trainer_state.json

{
  "best_global_step": 10,
  "best_metric": 0.6976040601730347,
  "best_model_checkpoint": "bert_base_km_20_v2_rte/checkpoint-10",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 60,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.4656991958618164,
      "learning_rate": 4.9e-05,
      "loss": 0.7038,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5234657039711191,
      "eval_loss": 0.6976040601730347,
      "eval_runtime": 0.1885,
      "eval_samples_per_second": 1469.145,
      "eval_steps_per_second": 10.608,
      "step": 10
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.0204411745071411,
      "learning_rate": 4.8e-05,
      "loss": 0.6567,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5234657039711191,
      "eval_loss": 0.7067984938621521,
      "eval_runtime": 0.1936,
      "eval_samples_per_second": 1430.579,
      "eval_steps_per_second": 10.329,
      "step": 20
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.1023492813110352,
      "learning_rate": 4.7e-05,
      "loss": 0.6219,
      "step": 30
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5234657039711191,
      "eval_loss": 0.7116214632987976,
      "eval_runtime": 0.1923,
      "eval_samples_per_second": 1440.339,
      "eval_steps_per_second": 10.4,
      "step": 30
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.1984957456588745,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.5515,
      "step": 40
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5270758122743683,
      "eval_loss": 0.7494421005249023,
      "eval_runtime": 0.2138,
      "eval_samples_per_second": 1295.442,
      "eval_steps_per_second": 9.353,
      "step": 40
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.3119920492172241,
      "learning_rate": 4.5e-05,
      "loss": 0.4533,
      "step": 50
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.5270758122743683,
      "eval_loss": 0.8514941930770874,
      "eval_runtime": 0.2007,
      "eval_samples_per_second": 1380.348,
      "eval_steps_per_second": 9.966,
      "step": 50
    },
    {
      "epoch": 6.0,
      "grad_norm": 4.073829174041748,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.3589,
      "step": 60
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.49458483754512633,
      "eval_loss": 1.037402868270874,
      "eval_runtime": 0.209,
      "eval_samples_per_second": 1325.151,
      "eval_steps_per_second": 9.568,
      "step": 60
    },
    {
      "epoch": 6.0,
      "step": 60,
      "total_flos": 1965439583539200.0,
      "train_loss": 0.5576799750328064,
      "train_runtime": 37.9004,
      "train_samples_per_second": 3284.925,
      "train_steps_per_second": 13.192
    }
  ],
  "logging_steps": 1,
  "max_steps": 500,
  "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": 1965439583539200.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}