trainer_state.json

{
  "best_global_step": 6,
  "best_metric": 0.7162193655967712,
  "best_model_checkpoint": "bert_base_rand_20_v1_wnli/checkpoint-6",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 21,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.692142367362976,
      "learning_rate": 4.9e-05,
      "loss": 0.7271,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5633802816901409,
      "eval_loss": 0.7196852564811707,
      "eval_runtime": 0.0623,
      "eval_samples_per_second": 1139.748,
      "eval_steps_per_second": 16.053,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.3788628578186035,
      "learning_rate": 4.8e-05,
      "loss": 0.719,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.323943661971831,
      "eval_loss": 0.7162193655967712,
      "eval_runtime": 0.057,
      "eval_samples_per_second": 1246.507,
      "eval_steps_per_second": 17.556,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.4024417400360107,
      "learning_rate": 4.7e-05,
      "loss": 0.708,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.43661971830985913,
      "eval_loss": 0.753108561038971,
      "eval_runtime": 0.0556,
      "eval_samples_per_second": 1277.205,
      "eval_steps_per_second": 17.989,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.5756946802139282,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.7018,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5352112676056338,
      "eval_loss": 0.7173195481300354,
      "eval_runtime": 0.0613,
      "eval_samples_per_second": 1157.297,
      "eval_steps_per_second": 16.3,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 2.2764294147491455,
      "learning_rate": 4.5e-05,
      "loss": 0.6982,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.43661971830985913,
      "eval_loss": 0.7290931344032288,
      "eval_runtime": 0.0569,
      "eval_samples_per_second": 1247.583,
      "eval_steps_per_second": 17.572,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.7415590286254883,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6947,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.38028169014084506,
      "eval_loss": 0.7643594145774841,
      "eval_runtime": 0.0641,
      "eval_samples_per_second": 1107.059,
      "eval_steps_per_second": 15.592,
      "step": 18
    },
    {
      "epoch": 7.0,
      "grad_norm": 1.2452402114868164,
      "learning_rate": 4.3e-05,
      "loss": 0.6962,
      "step": 21
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.19718309859154928,
      "eval_loss": 0.7731072902679443,
      "eval_runtime": 0.0615,
      "eval_samples_per_second": 1153.947,
      "eval_steps_per_second": 16.253,
      "step": 21
    },
    {
      "epoch": 7.0,
      "step": 21,
      "total_flos": 584764320537600.0,
      "train_loss": 0.7064262685321626,
      "train_runtime": 21.0951,
      "train_samples_per_second": 1505.085,
      "train_steps_per_second": 7.111
    }
  ],
  "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": 584764320537600.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}