      "eval_loss": 0.39194825291633606,
      "eval_precision": 0.8303047246295778,
      "eval_recall": 0.7286555446516192,
      "eval_runtime": 447.7517,
      "eval_samples_per_second": 21.811,
      "eval_steps_per_second": 2.727,
      "step": 684
    },
    {
      "epoch": 1.4619883040935673,
      "grad_norm": 2.7669053077697754,
      "learning_rate": 7.078947368421053e-06,
      "loss": 0.3398,
      "step": 1000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.830534507474913,
      "eval_f1": 0.792995622263915,
      "eval_loss": 0.4003181755542755,
      "eval_precision": 0.8088798162796632,
      "eval_recall": 0.7777232580961727,
      "eval_runtime": 384.6018,
      "eval_samples_per_second": 25.392,
      "eval_steps_per_second": 3.175,
      "step": 1368
    },
    {
      "epoch": 2.192982456140351,
      "grad_norm": 4.7090277671813965,
      "learning_rate": 5.616959064327486e-06,
      "loss": 0.3073,
      "step": 1500
    },
    {
      "epoch": 2.9239766081871346,
      "grad_norm": 2.2827835083007812,
      "learning_rate": 4.154970760233918e-06,
      "loss": 0.2834,
      "step": 2000
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8345279541265616,
      "eval_f1": 0.795959595959596,
      "eval_loss": 0.4070938229560852,
      "eval_precision": 0.8199791883454735,
      "eval_recall": 0.7733071638861629,
      "eval_runtime": 619.2538,
      "eval_samples_per_second": 15.771,
      "eval_steps_per_second": 1.972,
      "step": 2052
    },
    {
      "epoch": 3.654970760233918,
      "grad_norm": 3.448413848876953,
      "learning_rate": 2.692982456140351e-06,
      "loss": 0.251,
      "step": 2500
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8359614990784354,
      "eval_f1": 0.7973690867695421,
      "eval_loss": 0.43020206689834595,
      "eval_precision": 0.822976501305483,
      "eval_recall": 0.7733071638861629,
      "eval_runtime": 773.3763,
      "eval_samples_per_second": 12.628,
      "eval_steps_per_second": 1.579,
      "step": 2736
    },
    {
      "epoch": 4.385964912280702,
      "grad_norm": 2.972419500350952,
      "learning_rate": 1.2309941520467837e-06,
      "loss": 0.2379,
      "step": 3000
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8349375383985255,
      "eval_f1": 0.7986006996501749,
      "eval_loss": 0.45096608996391296,
      "eval_precision": 0.8136456211812627,
      "eval_recall": 0.7841020608439647,
      "eval_runtime": 774.6507,
      "eval_samples_per_second": 12.607,
      "eval_steps_per_second": 1.576,
      "step": 3420
    }
  ],
  "logging_steps": 500,
  "max_steps": 3420,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 3.982111352788608e+16,
  "train_batch_size": 100,
  "trial_name": null,
  "trial_params": null
}