Can't encode transcription

[zhoub]

Hi,

I'm using last main branch and trying to train Japanese, prepared some text from news but it generated the following error

If remove all Japanese Zenkaku characters, which means the symbol has same size as standard character, such as （７１） and ３０ , it would be able to train.

Is there anything special to train Japanese ?

Thanks a lot !

[zdenop]

Please provide a test case (all files and the exact steps you made) to reproduce the problem.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.