ValueError: [E024] Got invalid whitespace entity spans after using spacy debug data

[jacky0218]

Hi,
I got some error when training my Spacy NER model(in Japanese).
the error shows that:
ValueError: [E024] Could not find an optimal move to supervise the parser. Usually, this means that the model can't be updated in a way that's valid and satisfies the correct annotations specified in the GoldParse. For example, are all labels added to the model? If you're training a named entity recognizer, also make sure that none of your annotated entity spans have leading or trailing whitespace or punctuation. You can also use the debug data command to validate your JSON-formatted training data. For details, run:
python -m spacy debug data --help

After using
!python -m spacy debug data /content/model-best/config.cfg

I got the 3 invalid whitespace entity spans in Named Entity Recognition like this:

ℹ 8 label(s)
0 missing value(s) (tokens with '-' label)
✘ 3 invalid whitespace entity spans
⚠ Low number of examples for label 'e_1' (17)
⚠ Low number of examples for label 'e_7' (49)
⚠ Low number of examples for label 'e_9' (21)
⚠ Low number of examples for label 'e_4' (12)
⚠ Low number of examples for label 'e_3' (33)
⚠ Low number of examples for label 'e_8' (39)
✔ Examples without occurrences available for all labels
✔ No entities crossing sentence boundaries
Entity spans consisting of or starting/ending with whitespace characters are
considered invalid.

Is there any solution to deal with it?
Thx

[adrianeboyd]

Usually fixing the invalid whitespace spans will resolve this problem. No entity spans should start or end with whitespace, or consist entirely of whitespace. It's just a hard-coded limitation of the ner component.

[jacky0218]

Is there any solution that can fix the invalid whitespace spans?

[adrianeboyd]

There's not a built-in solution to fix it for you automatically, but it shouldn't be hard to check your data. Here's an example: #9029 (comment)

As Paul mentions, it would be nicer if there was better support for this.