In word/character embedding What is begin represent in vectors values of a single word ? the word "How" is represent in a vector of size 3072 . So what are these 3072 values denoting about the word "how"?

@spiel wrote:

from flair.embeddings import BertEmbeddings
Bertembedding = BertEmbeddings()
sentence2=Sentence("Hello good moring ")
Bertembedding.embed(sentence2)
for token in sentence2:
print(token.get_embedding)
print(token.embedding)
print(token.embedding.size())

o/p
bound method Token.get_embedding of Token: 1 Hello>
tensor([ 0.1705, 0.5013, 0.8154, …, -0.2748, -0.6512, -0.0866])
torch.Size([3072])
<bound method Token.get_embedding of Token: 2 good>
tensor([ 0.0079, 0.4349, 0.9639, …, -1.2404, -0.4342, -1.9663])
torch.Size([3072])
<bound method Token.get_embedding of Token: 3 moring>
tensor([ 0.0535, -0.1277, 0.2877, …, -0.3166, 0.0504, -0.2165])
torch.Size([3072])

Posts: 1

Participants: 1

Read full topic