def forward(self, x):
z = H.mean(x, axis=(2, 3))
z = self.fc1(z)
z = H.relu(z)
z = self.fc2(z)
s = H.sigmoid(z)
n, c = s.shape
s = s.reshape(n, c, 1, 1)
x = x * s
return x
def test_multi_axes():
a = [[[4., -10., 7., 24.], [-1., 4., -6., 11.], [8., 11., -8., 8.]],
[[3., -10., 2., -5.], [10., 18., -5., -15.], [-1., 10., -13., 17.]]]
a = np.array(a)
ha = gtensor(a)
# k = [
# [[ 8., 6., 8., 10.],
# [ 9., 7., 1., 8.],
# [ 3., 2., 6., 6.]],
# [[ 7., 7., 1., 5.],
# [ 3., 5., 5., 3.],
# [ 7., 9., 5., 10.]]
# ]
k = [1., 2., 4.]
k = np.array(k)
hk = gtensor(k)
(H.mean(ha, axis=(0, 2)) * hk).sum().backward()
grad = [[[0.125, 0.125, 0.125, 0.125], [0.25, 0.25, 0.25, 0.25],
[0.5, 0.5, 0.5, 0.5]],
[[0.125, 0.125, 0.125, 0.125], [0.25, 0.25, 0.25, 0.25],
[0.5, 0.5, 0.5, 0.5]]]
grad = np.array(grad)
np.testing.assert_allclose(ha.grad, grad)
def forward(self, x): m = H.mean(x, axis=0, keepdims=True) v = H.var(x, axis=0, keepdims=True) x = (x - m) / H.sqrt(v + self.eps) z = self.std * x + self.mean return z
def test_axis_none():
t1 = gtensor(dm('1 2 3; 4 5 6'))
k = 3
(H.mean(t1, axis=None) * k).backward()
grad = dm('.5 .5 .5; .5 .5 .5')
np.testing.assert_allclose(t1.grad, grad)
def test_axis():
t1 = gtensor(dm('1 2 3; 4 5 6'))
k = gtensor(dm('3 6'))
(H.mean(t1, axis=1) * k).sum().backward()
grad = dm('1 1 1; 2 2 2')
np.testing.assert_allclose(t1.grad, grad)
def BCELoss(logits, target): y = target z = logits return _H.mean(_H.maximum(z, 0) + _H.log(1 + _H.exp(-_H.abs(z))) - z*y)