Modelpublic

Cyber-ThreaD/CyBERT-AttackER

sourceHugging Faceupdated 2y agoView on Hugging Face

2likes16downloads

Model Card

our_data

This model is a fine-tuned version of SynamicTechnologies/CYBERT on an unknown dataset. It achieves the following results on the evaluation set:

—Loss: 1.6376
—Precision: 0.1972
—Recall: 0.3585
—F1: 0.2545
—Accuracy: 0.6637

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
—trainbatchsize: 2
—evalbatchsize: 2
—seed: 42
—optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
—lrschedulertype: linear
—num_epochs: 10.0

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
2.0809	0.4	500	2.0594	0.5	0.0066	0.0131	0.5298
1.8682	0.81	1000	1.8006	0.1043	0.0807	0.0910	0.5529
1.6332	1.21	1500	1.8356	0.1339	0.1495	0.1412	0.5748
1.468	1.61	2000	1.6261	0.1356	0.1706	0.1511	0.5891
1.401	2.01	2500	1.6943	0.1563	0.1693	0.1625	0.5986
1.1878	2.42	3000	1.6740	0.1194	0.2460	0.1608	0.5976
1.1182	2.82	3500	1.6201	0.1589	0.2196	0.1843	0.6227
0.9677	3.22	4000	1.6241	0.1393	0.2196	0.1704	0.6176
0.9055	3.63	4500	1.5932	0.1317	0.2646	0.1758	0.6158
0.8772	4.03	5000	1.5797	0.1654	0.2804	0.2080	0.6254
0.7224	4.43	5500	1.5723	0.1587	0.2976	0.2070	0.6413
0.7498	4.83	6000	1.5957	0.1794	0.2897	0.2215	0.6496
0.6632	5.24	6500	1.6825	0.1864	0.2751	0.2222	0.6427
0.6139	5.64	7000	1.5827	0.1769	0.3479	0.2345	0.6508
0.6212	6.04	7500	1.5537	0.1778	0.3413	0.2338	0.6526
0.5379	6.45	8000	1.5670	0.1792	0.3307	0.2325	0.6536
0.5376	6.85	8500	1.6113	0.1844	0.3386	0.2388	0.6530
0.5	7.25	9000	1.6432	0.1789	0.3214	0.2299	0.6600
0.4928	7.66	9500	1.6422	0.1881	0.3373	0.2415	0.6609
0.4877	8.06	10000	1.6851	0.2042	0.3360	0.254	0.6654
0.4339	8.46	10500	1.6376	0.1972	0.3585	0.2545	0.6637
0.4303	8.86	11000	1.6364	0.1835	0.3452	0.2397	0.6604
0.4509	9.27	11500	1.6448	0.1983	0.3413	0.2509	0.6664
0.4114	9.67	12000	1.6494	0.1956	0.3505	0.2511	0.6658

Framework versions

—Transformers 4.36.0.dev0
—Pytorch 2.1.0+cu118
—Datasets 2.15.0
—Tokenizers 0.15.0

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}
}