trainer_state.json

{
  "best_global_step": 820,
  "best_metric": 0.625216543674469,
  "best_model_checkpoint": "bert_base_rand_10_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": 2.0741770267486572,
      "learning_rate": 4.9e-05,
      "loss": 0.6627,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6258466044298004,
      "eval_loss": 0.6420552730560303,
      "eval_runtime": 3.4441,
      "eval_samples_per_second": 1586.194,
      "eval_steps_per_second": 6.388,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.8706566095352173,
      "learning_rate": 4.8e-05,
      "loss": 0.6209,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6447007138934652,
      "eval_loss": 0.625216543674469,
      "eval_runtime": 3.4085,
      "eval_samples_per_second": 1602.769,
      "eval_steps_per_second": 6.454,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.08432936668396,
      "learning_rate": 4.7e-05,
      "loss": 0.5518,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6317041918359876,
      "eval_loss": 0.6554394364356995,
      "eval_runtime": 3.4155,
      "eval_samples_per_second": 1599.462,
      "eval_steps_per_second": 6.441,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 4.187309741973877,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4506,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6507413509060955,
      "eval_loss": 0.6905505061149597,
      "eval_runtime": 3.4294,
      "eval_samples_per_second": 1593.007,
      "eval_steps_per_second": 6.415,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 5.3735127449035645,
      "learning_rate": 4.5e-05,
      "loss": 0.3408,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6582463847702728,
      "eval_loss": 0.7889552712440491,
      "eval_runtime": 3.4217,
      "eval_samples_per_second": 1596.586,
      "eval_steps_per_second": 6.43,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 6.6949076652526855,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.2419,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6496430532674354,
      "eval_loss": 0.9268649816513062,
      "eval_runtime": 3.4156,
      "eval_samples_per_second": 1599.444,
      "eval_steps_per_second": 6.441,
      "step": 2460
    },
    {
      "epoch": 7.0,
      "grad_norm": 11.023075103759766,
      "learning_rate": 4.3e-05,
      "loss": 0.1725,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6489108548416621,
      "eval_loss": 1.1216825246810913,
      "eval_runtime": 3.43,
      "eval_samples_per_second": 1592.703,
      "eval_steps_per_second": 6.414,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "step": 2870,
      "total_flos": 9.645664445050368e+16,
      "train_loss": 0.43446183686472395,
      "train_runtime": 1216.0331,
      "train_samples_per_second": 4306.749,
      "train_steps_per_second": 16.858
    }
  ],
  "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
}