Update README.md

README.md

@@ -10,5 +10,30 @@ pinned: false
 license: mit
 short_description: Zero-shot classification means no training data is needed.
 ---
+# 🔍 Zero-Shot Text Classification with BART and XLM-RoBERTa
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+This Hugging Face Space is inspired by the article:  
+🔗 [Zero-Shot Text Classification with BART and XLM-RoBERTa – C# Corner](https://www.c-sharpcorner.com/article/zero-shot-text-classification-with-bart-and-xlm-roberta/)
+
+## 💡 What this app does:
+- Takes any raw text input.
+- Accepts user-defined labels (comma-separated).
+- Uses Hugging Face's `pipeline("zero-shot-classification")` to predict the most relevant label(s) using:
+  - **facebook/bart-large-mnli** or
+  - **joeddav/xlm-roberta-large-xnli**
+
+## 📦 Models Supported
+- `facebook/bart-large-mnli` (English only)
+- `joeddav/xlm-roberta-large-xnli` (Multilingual)
+
+## ✅ Use Cases
+- Categorizing feedback, support tickets, news headlines, etc.
+- Works without any custom training — zero-shot!
+
+## 🛠 How it Works
+The model is prompted with your text and list of labels. It computes the probability of each label being appropriate, and returns scores.
+
+---
+
+Read the full article here:  
+👉 [https://www.c-sharpcorner.com/article/zero-shot-text-classification-with-bart-and-xlm-roberta/](https://www.c-sharpcorner.com/article/zero-shot-text-classification-with-bart-and-xlm-roberta/)