Sherman Xu
Sherman Xu
默认启动用的是fastchat,并发太差了。 后面用vllm 部署,并发提高了很多,并发能支持20-30,勉强能能用了 @CamDX3906 @truthsun22 > 用python安装版,简单好改代码
```php $proxy->filter(new RemoveEncodingFilter())->filter(function ($request, $response, $next) use ($path) { $username = 'admin'; $password = 'admin'; $request = $request->withHeader('Authorization', "Basic " . base64_encode("$username:$password")); $response = $next($request, $response); return $response; }); ```
遇到同样问题,偶尔会出现
@lycfight 代码仅供参考,第一版改的的确有点问题 这是第二版代码 ```python from langchain_community.vectorstores import FAISS from langchain_community.docstore import InMemoryDocstore from langchain_core.documents import Document from qanything_kernel.configs.model_config import VECTOR_SEARCH_TOP_K, FAISS_LOCATION, FAISS_CACHE_SIZE from typing import Optional, Union, Callable, Dict, Any,...