diff --git a/llm/docs/FastDeploy_usage_tutorial.md b/llm/docs/FastDeploy_usage_tutorial.md index 7cba22133..cda82dfeb 100644 --- a/llm/docs/FastDeploy_usage_tutorial.md +++ b/llm/docs/FastDeploy_usage_tutorial.md @@ -47,7 +47,7 @@ FastDeploy 为 PaddleNLP 静态图模型提供了高效的部署方案,模型 ``` cd /home/workspace/models_dir -# 导出的模型目录结构格式如下所示,理论上无缝支持 PaddleNLP 导出静态图模型,无需修改模型目录结构 +# 导出的模型目录结构格式如下所示,理论上无缝支持 PaddleNLP 导出的静态图模型,无需修改模型目录结构 # /opt/output/Serving/models # ├── config.json # 模型配置文件 # ├── xxxx.model # 词表模型文件 @@ -138,10 +138,12 @@ bash start_server.sh ``` # port为上面启动服务时候指定的HTTP_PORT + > 测试前请确保服务IP和端口正确 + live接口: (服务是否能正常接收请求) - http://{ip}:{HTTP_PORT}/v2/health/live + http://127.0.0.1:8110/v2/health/live health接口:(模型是否准备好推理) - http://{ip}:{HTTP_PORT}/v2/health/ready + http://127.0.0.1:8110/v2/health/ready ``` ## 服务测试