{
  "best_metric": 0.541095890410959,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-4/checkpoint-96",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 672,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.20222608745098114,
      "learning_rate": 0.00010108430383313451,
      "loss": 0.1454,
      "step": 96
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.541095890410959,
      "eval_f1": 0.20642978003384094,
      "eval_loss": 0.11932022124528885,
      "eval_precision": 0.7625,
      "eval_recall": 0.11937377690802348,
      "eval_runtime": 30.5181,
      "eval_samples_per_second": 33.488,
      "eval_steps_per_second": 1.049,
      "step": 96
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.3006563186645508,
      "learning_rate": 8.423691986094544e-05,
      "loss": 0.1209,
      "step": 192
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.538160469667319,
      "eval_f1": 0.1945392491467577,
      "eval_loss": 0.11518780887126923,
      "eval_precision": 0.76,
      "eval_recall": 0.11154598825831702,
      "eval_runtime": 30.8486,
      "eval_samples_per_second": 33.13,
      "eval_steps_per_second": 1.037,
      "step": 192
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.2538917064666748,
      "learning_rate": 6.738953588875635e-05,
      "loss": 0.1188,
      "step": 288
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5362035225048923,
      "eval_f1": 0.18835616438356165,
      "eval_loss": 0.11479847133159637,
      "eval_precision": 0.7534246575342466,
      "eval_recall": 0.10763209393346379,
      "eval_runtime": 31.7122,
      "eval_samples_per_second": 32.227,
      "eval_steps_per_second": 1.009,
      "step": 288
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.3130100667476654,
      "learning_rate": 5.0542151916567255e-05,
      "loss": 0.1176,
      "step": 384
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5332681017612525,
      "eval_f1": 0.17900172117039587,
      "eval_loss": 0.11518403142690659,
      "eval_precision": 0.7428571428571429,
      "eval_recall": 0.10176125244618395,
      "eval_runtime": 31.6985,
      "eval_samples_per_second": 32.241,
      "eval_steps_per_second": 1.01,
      "step": 384
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.2996502220630646,
      "learning_rate": 3.369476794437817e-05,
      "loss": 0.1174,
      "step": 480
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.5401174168297456,
      "eval_f1": 0.20068027210884357,
      "eval_loss": 0.11439184099435806,
      "eval_precision": 0.7662337662337663,
      "eval_recall": 0.11545988258317025,
      "eval_runtime": 30.5338,
      "eval_samples_per_second": 33.471,
      "eval_steps_per_second": 1.048,
      "step": 480
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.46337705850601196,
      "learning_rate": 1.6847383972189086e-05,
      "loss": 0.1172,
      "step": 576
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.538160469667319,
      "eval_f1": 0.1945392491467577,
      "eval_loss": 0.11425618827342987,
      "eval_precision": 0.76,
      "eval_recall": 0.11154598825831702,
      "eval_runtime": 31.6339,
      "eval_samples_per_second": 32.307,
      "eval_steps_per_second": 1.012,
      "step": 576
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.3762975037097931,
      "learning_rate": 0.0,
      "loss": 0.1171,
      "step": 672
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.5401174168297456,
      "eval_f1": 0.20068027210884357,
      "eval_loss": 0.11420020461082458,
      "eval_precision": 0.7662337662337663,
      "eval_recall": 0.11545988258317025,
      "eval_runtime": 31.0635,
      "eval_samples_per_second": 32.9,
      "eval_steps_per_second": 1.03,
      "step": 672
    }
  ],
  "logging_steps": 500,
  "max_steps": 672,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "save_steps": 500,
  "total_flos": 1649866380960.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.14091216322626776,
    "learning_rate": 0.00011793168780532361,
    "num_train_epochs": 7,
    "temperature": 26
  }
}