基于PyTorch的图像分割神经网络segmentation_models_pytorch
github地址:
环境配置
首先创建一个虚拟环境:
conda create -n segmentation python3.7
进入segmentation虚拟环境:
conda activate segmentation
segmentaion库安装:
pip install segmentation-models-pytorch
默认安装pytorch最新版本,卸载后重新安装:
pip install torch1.7.1+cu110 torchvision0.8.2+cu110 torchaudio0.7.2 -f https://download.pytorch.org/whl/torch_stable.html
安装相关依赖库:
pip install albumentations
pip install matplotlib
pip install imageio
pip install opencv-python