How to use non-default spacy model?

[betterthanever2]

I'm trying to use spacy with ru_core_news_lg model, but I'm not sure how to enable it. Could be that I'm adding extras in the non-regular way: I modified the Dockerfile to have this additional command:

RUN pip install -U pip setuptools wheel spacy[ru] \
	&& python -m spacy download ru_core_news_lg \
	&& python -m spacy download en_core_web_trf 

And I checked, the installed pipelines are present in the resulting container.

I create models like so:

CREATE MODEL spacy_mod
PREDICT recognition
USING
engine = 'spacy',
linguistic_feature = 'ner',
target_column = 'dbname.tablename.columnname';

I'm consistently getting

 OSError: [E050] Can't find model 'en_core_web_sm'. It doesn't seem to be a Python package or a valid path to a data directory., raised at: /mindsdb/./mindsdb/integrations/libs/ml_handler_process/learn_process.py#113 

How do I force specific model here? I haven't found anything about this in the docs.