test

环境

• 【FastDeploy版本】： 说明具体的版本，如fastdeploy-linux-gpu-1.0.7
• 【编译命令】
  git clone https://github.com/PaddlePaddle/FastDeploy.git
  cd FastDeploy/python
  export ENABLE_ORT_BACKEND=ON
  export ENABLE_PADDLE_BACKEND=ON
  export ENABLE_OPENVINO_BACKEND=ON
  export ENABLE_VISION=ON
  export ENABLE_TEXT=ON
  export ENABLE_TRT_BACKEND=ON
  export WITH_GPU=ON
  export TRT_DIRECTORY=/Paddle/TensorRT-8.6.1.6
  export CUDA_DIRECTORY=/usr/local/cuda
  export OPENCV_DIRECTORY=/usr/lib/x86_64-linux-gnu/cmake/opencv4 \

python setup.py build
python setup.py bdist_wheel

• 【系统平台】: Linux x64(Ubuntu 22.04)
• 【硬件】： 说明具体硬件型号，如 Nvidia GPU 4090TI， CUDA 11.7 CUDNN 8.9
• 【编译语言】：Python3.8

• 【性能问题】描述清楚对比的方式
  我对单张图片进行测试，每次换一张图片就会构建很长时间，我想知道如果我进行部署推理，如果每次目标（图像/视频）都不一样，那每次都会很长时间，不能只构建一次，后面换图片或者视频就迅速推理预测么？

Start with the build commands in python/setup.py and the reported Linux, CUDA, TensorRT, and OpenVINO configuration. Reproduce the behavior while changing only the input image or video, then determine whether the delay is compilation or inference. Done means the expected one-time build and subsequent deployment behavior is explained, with any required usage guidance identified.