run-4/checkpoint-240/trainer_state.json

{
  "best_metric": 0.5636007827788649,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-4/checkpoint-48",
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 240,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.5694383382797241,
      "learning_rate": 0.00021480966845398845,
      "loss": 0.2391,
      "step": 48
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5636007827788649,
      "eval_f1": 0.2829581993569132,
      "eval_loss": 0.21417579054832458,
      "eval_precision": 0.7927927927927928,
      "eval_recall": 0.17221135029354206,
      "eval_runtime": 31.6316,
      "eval_samples_per_second": 32.309,
      "eval_steps_per_second": 0.506,
      "step": 48
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.45431241393089294,
      "learning_rate": 0.00016110725134049132,
      "loss": 0.2131,
      "step": 96
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5547945205479452,
      "eval_f1": 0.26016260162601623,
      "eval_loss": 0.20607483386993408,
      "eval_precision": 0.7692307692307693,
      "eval_recall": 0.15655577299412915,
      "eval_runtime": 29.4377,
      "eval_samples_per_second": 34.717,
      "eval_steps_per_second": 0.544,
      "step": 96
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.9347383379936218,
      "learning_rate": 0.00010740483422699422,
      "loss": 0.2102,
      "step": 144
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5606653620352251,
      "eval_f1": 0.27697262479871176,
      "eval_loss": 0.2055068016052246,
      "eval_precision": 0.7818181818181819,
      "eval_recall": 0.16829745596868884,
      "eval_runtime": 29.2328,
      "eval_samples_per_second": 34.961,
      "eval_steps_per_second": 0.547,
      "step": 144
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.28971126675605774,
      "learning_rate": 5.370241711349711e-05,
      "loss": 0.2082,
      "step": 192
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5499021526418787,
      "eval_f1": 0.23841059602649006,
      "eval_loss": 0.20416943728923798,
      "eval_precision": 0.7741935483870968,
      "eval_recall": 0.14090019569471623,
      "eval_runtime": 29.0962,
      "eval_samples_per_second": 35.125,
      "eval_steps_per_second": 0.55,
      "step": 192
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.20556366443634033,
      "learning_rate": 0.0,
      "loss": 0.2071,
      "step": 240
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.5567514677103719,
      "eval_f1": 0.26580226904376014,
      "eval_loss": 0.20456695556640625,
      "eval_precision": 0.7735849056603774,
      "eval_recall": 0.16046966731898238,
      "eval_runtime": 30.0938,
      "eval_samples_per_second": 33.96,
      "eval_steps_per_second": 0.532,
      "step": 240
    }
  ],
  "logging_steps": 500,
  "max_steps": 240,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 1178475986400.0,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.27492253593123683,
    "learning_rate": 0.00026851208556748554,
    "num_train_epochs": 5,
    "temperature": 7
  }
}