|
--- |
|
license: mit |
|
library_name: transformers |
|
pipeline_tag: text-classification |
|
tags: |
|
- URL |
|
- BinaryCalssification |
|
- URL Classification |
|
- BERT URL calssifier |
|
|
|
widget: |
|
- text: "torgi.kz/help/id2.txt" |
|
|
|
- text: "vvps.ws/44/mothersdarlingcross.php?" |
|
|
|
- text: "torrentreactor.net/torrents/897644/ANGELCORPSE-4-(All)-Studio-Albums%5EWaPo" |
|
|
|
--- |
|
|
|
|
|
|
|
🔒🤖 **Binary URL Classifier with BERT** |
|
|
|
Welcome to our Binary URL Classifier! 🎉 |
|
|
|
### What it Does? |
|
|
|
This classifier categorizes URLs into two categories: Good (Non-Malicious) and Bad (Malicious). It uses BERT, a powerful language model, to understand the content and context of URLs for accurate predictions. |
|
|
|
### How it Works? |
|
|
|
1. **Data**: Trained on a curated dataset of good and bad URLs from this repo https://github.com/faizann24/Using-machine-learning-to-detect-malicious-URLs/tree/master?tab=readme-ov-file. |
|
|
|
2. **Training**: Fine-tuned a pre-trained BERT model using Hugging Face's Transformers library. |
|
|
|
3. **Evaluation**: Validation(15%) and Test(15%) set kept Aside. |
|
|
|
4. **Inference**: Provides probability scores for new URLs, indicating their likelihood of being good or bad. |
|
|
|
### Usage 🚀 |
|
|
|
1. **Input**: Provide a URL to classify. |
|
|
|
2. **Output**: Receive a probability score for its safety. |
|
|
|
### Why Trust Us? 🌟 |
|
|
|
- **Robust Training**: Trained on diverse and representative datasets. |
|
|
|
- **Evaluation**: Rigorously tested for accuracy and reliability. |
|
|
|
- **Simple & Effective**: Empowers users to make informed decisions about URL safety. |
|
|
|
### Let's Make the Web Safer! 🌐 |
|
|
|
Our Binary URL Classifier with BERT is here to enhance online safety. Whether you're a security professional, developer, or concerned user, it helps identify malicious URLs for proactive protection. |
|
|
|
Try it now and experience the power of BERT in URL classification! We're open to feedback and suggestions for continuous improvement. Stay safe online! 🔒🛡️ |
