trainer_state.json

{
  "best_global_step": 3,
  "best_metric": 0.7005391120910645,
  "best_model_checkpoint": "bert_base_rand_50_v2_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": 1.080239176750183,
      "learning_rate": 4.9e-05,
      "loss": 0.7156,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5633802816901409,
      "eval_loss": 0.7005391120910645,
      "eval_runtime": 0.0552,
      "eval_samples_per_second": 1285.802,
      "eval_steps_per_second": 18.11,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.0200259685516357,
      "learning_rate": 4.8e-05,
      "loss": 0.6922,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.43661971830985913,
      "eval_loss": 0.7250218987464905,
      "eval_runtime": 0.0667,
      "eval_samples_per_second": 1063.92,
      "eval_steps_per_second": 14.985,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.638716995716095,
      "learning_rate": 4.7e-05,
      "loss": 0.697,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5070422535211268,
      "eval_loss": 0.7091219425201416,
      "eval_runtime": 0.0597,
      "eval_samples_per_second": 1190.016,
      "eval_steps_per_second": 16.761,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.8990293741226196,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6964,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5211267605633803,
      "eval_loss": 0.7154490351676941,
      "eval_runtime": 0.0582,
      "eval_samples_per_second": 1220.019,
      "eval_steps_per_second": 17.183,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.9785750508308411,
      "learning_rate": 4.5e-05,
      "loss": 0.6878,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.2676056338028169,
      "eval_loss": 0.7476343512535095,
      "eval_runtime": 0.0592,
      "eval_samples_per_second": 1199.811,
      "eval_steps_per_second": 16.899,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.9305042028427124,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6949,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.22535211267605634,
      "eval_loss": 0.7614985704421997,
      "eval_runtime": 0.0563,
      "eval_samples_per_second": 1260.516,
      "eval_steps_per_second": 17.754,
      "step": 18
    },
    {
      "epoch": 6.0,
      "step": 18,
      "total_flos": 501226560460800.0,
      "train_loss": 0.6973049110836453,
      "train_runtime": 19.9196,
      "train_samples_per_second": 1593.907,
      "train_steps_per_second": 7.53
    }
  ],
  "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
}