Update app.py

app.py

@@ -359,13 +359,13 @@ def generate(
     formatted_prompt = format_prompt(formatted_prompt, history, max_history_turns=5)  # Truncated history
     logging.info(f"Formatted Prompt: {formatted_prompt}")
     stream = client.text_generation(formatted_prompt, temperature=temperature, max_new_tokens=max_new_tokens, top_p=top_p, repetition_penalty=repetition_penalty, stream=True, details=True, return_full_text=False)
-    resp = ""
+resp = ""
     for response in stream:
         resp += response.token.text
+        yield resp  # This allows for streaming the response
 
     if VERBOSE:
-        logging.info(LOG_RESPONSE.format(""))  # Log the response
-    return formatted_prompt # Return resp here
+        logging.info(LOG_RESPONSE.format(resp))  # Pass resp to format
 
 def generate_text_chunked(input_text, model, generation_parameters, max_tokens_to_generate):
     """Generates text in chunks to avoid token limit errors."""