Cyber-ThreaD/SecureBERT-CyNER
This model is a fine-tuned version of ehsanaghaei/SecureBERT on the CyNER dataset. It achieves the following results on the evaluation set:
- Loss: 0.0685
- Precision: 0.7334
- Recall: 0.8046
- F1: 0.7674
- Accuracy: 0.9836
It achieves the following results on the prediction set:
- Loss: 0.0993
- Precision: 0.7181
- Recall: 0.7564
- F1: 0.7367
- Accuracy: 0.9761
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10.0
Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|---|---|---|---|---|---|---|---|
| 0.1929 | 1.42 | 500 | 0.0685 | 0.7334 | 0.8046 | 0.7674 | 0.9836 |
| 0.048 | 2.84 | 1000 | 0.0745 | 0.8054 | 0.7931 | 0.7992 | 0.9837 |
| 0.0299 | 4.26 | 1500 | 0.0720 | 0.7936 | 0.8493 | 0.8205 | 0.9857 |
| 0.0199 | 5.68 | 2000 | 0.0846 | 0.8049 | 0.8327 | 0.8186 | 0.9848 |
| 0.014 | 7.1 | 2500 | 0.0878 | 0.7909 | 0.8455 | 0.8173 | 0.9847 |
| 0.0098 | 8.52 | 3000 | 0.0907 | 0.7830 | 0.8250 | 0.8035 | 0.9845 |
| 0.0073 | 9.94 | 3500 | 0.0917 | 0.7946 | 0.8301 | 0.8120 | 0.9852 |
Framework versions
- Transformers 4.36.0.dev0
- Pytorch 2.1.0+cu118
- Datasets 2.14.6
- Tokenizers 0.14.1
Citing & Authors
If you use the model kindly cite the following work
@inproceedings{deka2024attacker,
title={AttackER: Towards Enhancing Cyber-Attack Attribution with a Named Entity Recognition Dataset},
author={Deka, Pritam and Rajapaksha, Sampath and Rani, Ruby and Almutairi, Amirah and Karafili, Erisa},
booktitle={International Conference on Web Information Systems Engineering},
pages={255--270},
year={2024},
organization={Springer}
}