Megvii Engine Team ae8b38f634 fix(cmake/whl): reduce wheel size as some reason: some version of trt/cudnn need -Wl,--whole-archive attribute to fix cuda issue, target A(CMAKE OBJECT,eg megbrain/megdnn): -Wl,--whole-archive depend on trt/cudnn target B(SHARED LIBS,eg libmegengine.so) --> A with `PUBLIC` depends target C(SHARED LIBS,eg _impertive.so) --> B with `PUBLIC` depends, then ld will force link fatbin section into C, which will be undesired what`s more, attribute PUBLIC/PRIVATE do not take effect to OBJECT library(megbrain/megdnn) what`s more, megengine/megengine_export have to PUBLIC for mgb/imperative target, as SRC/include depends so we pull cudalib depend from megbrain/megdnn to megengine/megengine_export on linux os, to finall target at windows os, for example lar on windows GitOrigin-RevId: `b278a69e1c`		4 years ago
.github	ci(imperative): run imperative python test	4 years ago

ci	ci(imperative): run imperative python test	4 years ago

cmake	fix(mgb/jit): add cmake target MLIRTosa for latest llvm-project	4 years ago

dnn	fix(cmake/whl): reduce wheel size	4 years ago

imperative	fix(cmake/whl): reduce wheel size	4 years ago

scripts	fix(cmake/whl): reduce wheel size	4 years ago

sdk	fix(cmake/whl): reduce wheel size	4 years ago

src	fix(cmake/whl): reduce wheel size	4 years ago

test	fix(cmake/whl): reduce wheel size	4 years ago

third_party	chore(third_party): add range-v3 as third_party dependency	4 years ago

toolchains	fix(ios): fix cmake ios cross build	4 years ago

tools	feat(opr/nvof): add nvof operator	4 years ago

.gitattributes	feat(mge): tensorrt runtime opr	4 years ago

.gitignore	chroe(.gitignore): add more ignore rules for git	4 years ago

.gitmodules	chore(third_party): add range-v3 as third_party dependency	4 years ago

ACKNOWLEDGMENTS	chore(mge/license): remove multipledispatch from ACKNOWLEDGMENTS	4 years ago

CMakeLists.txt	fix(cmake/whl): reduce wheel size	4 years ago

CODE_OF_CONDUCT.md	MegEngine: Initial commit of MegEngine.	5 years ago

CONTRIBUTOR_LICENSE_AGREEMENT.md	MegEngine: Initial commit of MegEngine.	5 years ago

LICENSE	fix(license): correct and format the content of license to make it be detected by GitHub	4 years ago

README.md	Update README.md	4 years ago

README_CN.md	Update README.md	4 years ago

logo.png	MegEngine: Initial commit of MegEngine.	5 years ago

README.md

MegEngine

MegEngine

English | 中文

MegEngine is a fast, scalable and easy-to-use deep learning framework, with auto-differentiation.

Installation

NOTE: MegEngine now supports Linux-64bit/Windows-64bit/MacOS-10.14+ (CPU-Only) Platforms with Python from 3.5 to 3.8. On Windows 10 you can either install the Linux distribution through Windows Subsystem for Linux (WSL) or install the Windows distribution directly.

Binaries

Commands to install from binaries via pip wheels are as follows:

python3 -m pip install megengine -f https://megengine.org.cn/whl/mge.html

Build from Source

Prerequisites

Most of the dependencies of MegEngine are located in third_party directory, which can be prepared by executing:

./third_party/prepare.sh
./third_party/install-mkl.sh

But some dependencies need to be Installed manually:

CUDA(>=10.1), cuDNN(>=7.6)are required when building MegEngine with CUDA support.
TensorRT(>=5.1.5) is required when building with TensorRT support.
LLVM/Clang(>=6.0) is required when building with Halide JIT support.
Python(>=3.5), Numpy, are required to build Python modules.

Build

MegEngine uses CMake as the build tool.
We provide the following scripts to facilitate building.

host_build.sh is to build MegEngine targeted to run on the same host machine.
Please run the following command to get help information:
```
scripts/cmake-build/host_build.sh -h
```
cross_build_android_arm_inference.sh is to build MegEngine targeted to run at Android-ARM platforms.
Please run the following command to get help information:
```
scripts/cmake-build/cross_build_android_arm_inference.sh -h
```
cross_build_linux_arm_inference.sh is to build MegEngine targeted to run at Linux-ARM platforms.
Please run the following command to get help information:
```
scripts/cmake-build/cross_build_linux_arm_inference.sh -h
```
cross_build_ios_arm_inference.sh is to build MegEngine targeted to run iphone/iPad platforms.
Please run the following command to get help information:
```
scripts/cmake-build/cross_build_ios_arm_inference.sh
```

Please refer to BUILD_README.md for more details.

How to Contribute

MegEngine adopts Contributor Covenant to maintain our community. Please read the Code of Conduct to get more information.
Every contributor of MegEngine must sign a Contributor License Agreement (CLA) to clarify the intellectual property license granted with the contributions. For more details, please refer Contributor License Agreement
You can help MegEngine better in many ways:
- Write code.
- Improve documentation.
- Answer questions on MegEngine Forum, or Stack Overflow.
- Contribute new models in MegEngine Model Hub.
- Try a new idea on MegStudio.
- Report or investigate bugs and issues.
- Review Pull Requests.
- Star MegEngine repo.
- Reference MegEngine in your papers and articles.
- Recommend MegEngine to your friends.
- ...

We believe we can build an open and friendly community and power humanity with AI.

How to contact us

Issue: github.com/MegEngine/MegEngine/issues
Email: megengine-support@megvii.com
Forum: discuss.megengine.org.cn
QQ Group: 1029741705
OPENI: openi.org.cn/MegEngine

Resources

License

MegEngine is Licensed under the Apache License, Version 2.0

MegEngine 安装包中集成了使用 GPU 运行代码所需的 CUDA 环境，不用区分 CPU 和 GPU 版。如果想要运行 GPU 程序，请确保机器本身配有 GPU 硬件设备并安装好驱动。如果你想体验在云端 GPU 算力平台进行深度学习开发的感觉，欢迎访问 MegStudio 平台

模型部署模型训练

C++ Cuda Python C SVG other

megengine@megvii.com luzzyzhang@gmail.com acdoge.cao@gmail.com xxr@megvii.com wenjuan@megvii.com 1041563782@qq.com 44659878+flashrunrun@users.noreply.github.com qazazsw@gmail.com 61608041+megvii-mge@users.noreply.github.com 59949790+stoneMo@users.noreply.github.com wkcn@live.cn 942002795@qq.com