You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
At the moment, spacy_initialize use the default model name en. This would become a bit confusing as we are planning to use word-vectors that come with spaCy language models and vectors come only with _lg or _md (#171). Once this is implemented, we expect a number of spacyr users will have multiple models from the same language.
I think we should use a full model name as the default (e.g. the default will be en_core_web_sm) to avoid this confusion.
I agree, let's implement this in the next update. But as a transitional step, we should map the two-character language names to the "default" language models, if those exist, and print a warning.
At the moment,
spacy_initialize
use the default model nameen
. This would become a bit confusing as we are planning to use word-vectors that come with spaCy language models and vectors come only with_lg
or_md
(#171). Once this is implemented, we expect a number ofspacyr
users will have multiple models from the same language.I think we should use a full model name as the default (e.g. the default will be
en_core_web_sm
) to avoid this confusion.That's also what @ines recommends:
https://support.prodi.gy/t/spacy-load-en-download/1858
This will be implemented with #171 (unless vetoed by @kbenoit or someone else).
The text was updated successfully, but these errors were encountered: