{
  "best_metric": 0.5564356435643565,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-8/checkpoint-2079",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 2079,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.24534238874912262,
      "learning_rate": 0.0002919829387820306,
      "loss": 0.3074,
      "step": 297
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5148514851485149,
      "eval_f1": 0.16382252559726962,
      "eval_loss": 0.30149543285369873,
      "eval_precision": 0.5853658536585366,
      "eval_recall": 0.09523809523809523,
      "eval_runtime": 14.7271,
      "eval_samples_per_second": 34.291,
      "eval_steps_per_second": 1.086,
      "step": 297
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.5045191645622253,
      "learning_rate": 0.00025027109038459764,
      "loss": 0.3009,
      "step": 594
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5168316831683168,
      "eval_f1": 0.1643835616438356,
      "eval_loss": 0.2990603744983673,
      "eval_precision": 0.6,
      "eval_recall": 0.09523809523809523,
      "eval_runtime": 15.7792,
      "eval_samples_per_second": 32.004,
      "eval_steps_per_second": 1.014,
      "step": 594
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.6056041121482849,
      "learning_rate": 0.00020855924198716472,
      "loss": 0.2956,
      "step": 891
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5207920792079208,
      "eval_f1": 0.17687074829931973,
      "eval_loss": 0.2930928170681,
      "eval_precision": 0.6190476190476191,
      "eval_recall": 0.10317460317460317,
      "eval_runtime": 14.6946,
      "eval_samples_per_second": 34.366,
      "eval_steps_per_second": 1.089,
      "step": 891
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.4732127785682678,
      "learning_rate": 0.00016684739358973177,
      "loss": 0.2915,
      "step": 1188
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5207920792079208,
      "eval_f1": 0.16551724137931034,
      "eval_loss": 0.2868767976760864,
      "eval_precision": 0.631578947368421,
      "eval_recall": 0.09523809523809523,
      "eval_runtime": 14.9872,
      "eval_samples_per_second": 33.695,
      "eval_steps_per_second": 1.068,
      "step": 1188
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.4796178936958313,
      "learning_rate": 0.00012513554519229882,
      "loss": 0.2881,
      "step": 1485
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.5188118811881188,
      "eval_f1": 0.15916955017301038,
      "eval_loss": 0.2868916988372803,
      "eval_precision": 0.6216216216216216,
      "eval_recall": 0.09126984126984126,
      "eval_runtime": 14.7705,
      "eval_samples_per_second": 34.19,
      "eval_steps_per_second": 1.083,
      "step": 1485
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.5560742616653442,
      "learning_rate": 8.342369679486588e-05,
      "loss": 0.2859,
      "step": 1782
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.5544554455445545,
      "eval_f1": 0.25742574257425743,
      "eval_loss": 0.2852015793323517,
      "eval_precision": 0.7647058823529411,
      "eval_recall": 0.15476190476190477,
      "eval_runtime": 15.1062,
      "eval_samples_per_second": 33.43,
      "eval_steps_per_second": 1.059,
      "step": 1782
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.5270741581916809,
      "learning_rate": 4.171184839743294e-05,
      "loss": 0.2835,
      "step": 2079
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.5564356435643565,
      "eval_f1": 0.28205128205128205,
      "eval_loss": 0.28437066078186035,
      "eval_precision": 0.7333333333333333,
      "eval_recall": 0.1746031746031746,
      "eval_runtime": 14.6907,
      "eval_samples_per_second": 34.375,
      "eval_steps_per_second": 1.089,
      "step": 2079
    }
  ],
  "logging_steps": 500,
  "max_steps": 2376,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 8,
  "save_steps": 500,
  "total_flos": 5444902981980.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.37248229271478195,
    "learning_rate": 0.00033369478717946353,
    "num_train_epochs": 8,
    "per_device_train_batch_size": 32,
    "temperature": 11
  }
}