Hi! If you have your vectors exported in word2vec text format from gensim (save_word2vec_format), you can initialize a base model with spacy init vectors:
python -m spacy init vectors en /path/to/vectors.vec /path/to/spacy_vectors
Then use /path/to/spacy_vectors as the base model when training with prodigy:
prodigy docs (for --base-model): https://prodi.gy/docs/recipes#training
spacy docs (for spacy init vectors): https://spacy.io/usage/linguistic-features#adding-vectors