diff --git a/modules/text_generation.py b/modules/text_generation.py
index 7535d141..0d2f55c2 100644
--- a/modules/text_generation.py
+++ b/modules/text_generation.py
@@ -256,14 +256,14 @@ def generate_reply_HF(question, original_question, seed, state, eos_token=None,
         # This is based on the trick of using 'stopping_criteria' to create an iterator.
         else:
 
-            def generate_with_callback(callback=None, **kwargs):
+            def generate_with_callback(callback=None, *args, **kwargs):
                 kwargs['stopping_criteria'].append(Stream(callback_func=callback))
                 clear_torch_cache()
                 with torch.no_grad():
                     shared.model.generate(**kwargs)
 
             def generate_with_streaming(**kwargs):
-                return Iteratorize(generate_with_callback, kwargs, callback=None)
+                return Iteratorize(generate_with_callback, [], kwargs, callback=None)
 
             with generate_with_streaming(**generate_params) as generator:
                 for output in generator: