我也想能把你照亮，在你的生命中留下阳光
0.引子
在用pip或者conda安装一些包时有时会因为网络原因导致下载失败，进而无法安装。一般的解决方法就是换源，或者重复安装。
在本篇文章，将使用一个简单高效的方式来解决这个问题。整体思路很简单，把包下载到本地，在使用本地包进行安装。
这样有什么好处呢？其实这解决了直接使用安装时网络不通畅导致中断，又要重新开始下载的问题。因为很多下载工作都支持断点续传，不用重复操作。
1.实战
例子一：conda
当直接安装时：
代码语言：javascript复制conda install pytorch=1.4 torchvision cudatoolkit=10.0 cudnn=7 -y输出：
代码语言：javascript复制The following packages will be SUPERSEDED by a higher-priority channel:
pytorch anaconda/pkgs/main::pytorch-1.5.0-cpu~ --> anaconda/cloud//pytorch::pytorch-1.4.0-py3.7_cuda10.0.130_cudnn7.6.3_0
Downloading and Extracting Packages
pytorch-1.4.0 | 422.7 MB | ####################################################3 | 25%
CondaHTTPError: HTTP 000 CONNECTION FAILED for url <https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud//pytorch/linux-64/pytorch-1.4.0-py3.7_cuda10.0.130_cudnn7.6.3_0.tar.bz2>
Elapsed: -
An HTTP error occurred when trying to retrieve this URL.
HTTP errors are often intermittent, and a simple retry will get you on your way.可以看出是网络故障导致下载安装失败。细心的小伙伴同时也可以看到一个很有用信息：
CondaHTTPError: HTTP 000 CONNECTION FAILED for url < https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud//pytorch/linux-64/pytorch-1.4.0-py3.7_cuda10.0.130_cudnn7.6.3_0.tar.bz2>
这时可以把此链接拷贝到浏览器中下载到本地。笔者习惯使用wget：
代码语言：javascript复制wget https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud//pytorch/linux-64/pytorch-1.4.0-py3.7_cuda10.0.130_cudnn7.6.3_0.tar.bz2输出：
代码语言：javascript复制Saving to: 'pytorch-1.4.0-py3.7_cuda10.0.130_cudnn7.6.3_0.tar.bz2'
pytorch-1.4.0-py3.7_cuda10.0.130_cudnn7.6.3_0.tar.bz2 100%[ =========================================================>] 422.70M 2.46MB/s in 80s
2020-09-10 03:58:45 (2.06 MB/s) - 'pytorch-1.4.0-py3.7_cuda10.0.130_cudnn7.6.3_0.tar.bz2' saved [443230931/443230931]下载完后，仅需使用conda本地安装指令即可：conda install --use-local
代码语言：javascript复制conda install --use-local pytorch-1.4.0-py3.7_cuda10.0.130_cudnn7.6.3_0.tar.bz2输出：
代码语言：javascript复制Downloading and Extracting Packages
################################################################################################################################################################################################################################################## | 100%
Preparing transaction: done
Verifying transaction: done
Executing transaction: done例子二：pip
使用方法和conda类似，先直接安装一个包
代码语言：javascript复制pip install tensorflow==2.3输出：
代码语言：javascript复制Looking in indexes: https://pypi.tuna.tsinghua.edu.cn/simple
Collecting tensorflow==2.3
Downloading https://pypi.tuna.tsinghua.edu.cn/packages/16/89/f2d29c2eafc2eeafb17d5634340e06366af904d332341200a49d954bce85/tensorflow-2.3.0-cp37-cp37m-manylinux2010_x86_64.whl (320.4 MB)
|▏ | 1.4 MB 1.7 MB/s eta 0:03:06^C
ERROR: Operation cancelled by user关键信息如下：
Downloading https://pypi.tuna.tsinghua.edu.cn/packages/16/89/f2d29c2eafc2eeafb17d5634340e06366af904d332341200a49d954bce85/tensorflow-2.3.0-cp37-cp37m-manylinux2010_x86_64.whl (320.4 MB
下载到本地方法如例子一
代码语言：javascript复制wget https://pypi.tuna.tsinghua.edu.cn/packages/16/89/f2d29c2eafc2eeafb17d5634340e06366af904d332341200a49d954bce85/tensorflow-2.3.0-cp37-cp37m-manylinux2010_x86_64.whl输出：
代码语言：javascript复制--2020-09-10 04:11:44-- https://pypi.tuna.tsinghua.edu.cn/packages/16/89/f2d29c2eafc2eeafb17d5634340e06366af904d332341200a49d954bce85/tensorflow-2.3.0-cp37-cp37m-manylinux2010_x86_64.whl
Resolving pypi.tuna.tsinghua.edu.cn (pypi.tuna.tsinghua.edu.cn)... 101.6.8.193, 2402:f000:1:408:8100::1
Connecting to pypi.tuna.tsinghua.edu.cn (pypi.tuna.tsinghua.edu.cn)|101.6.8.193|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 320368291 (306M) [application/octet-stream]
Saving to: 'tensorflow-2.3.0-cp37-cp37m-manylinux2010_x86_64.whl'
tensorflow-2.3.0-cp37-cp37m-manylinux2010_x86_64.whl 100%[==================================================================================================================================================>] 305.53M 3.68MB/s in 2m 48s
2020-09-10 04:14:33 (1.82 MB/s) - 'tensorflow-2.3.0-cp37-cp37m-manylinux2010_x86_64.whl' saved [320368291/320368291]下载完成后，使用 pip install 即可
代码语言：javascript复制pip install tensorflow-2.3.0-cp37-cp37m-manylinux2010_x86_64.whl
