trainer_state.json

{
  "best_global_step": 2844,
  "best_metric": 0.38632383942604065,
  "best_model_checkpoint": "bert_base_rand_100_v2_qqp/checkpoint-2844",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 9954,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.6505345106124878,
      "learning_rate": 4.9e-05,
      "loss": 0.4754,
      "step": 1422
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7961167449913431,
      "eval_combined_score": 0.7469603778481664,
      "eval_f1": 0.6978040107049895,
      "eval_loss": 0.4311196208000183,
      "eval_runtime": 25.9058,
      "eval_samples_per_second": 1560.657,
      "eval_steps_per_second": 6.099,
      "step": 1422
    },
    {
      "epoch": 2.0,
      "grad_norm": 2.6791186332702637,
      "learning_rate": 4.8e-05,
      "loss": 0.3723,
      "step": 2844
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.817932228543161,
      "eval_combined_score": 0.7867106749757825,
      "eval_f1": 0.755489121408404,
      "eval_loss": 0.38632383942604065,
      "eval_runtime": 25.6452,
      "eval_samples_per_second": 1576.513,
      "eval_steps_per_second": 6.161,
      "step": 2844
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.180361747741699,
      "learning_rate": 4.7e-05,
      "loss": 0.2962,
      "step": 4266
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8258966114271581,
      "eval_combined_score": 0.8004574948336609,
      "eval_f1": 0.7750183782401636,
      "eval_loss": 0.3931578993797302,
      "eval_runtime": 25.6346,
      "eval_samples_per_second": 1577.166,
      "eval_steps_per_second": 6.164,
      "step": 4266
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.6064722537994385,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.2334,
      "step": 5688
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8353697749196142,
      "eval_combined_score": 0.8061118688196032,
      "eval_f1": 0.7768539627195923,
      "eval_loss": 0.4347355365753174,
      "eval_runtime": 25.6419,
      "eval_samples_per_second": 1576.718,
      "eval_steps_per_second": 6.162,
      "step": 5688
    },
    {
      "epoch": 5.0,
      "grad_norm": 3.1544950008392334,
      "learning_rate": 4.5e-05,
      "loss": 0.1823,
      "step": 7110
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8375216423447934,
      "eval_combined_score": 0.8071835544672297,
      "eval_f1": 0.776845466589666,
      "eval_loss": 0.45913636684417725,
      "eval_runtime": 25.6092,
      "eval_samples_per_second": 1578.729,
      "eval_steps_per_second": 6.17,
      "step": 7110
    },
    {
      "epoch": 6.0,
      "grad_norm": 4.33131217956543,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.1443,
      "step": 8532
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8384615384615385,
      "eval_combined_score": 0.806124191696219,
      "eval_f1": 0.7737868449308996,
      "eval_loss": 0.4977432191371918,
      "eval_runtime": 25.5656,
      "eval_samples_per_second": 1581.42,
      "eval_steps_per_second": 6.18,
      "step": 8532
    },
    {
      "epoch": 7.0,
      "grad_norm": 4.425964832305908,
      "learning_rate": 4.3e-05,
      "loss": 0.1156,
      "step": 9954
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8369527578530794,
      "eval_combined_score": 0.8105169947142894,
      "eval_f1": 0.7840812315754995,
      "eval_loss": 0.5663210153579712,
      "eval_runtime": 25.5977,
      "eval_samples_per_second": 1579.439,
      "eval_steps_per_second": 6.172,
      "step": 9954
    },
    {
      "epoch": 7.0,
      "step": 9954,
      "total_flos": 3.3506166766980096e+17,
      "train_loss": 0.25994441067092217,
      "train_runtime": 4309.9047,
      "train_samples_per_second": 4221.045,
      "train_steps_per_second": 16.497
    }
  ],
  "logging_steps": 1,
  "max_steps": 71100,
  "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": 3.3506166766980096e+17,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}