from transformers import pipeline

# Load the model
classifier = pipeline("text-classification", model="samurai9776/thought-classifier-menu-aware")

# Example usage
result = classifier("What sandwich would you like? [SEP] Chicken")
print(result)
# [{'label': 'COMPLETE', 'score': 0.98}]

result = classifier("Anything else? [SEP] Chicken")
print(result)  
# [{'label': 'INCOMPLETE', 'score': 0.95}]

Performance

Overall Accuracy: ~92%
Handles menu ambiguities (e.g., "chicken" appears in 75 menu items)
Context-aware predictions

Training Details

Fixed 3,000+ mislabeled examples from original dataset
Fine-tuned from samurai9776/thought-classifier
Optimized for restaurant/QSR ordering scenarios