Update app.py

app.py

@@ -1,7 +1,6 @@
 import gradio as gr
 from transformers import pipeline
 
-# Load model
 classifier = pipeline(
     "text-classification",
     model="WarTitan2077/Number-Classifier",
@@ -12,9 +11,7 @@ classifier = pipeline(
 labels = ["Symbolic", "Numeric", "Natural", "Integer", "Rational", "Irrational", "Real", "Prime", "Composite"]
 label_map = {f"LABEL_{i}": label for i, label in enumerate(labels)}
 
-# ✅ This function gets exposed at /run/predict
-@gr.predict()
-def classify_numbers(input1: str, input2: str, input3: str) -> dict:
+def classify_numbers(input1, input2, input3):
     inputs = {"Input1": input1, "Input2": input2, "Input3": input3}
     results = {}
     for name, value in inputs.items():
@@ -26,14 +23,22 @@ def classify_numbers(input1: str, input2: str, input3: str) -> dict:
             results[name] = {}
     return results
 
-# Optional web UI for users
-with gr.Blocks() as demo:
-    with gr.Row():
-        input1 = gr.Textbox(label="Input 1")
-        input2 = gr.Textbox(label="Input 2")
-        input3 = gr.Textbox(label="Input 3")
-    output = gr.JSON()
-    btn = gr.Button("Classify")
-    btn.click(fn=classify_numbers, inputs=[input1, input2, input3], outputs=output)
+# ✅ This creates /predict endpoint usable by Unity
+demo = gr.Interface(
+    fn=classify_numbers,
+    inputs=[
+        gr.Textbox(label="Input 1"),
+        gr.Textbox(label="Input 2"),
+        gr.Textbox(label="Input 3")
+    ],
+    outputs=gr.JSON(label="Predictions"),
+    title="Number Classifier",
+    description="Enter up to three inputs to classify them as Symbolic, Numeric, Prime, etc.",
+    allow_flagging="never",
+    examples=None,
+    cache_examples=False,
+    api_name="/predict"  # ✅ This exposes /predict endpoint
+)
 
-demo.launch()
+if __name__ == "__main__":
+    demo.launch()