run-34/checkpoint-856/trainer_state.json

{
  "best_metric": 0.7463386057410663,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-34/checkpoint-856",
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 856,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 4.8471832275390625,
      "learning_rate": 0.0003084590636066452,
      "loss": 0.4895,
      "step": 214
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7041593438781488,
      "eval_f1": 0.3631778058007566,
      "eval_loss": 0.4574507474899292,
      "eval_mcc": 0.25518418317792835,
      "eval_precision": 0.6428571428571429,
      "eval_recall": 0.2530755711775044,
      "eval_runtime": 3.1432,
      "eval_samples_per_second": 543.077,
      "eval_steps_per_second": 17.18,
      "step": 214
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.93764066696167,
      "learning_rate": 0.0002313442977049839,
      "loss": 0.4474,
      "step": 428
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7334504979496193,
      "eval_f1": 0.4291091593475534,
      "eval_loss": 0.4426566958427429,
      "eval_mcc": 0.3470390546850216,
      "eval_precision": 0.75,
      "eval_recall": 0.30052724077328646,
      "eval_runtime": 3.675,
      "eval_samples_per_second": 464.485,
      "eval_steps_per_second": 14.694,
      "step": 428
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.0922608375549316,
      "learning_rate": 0.0001542295318033226,
      "loss": 0.4323,
      "step": 642
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7223198594024605,
      "eval_f1": 0.35945945945945945,
      "eval_loss": 0.4437848925590515,
      "eval_mcc": 0.3145764348029479,
      "eval_precision": 0.7777777777777778,
      "eval_recall": 0.23374340949033393,
      "eval_runtime": 3.1441,
      "eval_samples_per_second": 542.917,
      "eval_steps_per_second": 17.175,
      "step": 642
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.645482301712036,
      "learning_rate": 7.71147659016613e-05,
      "loss": 0.4256,
      "step": 856
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7463386057410663,
      "eval_f1": 0.5162011173184358,
      "eval_loss": 0.4317722022533417,
      "eval_mcc": 0.3867634513662599,
      "eval_precision": 0.7085889570552147,
      "eval_recall": 0.40597539543058,
      "eval_runtime": 3.6592,
      "eval_samples_per_second": 466.501,
      "eval_steps_per_second": 14.758,
      "step": 856
    }
  ],
  "logging_steps": 500,
  "max_steps": 1070,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 2099102657760.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.7948524745832855,
    "learning_rate": 0.0003855738295083065,
    "num_train_epochs": 5,
    "temperature": 15
  }
}