docs/setup/mmeng_setup.md

Install the SegFormer++ without MMSegmentation/MMPose

Step 0. Prerequisites

• Pytorch: 2.3 (CUDA 12.1) (older versions should also work fine)

Step 1. Install MMCV using MIM.

pip install -U openmim
mim install mmengine
mim install "mmcv>=2.0.0"

Step 2. Install Segformer++

cd model
pip install .

Step 3. Run the SegFormer++

Troubleshooting There might be installation troubles with openmim, mmengine, and mmcv for new python versions. Thus, Step 1 might not work correctly. In this case, try the following alternative for step 1:

pip install torch torchvision numpy
pip install mmcv-full==1.2.7
pip install mmcv

If it is still not working, make sure to use a pip virtual environment (python -m venv "name of the environment"), not a conda environment.