# The second parameter(300) is the number of dimension of your embedding model
# padding_idx = 1: specify the <PAD>'s index is 1
# max_norm = 1: If given, will renormalize the embeddings to always have a norm lesser than the given number
embed = nn.Embedding(len(TEXT.vocab), 300, padding_idx=1, max_norm=1)

# Copy the word vectors from dataset to nn.Embedding object
embed.weight.data.copy_(TEXT.vocab.vectors)