run-3/checkpoint-16838/trainer_state.json

{
  "best_metric": 0.9013761467889908,
  "best_model_checkpoint": "Tech-oriented/best_model_bert_uncasedbert-base-uncased-finetuned-sst2/run-3/checkpoint-16838",
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 16838,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.25,
      "grad_norm": 295.93414306640625,
      "learning_rate": 9.536827410046812e-06,
      "loss": 0.6121,
      "step": 4209
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.9155740141868591,
      "learning_rate": 8.669955440089204e-06,
      "loss": 0.573,
      "step": 8418
    },
    {
      "epoch": 0.75,
      "grad_norm": 162.3487091064453,
      "learning_rate": 7.803083470131597e-06,
      "loss": 0.5403,
      "step": 12627
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.1670043170452118,
      "learning_rate": 6.936211500173989e-06,
      "loss": 0.4884,
      "step": 16836
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9013761467889908,
      "eval_loss": 0.4918152391910553,
      "eval_runtime": 2.4623,
      "eval_samples_per_second": 354.144,
      "eval_steps_per_second": 22.337,
      "step": 16838
    }
  ],
  "logging_steps": 4209,
  "max_steps": 50514,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 828836824376160.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 1.040369938000442e-05,
    "num_train_epochs": 3,
    "per_device_train_batch_size": 4,
    "seed": 7,
    "weight_decay": 0.2933024618852218
  }
}