chainer.functions.embed_id¶

chainer.functions.embed_id(x, W, ignore_label=None)[source]¶

Efficient linear function for one-hot input.

This function implements so called word embeddings. It takes two arguments: a set of IDs (words) x in $B$ dimensional integer vector, and a set of all ID (word) embeddings W in $V \times d$ float matrix. It outputs $B \times d$ matrix whose i-th column is the x[i]-th column of W.

This function is only differentiable on the input W.

Parameters

x (Variable or N-dimensional array) – Batch vectors of IDs. Each element must be signed integer.
W (Variable or N-dimensional array) – Distributed representation of each ID (a.k.a. word embeddings).
ignore_label (int or None) – If ignore_label is an int value, i-th column of return value is filled with 0.

Returns

Output variable.

Return type

Variable

See also

EmbedID

Example

>>> x = np.array([2, 1]).astype(np.int32)
>>> x
array([2, 1], dtype=int32)
>>> W = np.array([[0, 0, 0],
...               [1, 1, 1],
...               [2, 2, 2]]).astype(np.float32)
>>> W
array([[0., 0., 0.],
       [1., 1., 1.],
       [2., 2., 2.]], dtype=float32)
>>> F.embed_id(x, W).array
array([[2., 2., 2.],
       [1., 1., 1.]], dtype=float32)
>>> F.embed_id(x, W, ignore_label=1).array
array([[2., 2., 2.],
       [0., 0., 0.]], dtype=float32)