{
  "best_metric": 0.8297455968688845,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-17/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": 3.036878824234009,
      "learning_rate": 6.985166134218099e-05,
      "loss": 0.5738,
      "step": 766
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7876712328767124,
      "eval_f1": 0.8121212121212121,
      "eval_loss": 0.4637316167354584,
      "eval_precision": 0.7282608695652174,
      "eval_recall": 0.9178082191780822,
      "eval_runtime": 132.0721,
      "eval_samples_per_second": 7.738,
      "eval_steps_per_second": 1.938,
      "step": 766
    },
    {
      "epoch": 2.0,
      "grad_norm": 31.187244415283203,
      "learning_rate": 5.8209717785150834e-05,
      "loss": 0.5056,
      "step": 1532
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8043052837573386,
      "eval_f1": 0.82174688057041,
      "eval_loss": 0.4566289782524109,
      "eval_precision": 0.7545008183306056,
      "eval_recall": 0.9021526418786693,
      "eval_runtime": 132.5437,
      "eval_samples_per_second": 7.711,
      "eval_steps_per_second": 1.931,
      "step": 1532
    },
    {
      "epoch": 3.0,
      "grad_norm": 7.650017261505127,
      "learning_rate": 4.6567774228120664e-05,
      "loss": 0.5018,
      "step": 2298
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8189823874755382,
      "eval_f1": 0.829806807727691,
      "eval_loss": 0.5142984986305237,
      "eval_precision": 0.7829861111111112,
      "eval_recall": 0.8825831702544031,
      "eval_runtime": 132.4377,
      "eval_samples_per_second": 7.717,
      "eval_steps_per_second": 1.933,
      "step": 2298
    },
    {
      "epoch": 4.0,
      "grad_norm": 7.019213676452637,
      "learning_rate": 3.4925830671090495e-05,
      "loss": 0.4961,
      "step": 3064
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8238747553816047,
      "eval_f1": 0.832089552238806,
      "eval_loss": 0.5214771032333374,
      "eval_precision": 0.7950089126559715,
      "eval_recall": 0.87279843444227,
      "eval_runtime": 132.1793,
      "eval_samples_per_second": 7.732,
      "eval_steps_per_second": 1.937,
      "step": 3064
    },
    {
      "epoch": 5.0,
      "grad_norm": 63.91996383666992,
      "learning_rate": 2.3283887114060332e-05,
      "loss": 0.4823,
      "step": 3830
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8238747553816047,
      "eval_f1": 0.8366606170598911,
      "eval_loss": 0.49781981110572815,
      "eval_precision": 0.7800338409475466,
      "eval_recall": 0.9021526418786693,
      "eval_runtime": 132.1118,
      "eval_samples_per_second": 7.736,
      "eval_steps_per_second": 1.938,
      "step": 3830
    },
    {
      "epoch": 6.0,
      "grad_norm": 7.790257930755615,
      "learning_rate": 1.1641943557030166e-05,
      "loss": 0.4627,
      "step": 4596
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8297455968688845,
      "eval_f1": 0.8429602888086642,
      "eval_loss": 0.5300474762916565,
      "eval_precision": 0.7822445561139029,
      "eval_recall": 0.913894324853229,
      "eval_runtime": 132.3089,
      "eval_samples_per_second": 7.724,
      "eval_steps_per_second": 1.935,
      "step": 4596
    },
    {
      "epoch": 7.0,
      "grad_norm": 58.294677734375,
      "learning_rate": 0.0,
      "loss": 0.4654,
      "step": 5362
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8258317025440313,
      "eval_f1": 0.8378870673952642,
      "eval_loss": 0.5393737554550171,
      "eval_precision": 0.7836456558773425,
      "eval_recall": 0.9001956947162426,
      "eval_runtime": 132.364,
      "eval_samples_per_second": 7.721,
      "eval_steps_per_second": 1.934,
      "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": 0.9998794266799103,
    "learning_rate": 8.149360489921117e-05,
    "num_train_epochs": 7,
    "temperature": 5
  }
}