I created word vectors with fastText and tried to convert them into a spaCy model with the command python -m spacy init-model de C:\my\output\dir --vectors-loc de-vec.vec
But I get a ValueError: could not broadcast input array from shape (99) into shape (100):
What shall I do?
EDIT: I solved this problem, it was in my fastText vectors. There was one vector for a single space.
Thanks for updating with your solution! I remember that something similar came up a while ago on the spaCy issue tracker. Usually, a shape mismatch of 99 vs 100 is a pretty good indicator that there's something wrong in the data or that the header isn't parsed correctly (and read in like vector or something like that). Encoding issues are also a common problem, e.g. if the encoding isn't set correctly on the vectors file.