[update] support qwen3-vl

backend/llm_client_backend.py

@@ -106,7 +106,8 @@ class LlmClientBackend(BaseModelBackend):
                         "max_token_len": request.max_tokens,
                         "temperature": request.temperature,
                         "top_p": request.top_p,
-                        "prompt": parsed_prompt
+                        "prompt": parsed_prompt,
+                        "b_video": True
                     }
                 )
             )

services/model_list.py

@@ -93,6 +93,9 @@ class GetModelList:
                         if precompute_len is not None:
                             new_entry['max_context_length'] = precompute_len
 
+                        if '-vl-' in mode:
+                            new_entry['max_context_length'] = 1152
+
                     elif model_type == 'tts':
                         mode_param = model_data.get("mode_param", {})
                         precompute_len = None