Minimal example for generating word embeddings

[abdullahfurquan]

Hi,
In RepresentationModel I was trying to get word embedding. Below is my code snippet and I have taken one sentence- as input. My sentence has 3 word. By default 2 tokens are for [CLS] & [SEP]. So total tokens should be :3+2=5.

But the shape of wordvectors from code is : (1, 6, 768) .So it has 6 word embedding. Now I am little confused ,how do I get the embedding of each word of my sentence (i.e. "dont stop go" ). Is there a way to map word and corresponding embedding

code snippets:-

from simpletransformers.language_representation import RepresentationModel

model = RepresentationModel(
model_type="bert",
model_name="bert-base-uncased",
use_cuda=False
)

sentences=["dont stop go"]
wordvectors = model.encode_sentences(sentences, combine_strategy=None)
wordvectors.shape
(1,6,768)

below is a image of it :