Update app.py

app.py

@@ -17,17 +17,26 @@ Assistant:
 
 def respond(message, history, system_message, max_tokens, temperature, top_p):
     formatted_prompt = format_alpaca_prompt(history, message, system_message)
+    cleaned_response = ""
 
-    response = client.text_generation(
-        formatted_prompt,
-        max_new_tokens=max_tokens,
-        temperature=temperature,
-        top_p=top_p,
-    )
+    # Retry mechanism to regenerate response if it is empty
+    for _ in range(3):  # Retry up to 3 times
+        response = client.text_generation(
+            formatted_prompt,
+            max_new_tokens=max_tokens,
+            temperature=temperature,
+            top_p=top_p,
+        )
+
+        # Extract only the response
+        cleaned_response = response.split("Assistant:")[-1].strip()
+
+        if cleaned_response:
+            break  # Exit loop if a valid response is generated
+
+    if not cleaned_response:
+        cleaned_response = "I'm sorry, I couldn't generate a response. Please try again."
 
-    # Extract only the response
-    cleaned_response = response.split("Assistant:")[-1].strip()
-    
     yield cleaned_response  # Output only the answer
 
 demo = gr.ChatInterface(