文章目录
Ubuntu 部署 MMLab
构建基础环境
conda create -n mmlab python=3.7 -y
conda activate mmlab
pip install torch==1.9.0+cu111 torchvision==0.10.0+cu111 torchaudio==0.9.0 -f https://download.pytorch.org/whl/torch_stable.html
pip install mmcv-full -f https://download.openmmlab.com/mmcv/dist/cu111/torch1.9.0/index.html
安装 MMSegmentation
pip install mmsegmentation timm ipython attr
安装 MMDetection
pip install mmdet instaboostfast git+https://github.com/cocodataset/panopticapi.git git+https://github.com/lvis-dataset/lvis-api.git
使用 SegFormer
安装 SegFormer
git clone https://github.com/NVlabs/SegFormer.git
cd SegFormer
pip install -e . --user
下载 SegFormer 模型
pretrained_models :
https://drive.google.com/drive/folders/1b7bwrInTW4VLEm27YawHOAMSMikga2Ia
trained_models :
https://drive.google.com/drive/folders/1GAku0G0iR9DsBxCbfENWMJ27c5lYUeQA
修改源码使其依赖更新到支持最新版本
vi mmseg/__init__.py
修改为
MMCV_MAX = '1.3.0'
->MMCV_MAX = '1.3.10'
使用 SegFormer 进行推理
详见demo/image_demo.py
使用 SOLOv2
安装 SOLOv2
git clone https://github.com/WXinlong/SOLO.git
cd SOLO
pip install -r requirements/build.txt
pip install "git+https://github.com/cocodataset/cocoapi.git#subdirectory=PythonAPI"
pip install -v -e .
使用 SOLOv2 进行推理
详见demo/inference_demo.py