<!-- 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. -->
MALWARE-URL-DETECT
With this model, it detects harmful links created to harm people such as phishing in Turkey. Classifies url addresses as malware and benign.
Type the domain name of the url address in the text field for classification in API: Like this:
"huggingface.com"
To test the model, visit USOM. Harmful links used in Turkey are shared up-to-date on this site.
This model is a fine-tuned version of bert-base-multilingual-cased on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.2122
- Accuracy: 0.945
- Precision: 0.9611
- Recall: 0.9287
- F1: 0.9446
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: 5e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 |
|---|---|---|---|---|---|---|---|
| No log | 1.0 | 63 | 0.2153 | 0.921 | 0.9953 | 0.8475 | 0.9155 |
| No log | 2.0 | 126 | 0.1927 | 0.946 | 0.9669 | 0.9248 | 0.9453 |
| No log | 3.0 | 189 | 0.2122 | 0.945 | 0.9611 | 0.9287 | 0.9446 |
Framework versions
- Transformers 4.28.1
- Pytorch 2.0.0
- Datasets 2.1.0
- Tokenizers 0.13.3
