run-3/checkpoint-384/trainer_state.json

{
  "best_metric": 0.5459882583170255,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-3/checkpoint-384",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 384,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.7568773031234741,
      "learning_rate": 0.00039540069434609415,
      "loss": 0.3274,
      "step": 192
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5430528375733855,
      "eval_f1": 0.20442930153321975,
      "eval_loss": 0.30545827746391296,
      "eval_mcc": 0.1640968755879385,
      "eval_precision": 0.7894736842105263,
      "eval_recall": 0.11741682974559686,
      "eval_runtime": 66.7048,
      "eval_samples_per_second": 15.321,
      "eval_steps_per_second": 0.48,
      "step": 192
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.822717547416687,
      "learning_rate": 0.0003295005786217451,
      "loss": 0.3076,
      "step": 384
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5459882583170255,
      "eval_f1": 0.21355932203389827,
      "eval_loss": 0.2995118200778961,
      "eval_mcc": 0.17219814389174506,
      "eval_precision": 0.7974683544303798,
      "eval_recall": 0.1232876712328767,
      "eval_runtime": 66.8,
      "eval_samples_per_second": 15.299,
      "eval_steps_per_second": 0.479,
      "step": 384
    }
  ],
  "logging_steps": 500,
  "max_steps": 1344,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "save_steps": 500,
  "total_flos": 471390394560.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.4369516344275761,
    "learning_rate": 0.00046130081007044317,
    "num_train_epochs": 7,
    "per_device_train_batch_size": 16,
    "temperature": 15
  }
}