Unexpected argument for the fit method.

[gam-ba]

Hello!
Thanks for the awesome work.

I've tried to run the model on a list of very short documents (6 tokens on average per document), but I've been asked for a "vocab_size" argument when trying to fit it.

What's the role of that argument? Is it the vocabulary size as in most document-term matrices?

Thank you very much!

Guido