[GLM4.6v] Required changes for bumping up to transformer 5.x (#13229)

python/sglang/srt/entrypoints/openai/serving_chat.py

@@ -324,6 +324,7 @@ class OpenAIServingChat(OpenAIServingBase):
                 **(
                     request.chat_template_kwargs if request.chat_template_kwargs else {}
                 ),
+                return_dict=False,
             )
         except Exception:
             # This except branch will be triggered when the chosen model
@@ -343,6 +344,7 @@ class OpenAIServingChat(OpenAIServingBase):
                 **(
                     request.chat_template_kwargs if request.chat_template_kwargs else {}
                 ),
+                return_dict=False,
             )
 
         if assistant_prefix:

python/sglang/srt/model_loader/weight_utils.py

@@ -70,7 +70,8 @@ enable_hf_transfer()
 
 class DisabledTqdm(tqdm):
     def __init__(self, *args, **kwargs):
-        super().__init__(*args, **kwargs, disable=True)
+        kwargs["disable"] = True
+        super().__init__(*args, **kwargs)
 
 
 def get_lock(model_name_or_path: str, cache_dir: Optional[str] = None):