trainer_state.json

{
  "best_global_step": 23,
  "best_metric": 2.2505738735198975,
  "best_model_checkpoint": "bert_base_rand_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": 6.138053894042969,
      "learning_rate": 4.9e-05,
      "loss": 2.5065,
      "step": 23
    },
    {
      "epoch": 1.0,
      "eval_combined_score": 0.14165598564820725,
      "eval_loss": 2.2505738735198975,
      "eval_pearson": 0.15290276594334748,
      "eval_runtime": 0.9636,
      "eval_samples_per_second": 1556.616,
      "eval_spearmanr": 0.130409205353067,
      "eval_steps_per_second": 6.226,
      "step": 23
    },
    {
      "epoch": 2.0,
      "grad_norm": 8.916446685791016,
      "learning_rate": 4.8e-05,
      "loss": 1.8387,
      "step": 46
    },
    {
      "epoch": 2.0,
      "eval_combined_score": 0.20361793621346566,
      "eval_loss": 2.327321767807007,
      "eval_pearson": 0.21109381931307603,
      "eval_runtime": 1.0099,
      "eval_samples_per_second": 1485.269,
      "eval_spearmanr": 0.1961420531138553,
      "eval_steps_per_second": 5.941,
      "step": 46
    },
    {
      "epoch": 3.0,
      "grad_norm": 7.65189266204834,
      "learning_rate": 4.7e-05,
      "loss": 1.5726,
      "step": 69
    },
    {
      "epoch": 3.0,
      "eval_combined_score": 0.26004340706521684,
      "eval_loss": 2.3812336921691895,
      "eval_pearson": 0.26269662631010304,
      "eval_runtime": 1.0098,
      "eval_samples_per_second": 1485.478,
      "eval_spearmanr": 0.25739018782033063,
      "eval_steps_per_second": 5.942,
      "step": 69
    },
    {
      "epoch": 4.0,
      "grad_norm": 9.502486228942871,
      "learning_rate": 4.600000000000001e-05,
      "loss": 1.2262,
      "step": 92
    },
    {
      "epoch": 4.0,
      "eval_combined_score": 0.3058346282328328,
      "eval_loss": 2.30908465385437,
      "eval_pearson": 0.3068104941855512,
      "eval_runtime": 0.9592,
      "eval_samples_per_second": 1563.853,
      "eval_spearmanr": 0.30485876228011444,
      "eval_steps_per_second": 6.255,
      "step": 92
    },
    {
      "epoch": 5.0,
      "grad_norm": 10.076950073242188,
      "learning_rate": 4.5e-05,
      "loss": 0.9898,
      "step": 115
    },
    {
      "epoch": 5.0,
      "eval_combined_score": 0.3036459122427405,
      "eval_loss": 2.5820555686950684,
      "eval_pearson": 0.3065671873384344,
      "eval_runtime": 0.9802,
      "eval_samples_per_second": 1530.291,
      "eval_spearmanr": 0.3007246371470466,
      "eval_steps_per_second": 6.121,
      "step": 115
    },
    {
      "epoch": 6.0,
      "grad_norm": 22.569692611694336,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.8069,
      "step": 138
    },
    {
      "epoch": 6.0,
      "eval_combined_score": 0.28345622051244834,
      "eval_loss": 2.8197286128997803,
      "eval_pearson": 0.28319701039514467,
      "eval_runtime": 1.0012,
      "eval_samples_per_second": 1498.173,
      "eval_spearmanr": 0.2837154306297521,
      "eval_steps_per_second": 5.993,
      "step": 138
    },
    {
      "epoch": 6.0,
      "step": 138,
      "total_flos": 4537835628033024.0,
      "train_loss": 1.4901102176610974,
      "train_runtime": 74.5078,
      "train_samples_per_second": 3857.986,
      "train_steps_per_second": 15.435
    }
  ],
  "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
}