trainer_state.json

{
  "best_global_step": 3,
  "best_metric": 0.7447733879089355,
  "best_model_checkpoint": "bert_base_km_10_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": 1.30870521068573,
      "learning_rate": 4.9e-05,
      "loss": 0.7363,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.3380281690140845,
      "eval_loss": 0.7447733879089355,
      "eval_runtime": 0.0697,
      "eval_samples_per_second": 1018.042,
      "eval_steps_per_second": 14.339,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.9982154965400696,
      "learning_rate": 4.8e-05,
      "loss": 0.6975,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.2535211267605634,
      "eval_loss": 0.765625,
      "eval_runtime": 0.0585,
      "eval_samples_per_second": 1214.318,
      "eval_steps_per_second": 17.103,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.2340563535690308,
      "learning_rate": 4.7e-05,
      "loss": 0.6875,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.18309859154929578,
      "eval_loss": 0.790823221206665,
      "eval_runtime": 0.0757,
      "eval_samples_per_second": 937.369,
      "eval_steps_per_second": 13.202,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.3357309103012085,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6822,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.23943661971830985,
      "eval_loss": 0.8011114597320557,
      "eval_runtime": 0.0548,
      "eval_samples_per_second": 1295.084,
      "eval_steps_per_second": 18.241,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.4954864978790283,
      "learning_rate": 4.5e-05,
      "loss": 0.6762,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.1267605633802817,
      "eval_loss": 0.8423195481300354,
      "eval_runtime": 0.057,
      "eval_samples_per_second": 1246.366,
      "eval_steps_per_second": 17.554,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.42880117893219,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6704,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.16901408450704225,
      "eval_loss": 0.8818220496177673,
      "eval_runtime": 0.0558,
      "eval_samples_per_second": 1272.652,
      "eval_steps_per_second": 17.925,
      "step": 18
    },
    {
      "epoch": 6.0,
      "step": 18,
      "total_flos": 501226560460800.0,
      "train_loss": 0.691677451133728,
      "train_runtime": 18.2689,
      "train_samples_per_second": 1737.93,
      "train_steps_per_second": 8.211
    }
  ],
  "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
}