trainer_state.json

{
  "best_global_step": 820,
  "best_metric": 0.6328096985816956,
  "best_model_checkpoint": "bert_base_km_100_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": 1.716758370399475,
      "learning_rate": 4.9e-05,
      "loss": 0.6676,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6223686619073769,
      "eval_loss": 0.643682599067688,
      "eval_runtime": 3.4688,
      "eval_samples_per_second": 1574.875,
      "eval_steps_per_second": 6.342,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.7402069568634033,
      "learning_rate": 4.8e-05,
      "loss": 0.6285,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6415888705839282,
      "eval_loss": 0.6328096985816956,
      "eval_runtime": 3.4802,
      "eval_samples_per_second": 1569.745,
      "eval_steps_per_second": 6.322,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.697132110595703,
      "learning_rate": 4.7e-05,
      "loss": 0.5647,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6265788028555739,
      "eval_loss": 0.6712454557418823,
      "eval_runtime": 3.4613,
      "eval_samples_per_second": 1578.286,
      "eval_steps_per_second": 6.356,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 5.115638732910156,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4538,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6373787296357313,
      "eval_loss": 0.7106929421424866,
      "eval_runtime": 3.4666,
      "eval_samples_per_second": 1575.911,
      "eval_steps_per_second": 6.346,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 5.851642608642578,
      "learning_rate": 4.5e-05,
      "loss": 0.3318,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6384770272743914,
      "eval_loss": 0.8167354464530945,
      "eval_runtime": 3.4611,
      "eval_samples_per_second": 1578.403,
      "eval_steps_per_second": 6.356,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 11.106274604797363,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.2309,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6307889438037708,
      "eval_loss": 0.9983347654342651,
      "eval_runtime": 3.4752,
      "eval_samples_per_second": 1571.995,
      "eval_steps_per_second": 6.331,
      "step": 2460
    },
    {
      "epoch": 7.0,
      "grad_norm": 9.505951881408691,
      "learning_rate": 4.3e-05,
      "loss": 0.1638,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6284092989200073,
      "eval_loss": 1.1963454484939575,
      "eval_runtime": 3.4596,
      "eval_samples_per_second": 1579.087,
      "eval_steps_per_second": 6.359,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "step": 2870,
      "total_flos": 9.645664445050368e+16,
      "train_loss": 0.4344479072384718,
      "train_runtime": 1225.8627,
      "train_samples_per_second": 4272.216,
      "train_steps_per_second": 16.723
    }
  ],
  "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
}