trainer_state.json

{
  "best_global_step": 20,
  "best_metric": 0.7143642902374268,
  "best_model_checkpoint": "bert_base_km_20_v1_rte/checkpoint-20",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 70,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.6082864999771118,
      "learning_rate": 4.9e-05,
      "loss": 0.7172,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.45126353790613716,
      "eval_loss": 0.7190461158752441,
      "eval_runtime": 0.2048,
      "eval_samples_per_second": 1352.543,
      "eval_steps_per_second": 9.766,
      "step": 10
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.069560170173645,
      "learning_rate": 4.8e-05,
      "loss": 0.6658,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.48736462093862815,
      "eval_loss": 0.7143642902374268,
      "eval_runtime": 0.1966,
      "eval_samples_per_second": 1408.66,
      "eval_steps_per_second": 10.171,
      "step": 20
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.088363766670227,
      "learning_rate": 4.7e-05,
      "loss": 0.6169,
      "step": 30
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.48375451263537905,
      "eval_loss": 0.7388594150543213,
      "eval_runtime": 0.2029,
      "eval_samples_per_second": 1365.048,
      "eval_steps_per_second": 9.856,
      "step": 30
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.4970088005065918,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.5307,
      "step": 40
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.49458483754512633,
      "eval_loss": 0.7875006794929504,
      "eval_runtime": 0.1976,
      "eval_samples_per_second": 1401.985,
      "eval_steps_per_second": 10.123,
      "step": 40
    },
    {
      "epoch": 5.0,
      "grad_norm": 2.31677508354187,
      "learning_rate": 4.5e-05,
      "loss": 0.4009,
      "step": 50
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.4657039711191336,
      "eval_loss": 0.947138249874115,
      "eval_runtime": 0.2288,
      "eval_samples_per_second": 1210.554,
      "eval_steps_per_second": 8.74,
      "step": 50
    },
    {
      "epoch": 6.0,
      "grad_norm": 5.374250411987305,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.2719,
      "step": 60
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.48736462093862815,
      "eval_loss": 1.1685609817504883,
      "eval_runtime": 0.1915,
      "eval_samples_per_second": 1446.654,
      "eval_steps_per_second": 10.445,
      "step": 60
    },
    {
      "epoch": 7.0,
      "grad_norm": 3.844560384750366,
      "learning_rate": 4.3e-05,
      "loss": 0.153,
      "step": 70
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.49458483754512633,
      "eval_loss": 1.4269626140594482,
      "eval_runtime": 0.1935,
      "eval_samples_per_second": 1431.625,
      "eval_steps_per_second": 10.337,
      "step": 70
    },
    {
      "epoch": 7.0,
      "step": 70,
      "total_flos": 2293012847462400.0,
      "train_loss": 0.47949272564479284,
      "train_runtime": 44.6997,
      "train_samples_per_second": 2785.254,
      "train_steps_per_second": 11.186
    }
  ],
  "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": 2293012847462400.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}