run-16/checkpoint-1782/trainer_state.json

{
  "best_metric": 0.6633663366336634,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-16/checkpoint-1188",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 1782,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.6842290163040161,
      "learning_rate": 0.00013702102722124262,
      "loss": 0.6194,
      "step": 297
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5069306930693069,
      "eval_f1": 0.023529411764705882,
      "eval_loss": 0.6244187951087952,
      "eval_mcc": 0.07745844229442526,
      "eval_precision": 1.0,
      "eval_recall": 0.011904761904761904,
      "eval_runtime": 0.9461,
      "eval_samples_per_second": 533.75,
      "eval_steps_per_second": 16.911,
      "step": 297
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.26532781124115,
      "learning_rate": 0.0001096168217769941,
      "loss": 0.6076,
      "step": 594
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6,
      "eval_f1": 0.6366906474820144,
      "eval_loss": 0.6058114767074585,
      "eval_mcc": 0.20470740045924352,
      "eval_precision": 0.5822368421052632,
      "eval_recall": 0.7023809523809523,
      "eval_runtime": 0.946,
      "eval_samples_per_second": 533.844,
      "eval_steps_per_second": 16.914,
      "step": 594
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.4105544090270996,
      "learning_rate": 8.221261633274557e-05,
      "loss": 0.5931,
      "step": 891
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6138613861386139,
      "eval_f1": 0.6859903381642513,
      "eval_loss": 0.5950631499290466,
      "eval_mcc": 0.25770627027550275,
      "eval_precision": 0.5772357723577236,
      "eval_recall": 0.8452380952380952,
      "eval_runtime": 0.9327,
      "eval_samples_per_second": 541.416,
      "eval_steps_per_second": 17.154,
      "step": 891
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.0386664867401123,
      "learning_rate": 5.480841088849705e-05,
      "loss": 0.5675,
      "step": 1188
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6633663366336634,
      "eval_f1": 0.6953405017921147,
      "eval_loss": 0.5774483680725098,
      "eval_mcc": 0.3347533105713868,
      "eval_precision": 0.6339869281045751,
      "eval_recall": 0.7698412698412699,
      "eval_runtime": 0.9362,
      "eval_samples_per_second": 539.398,
      "eval_steps_per_second": 17.09,
      "step": 1188
    },
    {
      "epoch": 5.0,
      "grad_norm": 3.236018180847168,
      "learning_rate": 2.7404205444248525e-05,
      "loss": 0.557,
      "step": 1485
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6613861386138614,
      "eval_f1": 0.6984126984126985,
      "eval_loss": 0.576128363609314,
      "eval_mcc": 0.3336455858078349,
      "eval_precision": 0.6285714285714286,
      "eval_recall": 0.7857142857142857,
      "eval_runtime": 0.9461,
      "eval_samples_per_second": 533.775,
      "eval_steps_per_second": 16.912,
      "step": 1485
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.6593701839447021,
      "learning_rate": 0.0,
      "loss": 0.5524,
      "step": 1782
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6495049504950495,
      "eval_f1": 0.6752293577981651,
      "eval_loss": 0.5702252388000488,
      "eval_mcc": 0.3032544177266147,
      "eval_precision": 0.6279863481228669,
      "eval_recall": 0.7301587301587301,
      "eval_runtime": 0.9409,
      "eval_samples_per_second": 536.729,
      "eval_steps_per_second": 17.005,
      "step": 1782
    }
  ],
  "logging_steps": 500,
  "max_steps": 1782,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 6,
  "save_steps": 500,
  "total_flos": 4384207595880.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.8675008492554994,
    "learning_rate": 0.00016442523266549115,
    "num_train_epochs": 6,
    "temperature": 44
  }
}