run-0/checkpoint-144/trainer_state.json

{
  "best_metric": 0.5,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-0/checkpoint-48",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 144,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.7277214527130127,
      "learning_rate": 7.1080634981543736e-06,
      "loss": 0.4421,
      "step": 48
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5,
      "eval_f1": 0.0,
      "eval_loss": 0.4095905125141144,
      "eval_precision": 0.0,
      "eval_recall": 0.0,
      "eval_runtime": 28.1077,
      "eval_samples_per_second": 36.36,
      "eval_steps_per_second": 0.569,
      "step": 48
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.3527555465698242,
      "learning_rate": 3.5540317490771868e-06,
      "loss": 0.4039,
      "step": 96
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5,
      "eval_f1": 0.0,
      "eval_loss": 0.39591261744499207,
      "eval_precision": 0.0,
      "eval_recall": 0.0,
      "eval_runtime": 27.762,
      "eval_samples_per_second": 36.813,
      "eval_steps_per_second": 0.576,
      "step": 96
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.4053336381912231,
      "learning_rate": 0.0,
      "loss": 0.3971,
      "step": 144
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5,
      "eval_f1": 0.0,
      "eval_loss": 0.3941744863986969,
      "eval_precision": 0.0,
      "eval_recall": 0.0,
      "eval_runtime": 27.7669,
      "eval_samples_per_second": 36.806,
      "eval_steps_per_second": 0.576,
      "step": 144
    }
  ],
  "logging_steps": 500,
  "max_steps": 144,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 707085591840.0,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.48386139243976256,
    "learning_rate": 1.066209524723156e-05,
    "num_train_epochs": 3,
    "temperature": 23
  }
}