apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-10-06 17:17:14 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity
Files
d0fa2fcf5a94e530453d16f46fd8dbbdecb1bfbc
FastDeploy/docs/en/build_and_install/jetson.md
Jason dffcda513a polish build docs
2022-11-14 20:50:01 +08:00

1.6 KiB
Raw Blame History

How to Build FastDeploy Library on Nvidia Jetson Platform

FastDeploy supports CPU inference with ONNX Runtime and GPU inference with Nvidia TensorRT on Nvidia Jetson platform

How to Build and Install FastDeploy C++ Library

Prerequisite for Compiling on NVIDIA Jetson:

  • gcc/g++ >= 5.4 (8.2 is recommended)
  • cmake >= 3.10.0
  • jetpack >= 4.6.1
git clone https://github.com/PaddlePaddle/FastDeploy.git
cd FastDeploy
mkdir build && cd build
cmake .. -DBUILD_ON_JETSON=ON \
         -DENABLE_VISION=ON \
         -DCMAKE_INSTALL_PREFIX=${PWD}/installed_fastdeploy
make -j8
make install

Once compiled, the C++ inference library is generated in the directory specified by CMAKE_INSTALL_PREFIX

How to Build and Install FastDeploy Python Library

Prerequisite for Compiling on NVIDIA Jetson:

  • gcc/g++ >= 5.4 (8.2 is recommended)
  • cmake >= 3.10.0
  • jetpack >= 4.6.1
  • python >= 3.6

Notice the wheel is required if you need to pack a wheel, execute pip install wheel first.

All compilation options are imported via environment variables

git clone https://github.com/PaddlePaddle/FastDeploy.git
cd FastDeploy/python
export BUILD_ON_JETSON=ON
export ENABLE_VISION=ON

python setup.py build
python setup.py bdist_wheel

The compiled wheel package will be generated in the FastDeploy/python/dist directory once finished. Users can pip-install it directly.

During the compilation, if developers want to change the compilation parameters, it is advisable to delete the build and .setuptools-cmake-build subdirectories in the FastDeploy/python to avoid the possible impact from cache, and then recompile.