UnicodeDecodeError: 'utf-8' codec can't decode byte 0xd3 in position 0: invalid continuation byte

在windows11上面，我使用pip install mmcv-full -f https://download.openmmlab.com/mmcv/dist/cu113/torch1.11.0/index.html安装了mmcv-full 1.5.0。 之后使用测试脚本测试是否安装成功，出现了如下错误:

C:\Users\ZhangQu\.conda\envs\mmlab\python.exe D:/masterResearch/envtest/check_mmcv.py
Start checking the installation of mmcv-full ...
CPU ops were compiled successfully.
CUDA ops were compiled successfully.
mmcv-full has been installed successfully.

Traceback (most recent call last):
  File "D:\masterResearch\envtest\check_mmcv.py", line 40, in <module>
    env_info_dict = collect_env()
  File "C:\Users\ZhangQu\.conda\envs\mmlab\lib\site-packages\mmcv\utils\env.py", line 91, in collect_env
    env_info['MSVC'] = cc.decode(encoding).partition('\n')[0].strip()
UnicodeDecodeError: 'utf-8' codec can't decode byte 0xd3 in position 0: invalid continuation byte

Process finished with exit code 1

[creater-zq]

@creater-zq 请帮忙测试一下 #1905 是否可以解决该问题，只用修改本地环境对应的代码即可，谢谢

• 正常工作，谢谢

[creater-zq]

@SuTanTank

import sys, os, locale

print(locale.getdefaultlocale())
print(sys.stdout.fileno())
print(os.device_encoding(sys.stdout.fileno()))
print(sys.getdefaultencoding())

• 我在Ubuntu20.04测试结果：

('zh_CN', 'UTF-8')
1
UTF-8
utf-8