def test_sum():
"""Tests sum"""
# Should sum along the last dimension by default
ones = tf.ones([5, 4, 3])
val = ops.sum(ones)
assert isinstance(val, tf.Tensor)
assert val.ndim == 2
assert val.shape[0] == 5
assert val.shape[1] == 4
assert np.all(val.numpy() == 3.0)
# But can change that w/ the axis kwarg
ones = tf.ones([5, 4, 3])
val = ops.sum(ones, axis=1)
assert isinstance(val, tf.Tensor)
assert val.ndim == 2
assert val.shape[0] == 5
assert val.shape[1] == 3
assert np.all(val.numpy() == 4.0)
# Should sum along all dimensions w/ axis=None
ones = tf.ones([5, 4, 3])
val = ops.sum(ones, axis=None)
assert isinstance(val, tf.Tensor)
assert val.ndim == 0
assert val.numpy() == 60
# Actually test values
val = ops.sum(tf.constant([1.1, 2.0, 3.3]))
assert is_close(val.numpy(), 6.4)
def log_likelihood(self, x_data, y_data):
"""Compute the sum log likelihood of the model given a batch of data"""
if x_data is None:
log_likelihoods = self().log_prob(y_data)
else:
log_likelihoods = self(x_data).log_prob(y_data)
return O.sum(log_likelihoods, axis=None)
def kl_loss(self):
"""Compute the sum of the Kullback–Leibler divergences between this
parameter's priors and its variational posteriors."""
if self.prior is None:
return O.zeros([])
else:
return O.sum(O.kl_divergence(self.posterior, self.prior),
axis=None)
def __call__(self, x):
return O.sum(self.p2(), axis=None) + x * self.p1()
def __call__(self, x):
return self.mod(x) + O.sum(self.p3(), axis=None)
def add_kl_loss(self, loss, d2=None):
"""Add additional loss due to KL divergences."""
if d2 is None:
self._kl_losses += [O.sum(loss, axis=None)]
else:
self._kl_losses += [O.sum(O.kl_divergence(loss, d2), axis=None)]