File size: 2,754 Bytes
d03a9bf
 
 
 
 
 
 
 
 
 
571bbbf
d03a9bf
 
 
 
 
 
ef8dc1b
d03a9bf
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ef8dc1b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
---
base_model: SynamicTechnologies/CYBERT
tags:
- generated_from_trainer
metrics:
- precision
- recall
- f1
- accuracy
model-index:
- name: Cyber-ThreaD/CyBERT-CyNER
  results: []
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# Cyber-ThreaD/CyBERT-CyNER

This model is a fine-tuned version of [SynamicTechnologies/CYBERT](https://huggingface.co/SynamicTechnologies/CYBERT) on the [CyNER](https://github.com/aiforsec/CyNER) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.2405
- Precision: 0.4671
- Recall: 0.2810
- F1: 0.3509
- Accuracy: 0.9568

It achieves the following results on the prediction set:
- Loss: 0.2747
- Precision: 0.5442
- Recall: 0.3483
- F1: 0.4248
- Accuracy: 0.9471


## 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 |
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
| 0.2304        | 1.42  | 500  | 0.2405          | 0.4671    | 0.2810 | 0.3509 | 0.9568   |
| 0.1092        | 2.84  | 1000 | 0.2575          | 0.5426    | 0.2848 | 0.3735 | 0.9601   |
| 0.0797        | 4.26  | 1500 | 0.2454          | 0.4701    | 0.3308 | 0.3883 | 0.9576   |
| 0.0615        | 5.68  | 2000 | 0.2669          | 0.4902    | 0.3180 | 0.3857 | 0.9586   |
| 0.0504        | 7.1   | 2500 | 0.2687          | 0.4885    | 0.3525 | 0.4095 | 0.9580   |
| 0.0379        | 8.52  | 3000 | 0.2752          | 0.4656    | 0.3627 | 0.4078 | 0.9573   |
| 0.0339        | 9.94  | 3500 | 0.2828          | 0.4991    | 0.3499 | 0.4114 | 0.9586   |


### 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}
}

```