trainer_state.json

{
  "best_global_step": 6,
  "best_metric": 0.720895528793335,
  "best_model_checkpoint": "bert_base_rand_10_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": 2.869922399520874,
      "learning_rate": 4.9e-05,
      "loss": 0.7306,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.28169014084507044,
      "eval_loss": 0.7360255718231201,
      "eval_runtime": 0.056,
      "eval_samples_per_second": 1267.62,
      "eval_steps_per_second": 17.854,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 2.0211639404296875,
      "learning_rate": 4.8e-05,
      "loss": 0.7091,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5492957746478874,
      "eval_loss": 0.720895528793335,
      "eval_runtime": 0.0559,
      "eval_samples_per_second": 1271.066,
      "eval_steps_per_second": 17.902,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.0512336492538452,
      "learning_rate": 4.7e-05,
      "loss": 0.6988,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.352112676056338,
      "eval_loss": 0.7334397435188293,
      "eval_runtime": 0.0591,
      "eval_samples_per_second": 1202.17,
      "eval_steps_per_second": 16.932,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.5780447721481323,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.7021,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5352112676056338,
      "eval_loss": 0.7260123491287231,
      "eval_runtime": 0.0575,
      "eval_samples_per_second": 1234.058,
      "eval_steps_per_second": 17.381,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.5089037418365479,
      "learning_rate": 4.5e-05,
      "loss": 0.6924,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.2112676056338028,
      "eval_loss": 0.7530258893966675,
      "eval_runtime": 0.0582,
      "eval_samples_per_second": 1219.185,
      "eval_steps_per_second": 17.172,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.8330632448196411,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.692,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.2112676056338028,
      "eval_loss": 0.791263222694397,
      "eval_runtime": 0.0573,
      "eval_samples_per_second": 1239.245,
      "eval_steps_per_second": 17.454,
      "step": 18
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.7287513613700867,
      "learning_rate": 4.3e-05,
      "loss": 0.6935,
      "step": 21
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.2112676056338028,
      "eval_loss": 0.8150032758712769,
      "eval_runtime": 0.0574,
      "eval_samples_per_second": 1237.535,
      "eval_steps_per_second": 17.43,
      "step": 21
    },
    {
      "epoch": 7.0,
      "step": 21,
      "total_flos": 584764320537600.0,
      "train_loss": 0.7026454721178327,
      "train_runtime": 20.9838,
      "train_samples_per_second": 1513.074,
      "train_steps_per_second": 7.148
    }
  ],
  "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
}