import numpy
from chainer import cuda
from chainer.functions.loss import negative_sampling
from chainer import link
from chainer.utils import walker_alias
[docs]class NegativeSampling(link.Link):
"""Negative sampling loss layer.
This link wraps the :func:`~chainer.functions.negative_sampling` function.
It holds the weight matrix as a parameter. It also builds a sampler
internally given a list of word counts.
Args:
in_size (int): Dimension of input vectors.
counts (int list): Number of each identifiers.
sample_size (int): Number of negative samples.
power (float): Power factor :math:`\\alpha`.
.. seealso:: :func:`~chainer.functions.negative_sampling` for more detail.
Attributes:
W (~chainer.Variable): Weight parameter matrix.
"""
def __init__(self, in_size, counts, sample_size, power=0.75):
vocab_size = len(counts)
super(NegativeSampling, self).__init__(W=(vocab_size, in_size))
self.W.data.fill(0)
self.sample_size = sample_size
power = numpy.float32(power)
p = numpy.array(counts, power.dtype)
numpy.power(p, power, p)
self.sampler = walker_alias.WalkerAlias(p)
def to_cpu(self):
super(NegativeSampling, self).to_cpu()
self.sampler.to_cpu()
def to_gpu(self, device=None):
with cuda._get_device(device):
super(NegativeSampling, self).to_gpu()
self.sampler.to_gpu()
[docs] def __call__(self, x, t, reduce='sum'):
"""Computes the loss value for given input and ground truth labels.
Args:
x (~chainer.Variable): Input of the weight matrix multiplication.
t (~chainer.Variable): Batch of ground truth labels.
reduce (str): Reduction option. Its value must be either
``'sum'`` or ``'no'``. Otherwise, :class:`ValueError` is
raised.
Returns:
~chainer.Variable: Loss value.
"""
return negative_sampling.negative_sampling(
x, t, self.W, self.sampler.sample, self.sample_size,
reduce=reduce)