ERROR: 'gbk' codec can't decode byte 0xa1 in position 82: illegal multibyte sequence

[HGladiator]

Environment:

• NNI version:1.8
• NNI mode (local|remote|pai):local
• Client OS:windows10.1809
• Server OS (for remote mode only):
• Python version:3.7.6
• PyTorch/TensorFlow version:tensorflow. 2.0.0
• Is conda/virtualenv/venv used?:conda
• Is running in Docker?:no

Log message:

• nnictl stdout and stderr:
  ERROR: 'gbk' codec can't decode byte 0xa1 in position 82: illegal multibyte sequence

What issue meet, what's expected?:
when i run cmdnnictl create --config config_windows.yml,then stdout decode error,nii cann`t start.

How to reproduce it?:
only remove chinese Hanzi Comments from config_windows.yml, nni can normal start

Additional information:
eg. maxTrialNum: #可选项: local, remote, pai, kubeflow, remove "可选项"

[JunweiSUN]

@HGladiator this issue may related to the encoding of your cmd/powershell, please change your command line's encoding to UTF-8 and try again.

[HGladiator]

my code file is utf-8, use git bash.

[liuzhe-lz]

We are removing non-ascii characters from example yaml scripts. This needs to go through the translation process so it may take some time.

[JunweiSUN]

my code file is utf-8, use git bash.

you may want to change your git bash encoding to UTF-8: right click on git bash -> options -> Text -> change Character set to UTF-8 -> Save, then try again.

[HGladiator]

my code file is utf-8, use git bash.

you may want to change your git bash encoding to UTF-8: right click on git bash -> options -> Text -> change Character set to UTF-8 -> Save, then try again.

yes. it`s utf-8, the gbk ERROR reappeared.

[5wuuy5]

I meet this problem also. Have you fixed it?

[liuzhe-lz]

I meet this problem also. Have you fixed it?

The Chinese characters inside example config file are merely comments. You can safely remove them.

[Jia-py]

I met this problem today and found no way to solve it.
However, after turning off my computer for a while. It works. Hope this help.