trainer_state.json

{
  "best_global_step": 50,
  "best_metric": 0.5587511111365233,
  "best_model_checkpoint": "models/electra_lora_cls_globalthr_multirun_200_50_50/seed_44/checkpoint-50",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 50,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_f1_combined": 0.5452116232628655,
      "eval_f1_macro": 0.3379034785090686,
      "eval_f1_micro": 0.6340579710144928,
      "eval_loss": 0.8292089700698853,
      "eval_runtime": 15.7432,
      "eval_samples_per_second": 3.176,
      "eval_steps_per_second": 0.445,
      "step": 25
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.5471842288970947,
      "learning_rate": 2.0357142857142858e-05,
      "loss": 0.8112,
      "step": 50
    },
    {
      "epoch": 2.0,
      "eval_f1_combined": 0.5587511111365233,
      "eval_f1_macro": 0.3246249159096231,
      "eval_f1_micro": 0.6590909090909091,
      "eval_loss": 0.8023490905761719,
      "eval_runtime": 15.6165,
      "eval_samples_per_second": 3.202,
      "eval_steps_per_second": 0.448,
      "step": 50
    }
  ],
  "logging_steps": 50,
  "max_steps": 125,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 2,
        "early_stopping_threshold": 0.0001
      },
      "attributes": {
        "early_stopping_patience_counter": 0
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 53359742976000.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}