命令使用
- 创建新的虚拟环境:
conda create -n env_name
- 创建新的虚拟环境并指定python版本:
conda create -n env_name python=3.6
- 激活虚拟环境:
conda activate env_name
- 退出虚拟环境:
conda deactivate
- 安装软件包(指定版本用==x.y.z):
conda install tensorflow==1.12.0
- 克隆已存在的环境:
conda create -n env_name2 --clone env_name
- 移除环境:
conda remove -n env_name2 --all
- 列出所有虚拟环境
conda env list
- 列出已安装的包
conda list
- 配置清华源
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main/
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/pytorch/
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge/
conda config --set show_channel_urls yes
或者修改用户目录下的 .condarc
文件:
channels:
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main/
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge/
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/pytorch/
- defaults
show_channel_urls: true
conda config --show channels
显示所有源。
- 移除清华源
以conda-forge为例:
conda config --remove channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge/
- 卸载包
conda uninstall tensorflow
- 显示当前conda环境信息
conda info
- conda升级PyTorch
conda update pytorch torchvision
MAC环境下升级CPU版PyTorch:
conda install pytorch torchvision -c pytorch
- conda离线安装pytorch包
conda install --offline pytorch-1.7.0.dev20200727-py3.7_cuda10.1.243_cudnn7.6.3_0.tar.bz2
pytorch nightly版本离线包下载可以参考[7]。
其他软件离线包可以参考[8]。
- conda虚拟环境下python版本降级
conda uninstall python=3.8
conda install python=3.7
- conda环境打包迁移(见参考资料[9])
conda 可以很方便地创建环境的快照或备份:
conda create --name snapshot --clone env_name
或者使用Conda Pack工具打包和还原:
安装Conda Pack方式:
conda install -c conda-forge conda-pack
或
pip install conda-pack
打包一个环境:
# Pack environment env_name into env_name.tar.gz
conda pack -n env_name
# Pack environment env_name into out_name.tar.gz
conda pack -n env_name -o out_name.tar.gz
# Pack environment located at path into env_name.tar.gz
conda pack -p /home/root/path/env_name
重现环境:
# Unpack environment into directory `env_name`
mkdir -p env_name
tar -xzf env_name.tar.gz -C env_name
# Use Python without activating or fixing the prefixes. Most Python libraries will work fine, but things that require prefix cleanups will fail.
./env_name/bin/python
# Activate the environment. This adds `env_name/bin` to your path
source env_name/bin/activate
# Run Python from in the environment
(env_name) $ python
# Cleanup prefixes from in the active environment.
# Note that this command can also be run without activating the environment as long as some version of Python is already installed on the machine.
(env_name) $ conda-unpack
如果是目标机器配置一致的情况下,可以直接将env_name.tar.gz解压出来的env_name文件夹移到/path/miniconda3/envs下,使用
conda env list
即可看到新环境。
使用
conda activate env_name
即可切换到新环境。
- 关闭终端启动时自动打开base环境
conda config --set auto_activate_base false
参考资料
[1] Anaconda介绍、安装及使用教程
[2] Anaconda Individual Edition Docs
[3] Unofficial Windows Binaries for Python Extension Packages
[4] 清华大学开源软件镜像站 - Anaconda 镜像使用帮助
[5] 清华大学开源软件镜像站 Index of /anaconda/cloud/pytorch/linux-64/
[6] Anaconda Cloud - pytorch / packages / pytorch
[7] Anaconda Cloud - pytorch / packages / pytorch-nightly
[8] Anaconda Packages Repo
[9] Conda 环境迁移