[BugFix] Fix ernie4_5_vl_processor.py and qwen_vl_processor.py can not disable thinking (#4762)

* fix ernie4_5_vl_processor.py and qwen_vl_processor.py

* add unit test

fastdeploy/input/ernie4_5_vl_processor/ernie4_5_vl_processor.py

@@ -233,7 +233,7 @@ class Ernie4_5_VLProcessor(Ernie4_5Processor):
             if chat_template_kwargs:
                 if isinstance(chat_template_kwargs, dict):
                     for k, v in chat_template_kwargs.items():
-                        if k not in request:
+                        if k not in request or request[k] is None:
                             request[k] = v
                 else:
                     raise ValueError("Invalid input: chat_template_kwargs must be a dict")

fastdeploy/input/qwen_vl_processor/qwen_vl_processor.py

@@ -237,7 +237,7 @@ class QwenVLProcessor(TextProcessor):
             if chat_template_kwargs:
                 if isinstance(chat_template_kwargs, dict):
                     for k, v in chat_template_kwargs.items():
-                        if k not in request:
+                        if k not in request or request[k] is None:
                             request[k] = v
                 else:
                     raise ValueError("Invalid input: chat_template_kwargs must be a dict")

tests/input/test_qwen_vl_processor.py

@@ -204,6 +204,26 @@ class TestQwenVLProcessor(unittest.TestCase):
             result["multimodal_inputs"]["image_type_ids"].shape[0], result["multimodal_inputs"]["grid_thw"][:, 0].sum()
         )
 
+    def test_process_request_dict_enable_thinking(self):
+        num_completion_token_ids = 10
+        request = {
+            "request_id": "12345",
+            "completion_token_ids": [1] * num_completion_token_ids,
+            "stop": ["stop", "eof"],
+            "messages": [
+                {
+                    "role": "user",
+                    "content": [
+                        {"type": "text", "text": "Hello"},
+                    ],
+                }
+            ],
+            "chat_template_kwargs": {"enable_thinking": True},
+        }
+
+        result = self.processor.process_request_dict(request, 100)
+        self.assertEqual(result.get("enable_thinking"), False)
+
     def test_prompt(self):
         """
         Test processing of prompt with image and video placeholders