{
  "best_metric": 0.8365949119373777,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-12/checkpoint-672",
  "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": 1.2145909070968628,
      "learning_rate": 0.00025674348589389584,
      "loss": 0.5177,
      "step": 96
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.773972602739726,
      "eval_f1": 0.7867036011080332,
      "eval_loss": 0.45424696803092957,
      "eval_precision": 0.7447552447552448,
      "eval_recall": 0.8336594911937377,
      "eval_runtime": 25.7193,
      "eval_samples_per_second": 39.737,
      "eval_steps_per_second": 1.244,
      "step": 96
    },
    {
      "epoch": 2.0,
      "grad_norm": 4.175304412841797,
      "learning_rate": 0.0002282164319056852,
      "loss": 0.4531,
      "step": 192
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7886497064579256,
      "eval_f1": 0.8021978021978021,
      "eval_loss": 0.42793330550193787,
      "eval_precision": 0.7538726333907056,
      "eval_recall": 0.8571428571428571,
      "eval_runtime": 25.2863,
      "eval_samples_per_second": 40.417,
      "eval_steps_per_second": 1.266,
      "step": 192
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.639939069747925,
      "learning_rate": 0.00019968937791747451,
      "loss": 0.4311,
      "step": 288
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.812133072407045,
      "eval_f1": 0.8309859154929577,
      "eval_loss": 0.42055240273475647,
      "eval_precision": 0.7552,
      "eval_recall": 0.923679060665362,
      "eval_runtime": 24.797,
      "eval_samples_per_second": 41.215,
      "eval_steps_per_second": 1.29,
      "step": 288
    },
    {
      "epoch": 4.0,
      "grad_norm": 4.02768087387085,
      "learning_rate": 0.00017116232392926388,
      "loss": 0.408,
      "step": 384
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.821917808219178,
      "eval_f1": 0.8263358778625954,
      "eval_loss": 0.4028782844543457,
      "eval_precision": 0.8063314711359404,
      "eval_recall": 0.8473581213307241,
      "eval_runtime": 25.5911,
      "eval_samples_per_second": 39.936,
      "eval_steps_per_second": 1.25,
      "step": 384
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.4692848920822144,
      "learning_rate": 0.00014263526994105324,
      "loss": 0.397,
      "step": 480
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.821917808219178,
      "eval_f1": 0.8289473684210525,
      "eval_loss": 0.40207499265670776,
      "eval_precision": 0.7974683544303798,
      "eval_recall": 0.863013698630137,
      "eval_runtime": 25.9056,
      "eval_samples_per_second": 39.451,
      "eval_steps_per_second": 1.235,
      "step": 480
    },
    {
      "epoch": 6.0,
      "grad_norm": 2.9322688579559326,
      "learning_rate": 0.0001141082159528426,
      "loss": 0.393,
      "step": 576
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.821917808219178,
      "eval_f1": 0.840630472854641,
      "eval_loss": 0.4070335328578949,
      "eval_precision": 0.7606973058637084,
      "eval_recall": 0.9393346379647749,
      "eval_runtime": 26.253,
      "eval_samples_per_second": 38.929,
      "eval_steps_per_second": 1.219,
      "step": 576
    },
    {
      "epoch": 7.0,
      "grad_norm": 1.5300757884979248,
      "learning_rate": 8.558116196463194e-05,
      "loss": 0.3836,
      "step": 672
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8365949119373777,
      "eval_f1": 0.8512911843276937,
      "eval_loss": 0.3935915529727936,
      "eval_precision": 0.7810457516339869,
      "eval_recall": 0.9354207436399217,
      "eval_runtime": 25.6127,
      "eval_samples_per_second": 39.902,
      "eval_steps_per_second": 1.249,
      "step": 672
    }
  ],
  "logging_steps": 500,
  "max_steps": 960,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 1649866380960.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.7823972938256231,
    "learning_rate": 0.0002852705398821065,
    "num_train_epochs": 10,
    "temperature": 9
  }
}