Zero shot - information extraction with the Gliner Method

[CP500]

Hi,
Ive found model2vec extremely performant for embedding. Wondering if we can apply it for fast information extraction. If you embed a query and compare it to all the spans up to k in a sentence, then you can pull out the closest spans and rank the best candidate. Ive just been playing around with the method Gliner uses but adapted to model2vec. The speed ia there, I reckon we can tune the embedding and distill it for more performance.

[Pringled]

Hi @CP500,

Interesting idea! I'm not too familiar with the Glinder method unfortunately, but I'd be happy to help out with any questions on how to integrate model2vec with your idea. Let me know if you need any help.

[stephantul]

Closing this because of lack of activity, feel free to reopen this