32 lines
932 B
Python
32 lines
932 B
Python
|
# coding=utf-8
|
|||
|
|
"""
|
|||
|
|
@project: maxkb
|
|||
|
|
@Author：虎
|
|||
|
|
@file： llm.py
|
|||
|
|
@date：2024/4/28 11:44
|
|||
|
|
@desc:
|
|||
|
|
"""
|
|||
|
|
from typing import Dict
|
|||
|
|
|
|||
|
|
from models_provider.base_model_provider import MaxKBBaseModel
|
|||
|
|
from models_provider.impl.base_chat_open_ai import BaseChatOpenAI
|
|||
|
|
|
|||
|
|
|
|||
|
|
class QwenChatModel(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 = MaxKBBaseModel.filter_optional_params(model_kwargs)
|
|||
|
|
chat_tong_yi = QwenChatModel(
|
|||
|
|
model_name=model_name,
|
|||
|
|
openai_api_key=model_credential.get('api_key'),
|
|||
|
|
openai_api_base='https://dashscope.aliyuncs.com/compatible-mode/v1',
|
|||
|
|
streaming=True,
|
|||
|
|
stream_usage=True,
|
|||
|
|
**optional_params,
|
|||
|
|
)
|
|||
|
|
return chat_tong_yi
|