| Property | Value |
|---|---|
| Parameter Count | 4.39M |
| License | Apache 2.0 |
| Accuracy | 98.6% |
| F1 Score | 0.9861 |
| Dataset | SetFit/enron_spam |
What is bert-tiny-finetuned-enron-spam-detection?
This model is a lightweight implementation of BERT-Tiny specifically fine-tuned for spam detection using the Enron spam dataset. It represents a highly efficient approach to email classification, achieving remarkable accuracy while maintaining a small parameter footprint.
Implementation Details
Built on the google/bert_uncased_L-2_H-128_A-2 architecture, this model was trained using a carefully optimized process with Adam optimizer, linear learning rate scheduling, and a batch size of 16. The training spanned 4 epochs with a learning rate of 2e-05.
- Precision: 98.51%
- Recall: 98.71%
- F1 Score: 98.61%
- Training Loss: 0.048
Core Capabilities
- High-accuracy spam detection
- Efficient processing with minimal computational requirements
- Optimized for production deployment
- Robust performance on email content analysis
Frequently Asked Questions
Q: What makes this model unique?
The model's standout feature is its ability to achieve near-99% accuracy in spam detection while using only 4.39M parameters, making it highly efficient for deployment in resource-constrained environments.
Q: What are the recommended use cases?
This model is ideal for email filtering systems, content moderation platforms, and any application requiring efficient spam detection with minimal computational overhead.