docs/setup/noopenmmlab_setup.md

# Install the SegFormer++ without OpenMMLab

**Step 0.** Prerequisites

- Pytorch: 2.3 (CUDA 12.1) (older versions should also work fine)
- Setup a virtual enviroment (optional, python -m venv "name of the environment")

**Step 1.** Install Segformer++

```shell
cd model_without_OpenMMLab
pip install .
```

**Step 2.** [Run the SegFormer++](../run/run_noopenmmlab.md)