Closed AngledLuffa closed 8 months ago
also, in the test set, there's reciept
Thanks for pointing these out, let's fix after the data freeze
Oct
as an abbreviation when not treated as a date (not sure why, honestly, usually lemmatized asOctober
). Happens 3x:
Just in filenames, when tagged as X
, maybe? We should really figure out a consistent policy for filenames; I'm not sure we want to analyze them as containing individual "words" in the first place (UniversalDependencies/docs#666).
this should probably be a typo for
patience
?
Yes, it is spelled correctly 3 sentences later
this should probably be a typo for patience?
Yes, it is spelled correctly 3 sentences later
You could say I wasn't ... determined? ... enough to find that
Found some cases of inconsistency in the lemmas for typos
Similarly,
recieved
is sometimes corrected, sometimes notOct
as an abbreviation when not treated as a date (not sure why, honestly, usually lemmatized asOctober
). Happens 3x:this should probably be a typo for
patience
? or maybe it's meant to be a pun, in which case the lemma should bepatient
and then this
enchiladas
is anNNS
and should be singular in the lemmaI would send in a PR, but it's the data freeze, after all