Create app.py

app.py

@@ -0,0 +1,37 @@
+import gradio as gr
+from transformers import AutoTokenizer, AutoModelForSequenceClassification, pipeline
+
+# 1. Load model & tokenizer
+MODEL = "microsoft/deberta-v3-small"  # you can fine-tune later on BiScope_Data
+tokenizer = AutoTokenizer.from_pretrained(MODEL)
+model = AutoModelForSequenceClassification.from_pretrained(MODEL, num_labels=2)
+
+# 2. Build pipeline
+detector = pipeline("text-classification", model=model, tokenizer=tokenizer, return_all_scores=True)
+
+# 3. Detection function
+def detect_ai(text):
+    results = detector(text)[0]
+    # Assuming label 0 = Human, 1 = AI
+    human_score = [r["score"] for r in results if r["label"] in ["LABEL_0", "0"]][0]
+    ai_score = [r["score"] for r in results if r["label"] in ["LABEL_1", "1"]][0]
+
+    prediction = "🧑 Human" if human_score > ai_score else "🤖 AI"
+    return {
+        "Prediction": prediction,
+        "Human Probability": round(human_score * 100, 2),
+        "AI Probability": round(ai_score * 100, 2)
+    }
+
+# 4. Gradio UI
+demo = gr.Interface(
+    fn=detect_ai,
+    inputs=gr.Textbox(lines=5, placeholder="Enter text here..."),
+    outputs="json",
+    title="AI vs Human Text Detector",
+    description="Demo AI text detection using Hugging Face Transformers.\n"
+                "Trained/Fine-tuned models can be swapped in for better accuracy on BiScope_Data."
+)
+
+if __name__ == "__main__":
+    demo.launch()