import numpy
from chainer import cuda
from chainer import function
from chainer.utils import type_check
class MeanAbsoluteError(function.Function):
"""Mean absolute error function."""
def check_type_forward(self, in_types):
type_check.expect(in_types.size() == 2)
type_check.expect(
in_types[0].dtype == numpy.float32,
in_types[1].dtype == numpy.float32,
in_types[0].shape == in_types[1].shape
)
def forward_cpu(self, inputs):
x0, x1 = inputs
self.diff = x0 - x1
diff = self.diff.ravel()
return numpy.array(abs(diff).sum() / diff.size, dtype=diff.dtype),
def forward_gpu(self, inputs):
x0, x1 = inputs
self.diff = x0 - x1
diff = self.diff.ravel()
return abs(diff).sum() / diff.dtype.type(diff.size),
def backward(self, inputs, gy):
xp = cuda.get_array_module(*inputs)
coeff = gy[0] * gy[0].dtype.type(1. / self.diff.size)
gx0 = coeff * xp.sign(self.diff)
return gx0, -gx0
[docs]def mean_absolute_error(x0, x1):
"""Mean absolute error function.
This function computes mean absolute error between two variables. The mean
is taken over the minibatch.
"""
return MeanAbsoluteError()(x0, x1)