---
base_model: SynamicTechnologies/CYBERT
tags:
- generated_from_trainer
metrics:
- precision
- recall
- f1
- accuracy
model-index:
- name: anonymouspd/CyBERT-DNRTI
  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-DNRTI

This model is a fine-tuned version of [SynamicTechnologies/CYBERT](https://huggingface.co/SynamicTechnologies/CYBERT) on the [DNRTI](https://github.com/SCreaMxp/DNRTI-A-Large-scale-Dataset-for-Named-Entity-Recognition-in-Threat-Intelligence) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3378
- Precision: 0.5628
- Recall: 0.6439
- F1: 0.6006
- Accuracy: 0.9077

It achieves the following results on the prediction set:
- Loss: 0.2841
- Precision: 0.6301
- Recall: 0.6926
- F1: 0.6599
- Accuracy: 0.9201

## 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 |
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
| 0.8529        | 0.76  | 500  | 0.5937          | 0.4470    | 0.3593 | 0.3984 | 0.8508   |
| 0.5566        | 1.52  | 1000 | 0.5027          | 0.4669    | 0.4196 | 0.4420 | 0.8636   |
| 0.4678        | 2.28  | 1500 | 0.4671          | 0.4706    | 0.4832 | 0.4768 | 0.8694   |
| 0.4038        | 3.04  | 2000 | 0.4320          | 0.4629    | 0.5371 | 0.4972 | 0.8739   |
| 0.3572        | 3.81  | 2500 | 0.4002          | 0.5134    | 0.5394 | 0.5261 | 0.8858   |
| 0.3167        | 4.57  | 3000 | 0.4047          | 0.4691    | 0.6094 | 0.5302 | 0.8826   |
| 0.2987        | 5.33  | 3500 | 0.3761          | 0.5158    | 0.5854 | 0.5484 | 0.8948   |
| 0.2706        | 6.09  | 4000 | 0.3558          | 0.5362    | 0.6066 | 0.5693 | 0.9001   |
| 0.2461        | 6.85  | 4500 | 0.3493          | 0.5511    | 0.5735 | 0.5621 | 0.9028   |
| 0.2311        | 7.61  | 5000 | 0.3526          | 0.5334    | 0.6518 | 0.5867 | 0.9024   |
| 0.2171        | 8.37  | 5500 | 0.3418          | 0.5586    | 0.6407 | 0.5969 | 0.9071   |
| 0.2062        | 9.13  | 6000 | 0.3378          | 0.5628    | 0.6439 | 0.6006 | 0.9077   |
| 0.1972        | 9.89  | 6500 | 0.3384          | 0.5648    | 0.6527 | 0.6056 | 0.9087   |


### 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}
}

```