Add support for word2vec keyed vector

[dkuku]

I tried to use this library with a pretrained models from https://github.com/sdadas/polish-nlp-resources?tab=readme-ov-file#word2vec and I found out that these are in keyed vector format which is currently not supported.

[danieldk]

Unless I misremember, the keyed vector format relies heavily on Python pickling, so I don't think that something we would want to support.

[dkuku]

Thanks for the answer.

[sebpuetz]

You might be able to build a finalfusion model in finalfusion-python by loading the KeyedVector model via gensim and constructing the necessary objects. Writing to finalfusion is supported on the Embeddings class through its write() method. You'd then be able to use it in the Rust crate.

Conversion might not be lossless, depending on what kind of "extra" information the gensim model has

[dkuku]

Thanks, it may come useful at some point. I'm trying to polish my rust and nlp skills by creating elixir bindings for finalfusion.