trainer_state.json

{
  "best_global_step": 3,
  "best_metric": 0.7008141279220581,
  "best_model_checkpoint": "bert_base_rand_5_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": 1.5306988954544067,
      "learning_rate": 4.9e-05,
      "loss": 0.7385,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5352112676056338,
      "eval_loss": 0.7008141279220581,
      "eval_runtime": 0.0586,
      "eval_samples_per_second": 1210.6,
      "eval_steps_per_second": 17.051,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.2955060005187988,
      "learning_rate": 4.8e-05,
      "loss": 0.6977,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.4225352112676056,
      "eval_loss": 0.7293685078620911,
      "eval_runtime": 0.0915,
      "eval_samples_per_second": 776.193,
      "eval_steps_per_second": 10.932,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.6576025485992432,
      "learning_rate": 4.7e-05,
      "loss": 0.6963,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5211267605633803,
      "eval_loss": 0.7183099985122681,
      "eval_runtime": 0.0571,
      "eval_samples_per_second": 1244.185,
      "eval_steps_per_second": 17.524,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.8121181726455688,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.699,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.28169014084507044,
      "eval_loss": 0.742902934551239,
      "eval_runtime": 0.0636,
      "eval_samples_per_second": 1116.824,
      "eval_steps_per_second": 15.73,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.8144534826278687,
      "learning_rate": 4.5e-05,
      "loss": 0.6899,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.2676056338028169,
      "eval_loss": 0.7939040064811707,
      "eval_runtime": 0.0558,
      "eval_samples_per_second": 1271.354,
      "eval_steps_per_second": 17.906,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.4151126146316528,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6944,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.2112676056338028,
      "eval_loss": 0.815746009349823,
      "eval_runtime": 0.0592,
      "eval_samples_per_second": 1198.715,
      "eval_steps_per_second": 16.883,
      "step": 18
    },
    {
      "epoch": 6.0,
      "step": 18,
      "total_flos": 501226560460800.0,
      "train_loss": 0.7026170624627007,
      "train_runtime": 18.1785,
      "train_samples_per_second": 1746.571,
      "train_steps_per_second": 8.252
    }
  ],
  "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
}