prodigy / spacy / sense2vec version mismatches, requesting `pip freeze` or requirements.txt , thanks!

,
1

I've been trying to follow some of the tutorials and it has been a nightmare getting compatible library versions .
pip install spacy, installs spacy 3, which isn't compatible with prodigy.
installing sense2vec without specifying a version, installs a version that requires spacy 3

Could you please post a pip freeze of your working environment so that i can match the versions ?

thanks

2

Hi! Prodigy should always know which spaCy version its compatible with and it will install this automatically.

It seems like installing sense2vec via pip will automatically upgrade spaCy, which is kind of unfortunate (I was hoping that the new dependency resolver in pip could handle this better and give you the best-matching version for your environment). But the versioning here is actually pretty simple: sense2vec==1.0.3 is the latest version for spaCy v2, and sense2vec==2.0.0 is the latest version for spaCy v3. So if you install 1.0.3 explicitly in your environment, pip shouldn't be upgrading anything under th hood.