Good keyword extraction models?


Was wondering if anyone knew of any good keyword extraction models for the english language hosted on hugging face or even else where? Have been webscraping ‘about us’ pages and need to extract the keywords that best describe the business function from those pages? Any help would be greatly appreciated!

Thank you

1 Like

T5, bart are all good models, the general class is seq2seq. There are plenty here.

My concern would be finding a suitable set to train on. Pile has a big data, but sometimes random big data did not help me with keyword generation/extraction.