trainer_state.json

{
  "best_global_step": 23,
  "best_metric": 2.231001615524292,
  "best_model_checkpoint": "tiny_bert_km_10_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": 3.4998087882995605,
      "learning_rate": 4.9e-05,
      "loss": 3.4768,
      "step": 23
    },
    {
      "epoch": 1.0,
      "eval_combined_score": 0.10647189293092346,
      "eval_loss": 2.231001615524292,
      "eval_pearson": 0.1085755656962291,
      "eval_runtime": 0.4473,
      "eval_samples_per_second": 3353.208,
      "eval_spearmanr": 0.10436822016561782,
      "eval_steps_per_second": 13.413,
      "step": 23
    },
    {
      "epoch": 2.0,
      "grad_norm": 5.5359883308410645,
      "learning_rate": 4.8e-05,
      "loss": 2.0479,
      "step": 46
    },
    {
      "epoch": 2.0,
      "eval_combined_score": 0.16415000302973756,
      "eval_loss": 2.444279432296753,
      "eval_pearson": 0.16666248518393506,
      "eval_runtime": 0.5098,
      "eval_samples_per_second": 2942.467,
      "eval_spearmanr": 0.16163752087554006,
      "eval_steps_per_second": 11.77,
      "step": 46
    },
    {
      "epoch": 3.0,
      "grad_norm": 6.075698375701904,
      "learning_rate": 4.7e-05,
      "loss": 1.9049,
      "step": 69
    },
    {
      "epoch": 3.0,
      "eval_combined_score": 0.20889157319674626,
      "eval_loss": 2.2881205081939697,
      "eval_pearson": 0.21778903054526308,
      "eval_runtime": 0.439,
      "eval_samples_per_second": 3417.045,
      "eval_spearmanr": 0.19999411584822943,
      "eval_steps_per_second": 13.668,
      "step": 69
    },
    {
      "epoch": 4.0,
      "grad_norm": 4.910608291625977,
      "learning_rate": 4.600000000000001e-05,
      "loss": 1.7247,
      "step": 92
    },
    {
      "epoch": 4.0,
      "eval_combined_score": 0.27839785784759497,
      "eval_loss": 2.4856960773468018,
      "eval_pearson": 0.28412465271234644,
      "eval_runtime": 0.439,
      "eval_samples_per_second": 3416.509,
      "eval_spearmanr": 0.27267106298284355,
      "eval_steps_per_second": 13.666,
      "step": 92
    },
    {
      "epoch": 5.0,
      "grad_norm": 7.410452365875244,
      "learning_rate": 4.5e-05,
      "loss": 1.4995,
      "step": 115
    },
    {
      "epoch": 5.0,
      "eval_combined_score": 0.28896417503149296,
      "eval_loss": 2.299758195877075,
      "eval_pearson": 0.29489535524400456,
      "eval_runtime": 0.5114,
      "eval_samples_per_second": 2933.009,
      "eval_spearmanr": 0.28303299481898136,
      "eval_steps_per_second": 11.732,
      "step": 115
    },
    {
      "epoch": 6.0,
      "grad_norm": 6.83831787109375,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 1.2606,
      "step": 138
    },
    {
      "epoch": 6.0,
      "eval_combined_score": 0.323121440564714,
      "eval_loss": 2.3139901161193848,
      "eval_pearson": 0.326728498013101,
      "eval_runtime": 0.453,
      "eval_samples_per_second": 3311.54,
      "eval_spearmanr": 0.31951438311632696,
      "eval_steps_per_second": 13.246,
      "step": 138
    },
    {
      "epoch": 6.0,
      "step": 138,
      "total_flos": 904528923227136.0,
      "train_loss": 1.9857357688572095,
      "train_runtime": 25.6978,
      "train_samples_per_second": 11185.796,
      "train_steps_per_second": 44.751
    }
  ],
  "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
}