trainer_state.json

{
  "best_global_step": 820,
  "best_metric": 0.636478841304779,
  "best_model_checkpoint": "bert_base_rand_20_v1_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": 2.3261513710021973,
      "learning_rate": 4.9e-05,
      "loss": 0.6633,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6207212154493867,
      "eval_loss": 0.6409383416175842,
      "eval_runtime": 3.4445,
      "eval_samples_per_second": 1585.995,
      "eval_steps_per_second": 6.387,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.662063479423523,
      "learning_rate": 4.8e-05,
      "loss": 0.6208,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6366465312099578,
      "eval_loss": 0.636478841304779,
      "eval_runtime": 3.4384,
      "eval_samples_per_second": 1588.825,
      "eval_steps_per_second": 6.398,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.498321056365967,
      "learning_rate": 4.7e-05,
      "loss": 0.5518,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6265788028555739,
      "eval_loss": 0.6709644794464111,
      "eval_runtime": 3.4177,
      "eval_samples_per_second": 1598.45,
      "eval_steps_per_second": 6.437,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 4.980870723724365,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4499,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6492769540545488,
      "eval_loss": 0.7233583927154541,
      "eval_runtime": 3.4273,
      "eval_samples_per_second": 1593.971,
      "eval_steps_per_second": 6.419,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 6.657159805297852,
      "learning_rate": 4.5e-05,
      "loss": 0.3318,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6496430532674354,
      "eval_loss": 0.8131396174430847,
      "eval_runtime": 3.4156,
      "eval_samples_per_second": 1599.424,
      "eval_steps_per_second": 6.441,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 7.397583484649658,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.2331,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6494600036609921,
      "eval_loss": 0.9699184894561768,
      "eval_runtime": 3.4124,
      "eval_samples_per_second": 1600.942,
      "eval_steps_per_second": 6.447,
      "step": 2460
    },
    {
      "epoch": 7.0,
      "grad_norm": 10.258642196655273,
      "learning_rate": 4.3e-05,
      "loss": 0.1686,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6490939044481054,
      "eval_loss": 1.1583185195922852,
      "eval_runtime": 3.4274,
      "eval_samples_per_second": 1593.928,
      "eval_steps_per_second": 6.419,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "step": 2870,
      "total_flos": 9.645664445050368e+16,
      "train_loss": 0.4313258712715388,
      "train_runtime": 1217.4796,
      "train_samples_per_second": 4301.633,
      "train_steps_per_second": 16.838
    }
  ],
  "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
}