trainer_state.json

{
  "best_global_step": 20793,
  "best_metric": 0.9469405010418878,
  "best_model_checkpoint": "./ner_output_tinybert6L/combined_300percent_TinyBERT_General_6L_768D_20250619_074923/checkpoints/checkpoint-20793",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 20793,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.24989179050642044,
      "grad_norm": 0.6586679816246033,
      "learning_rate": 4.5837541480305874e-05,
      "loss": 0.3426,
      "step": 1732
    },
    {
      "epoch": 0.49978358101284087,
      "grad_norm": 0.5292770266532898,
      "learning_rate": 4.1672678305198864e-05,
      "loss": 0.2178,
      "step": 3464
    },
    {
      "epoch": 0.7496753715192613,
      "grad_norm": 1.1701136827468872,
      "learning_rate": 3.751262444091762e-05,
      "loss": 0.1982,
      "step": 5196
    },
    {
      "epoch": 0.9995671620256817,
      "grad_norm": 3.1348302364349365,
      "learning_rate": 3.3347761265810614e-05,
      "loss": 0.1881,
      "step": 6928
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9393524410059049,
      "eval_f1": 0.9379816398725478,
      "eval_loss": 0.17937302589416504,
      "eval_precision": 0.9389823871991911,
      "eval_recall": 0.9393524410059049,
      "eval_runtime": 26.2651,
      "eval_samples_per_second": 745.056,
      "eval_steps_per_second": 46.602,
      "step": 6931
    },
    {
      "epoch": 1.2494589525321023,
      "grad_norm": 3.0484063625335693,
      "learning_rate": 2.9182898090703604e-05,
      "loss": 0.1595,
      "step": 8660
    },
    {
      "epoch": 1.4993507430385224,
      "grad_norm": 0.3237079679965973,
      "learning_rate": 2.5020439571009476e-05,
      "loss": 0.1524,
      "step": 10392
    },
    {
      "epoch": 1.749242533544943,
      "grad_norm": 1.4062304496765137,
      "learning_rate": 2.0857981051315347e-05,
      "loss": 0.1491,
      "step": 12124
    },
    {
      "epoch": 1.9991343240513635,
      "grad_norm": 3.2153682708740234,
      "learning_rate": 1.6693117876208337e-05,
      "loss": 0.1402,
      "step": 13856
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9458694225866252,
      "eval_f1": 0.9456418383157583,
      "eval_loss": 0.15713337063789368,
      "eval_precision": 0.9463121752269362,
      "eval_recall": 0.9458694225866252,
      "eval_runtime": 27.0546,
      "eval_samples_per_second": 723.316,
      "eval_steps_per_second": 45.242,
      "step": 13862
    },
    {
      "epoch": 2.249026114557784,
      "grad_norm": 0.006557302549481392,
      "learning_rate": 1.2528254701101333e-05,
      "loss": 0.115,
      "step": 15588
    },
    {
      "epoch": 2.4989179050642045,
      "grad_norm": 1.2286018133163452,
      "learning_rate": 8.365796181407206e-06,
      "loss": 0.1167,
      "step": 17320
    },
    {
      "epoch": 2.7488096955706247,
      "grad_norm": 0.032648004591464996,
      "learning_rate": 4.200933006300197e-06,
      "loss": 0.1112,
      "step": 19052
    },
    {
      "epoch": 2.998701486077045,
      "grad_norm": 4.031618118286133,
      "learning_rate": 3.847448660606935e-08,
      "loss": 0.1071,
      "step": 20784
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9477408531081158,
      "eval_f1": 0.9469405010418878,
      "eval_loss": 0.15903286635875702,
      "eval_precision": 0.946831252072933,
      "eval_recall": 0.9477408531081158,
      "eval_runtime": 26.9292,
      "eval_samples_per_second": 726.682,
      "eval_steps_per_second": 45.452,
      "step": 20793
    }
  ],
  "logging_steps": 1732,
  "max_steps": 20793,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 3,
        "early_stopping_threshold": 0.001
      },
      "attributes": {
        "early_stopping_patience_counter": 0
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.0869727318769664e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}