CODEWITHSUNDEEP
pythonmachine-learningnlpfasttext
Bhaskar
Bhaskar

Reputation: 343

What is happening under the hood of fasttext supervised learning model?

We can train a supervised model in fasttext using

import fasttext
model = fasttext.train_supervised(input="cooking.train")

My question is how it is representing the features(Bag of words or tf/idf or word embedding) and what algorithm is it using for text classification?

Upvotes: 0

Views: 747

Answers (1)

gojomo
gojomo

Reputation: 54243

It's still learning word-vectors for the input text - but then averaging them all together, a bit like an infinite-window CBOW mode – then using that to predict the the labels, as if the labels were the (word2vec-style) predicted-word.

Upvotes: 1

Related Questions