Using my own embedding model

[mehrdad-dev]

Can anyone provide an example of how to use our custom-trained embedding model with Top2Vec? Thanks.

[mirorac]

haven't tried this but embedding_model argument of Top2Vec class accepts a callable

[hedgeho]

I would also like a possibility to use other embedding models. It is very unfortunate that contextual_top2vec flag restricts embedding models to only all-MiniLM-L6-v2 and all-mpnet-base-v2.

My corpus consists of documents in several non-English languages, and I would like to use a multilingual embedding model.

To devs: please enable a way to pass a callable or a huggingface url, or a list of pre-embedded docs 🙏