trainer_state.json

{
  "best_global_step": 23,
  "best_metric": 2.251096725463867,
  "best_model_checkpoint": "tiny_bert_km_20_v2_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": 4.101377487182617,
      "learning_rate": 4.9e-05,
      "loss": 3.587,
      "step": 23
    },
    {
      "epoch": 1.0,
      "eval_combined_score": 0.08975404205501787,
      "eval_loss": 2.251096725463867,
      "eval_pearson": 0.09287681409862332,
      "eval_runtime": 0.4548,
      "eval_samples_per_second": 3297.889,
      "eval_spearmanr": 0.08663127001141242,
      "eval_steps_per_second": 13.192,
      "step": 23
    },
    {
      "epoch": 2.0,
      "grad_norm": 4.866882801055908,
      "learning_rate": 4.8e-05,
      "loss": 2.048,
      "step": 46
    },
    {
      "epoch": 2.0,
      "eval_combined_score": 0.1328824379643906,
      "eval_loss": 2.480339765548706,
      "eval_pearson": 0.1372525837029327,
      "eval_runtime": 0.4811,
      "eval_samples_per_second": 3118.032,
      "eval_spearmanr": 0.12851229222584845,
      "eval_steps_per_second": 12.472,
      "step": 46
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.108093500137329,
      "learning_rate": 4.7e-05,
      "loss": 1.919,
      "step": 69
    },
    {
      "epoch": 3.0,
      "eval_combined_score": 0.15576252510849337,
      "eval_loss": 2.369476318359375,
      "eval_pearson": 0.15979532124431572,
      "eval_runtime": 0.4988,
      "eval_samples_per_second": 3007.003,
      "eval_spearmanr": 0.15172972897267104,
      "eval_steps_per_second": 12.028,
      "step": 69
    },
    {
      "epoch": 4.0,
      "grad_norm": 5.165529727935791,
      "learning_rate": 4.600000000000001e-05,
      "loss": 1.8084,
      "step": 92
    },
    {
      "epoch": 4.0,
      "eval_combined_score": 0.2241098720221349,
      "eval_loss": 2.482461929321289,
      "eval_pearson": 0.22656618416233015,
      "eval_runtime": 0.4453,
      "eval_samples_per_second": 3368.187,
      "eval_spearmanr": 0.2216535598819396,
      "eval_steps_per_second": 13.473,
      "step": 92
    },
    {
      "epoch": 5.0,
      "grad_norm": 9.7904634475708,
      "learning_rate": 4.5e-05,
      "loss": 1.6039,
      "step": 115
    },
    {
      "epoch": 5.0,
      "eval_combined_score": 0.20435892402954994,
      "eval_loss": 2.5223915576934814,
      "eval_pearson": 0.20516208329641367,
      "eval_runtime": 0.5032,
      "eval_samples_per_second": 2981.028,
      "eval_spearmanr": 0.20355576476268622,
      "eval_steps_per_second": 11.924,
      "step": 115
    },
    {
      "epoch": 6.0,
      "grad_norm": 7.459550380706787,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 1.3938,
      "step": 138
    },
    {
      "epoch": 6.0,
      "eval_combined_score": 0.26507947253641295,
      "eval_loss": 2.42105770111084,
      "eval_pearson": 0.26619993340482795,
      "eval_runtime": 0.4538,
      "eval_samples_per_second": 3305.359,
      "eval_spearmanr": 0.2639590116679979,
      "eval_steps_per_second": 13.221,
      "step": 138
    },
    {
      "epoch": 6.0,
      "step": 138,
      "total_flos": 904528923227136.0,
      "train_loss": 2.060001290362814,
      "train_runtime": 25.4652,
      "train_samples_per_second": 11287.959,
      "train_steps_per_second": 45.16
    }
  ],
  "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": 904528923227136.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}