How to use other models ?

[Skylixia]

Which files should be changed if loading another pretrained bert with pytorch and other fasttext embeddings ?

[luxinyu1]

Check the newest version of transformers document and you can easily solve the problem https://huggingface.co/transformers/serialization.html#from-pretrained-method and if you want to use other embeddings, you may use a package called genism https://radimrehurek.com/gensim/

[Skylixia]

I have already acquired the model and embeddings I'd like to use and I have read the commands to load them. My question was about the necessary substitutions in this repo to make it work when using different ressources. In other words, which files are dependent on the Bert model and should be adapted, same goes for fasttext.

[luxinyu1]

You could try to adapt cache_dir and wordVecPath in LSBert1.py

[qiang2100]

you can change the path of your file in "run_LSBert1.sh" or "run_LSBert2.sh" file.