{
  "best_metric": 0.7211350293542075,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-2/checkpoint-4596",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 5362,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 2.795396089553833,
      "learning_rate": 2.0226645656833037e-05,
      "loss": 0.3457,
      "step": 766
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5870841487279843,
      "eval_f1": 0.352760736196319,
      "eval_loss": 0.3146859109401703,
      "eval_precision": 0.8156028368794326,
      "eval_recall": 0.22504892367906065,
      "eval_runtime": 132.813,
      "eval_samples_per_second": 7.695,
      "eval_steps_per_second": 1.928,
      "step": 766
    },
    {
      "epoch": 2.0,
      "grad_norm": 7.475469589233398,
      "learning_rate": 1.6855538047360867e-05,
      "loss": 0.3142,
      "step": 1532
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5978473581213307,
      "eval_f1": 0.3763277693474962,
      "eval_loss": 0.3066306710243225,
      "eval_precision": 0.8378378378378378,
      "eval_recall": 0.24266144814090018,
      "eval_runtime": 132.6826,
      "eval_samples_per_second": 7.703,
      "eval_steps_per_second": 1.929,
      "step": 1532
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.5962014198303223,
      "learning_rate": 1.3484430437888691e-05,
      "loss": 0.3086,
      "step": 2298
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6438356164383562,
      "eval_f1": 0.5172413793103448,
      "eval_loss": 0.30034366250038147,
      "eval_precision": 0.8024691358024691,
      "eval_recall": 0.3816046966731898,
      "eval_runtime": 133.6957,
      "eval_samples_per_second": 7.644,
      "eval_steps_per_second": 1.915,
      "step": 2298
    },
    {
      "epoch": 4.0,
      "grad_norm": 5.574785232543945,
      "learning_rate": 1.0113322828416519e-05,
      "loss": 0.3062,
      "step": 3064
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6663405088062623,
      "eval_f1": 0.5742821473158551,
      "eval_loss": 0.29888635873794556,
      "eval_precision": 0.7931034482758621,
      "eval_recall": 0.4500978473581213,
      "eval_runtime": 131.8312,
      "eval_samples_per_second": 7.752,
      "eval_steps_per_second": 1.942,
      "step": 3064
    },
    {
      "epoch": 5.0,
      "grad_norm": 5.6824212074279785,
      "learning_rate": 6.7422152189443455e-06,
      "loss": 0.3046,
      "step": 3830
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.7084148727984344,
      "eval_f1": 0.6526806526806528,
      "eval_loss": 0.29940325021743774,
      "eval_precision": 0.8069164265129684,
      "eval_recall": 0.547945205479452,
      "eval_runtime": 132.0554,
      "eval_samples_per_second": 7.739,
      "eval_steps_per_second": 1.939,
      "step": 3830
    },
    {
      "epoch": 6.0,
      "grad_norm": 6.931075572967529,
      "learning_rate": 3.3711076094721728e-06,
      "loss": 0.3016,
      "step": 4596
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.7211350293542075,
      "eval_f1": 0.6772366930917327,
      "eval_loss": 0.29910269379615784,
      "eval_precision": 0.803763440860215,
      "eval_recall": 0.5851272015655578,
      "eval_runtime": 132.5392,
      "eval_samples_per_second": 7.711,
      "eval_steps_per_second": 1.932,
      "step": 4596
    },
    {
      "epoch": 7.0,
      "grad_norm": 2.585300922393799,
      "learning_rate": 0.0,
      "loss": 0.3018,
      "step": 5362
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6947162426614482,
      "eval_f1": 0.625,
      "eval_loss": 0.29733312129974365,
      "eval_precision": 0.8099688473520249,
      "eval_recall": 0.5088062622309197,
      "eval_runtime": 133.4769,
      "eval_samples_per_second": 7.657,
      "eval_steps_per_second": 1.918,
      "step": 5362
    }
  ],
  "logging_steps": 500,
  "max_steps": 5362,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "save_steps": 500,
  "total_flos": 1649866380960.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.43980036163769065,
    "learning_rate": 2.359775326630521e-05,
    "num_train_epochs": 7,
    "temperature": 17
  }
}