Installation

Prerequisites

• Python 3.7+

• PyTorch 1.6+

• CUDA 9.2+

• GCC 5.4+

Prepare the Environment

1. Use conda and activate the environment:

   conda create -n open-mmlab python=3.7 -y
   conda activate open-mmlab
   

2. Install PyTorch

   Before installing MMEngine, please make sure that PyTorch has been successfully installed in the environment. You can refer to PyTorch official installation documentation. Verify the installation with the following command:

   python -c 'import torch;print(torch.__version__)'
   

Install MMEngine

Note

If you only want to use the fileio, registry, and config modules in MMEngine, you can install mmengine-lite, which will only install the few third-party library dependencies that are necessary (e.g., it will not install opencv, matplotlib):

pip install mmengine-lite

Install with mim

mim is a package management tool for OpenMMLab projects, which can be used to install the OpenMMLab project easily.

pip install -U openmim
mim install mmengine

Install with pip

pip install mmengine

Use docker images

1. Build the image

   docker build -t mmengine https://github.com/open-mmlab/mmengine.git#main:docker/release
   

   More information can be referred from mmengine/docker.

2. Run the image

   docker run --gpus all --shm-size=8g -it mmengine
   

Build from source

Build mmengine

# if cloning speed is too slow, you can switch the source to https://gitee.com/open-mmlab/mmengine.git
git clone https://github.com/open-mmlab/mmengine.git
cd mmengine
pip install -e . -v

Build mmengine-lite

# if cloning speed is too slow, you can switch the source to https://gitee.com/open-mmlab/mmengine.git
git clone https://github.com/open-mmlab/mmengine.git
cd mmengine
MMENGINE_LITE=1 pip install -e . -v

Verify the Installation

To verify if MMEngine and the necessary environment are successfully installed, we can run this command:

python -c 'import mmengine;print(mmengine.__version__)'