search

CODAIT / Identifying-Incorrect-Labels-In-CoNLL-2003

Research into identifying and correcting incorrect labels in the CoNLL-2003 corpus.
Apache License 2.0
12 stars 2 forks source link

Delete all specific skipping messages #27

Closed xuhdev closed 4 years ago

BryanCutler commented 4 years ago

So we currently did not apply these fixes manually?

I don't think we should print out these messages, if it's something we want to do in the future we have issue #4 , do you agree @frreiss ?

xuhdev commented 4 years ago

No we don't actually manually delete them, but I think it makes sense to inform users that these changes are not applied, although they are in the correction file.

BryanCutler commented 4 years ago

@xuhdev I think it would be better to have a column in all_conll_corrections_combined.csv called "Skip Reason" that gives a reason for the correction to be skipped. The script could just look at that column and print the reason. If that's a bit too much work to do right now, I think we should at least change this printout to give a reason why it is not being applied.

xuhdev commented 4 years ago

@BryanCutler Makes sense. I removed all these specific skipping messages now and add more details on what's happening. Now it prints

[WARNING] Could not find [16, 22): 'S Minn': No span begins with 16
[WARNING] Could not find [16, 22): 'S Minn': No span begins with 16