trainer_state.json

{
  "best_global_step": 820,
  "best_metric": 0.6290686726570129,
  "best_model_checkpoint": "bert_base_rand_50_v2_qnli/checkpoint-820",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 2870,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 2.063289165496826,
      "learning_rate": 4.9e-05,
      "loss": 0.6614,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6243822075782537,
      "eval_loss": 0.6421895027160645,
      "eval_runtime": 3.4927,
      "eval_samples_per_second": 1564.117,
      "eval_steps_per_second": 6.299,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.5447287559509277,
      "learning_rate": 4.8e-05,
      "loss": 0.6209,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6423210690097017,
      "eval_loss": 0.6290686726570129,
      "eval_runtime": 3.4716,
      "eval_samples_per_second": 1573.614,
      "eval_steps_per_second": 6.337,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.6968469619750977,
      "learning_rate": 4.7e-05,
      "loss": 0.5531,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6346329855390811,
      "eval_loss": 0.6537161469459534,
      "eval_runtime": 3.4779,
      "eval_samples_per_second": 1570.782,
      "eval_steps_per_second": 6.326,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 5.301610469818115,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4498,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6456159619256818,
      "eval_loss": 0.7286065220832825,
      "eval_runtime": 3.4811,
      "eval_samples_per_second": 1569.347,
      "eval_steps_per_second": 6.32,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 5.279804706573486,
      "learning_rate": 4.5e-05,
      "loss": 0.3341,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6467142595643419,
      "eval_loss": 0.7746986746788025,
      "eval_runtime": 3.4866,
      "eval_samples_per_second": 1566.845,
      "eval_steps_per_second": 6.31,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 7.716037273406982,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.2345,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6408566721581549,
      "eval_loss": 1.0803108215332031,
      "eval_runtime": 3.4654,
      "eval_samples_per_second": 1576.447,
      "eval_steps_per_second": 6.348,
      "step": 2460
    },
    {
      "epoch": 7.0,
      "grad_norm": 11.980489730834961,
      "learning_rate": 4.3e-05,
      "loss": 0.1664,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6448837634999085,
      "eval_loss": 1.203567385673523,
      "eval_runtime": 3.4862,
      "eval_samples_per_second": 1567.044,
      "eval_steps_per_second": 6.311,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "step": 2870,
      "total_flos": 9.645664445050368e+16,
      "train_loss": 0.43145642164276865,
      "train_runtime": 1229.703,
      "train_samples_per_second": 4258.874,
      "train_steps_per_second": 16.671
    }
  ],
  "logging_steps": 1,
  "max_steps": 20500,
  "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": 9.645664445050368e+16,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}