refactor: update model params

apps/setting/models_provider/impl/ollama_model_provider/model/llm.py

@@ -25,6 +25,10 @@ def get_base_url(url: str):
 
 
 class OllamaChatModel(MaxKBBaseModel, ChatOpenAI):
+    @staticmethod
+    def is_cache_model():
+        return False
+
     @staticmethod
     def new_instance(model_type, model_name, model_credential: Dict[str, object], **model_kwargs):
         api_base = model_credential.get('api_base', '')

apps/setting/models_provider/impl/openai_model_provider/model/llm.py

@@ -13,6 +13,11 @@ from setting.models_provider.base_model_provider import MaxKBBaseModel
 
 
 class OpenAIChatModel(MaxKBBaseModel, ChatOpenAI):
+
+    @staticmethod
+    def is_cache_model():
+        return False
+
     @staticmethod
     def new_instance(model_type, model_name, model_credential: Dict[str, object], **model_kwargs):
         optional_params = {}

apps/setting/models_provider/impl/qwen_model_provider/credential/llm.py

@@ -57,4 +57,13 @@ class OpenAILLMModelCredential(BaseForm, BaseModelCredential):
                 'precision': 2,
                 'tooltip': '较高的数值会使输出更加随机，而较低的数值会使其更加集中和确定'
             },
+            'max_tokens': {
+                'value': 800,
+                'min': 1,
+                'max': 2048,
+                'step': 1,
+                'label': '输出最大Tokens',
+                'precision': 0,
+                'tooltip': '指定模型可生成的最大token个数'
+            }
         }

apps/setting/models_provider/impl/qwen_model_provider/model/llm.py

@@ -19,6 +19,10 @@ from setting.models_provider.base_model_provider import MaxKBBaseModel
 
 
 class QwenChatModel(MaxKBBaseModel, ChatTongyi):
+    @staticmethod
+    def is_cache_model():
+        return False
+
     @staticmethod
     def new_instance(model_type, model_name, model_credential: Dict[str, object], **model_kwargs):
         optional_params = {}
@@ -29,7 +33,7 @@ class QwenChatModel(MaxKBBaseModel, ChatTongyi):
         chat_tong_yi = QwenChatModel(
             model_name=model_name,
             dashscope_api_key=model_credential.get('api_key'),
-            **optional_params,
+            model_kwargs=optional_params,
         )
         return chat_tong_yi
 
@@ -61,7 +65,7 @@ class QwenChatModel(MaxKBBaseModel, ChatTongyi):
             if (
                     choice["finish_reason"] == "stop"
                     and message["content"] == ""
-            ):
+            ) or (choice["finish_reason"] == "length"):
                 token_usage = stream_resp["usage"]
                 self.__dict__.setdefault('_last_generation_info', {}).update(token_usage)
             if (

apps/setting/models_provider/impl/tencent_model_provider/model/llm.py

@@ -9,6 +9,9 @@ from setting.models_provider.impl.tencent_model_provider.model.hunyuan import Ch
 
 
 class TencentModel(MaxKBBaseModel, ChatHunyuan):
+    @staticmethod
+    def is_cache_model():
+        return False
 
     def __init__(self, model_name: str, credentials: Dict[str, str], streaming: bool = False, **kwargs):
         hunyuan_app_id = credentials.get('hunyuan_app_id')
@@ -25,7 +28,7 @@ class TencentModel(MaxKBBaseModel, ChatHunyuan):
 
         super().__init__(model=model_name, hunyuan_app_id=hunyuan_app_id, hunyuan_secret_id=hunyuan_secret_id,
                          hunyuan_secret_key=hunyuan_secret_key, streaming=streaming,
-                         temperature=optional_params.get('temperature', None)
+                         temperature=optional_params.get('temperature', 1.0)
                          )
 
     @staticmethod

apps/setting/models_provider/impl/volcanic_engine_model_provider/model/llm.py

@@ -6,6 +6,10 @@ from setting.models_provider.impl.base_chat_open_ai import BaseChatOpenAI
 
 
 class VolcanicEngineChatModel(MaxKBBaseModel, BaseChatOpenAI):
+    @staticmethod
+    def is_cache_model():
+        return False
+
     @staticmethod
     def new_instance(model_type, model_name, model_credential: Dict[str, object], **model_kwargs):
         optional_params = {}