DistilBERT for Zero Shot Classification

Model Details

• Architecture: DistilBERT
• Training Data: CNN articles
• Accuracy: 0.956
• F1 Score: 0.955

Usage

1. Load the trained model:
   python

   1 2from transformers import AutoTokenizer, AutoModelForSequenceClassification 3 4 tokenizer = AutoTokenizer.from_pretrained("AyoubChLin/DistilBERT_ZeroShot") 5 6 model = AutoModelForSequenceClassification.from_pretrained("AyoubChLin/DistilBERT_ZeroShot") 7
2. Classify text using zero-shot classification:
   python

   1 2 3 from transformers import pipeline 4 5 # Create a zero-shot classification pipeline 6 classifier = pipeline("zero-shot-classification", model=model, tokenizer=tokenizer) 7 8 # Classify a sentence 9 sentence = "The latest scientific breakthroughs in medicine" 10 candidate_labels = ["politics", "sports", "technology", "business"] 11 12 result = classifier(sentence, candidate_labels) 13 14 print(result) 15
   The output will be a dictionary containing the classified label and the corresponding classification score.

About the Author