Classic word embeddings in Flair
The simplest type of embedding in Flair that still contains semantic information about the word is called classic word embeddings. These embeddings are pre-trained and non-contextual and require no preprocessing. The non-contextual nature means one word always maps to one precomputed embedding regardless of the context. To use them, we simply instantiate a WordEmbeddings class by passing in the ID of the embedding of our choice. Then, we simply wrap our text into a Sentence object and call the embed(sentence) method on our WordEmbeddings class.
Let's obtain embeddings for a few random words using the FastText "crawl" embeddings:
from flair.data import Sentence
from flair.embeddings import WordEmbeddings
embedding = WordEmbeddings('crawl')
sentence = Sentence("one two three one")
embedding.embed(sentence)
for token in sentence:
print(token.embedding)
The preceding code will print out...