Load computed embeddings data

[kklot]

This is not an issue but rather a question/feature request. Could Top2Vec be used to load an existing data with computed embedding vector for further analyses? I could not find something in the API guide. Thanks!

[ddangelov]

Can you elaborate a little more? Do you mean you have pre-trained document embeddings?

Currently the closes thing is to pass a callable to embedding_model which could be any model.

[kklot]

thanks for the reply, I meant I already have the embedding vectors calculated from other program for each sentence in my data. I wanted to use that and check out other things like clustering, visualization... from your program.