Word embeddings

How can I use a sent2vec based word embedding to convert a document to pytorch tensor, my word embedding is in binary format. I do not know the input or output dimension for the embedding.