Get Similarity of two sentences based gensim word2vec trained model.
- A working Example exist on example directory.
- Docs available here
For saving your model from Gensim to our Module use this steps:
1- Train your Word2Vec model in Python Gensim.
2- Then store it as a plain text Google Word2Vec format.
model.wv.save_word2vec_format('./data/text.model', binary = False)
model.intersect_word2vec_format('./data/text.model', lockf=0.0, binary=False, encoding='utf8', unicode_errors='strict')
Aether: Amin Aghabeiki.
License MIT.
If you updated/improved it, please keep the code open source as is it! let make Open Source Greater!