trainer_state.json

{
  "best_global_step": 820,
  "best_metric": 0.6276963949203491,
  "best_model_checkpoint": "bert_base_rand_5_v2_qnli/checkpoint-820",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 2870,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 2.4627881050109863,
      "learning_rate": 4.9e-05,
      "loss": 0.6613,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6249313563975838,
      "eval_loss": 0.6414567828178406,
      "eval_runtime": 3.4899,
      "eval_samples_per_second": 1565.353,
      "eval_steps_per_second": 6.304,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.6617871522903442,
      "learning_rate": 4.8e-05,
      "loss": 0.6213,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6414058209774849,
      "eval_loss": 0.6276963949203491,
      "eval_runtime": 3.4603,
      "eval_samples_per_second": 1578.747,
      "eval_steps_per_second": 6.358,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.3013253211975098,
      "learning_rate": 4.7e-05,
      "loss": 0.5509,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6364634816035145,
      "eval_loss": 0.6465625166893005,
      "eval_runtime": 3.4569,
      "eval_samples_per_second": 1580.295,
      "eval_steps_per_second": 6.364,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 5.9752068519592285,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4463,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6511074501189822,
      "eval_loss": 0.7135743498802185,
      "eval_runtime": 3.4569,
      "eval_samples_per_second": 1580.321,
      "eval_steps_per_second": 6.364,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 6.846456527709961,
      "learning_rate": 4.5e-05,
      "loss": 0.3322,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6624565257184697,
      "eval_loss": 0.7906157374382019,
      "eval_runtime": 3.4605,
      "eval_samples_per_second": 1578.654,
      "eval_steps_per_second": 6.357,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 8.928763389587402,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.2362,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6468973091707853,
      "eval_loss": 0.9955342411994934,
      "eval_runtime": 3.4712,
      "eval_samples_per_second": 1573.806,
      "eval_steps_per_second": 6.338,
      "step": 2460
    },
    {
      "epoch": 7.0,
      "grad_norm": 9.10348892211914,
      "learning_rate": 4.3e-05,
      "loss": 0.1685,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6545853926414058,
      "eval_loss": 1.0699700117111206,
      "eval_runtime": 3.4558,
      "eval_samples_per_second": 1580.809,
      "eval_steps_per_second": 6.366,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "step": 2870,
      "total_flos": 9.645664445050368e+16,
      "train_loss": 0.4309434897392884,
      "train_runtime": 1225.0183,
      "train_samples_per_second": 4275.161,
      "train_steps_per_second": 16.734
    }
  ],
  "logging_steps": 1,
  "max_steps": 20500,
  "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": 9.645664445050368e+16,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}