trainer_state.json

{
  "best_global_step": 23,
  "best_metric": 2.187478542327881,
  "best_model_checkpoint": "bert_base_km_10_v1_stsb/checkpoint-23",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 138,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 3.9459848403930664,
      "learning_rate": 4.9e-05,
      "loss": 2.7027,
      "step": 23
    },
    {
      "epoch": 1.0,
      "eval_combined_score": 0.21406400269142756,
      "eval_loss": 2.187478542327881,
      "eval_pearson": 0.2207100243077243,
      "eval_runtime": 0.9709,
      "eval_samples_per_second": 1544.887,
      "eval_spearmanr": 0.20741798107513082,
      "eval_steps_per_second": 6.18,
      "step": 23
    },
    {
      "epoch": 2.0,
      "grad_norm": 10.513463973999023,
      "learning_rate": 4.8e-05,
      "loss": 1.9228,
      "step": 46
    },
    {
      "epoch": 2.0,
      "eval_combined_score": 0.2302906062932306,
      "eval_loss": 2.1904146671295166,
      "eval_pearson": 0.23344625218227338,
      "eval_runtime": 0.9664,
      "eval_samples_per_second": 1552.129,
      "eval_spearmanr": 0.2271349604041878,
      "eval_steps_per_second": 6.209,
      "step": 46
    },
    {
      "epoch": 3.0,
      "grad_norm": 6.263645172119141,
      "learning_rate": 4.7e-05,
      "loss": 1.6984,
      "step": 69
    },
    {
      "epoch": 3.0,
      "eval_combined_score": 0.28164812059092836,
      "eval_loss": 2.2391693592071533,
      "eval_pearson": 0.28053417859824636,
      "eval_runtime": 0.9545,
      "eval_samples_per_second": 1571.51,
      "eval_spearmanr": 0.28276206258361036,
      "eval_steps_per_second": 6.286,
      "step": 69
    },
    {
      "epoch": 4.0,
      "grad_norm": 9.83980655670166,
      "learning_rate": 4.600000000000001e-05,
      "loss": 1.3717,
      "step": 92
    },
    {
      "epoch": 4.0,
      "eval_combined_score": 0.28480233161198865,
      "eval_loss": 2.377173662185669,
      "eval_pearson": 0.28509458425166595,
      "eval_runtime": 0.982,
      "eval_samples_per_second": 1527.5,
      "eval_spearmanr": 0.28451007897231134,
      "eval_steps_per_second": 6.11,
      "step": 92
    },
    {
      "epoch": 5.0,
      "grad_norm": 11.876545906066895,
      "learning_rate": 4.5e-05,
      "loss": 1.044,
      "step": 115
    },
    {
      "epoch": 5.0,
      "eval_combined_score": 0.3106563828141037,
      "eval_loss": 2.494642972946167,
      "eval_pearson": 0.30931274369904616,
      "eval_runtime": 0.9476,
      "eval_samples_per_second": 1582.949,
      "eval_spearmanr": 0.31200002192916115,
      "eval_steps_per_second": 6.332,
      "step": 115
    },
    {
      "epoch": 6.0,
      "grad_norm": 9.593144416809082,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.7602,
      "step": 138
    },
    {
      "epoch": 6.0,
      "eval_combined_score": 0.27269630445966425,
      "eval_loss": 2.5727880001068115,
      "eval_pearson": 0.27378787732681964,
      "eval_runtime": 1.0021,
      "eval_samples_per_second": 1496.789,
      "eval_spearmanr": 0.27160473159250886,
      "eval_steps_per_second": 5.987,
      "step": 138
    },
    {
      "epoch": 6.0,
      "step": 138,
      "total_flos": 4537835628033024.0,
      "train_loss": 1.5832905009172964,
      "train_runtime": 74.6338,
      "train_samples_per_second": 3851.471,
      "train_steps_per_second": 15.409
    }
  ],
  "logging_steps": 1,
  "max_steps": 1150,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 50,
  "save_steps": 500,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 5,
        "early_stopping_threshold": 0.0
      },
      "attributes": {
        "early_stopping_patience_counter": 5
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 4537835628033024.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}