trainer_state.json

{
  "best_global_step": 23,
  "best_metric": 2.2702600955963135,
  "best_model_checkpoint": "tiny_bert_km_50_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": 3.5258214473724365,
      "learning_rate": 4.9e-05,
      "loss": 3.769,
      "step": 23
    },
    {
      "epoch": 1.0,
      "eval_combined_score": 0.061850741748472673,
      "eval_loss": 2.2702600955963135,
      "eval_pearson": 0.06325951085085163,
      "eval_runtime": 0.4573,
      "eval_samples_per_second": 3280.143,
      "eval_spearmanr": 0.060441972646093724,
      "eval_steps_per_second": 13.121,
      "step": 23
    },
    {
      "epoch": 2.0,
      "grad_norm": 8.438042640686035,
      "learning_rate": 4.8e-05,
      "loss": 2.1015,
      "step": 46
    },
    {
      "epoch": 2.0,
      "eval_combined_score": 0.09490472048138736,
      "eval_loss": 2.463663339614868,
      "eval_pearson": 0.09819772263897679,
      "eval_runtime": 0.4785,
      "eval_samples_per_second": 3135.109,
      "eval_spearmanr": 0.09161171832379793,
      "eval_steps_per_second": 12.54,
      "step": 46
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.280874729156494,
      "learning_rate": 4.7e-05,
      "loss": 1.9322,
      "step": 69
    },
    {
      "epoch": 3.0,
      "eval_combined_score": 0.13107964907904068,
      "eval_loss": 2.4901719093322754,
      "eval_pearson": 0.13888928882063417,
      "eval_runtime": 0.4572,
      "eval_samples_per_second": 3280.511,
      "eval_spearmanr": 0.12327000933744717,
      "eval_steps_per_second": 13.122,
      "step": 69
    },
    {
      "epoch": 4.0,
      "grad_norm": 10.058952331542969,
      "learning_rate": 4.600000000000001e-05,
      "loss": 1.7867,
      "step": 92
    },
    {
      "epoch": 4.0,
      "eval_combined_score": 0.2506039361178365,
      "eval_loss": 2.272320032119751,
      "eval_pearson": 0.25382121009799424,
      "eval_runtime": 0.439,
      "eval_samples_per_second": 3416.855,
      "eval_spearmanr": 0.2473866621376787,
      "eval_steps_per_second": 13.667,
      "step": 92
    },
    {
      "epoch": 5.0,
      "grad_norm": 6.998331546783447,
      "learning_rate": 4.5e-05,
      "loss": 1.5685,
      "step": 115
    },
    {
      "epoch": 5.0,
      "eval_combined_score": 0.20585434593567933,
      "eval_loss": 2.7301976680755615,
      "eval_pearson": 0.20943604996132886,
      "eval_runtime": 0.4682,
      "eval_samples_per_second": 3203.619,
      "eval_spearmanr": 0.2022726419100298,
      "eval_steps_per_second": 12.814,
      "step": 115
    },
    {
      "epoch": 6.0,
      "grad_norm": 7.808454990386963,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 1.3328,
      "step": 138
    },
    {
      "epoch": 6.0,
      "eval_combined_score": 0.26494138305347625,
      "eval_loss": 2.6151115894317627,
      "eval_pearson": 0.2652234653348802,
      "eval_runtime": 0.4829,
      "eval_samples_per_second": 3105.973,
      "eval_spearmanr": 0.26465930077207234,
      "eval_steps_per_second": 12.424,
      "step": 138
    },
    {
      "epoch": 6.0,
      "step": 138,
      "total_flos": 904528923227136.0,
      "train_loss": 2.081795498944711,
      "train_runtime": 25.6098,
      "train_samples_per_second": 11224.203,
      "train_steps_per_second": 44.905
    }
  ],
  "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
}