Adriane Boyd

The problem is a combination of too-old-torch + too-new-numpy. numpy v1.23 works with torch v1.10+, but not torch v1.9.

There is no particularly good way to specify this in the package dependencies.

Daniël said that it's because the older version does not have `__dlpack__` on arrays, so older numpy does not hit this code.

We haven't had any external reports of this so far, so my guess is that it's a pretty unusual version combination to have. If we add something, only as a...

Hi, this code looks like it was written for spaCy v2.x. The version of the course at https://course.spacy.io should be compatible with `spacy==2.3.5`. It's possible this is a slightly older...

Sorry, some of the tag maps haven't been tested well for unsupported morphological features, in particular for languages where spacy doesn't have provided models, since we don't train a tagger...

The background is that we originally developed `spacy-stanza` before NER components were added, so we focused on providing access to the morpho-syntactic annotation, which is annotated on the expanded multi-word...

I think it's fine to leave it open. It's not going to be a high priority for us to work on right now, but since I think it should be...

Hi, you can use python `warnings` filters to manage how these warnings are handled: https://docs.python.org/3/library/warnings.html#the-warnings-filter

Concerns: * ~~`use_fast` should be saved correctly in the tokenizer settings (as a separate PR, #339)~~ * ~~naming of `_align` and new methods~~ * whether we (initially) want an automatic...