{
  "best_metric": 0.821917808219178,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-20/checkpoint-3830",
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 3830,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 2.1449050903320312,
      "learning_rate": 0.00010427761773013143,
      "loss": 0.5295,
      "step": 766
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8043052837573386,
      "eval_f1": 0.82174688057041,
      "eval_loss": 0.44088178873062134,
      "eval_precision": 0.7545008183306056,
      "eval_recall": 0.9021526418786693,
      "eval_runtime": 133.3851,
      "eval_samples_per_second": 7.662,
      "eval_steps_per_second": 1.919,
      "step": 766
    },
    {
      "epoch": 2.0,
      "grad_norm": 32.69517517089844,
      "learning_rate": 7.820821329759857e-05,
      "loss": 0.4623,
      "step": 1532
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.815068493150685,
      "eval_f1": 0.832,
      "eval_loss": 0.4221538305282593,
      "eval_precision": 0.762214983713355,
      "eval_recall": 0.9158512720156555,
      "eval_runtime": 131.8748,
      "eval_samples_per_second": 7.75,
      "eval_steps_per_second": 1.941,
      "step": 1532
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.419471263885498,
      "learning_rate": 5.2138808865065714e-05,
      "loss": 0.4443,
      "step": 2298
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8180039138943248,
      "eval_f1": 0.825515947467167,
      "eval_loss": 0.4691775441169739,
      "eval_precision": 0.7927927927927928,
      "eval_recall": 0.8610567514677103,
      "eval_runtime": 133.0186,
      "eval_samples_per_second": 7.683,
      "eval_steps_per_second": 1.925,
      "step": 2298
    },
    {
      "epoch": 4.0,
      "grad_norm": 36.411842346191406,
      "learning_rate": 2.6069404432532857e-05,
      "loss": 0.4368,
      "step": 3064
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8189823874755382,
      "eval_f1": 0.8243114909781577,
      "eval_loss": 0.45713627338409424,
      "eval_precision": 0.8007380073800738,
      "eval_recall": 0.8493150684931506,
      "eval_runtime": 133.4653,
      "eval_samples_per_second": 7.657,
      "eval_steps_per_second": 1.918,
      "step": 3064
    },
    {
      "epoch": 5.0,
      "grad_norm": 17.702917098999023,
      "learning_rate": 0.0,
      "loss": 0.4211,
      "step": 3830
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.821917808219178,
      "eval_f1": 0.8357400722021661,
      "eval_loss": 0.4328462481498718,
      "eval_precision": 0.7755443886097152,
      "eval_recall": 0.9060665362035225,
      "eval_runtime": 132.5362,
      "eval_samples_per_second": 7.711,
      "eval_steps_per_second": 1.932,
      "step": 3830
    }
  ],
  "logging_steps": 500,
  "max_steps": 3830,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 1178475986400.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.8705195866732912,
    "learning_rate": 0.00013034702216266429,
    "num_train_epochs": 5,
    "temperature": 7
  }
}