Get Started: Install and Run MMEditing

In this section, you will know about:


We recommend that users follow our Best practices to install MMEditing 1.x. However, the whole process is highly customizable. See Customize installation section for more information.


In this section, we demonstrate how to prepare an environment with PyTorch.

MMEditing works on Linux, Windows, and macOS. It requires:

If you are experienced with PyTorch and have already installed it, just skip this part and jump to the next section. Otherwise, you can follow these steps for the preparation.

Step 0. Download and install Miniconda from official website.

Step 1. Create a conda environment and activate it

conda create --name mmedit python=3.8 -y
conda activate mmedit

Step 2. Install PyTorch following official instructions, e.g.

  • On GPU platforms:

    conda install pytorch torchvision cudatoolkit=11.3 -c pytorch
  • On CPU platforms:

    conda install pytorch=1.10 torchvision cpuonly -c pytorch

Best practices

Step 0. Install MMCV using MIM.

pip install -U openmim
mim install 'mmcv>=2.0.0rc1'

Step 1. Install MMEngine.

pip install git+

Step 2. Install MMEditing 1.x . Install MMEditing from the source code.

git clone -b 1.x
cd mmediting
pip3 install -e . -v

Step 5. Verification.

cd ~
python -c "import mmedit; print(mmedit.__version__)"
# Example output: 1.0.0rc1

The installation is successful if the version number is output correctly.


You may be curious about what -e . means when supplied with pip install. Here is the description:

  • -e means editable mode. When import mmedit, modules under the cloned directory are imported. If pip install without -e, pip will copy cloned codes to somewhere like lib/python/site-package. Consequently, modified code under the cloned directory takes no effect unless pip install again. Thus, pip install with -e is particularly convenient for developers. If some codes are modified, new codes will be imported next time without reinstallation.

  • . means code in this directory

You can also use pip install -e .[all], which will install more dependencies, especially for pre-commit hooks and unittests.

Customize installation

CUDA Version

When installing PyTorch, you need to specify the version of CUDA. If you are not clear on which to choose, follow our recommendations:

  • For Ampere-based NVIDIA GPUs, such as GeForce 30 series and NVIDIA A100, CUDA 11 is a must.

  • For older NVIDIA GPUs, CUDA 11 is backward compatible, but CUDA 10.2 offers better compatibility and is more lightweight.

Please make sure the GPU driver satisfies the minimum version requirements. See this table for more information.

note Installing CUDA runtime libraries is enough if you follow our best practices, because no CUDA code will be compiled locally. However, if you hope to compile MMCV from source or develop other CUDA operators, you need to install the complete CUDA toolkit from NVIDIA’s website, and its version should match the CUDA version of PyTorch. i.e., the specified version of cudatoolkit in conda install command.

Install MMCV without MIM

MMCV contains C++ and CUDA extensions, thus depending on PyTorch in a complex way. MIM solves such dependencies automatically and makes the installation easier. However, it is not a must.

To install MMCV with pip instead of MIM, please follow MMCV installation guides. This requires manually specifying a find-url based on PyTorch version and its CUDA version.

For example, the following command install mmcv-full built for PyTorch 1.10.x and CUDA 11.3.

pip install 'mmcv>=2.0.0rc1' -f

Using MMEditing with Docker

We provide a Dockerfile to build an image. Ensure that your docker version >=19.03.

# build an image with PyTorch 1.8, CUDA 11.1
# If you prefer other versions, just modified the Dockerfile
docker build -t mmediting docker/

Run it with

docker run --gpus all --shm-size=8g -it -v {DATA_DIR}:/mmediting/data mmediting

Trouble shooting

If you have some issues during the installation, please first view the FAQ page. You may open an issue on GitHub if no solution is found.

Developing with multiple MMEditing versions

The train and test scripts already modify the PYTHONPATH to ensure the script uses the MMEditing in the current directory.

To use the default MMEditing installed in the environment rather than that you are working with, you can remove the following line in those scripts

PYTHONPATH="$(dirname $0)/..":$PYTHONPATH

Quick run

After installing MMEditing successfully, now you are able to play with MMEditing!

To synthesize an image of a church, you only need several lines of codes by MMEditing!

from mmedit.apis import init_model, sample_unconditional_model

config_file = 'configs/styleganv2/'
# you can download this checkpoint in advance and use a local file path.
checkpoint_file = ''
device = 'cuda:0'
# init a generative model
model = init_model(config_file, checkpoint_file, device=device)
# sample images
fake_imgs = sample_unconditional_model(model, 4)

Or you can just run the following command.

python demo/ \
configs/styleganv2/ \

You will see a new image unconditional_samples.png in folder work_dirs/demos/, which contained generated samples.

What’s more, if you want to make these photos much more clear, you only need several lines of codes for image super-resolution by MMEditing!

import mmcv
from mmedit.apis import init_model, restoration_inference
from mmedit.engine.misc import tensor2img

config = 'configs/esrgan/'
checkpoint = ''
img_path = 'tests/data/image/lq/baboon_x4.png'
model = init_model(config, checkpoint)
output = restoration_inference(model, img_path)
output = tensor2img(output)
mmcv.imwrite(output, 'output.png')

Now, you can check your fancy photos in output.png.

Read the Docs v: zyh/doc-notfound-extend
On Read the Docs
Project Home

Free document hosting provided by Read the Docs.