trainer_state.json

{
  "best_global_step": 3,
  "best_metric": 0.6973482966423035,
  "best_model_checkpoint": "bert_base_rand_50_v1_wnli/checkpoint-3",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 18,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.9846983551979065,
      "learning_rate": 4.9e-05,
      "loss": 0.7188,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5633802816901409,
      "eval_loss": 0.6973482966423035,
      "eval_runtime": 0.0558,
      "eval_samples_per_second": 1271.805,
      "eval_steps_per_second": 17.913,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.0693150758743286,
      "learning_rate": 4.8e-05,
      "loss": 0.6935,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.43661971830985913,
      "eval_loss": 0.7157790064811707,
      "eval_runtime": 0.0591,
      "eval_samples_per_second": 1201.36,
      "eval_steps_per_second": 16.921,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.6289031505584717,
      "learning_rate": 4.7e-05,
      "loss": 0.6958,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5211267605633803,
      "eval_loss": 0.7030148506164551,
      "eval_runtime": 0.0581,
      "eval_samples_per_second": 1222.854,
      "eval_steps_per_second": 17.223,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.810719609260559,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6966,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5070422535211268,
      "eval_loss": 0.7106072902679443,
      "eval_runtime": 0.0642,
      "eval_samples_per_second": 1105.239,
      "eval_steps_per_second": 15.567,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.7822015285491943,
      "learning_rate": 4.5e-05,
      "loss": 0.688,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.36619718309859156,
      "eval_loss": 0.741142213344574,
      "eval_runtime": 0.0694,
      "eval_samples_per_second": 1022.871,
      "eval_steps_per_second": 14.407,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.7482801675796509,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6948,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.2112676056338028,
      "eval_loss": 0.7464239597320557,
      "eval_runtime": 0.0552,
      "eval_samples_per_second": 1286.774,
      "eval_steps_per_second": 18.124,
      "step": 18
    },
    {
      "epoch": 6.0,
      "step": 18,
      "total_flos": 501226560460800.0,
      "train_loss": 0.6979075272878011,
      "train_runtime": 18.3803,
      "train_samples_per_second": 1727.394,
      "train_steps_per_second": 8.161
    }
  ],
  "logging_steps": 1,
  "max_steps": 150,
  "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": 501226560460800.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}