vseledkin / torch-word-emb

load word embeddings to Torch.Tensor

#torch-word-emb Load your word embeddings (Word2Vec and GloVe) into Torch tensor.

Install

Torch 7 required.

luarocks install torch-word-emb

or

git clone https://github.com/iamalbert/torch-word-emb
cd torch-word-emb && luarocks make

Usage

local wordemb = require 'wordemb'

wordemb.load_word2vec_bin(path)

read word2vec binary-format model from path.

returns a table of two fields.

• vec is a torch FloatTensor of size V x D, where V is the vocabulary size and D is the dimension of word2vec.
• words is a table of (word,offset) pairs, each of which represents the corresponding index of such word in vec.

local w2v = wordemb.load_word2vec_bin("/path/to/word2vec/model.bin")
print(w2v.vec:size())
print(w2v.vec[ w2v.words["apple"] ] )

wordemb.load_word2vec_text(path)

read word2vec text-format model from path.

wordemb.load_glove_text(path)

read GloVe text-format model from path.