README.md

Model Information

• Model Type: deberta
• Base Model: microsoft/deberta-v3-base
• Dataset: yelp
• Full Dataset Path: yelp_polarity
• Number of Classes: 2
• Classes: negative, positive
• Training Date: 2025-05-09 05:56:06
• Best Validation Loss: 0.0875
• Test Accuracy: 0.9671