run-1/checkpoint-800/trainer_state.json

{
  "best_metric": 0.6666666666666666,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-1/checkpoint-160",
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 800,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.6969084143638611,
      "learning_rate": 2.475738136709632e-05,
      "loss": 0.064,
      "step": 160
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6666666666666666,
      "eval_f1": 0.0,
      "eval_loss": 0.04275359958410263,
      "eval_mcc": 0.0,
      "eval_precision": 0.0,
      "eval_recall": 0.0,
      "eval_runtime": 1.9263,
      "eval_samples_per_second": 661.882,
      "eval_steps_per_second": 20.765,
      "step": 160
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.23619167506694794,
      "learning_rate": 1.8568036025322237e-05,
      "loss": 0.0425,
      "step": 320
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6666666666666666,
      "eval_f1": 0.0,
      "eval_loss": 0.038055840879678726,
      "eval_mcc": 0.0,
      "eval_precision": 0.0,
      "eval_recall": 0.0,
      "eval_runtime": 1.8737,
      "eval_samples_per_second": 680.488,
      "eval_steps_per_second": 21.349,
      "step": 320
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.1795702576637268,
      "learning_rate": 1.237869068354816e-05,
      "loss": 0.0382,
      "step": 480
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6666666666666666,
      "eval_f1": 0.009324009324009322,
      "eval_loss": 0.03556740656495094,
      "eval_mcc": 0.01983408759395048,
      "eval_precision": 0.5,
      "eval_recall": 0.004705882352941176,
      "eval_runtime": 1.8538,
      "eval_samples_per_second": 687.768,
      "eval_steps_per_second": 21.577,
      "step": 480
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.33438435196876526,
      "learning_rate": 6.18934534177408e-06,
      "loss": 0.0364,
      "step": 640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6658823529411765,
      "eval_f1": 0.013888888888888888,
      "eval_loss": 0.0341472327709198,
      "eval_mcc": 0.015010886840871084,
      "eval_precision": 0.42857142857142855,
      "eval_recall": 0.007058823529411765,
      "eval_runtime": 1.8431,
      "eval_samples_per_second": 691.757,
      "eval_steps_per_second": 21.702,
      "step": 640
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.3762623369693756,
      "learning_rate": 0.0,
      "loss": 0.0359,
      "step": 800
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6643137254901961,
      "eval_f1": 0.013824884792626727,
      "eval_loss": 0.033982668071985245,
      "eval_mcc": 0.0,
      "eval_precision": 0.3333333333333333,
      "eval_recall": 0.007058823529411765,
      "eval_runtime": 1.8546,
      "eval_samples_per_second": 687.49,
      "eval_steps_per_second": 21.568,
      "step": 800
    }
  ],
  "logging_steps": 500,
  "max_steps": 800,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 1458777189600.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.031053442748151472,
    "learning_rate": 3.09467267088704e-05,
    "num_train_epochs": 5,
    "temperature": 3
  }
}