Serving icon indicating copy to clipboard operation
Serving copied to clipboard

A flexible, high-performance carrier for machine learning models(『飞桨』服务化部署框架)

Results 218 Serving issues
Sort by recently updated
recently updated
newest added

在python server的config文件里同时配置了http_port=30100,prometheus_port=30100,启动时报错:Failed to run grpc-gateway: prot 30100 is already used,请问有办法通过同一个端口实现服务和监控么

![1690340208326](https://github.com/PaddlePaddle/Serving/assets/5021119/f8d7aa85-86ca-46f8-bac3-b0f324b35b54)

python -m paddle_serving_server. serve --model ./model/mode_serving_rec/ --thread 10 --port 9201 --ir_optim True --use_mkl 请问是这样使用吗,为什么没有一点加速

[input]: python -m pip install --user -U paddle_serving_server paddle_serving_client paddle_serving_app [end of output]: note: This error originates from a subprocess, and is likely not a problem with pip. ERROR: Failed...

mac os : 10.15.7 python: 3.9.17 paddlepaddle: 2.5.0 paddle-serving-app 0.8.3 paddle-serving-client 0.8.3 paddle-serving-server 0.8.3 安装的是cpu版本的,按照https://github.com/PaddlePaddle/PaddleOCR/blob/release/2.6/deploy/pdserving/README_CN.md操作,执行到python3 web_service.py --config=config.yml &>log.txt &的时候报下面的错误 [PipelineServicer] succ init Traceback (most recent call last): File "/Users/xxxxx/projectPath/PaddleOCR/deploy/pdserving/web_service.py", line...

微信群二维码过期了

改造了Serving/core/general-server/op/general_detection_op.cpp增加了输出boxes的代码。但是client端获取不到增加的输出。请问应该怎么改prototxt或者client端实现获取多个输出? 增加部分如下: ```//generate boxes output int box_size_out = box_num*4*2*sizeof(int); void *box_data_out = MempoolWrapper::instance().malloc(box_size_out); if (!box_data_out) { LOG(ERROR)

环境:ubuntu20.04;python3.8;paddlepaddle==2.3.2;PaddleClas2.5;cmake:3.16.3;gcc:9.4.0 使用CPU,与原build_server.sh相比,删除了CUDA相关命令,并对一些命令做出调整 cmake命令如下, ![cmake命令](https://github.com/PaddlePaddle/Serving/assets/101171150/50ba0815-dd8b-4d63-bda3-705e5139889c) 执行后的输出, ![cmake_1](https://github.com/PaddlePaddle/Serving/assets/101171150/269809ae-2412-4e39-8415-5536da13d599) ![cmake_2](https://github.com/PaddlePaddle/Serving/assets/101171150/97d89d04-d235-4f0d-967f-6f8258e0696f) ![cmake_3](https://github.com/PaddlePaddle/Serving/assets/101171150/2a7778b9-7208-4cfe-b391-5373ad24a58d) ![cmake_4](https://github.com/PaddlePaddle/Serving/assets/101171150/957c4ad3-b42d-4138-aadc-2a708a066df0) ![cmake_5](https://github.com/PaddlePaddle/Serving/assets/101171150/22f656cf-d9b2-4a17-9665-052754c11f1f) 存在几处Failed和not found,忽略。 make后的部分输出, ![make_1](https://github.com/PaddlePaddle/Serving/assets/101171150/fcd9f826-a9f2-4c5b-91c9-bec462e7dbff) ![make_2](https://github.com/PaddlePaddle/Serving/assets/101171150/981632e3-cf30-49d7-85ab-39a1613f4c23) ![make_3](https://github.com/PaddlePaddle/Serving/assets/101171150/e6e5d138-2396-4aad-854f-af96ae4a1744) ![make_4](https://github.com/PaddlePaddle/Serving/assets/101171150/eb84a17c-459e-4dbf-9cae-274a43ba8f55) ![make_5](https://github.com/PaddlePaddle/Serving/assets/101171150/8d008072-81a4-4a05-b6ca-dc3deb097269) ![make_6](https://github.com/PaddlePaddle/Serving/assets/101171150/af6497c4-517b-4ae2-ac31-c14539408895) 开启服务后,生成的logPPShiTu.txt如下, ![log_PPShiTu](https://github.com/PaddlePaddle/Serving/assets/101171150/ad9d9ce1-adbb-4efc-b02b-032a9efc2b8f) 执行test_cpp_serving_client.py发送识别请求后, logPPShiTu.txt如下, ![log_PPShiTu_发送请求后](https://github.com/PaddlePaddle/Serving/assets/101171150/b5a66c54-ce92-46af-a951-f27d91b09470) 终端显示如下, ![发送请求后的终端_1](https://github.com/PaddlePaddle/Serving/assets/101171150/261eaefb-9176-4531-8111-6f4ec601978e) ![发送请求后的终端_2](https://github.com/PaddlePaddle/Serving/assets/101171150/40741117-d0e9-453c-a554-faea16f5c033) 报错如下:inference call failed,...

paddleserving部署paddleocr后,运行web_service.py, 使用命令行请求可以访问,但是前端页面发送post请求报错,不能跨域请求,访问失败。 这个web_service怎么添加CORS跨域配置,让前端能够访问?

服务器系统:Ubuntu16.04 CPU核数:16 内存:64G 镜像:registry.baidubce.com/paddlepaddle/serving:0.9.0-cuda10.1-cudnn7-runtime OCR模型:PPOCR-V3 配置文件如下: ``` #rpc端口, rpc_port和http_port不允许同时为空。当rpc_port为空且http_port不为空时,会自动将rpc_port设置为http_port+1 rpc_port: 18090 #http端口, rpc_port和http_port不允许同时为空。当rpc_port可用且http_port为空时,不自动生成http_port http_port: 9999 #worker_num, 最大并发数。当build_dag_each_worker=True时, 框架会创建worker_num个进程,每个进程内构建grpcSever和DAG ##当build_dag_each_worker=False时,框架会设置主线程grpc线程池的max_workers=worker_num worker_num: 20 #build_dag_each_worker, False,框架在进程内创建一条DAG;True,框架会每个进程内创建多个独立的DAG build_dag_each_worker: False dag: #op资源类型, True, 为线程模型;False,为进程模型 is_thread_op:...