conda创建pytorch环境报错

昨天训练数据的时候，发现Anaconda占用C盘达到了20G（暑假在cmd状态下安装的，默认下载到了C盘），心道再创建几个环境，C盘就要爆红了，于是重装Anaconda到了D盘，不过之后的初始化并不顺利--创建pytorch环境总是失败，网上查阅了许多资料：配置path环境变量，关闭VPN，.condarc文件删除或添加路径，更换清华源等等都无果。报错如下：

>>>>>>>>>>>>>>>>>>>>>> ERROR REPORT <<<<<<<<<<<<<<<<<<<<<<Traceback (most recent call last):File "C:\Users\86150\anaconda3\Lib\site-packages\conda\exception_handler.py", line 17, in __call__return func(*args, **kwargs)^^^^^^^^^^^^^^^^^^^^^File "C:\Users\86150\anaconda3\Lib\site-packages\conda\cli\main.py", line 54, in main_subshellparser = generate_parser(add_help=True)^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^File "C:\Users\86150\anaconda3\Lib\site-packages\conda\cli\conda_argparse.py", line 127, in generate_parserconfigure_parser_plugins(sub_parsers)File "C:\Users\86150\anaconda3\Lib\site-packages\conda\cli\conda_argparse.py", line 354, in configure_parser_pluginselse set(find_commands()).difference(plugin_subcommands)^^^^^^^^^^^^^^^File "C:\Users\86150\anaconda3\Lib\site-packages\conda\cli\find_commands.py", line 71, in find_commandsfor entry in os.scandir(dir_path):^^^^^^^^^^^^^^^^^^^^OSError: [WinError 123] 文件名、目录名或卷标语法不正确。: 'D:\\1瀵邦喕淇?缁嬪\ue0c5绨璡瀵邦喕淇妛eb瀵\ue1b9偓閸欐垼鈧\ue100懎浼愰崗绌僤ll'`$ C:\Users\86150\anaconda3\Scripts\conda-script.py config --set show_channel_urls yes`environment variables:CIO_TEST=<not set>CLASSPATH=.;C:\Program Files\Java\jdk1.8.0_221\bin;C:\ProgramFiles\Java\jdk1.8.0_221\lib\dt.jar;C:\ProgramFiles\Java\jdk1.8.0_221\lib\tools.jar;D:\sqljdbc4\sqljdbc4.jarCONDA_DEFAULT_ENV=pytorchCONDA_EXE=C:\Users\86150\anaconda3\condabin\..\Scripts\conda.exeCONDA_EXES="C:\Users\86150\anaconda3\condabin\..\Scripts\conda.exe"CONDA_PREFIX=C:\Users\86150\anaconda3\envs\pytorchCONDA_PREFIX_1=C:\Users\86150\anaconda3CONDA_PROMPT_MODIFIER=(pytorch)CONDA_PYTHON_EXE=C:\Users\86150\anaconda3\python.exeCONDA_ROOT=C:\Users\86150\anaconda3CONDA_SHLVL=2CURL_CA_BUNDLE=<not set>HOMEPATH=\Users\86150LD_PRELOAD=<not set>NODE_PATH=D:\1微信小程序\node\node_global\node_modulesPATH=C:\Users\86150\anaconda3\envs\pytorch;C:\Users\86150\anaconda3\envs\pytorch\Library\mingw-w64\bin;C:\Users\86150\anaconda3\envs\pytorch\Library\usr\bin;C:\Users\86150\anaconda3\envs\pytorch\Library\bin;C:\Users\86150\anaconda3\envs\pytorch\Scripts;C:\Users\86150\anaconda3\envs\pytorch\bin;C:\Users\86150\anaconda3\condabin;E:\;E:\VM\bin;E:\app\Qracleinstall\product\12.2.0\dbhome_1\bin;C:\Program Files\Microsoft MPI\Bin;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;E:\bin;C:\ProgramFiles\Microsoft SQL Server\Client SDK\ODBC\130\Tools\Binn;C:\ProgramFiles (x86)\MySQL\MySQL Server 5.5\bin;C:\ProgramFiles\Java\jdk1.8.0_221\bin;D:\apache-tomcat-9.0.37-windows-x64\apache-tomcat-9.0.37\bin;C:\ProgramFiles\Java\jdk1.8.0_221\jre\bin;C:\Program Files\dotnet;C:\ProgramFiles\Microsoft SQL Server\Client SDK\ODBC\170\Tools\Binn;D:\x64\瀵邦喕淇妛eb瀵偓閸欐垼鈧懎浼愰崗绌僤ll;D:\1瀵邦喕淇?缁嬪绨璡瀵邦喕淇妛eb瀵偓閸欐垼鈧懎浼愰崗绌僤ll;D:\1瀵邦喕淇婄亸蹇曗柤;\node;C:\Program Files (x86)\MySQL;D:\Git\Git\cmd;E:\Xshell;C:\Users\86150\anaconda3;C:\Users\86150\anaconda3\Library\mingw-w64\bin;C:\Users\86150\anaconda3\Library\usr\bin;C:\Users\86150\anaconda3\Library\bin;C:\Users\86150\anaconda3\Scripts;D:\Python\python3.11.5\Scripts;D:\Python\python3.11.5;C:\Users\MySQL\MySQL Server 8.0\bin;C:\Users\86150\AppData\Local\Microsoft\WindowsApps;.;C:\Users\86150\.dotnet\tools;C:\Users\86150\AppData\Local\Microsoft\WindowsApps;D:\1瀵邦喕淇婄亸蹇曗柤鎼村猾node\node_global;E:\Python\PyCharm Community Edition2023.1.3\bin;.;C:\Users\86150\Python 3.11;D:\Microsoft VS Code\binPSMODULEPATH=C:\Program Files\WindowsPowerShell\Modules;C:\WINDOWS\system32\WindowsPowerShell\v1.0\Modules;C:\Users\86150\Desktop\编程;REQUESTS_CA_BUNDLE=<not set>SSL_CERT_FILE=C:\Users\86150\anaconda3\Library\ssl\cacert.pemactive environment : pytorchactive env location : C:\Users\86150\anaconda3\envs\pytorchshell level : 2user config file : C:\Users\86150\.condarcpopulated config files : C:\Users\86150\.condarcconda version : 23.7.4conda-build version : 3.26.1python version : 3.11.5.final.0virtual packages : __archspec=1=x86_64__win=0=0base environment : C:\Users\86150\anaconda3  (writable)conda av data dir : C:\Users\86150\anaconda3\etc\condaconda av metadata url : Nonechannel URLs : https://repo.anaconda.com/pkgs/main/win-64https://repo.anaconda.com/pkgs/main/noarchhttps://repo.anaconda.com/pkgs/r/win-64https://repo.anaconda.com/pkgs/r/noarchhttps://repo.anaconda.com/pkgs/msys2/win-64https://repo.anaconda.com/pkgs/msys2/noarchpackage cache : C:\Users\86150\anaconda3\pkgsC:\Users\86150\.conda\pkgsC:\Users\86150\AppData\Local\conda\conda\pkgsenvs directories : C:\Users\86150\anaconda3\envsC:\Users\86150\.conda\envsC:\Users\86150\AppData\Local\conda\conda\envsplatform : win-64user-agent : conda/23.7.4 requests/2.31.0 CPython/3.11.5 Windows/10 Windows/10.0.19045administrator : Falsenetrc file : Noneoffline mode : FalseAn unexpected error has occurred. Conda has prepared the above report.
If you suspect this error is being caused by a malfunctioning plugin,
consider using the --no-plugins option to turn off plugins.Example: conda --no-plugins install <package>Alternatively, you can set the CONDA_NO_PLUGINS environment variable on
the command line to run the command without plugins enabled.Example: CONDA_NO_PLUGINS=true conda install <package>If submitted, this report will be used by core maintainers to improve
future releases of conda.
Would you like conda to send this report to the core maintainers? [y/N]:No report sent. To permanently opt-out, use$ conda config --set report_errors false

又看到安装Anaconda时，选择JUST ME默认环境路径在annconda/env下，而选择ALL USERS默认环境路径则在C盘.anaconda/env文件夹下。卸载后又重装了Anaconda，先是选择JUST ME，报相同的错误，又重装后选择ALL USERS，创建pytorch还是相同的错误。

无奈之下，再次把Anaconda安装到C盘，这时候执行conda create -n pytorch python=3.8命令是没问题的，环境可以成功创建。conda activate pytorch激活环境也没问题，不过最重要的一步--安装torch又出错了，pytorch官网的conda代码，清华源，华科源都报相同的错误（上述代码）。Google/GitHub/CSDN/Stack Overflow/博客园等等都查了一遍，都没出现过类似的问题。晚上睡觉的时候心想难道要重装下系统？

还好山重水复疑无路，早上在知乎浏览技术贴的时候看到一句话“pip 或者 conda的时候经常被混合使用，这两者也通常被认为是几乎相同的，尽管他们的很多功能是重叠的。“

突然意识到pip install 和conda install的功能应该是相同的。兴致冲冲在Anaconda Prompt命令行键入：pip3 install torch torchvision torchaudio，等待一段时间后torch终于下载成功。又进入python环境中导入的torch包，输出了torch的版本号

>>> import torch
>>> torch.__version__
'2.1.1+cpu'

这下终于尘埃落定，解决掉这个难题了。