mirrors/mmdeploy
1
0
Fork 0
mirror of https://github.com/open-mmlab/mmdeploy.git synced 2025-01-14 08:09:43 +08:00
Code Issues Projects Releases Wiki Activity
mmdeploy/tools/package_tools/README.md
lvhan028 a15365e93f
[Enhancement] build sdk python api in standard-alone manner (#810) 
* build sdk python api in standard-alone manner

* enable MMDEPLOY_BUILD_SDK_MONOLITHIC and MMDEPLOY_BUILD_EXAMPLES in prebuild config

* link mmdeploy to python target when monolithic option is on

* checkin README to describe precompiled package build procedure

* use packaging.version.parse(python_version) instead of list(python_version)

* fix according to review results

* rebase master

* rollback cmake.in and apis/python/CMakeLists.txt

* reorganize files in install/example

* let cmake detect visual studio instead of specifying 2019

* rename whl name of precompiled package

* fix according to review results
2022-08-02 10:23:48 +08:00

1.8 KiB
Raw Blame History

Precompiled package

This document is going to describe the way to build MMDeploy package.

Prerequisites

  • Download and install Miniconda from the official website.

  • Create conda environments for python 3.6, 3.7, 3.8 and 3.9, respectively.

    for PYTHON_VERSION in 3.6 3.7 3.8 3.9
do
  conda create --name mmdeploy-$PYTHON_VERSION python=$PYTHON_VERSION -y
done

  • Prepare MMDeploy dependencies

    Please follow the build-on-Linux guide or build-on-Windows guide to install dependencies of MMDeploy, including PyTorch, MMCV, OpenCV, ppl.cv, ONNX Runtime and TensorRT.

    Make sure the environment variables pplcv_DIR, ONNXRUNTIME_DIR, TENSORRT_DIR, CUDNN_DIR and CUDA_TOOLKIT_ROOT_DIR are exported.

Run precompiled command

  • On Linux platform,

    conda activate mmdeploy-3.6
pip install pyyaml
cd the/root/path/of/mmdeploy
python tools/package_tools/mmdeploy_builder.py tools/package_tools/configs/linux_x64.yaml .

    You will get the precompiled packages mmdeploy-{version}-linux-x86_64-cuda11.1-tensorrt8.2.3.0 and mmdeploy-{version}-linux-x86_64-onnxruntime1.8.1 in the current directory if everything's going well.

  • On Windows platform, open Anaconda Powershell Prompt from the start menu and execute:

    conda activate mmdeploy-3.6
pip install pyyaml
cd the/root/path/of/MMDeploy
python tools/package_tools/mmdeploy_builder.py tools/package_tools/configs/windows_x64.yaml .

    When the build procedure finishes successfully, you will find mmdeploy-{version}-windows-amd64-cuda11.1-tensorrt8.2.3.0 and mmdeploy-{version}-windows-amd64-onnxruntime1.8.1 precompiled packages in the current directory.