BERT-Banking77
| Property | Value |
|---|---|
| Author | philschmid |
| Task Type | Text Classification |
| Accuracy | 92.76% |
| CO2 Emissions | 0.033g |
What is BERT-Banking77?
BERT-Banking77 is a specialized BERT-based model fine-tuned for banking-related text classification tasks. Trained on the Banking77 dataset, it excels at categorizing customer service queries and banking-related text with high precision and recall scores.
Implementation Details
The model achieves impressive performance metrics, including 92.76% accuracy and F1 score, with a notably low loss of 0.32. It's implemented using PyTorch and the Transformers library, making it easily accessible for deployment in production environments.
- Macro Precision: 93.05%
- Macro Recall: 92.76%
- Weighted F1: 92.76%
- Environmental Impact: Minimal with only 0.033g CO2 emissions during training
Core Capabilities
- Banking-specific text classification
- Customer service query categorization
- Multi-class classification across banking categories
- Real-time inference through API endpoints
Frequently Asked Questions
Q: What makes this model unique?
This model combines BERT's powerful language understanding capabilities with specialized training on banking-sector data, achieving over 92% accuracy while maintaining a minimal environmental footprint.
Q: What are the recommended use cases?
The model is ideal for banking customer service automation, query routing, and financial text classification tasks. It's particularly effective for categorizing customer inquiries and automating response systems in banking applications.
