trainer_state.json

{
  "best_metric": 0.7026298642158508,
  "best_model_checkpoint": "bert_uncased_L-2_H-512_A-8_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": 1.546919584274292,
      "learning_rate": 4.9e-05,
      "loss": 0.709,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.4507042253521127,
      "eval_loss": 0.7026298642158508,
      "eval_runtime": 0.0333,
      "eval_samples_per_second": 2131.954,
      "eval_steps_per_second": 30.028,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.9109190106391907,
      "learning_rate": 4.8e-05,
      "loss": 0.6989,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.4084507042253521,
      "eval_loss": 0.7127256989479065,
      "eval_runtime": 0.0349,
      "eval_samples_per_second": 2034.693,
      "eval_steps_per_second": 28.658,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.8286433219909668,
      "learning_rate": 4.7e-05,
      "loss": 0.6959,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.323943661971831,
      "eval_loss": 0.7356405854225159,
      "eval_runtime": 0.0345,
      "eval_samples_per_second": 2060.313,
      "eval_steps_per_second": 29.018,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.0639235973358154,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.692,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.30985915492957744,
      "eval_loss": 0.7548967003822327,
      "eval_runtime": 0.0352,
      "eval_samples_per_second": 2014.31,
      "eval_steps_per_second": 28.371,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.0856434106826782,
      "learning_rate": 4.5e-05,
      "loss": 0.6833,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.28169014084507044,
      "eval_loss": 0.7648547291755676,
      "eval_runtime": 0.0343,
      "eval_samples_per_second": 2072.14,
      "eval_steps_per_second": 29.185,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.3245850801467896,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6881,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.29577464788732394,
      "eval_loss": 0.7724470496177673,
      "eval_runtime": 0.0331,
      "eval_samples_per_second": 2141.875,
      "eval_steps_per_second": 30.167,
      "step": 18
    },
    {
      "epoch": 6.0,
      "step": 18,
      "total_flos": 38445612963840.0,
      "train_loss": 0.6945502228207059,
      "train_runtime": 6.1777,
      "train_samples_per_second": 5139.486,
      "train_steps_per_second": 24.281
    }
  ],
  "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": 38445612963840.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}