facebookresearch / fastText

Library for fast text representation and classification.
https://fasttext.cc/
MIT License
25.93k stars 4.72k forks source link

Use pretrained chinese word-vectors for training supervised classifier #1201

Open azamatolegen opened 3 years ago

azamatolegen commented 3 years ago

Hello I want to train supervised classifier, but my dataset is very small, can I use pre-trained chinese word vectors during the training? If yes how to do that? I saw in documentation pretrainedVectors argument, what we should provide to this .bin or text file, or can we just input something like 'en' for english or 'zh' for chinese so the model would download the corresponding pre-trained vectors?

miso-belica commented 3 years ago

I think this should work fastext supervised -input your-file.txt -pretrainedVectors pretained-vectors.vec -output model.

You have to download the correct pre-trained vectors yourself.