Get word rank from fastText model?

[Adam M B]

Words are sorted in the model, so we can load with limited=n, and then 'apple' in model implies that 'apple' is in the top n words.

Is there a straightforward way to get the exact rank explicitly?

[Gordon Mohr]

As with other KeyedVectors, there's a property `vocab` that's a dictionary of word-keys to `Vocab` object values. The `index` property of the Vocab tells which int-index in the full array holds that word's vector. In the usual case, those slot indexes also match the order the words were read, which also (by convention but not necessarily guaranteed) is usually from most-frequent to least-frequent. 

(There's also the `index2word` property – a list of words in their slot order.) 

- Gordon

[Adam Mathias Bittlingmayer]

Thank you

For others: model.vocab['apple'].index

And to get the word at an index: model.index2word[42]

--
You received this message because you are subscribed to a topic in the Google Groups "gensim" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/gensim/hE8jr0hKYa0/unsubscribe.
To unsubscribe from this group and all its topics, send an email to gensim+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.