Update app.py

app.py

@@ -32,10 +32,14 @@ def generate(prompt, history, temperature=0.9, max_new_tokens=9048, top_p=0.95,
 
     formatted_prompt = format_prompt(f"{system_prompt}, {prompt}", history)
     stream = client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=False)
+
     output = ""
     for response in stream:
         output += response.token.text
-    return output
+        yield output
+
+def chat(prompt, history, temperature, max_new_tokens, top_p, repetition_penalty):
+    return generate(prompt, history, temperature, max_new_tokens, top_p, repetition_penalty)
 
 additional_inputs = [
     gr.Slider(label="Temperature", value=0.9, minimum=0.0, maximum=1.0, step=0.05, interactive=True, info="Higher values produce more diverse outputs"),
@@ -47,7 +51,7 @@ additional_inputs = [
 avatar_images = ("https://i.postimg.cc/pXjKKVXG/user-circle.png", "https://i.postimg.cc/qq04Yz93/CL3.png")
 
 gr.ChatInterface(
-    fn=generate,
+    fn=chat,
     chatbot=gr.Chatbot(show_label=True, show_share_button=False, show_copy_button=True, likeable=True, layout="panel", height="auto", avatar_images=avatar_images),
     additional_inputs=additional_inputs,
     title="ConvoLite",