|
|
--- |
|
|
license: bigscience-openrail-m |
|
|
base_model: ehsanaghaei/SecureBERT |
|
|
tags: |
|
|
- generated_from_trainer |
|
|
metrics: |
|
|
- precision |
|
|
- recall |
|
|
- f1 |
|
|
- accuracy |
|
|
model-index: |
|
|
- name: Cyber-ThreaD/SecureBERT-APTNER |
|
|
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/SecureBERT-APTNER |
|
|
|
|
|
This model is a fine-tuned version of [ehsanaghaei/SecureBERT](https://huggingface.co/ehsanaghaei/SecureBERT) on the [APTNER](https://github.com/wangxuren/APTNER) dataset. |
|
|
It achieves the following results on the evaluation set: |
|
|
- Loss: 0.2915 |
|
|
- Precision: 0.5392 |
|
|
- Recall: 0.5871 |
|
|
- F1: 0.5621 |
|
|
- Accuracy: 0.9211 |
|
|
|
|
|
It achieves the following results on the prediction set: |
|
|
- Loss: 0.2404 |
|
|
- Precision: 0.6277 |
|
|
- Recall: 0.6450 |
|
|
- F1: 0.6362 |
|
|
- Accuracy: 0.9367 |
|
|
|
|
|
## 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 | |
|
|
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:| |
|
|
| 0.8252 | 0.59 | 500 | 0.3771 | 0.4383 | 0.4413 | 0.4398 | 0.9112 | |
|
|
| 0.3593 | 1.19 | 1000 | 0.2915 | 0.5392 | 0.5871 | 0.5621 | 0.9211 | |
|
|
| 0.2704 | 1.78 | 1500 | 0.2949 | 0.5480 | 0.6201 | 0.5818 | 0.9203 | |
|
|
| 0.2308 | 2.37 | 2000 | 0.2988 | 0.5524 | 0.6269 | 0.5873 | 0.9187 | |
|
|
| 0.1934 | 2.97 | 2500 | 0.3123 | 0.5365 | 0.6515 | 0.5884 | 0.9152 | |
|
|
| 0.1567 | 3.56 | 3000 | 0.3128 | 0.5702 | 0.6404 | 0.6033 | 0.9210 | |
|
|
| 0.1471 | 4.15 | 3500 | 0.3651 | 0.5379 | 0.6243 | 0.5779 | 0.9117 | |
|
|
| 0.1249 | 4.74 | 4000 | 0.3771 | 0.5363 | 0.6566 | 0.5904 | 0.9125 | |
|
|
| 0.1106 | 5.34 | 4500 | 0.3866 | 0.5624 | 0.6341 | 0.5961 | 0.9156 | |
|
|
| 0.1063 | 5.93 | 5000 | 0.3754 | 0.5731 | 0.6371 | 0.6034 | 0.9191 | |
|
|
| 0.0835 | 6.52 | 5500 | 0.4015 | 0.5551 | 0.6428 | 0.5957 | 0.9165 | |
|
|
| 0.0854 | 7.12 | 6000 | 0.4325 | 0.5461 | 0.6425 | 0.5904 | 0.9138 | |
|
|
| 0.0743 | 7.71 | 6500 | 0.4184 | 0.5642 | 0.6473 | 0.6029 | 0.9179 | |
|
|
| 0.0704 | 8.3 | 7000 | 0.4315 | 0.5613 | 0.6323 | 0.5947 | 0.9172 | |
|
|
| 0.06 | 8.9 | 7500 | 0.4354 | 0.5635 | 0.6401 | 0.5994 | 0.9176 | |
|
|
| 0.0612 | 9.49 | 8000 | 0.4452 | 0.5643 | 0.6452 | 0.6020 | 0.9179 | |
|
|
|
|
|
|
|
|
### 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} |
|
|
} |
|
|
|
|
|
``` |
|
|
|
