monailabel Installation

 一、安装Anaconda，可参考ubuntu安装conda

二、Creat environment

~/Documents/workspace/下的终端

$conda --version

查看一下conda的版本,若没有出现conda版本信息，则conda安装失败；

$conda create -n monailabel-env python=3.9

$conda activate monailabel-env

如果出现CommandNotFoundError: Your shell has not been properly configured to use 'conda activate'相关的错误，键入source activate之后再次键入conda activate monailabel-env
三、install monailabel

   ~/Documents/workspace/下的终端

$conda activate monailabel-env

$pip install monailabel

$pip freeze

 四、download sample application

      ~/Documents/workspace/下的终端

$monailabel -h
$monailabel apps --download --name radiology --output apps

五、download image study dataset 

         ~/Documents/workspace/下的终端

$monailabel datasets
$monailabel datasets --download --name Task09_Spleen --output datasets

 六、start server

        ~/Documents/workspace/下的终端

$monailabel start_server --app apps/radiology --studies datasets/Task09_Spleen/imagesTr --conf models deepedit

七、install in docker container    

条件：

1.   安装docker，请参考Ubuntu 20.04 上安装 Docker
2.   安装nvidia-dicker2，请参考nvidia-docker2安装
3.   安装pytorch,pytorch与cuda的安装（$nvcc -V可查看当前Cuda的版本信息，注意pytorch与    cuda的版本适配关系）

~/Documents/workspace/下的终端

$nvidia-docker --version 
$docker --version
$sudo docker run --gpus all --rm -ti --ipc=host --net=host projectmonai/monailabel:latest

 

 八、install monailabel-weekly

 ~/Documents/下打开终端

$conda create -n monailabel-weekly-env python=3.9

$conda activate monailabel-weekly-env

$pip install monailabel-weekly

$pip freeze
$monailabel -h
$monailabel apps

九、install monailabel github

        ~/Documents/下打开终端

$conda create -n monailabel-env python=3.9
$conda activate monailabel-env

$git --version
$git clone https://github.com/project-MONAI/MONAILabel.git

 

$cd MONAILabel/
$ll

$pip install -r requirements.txt
$ll

$./monailabel/scripts/monailabel apps
$./monailabel/scripts/monailabel -h