conda镜像源配置
镜像源
华为云
# 1. 添加华为云的 Anaconda 仓库
conda config --add channels https://repo.huaweicloud.com/anaconda/pkgs/free/
conda config --add channels https://repo.huaweicloud.com/anaconda/pkgs/main/
# 2. 添加华为云的 PyTorch 仓库
conda config --add channels https://repo.huaweicloud.com/anaconda/cloud/pytorch/
conda config --add channels https://repo.huaweicloud.com/anaconda/cloud/conda-forge/
conda config --add channels https://repo.huaweicloud.com/anaconda/cloud/msys2/清华源
conda config --set show_channel_urls yes
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/实践示例
下载miniconda
使用华为镜像源下载 x86 版本
wget https://repo.huaweicloud.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
bash Miniconda3-latest-Linux-x86_64.sh配置 pip 全局镜像
使用华为云作为全局的 pip 镜像
# 全局配置
pip config set global.index-url https://mirrors.huaweicloud.com/repository/pypi/simple/激活环境
# 激活环境
source ~/.bashrc创建conda环境
conda create -n dl python=3.10 -y设置搜索时显示通道地址
conda config --set show_channel_urls yes配置 conda 镜像
conda config --add channels https://repo.huaweicloud.com/anaconda/pkgs/free/
conda config --add channels https://repo.huaweicloud.com/anaconda/pkgs/main/
conda config --add channels https://repo.huaweicloud.com/anaconda/cloud/pytorch/
conda config --add channels https://repo.huaweicloud.com/anaconda/cloud/conda-forge/
conda config --add channels https://repo.huaweicloud.com/anaconda/cloud/msys2/删除镜像配置（可选）
conda config --remove-key channels
conda config --remove-key show_channel_urls
conda config --show channels # 如果成功恢复默认，你会看到输出里只有 defaults，或者什么都不显示激活环境
conda activate dl安装环境依赖
conda install pytorch torchvision torchaudio pytorch-cuda=11.8 -c pytorch -c nvidia
uv pip install matplotlib tensorboard scikit-learn numpy pandas seaborn kagglehub