Results BERT base allennlp

2019-10-22 00:15:51,049 - INFO - allennlp.models.archival - archiving weights and vocabulary to /tmp/bert-mal-detect/model.tar.gz
2019-10-22 00:16:20,067 - INFO - allennlp.common.util - Metrics: {
  "best_epoch": 28,
  "peak_cpu_memory_MB": 3421.912,
  "peak_gpu_0_memory_MB": 12,
  "peak_gpu_1_memory_MB": 2528,
  "training_duration": "0:19:41.596926",
  "training_start_epoch": 0,
  "training_epochs": 29,
  "epoch": 29,
  "training_accuracy": 0.7378208300926253,
  "training_loss": 0.5308393001960953,
  "training_cpu_memory_MB": 3421.912,
  "training_gpu_0_memory_MB": 12,
  "training_gpu_1_memory_MB": 2528,
  "validation_accuracy": 0.7391857506361323,
  "validation_loss": 0.5321029019355774,
  "best_validation_accuracy": 0.7512722646310432,
  "best_validation_loss": 0.530970630645752
}

              precision    recall  f1-score   support

           1  0.68041237 0.39919355 0.50317662       496
           2  0.76736924 0.91356877 0.83411116      1076

    accuracy                      0.75127226      1572
   macro avg  0.72389081 0.65638116 0.66864389      1572
weighted avg  0.73993247 0.75127226 0.72969415      1572