Update app.py

Updated to include both UI and Publicly callable API

app.py

@@ -1,24 +1,31 @@
 import gradio as gr
 from transformers import T5ForConditionalGeneration, T5Tokenizer
 
-# Load your fine-tuned model
+# Load model and tokenizer
 model = T5ForConditionalGeneration.from_pretrained("gcuomo/open-source-ai-t5-liar-lens")
 tokenizer = T5Tokenizer.from_pretrained("gcuomo/open-source-ai-t5-liar-lens")
 
-def classify_statement(statement):
+# Shared prediction function
+def classify(statement):
     prompt = f"summarize: {statement}"
     inputs = tokenizer(prompt, return_tensors="pt", padding=True, truncation=True, max_length=128)
     output = model.generate(**inputs, max_new_tokens=8)
-    label = tokenizer.decode(output[0], skip_special_tokens=True).strip().lower()
-    return label
+    return tokenizer.decode(output[0], skip_special_tokens=True).strip().lower()
 
-# Build the UI
-demo = gr.Interface(
-    fn=classify_statement,
-    inputs=gr.Textbox(lines=2, label="Enter a short statement"),
-    outputs=gr.Textbox(label="Predicted Label"),
-    title="Open Source AI – LIAR Lens",
-    description="A T5 model fine-tuned on the LIAR dataset and Open Source AI examples. Try a claim and see how the model labels it (e.g. true, half-true, pants-fire)."
-)
+# Build UI with Blocks
+with gr.Blocks() as demo:
+    gr.Markdown("## 🤥 Open Source AI – LIAR Lens")
+    
+    with gr.Row():
+        inp = gr.Textbox(label="Enter a statement", lines=2, placeholder="e.g. The Eiffel Tower is in Berlin.")
+        out = gr.Textbox(label="Predicted label")
 
+    btn = gr.Button("Classify")
+    btn.click(fn=classify, inputs=inp, outputs=out)
+
+    # Register for remote access via gradio_client
+    demo.predict = classify  # 👈 this makes remote .predict(...) possible
+
+# Enable queueing and launch
+demo.queue()
 demo.launch()