def _build_loss(self, results, features, labels, loss_config, **kwargs):
losses, loss = getters.get_loss(loss_config.module, results, features, **loss_config.params)
with get_name_scope('latent_loss'):
z_mean = kwargs['z_mean']
z_log_sigma = kwargs['z_log_sigma']
latent_losses = -0.5 * tf.reduce_sum(
1 + z_log_sigma - tf.square(z_mean) - tf.exp(z_log_sigma))
latent_loss = tf.losses.compute_weighted_loss(latent_losses)
losses += latent_losses
loss += latent_loss
return losses, loss
def _build_loss(self, results, features, labels):
"""Creates the loss operation
Returns:
tuple `(losses, loss)`:
`losses` are the per-batch losses.
`loss` is a single scalar tensor to minimize.
"""
losses, loss = getters.get_loss(
self.loss.IDENTIFIER, results, labels, **self.loss.to_dict())
self._loss = loss
self._losses = losses
other_losses = get_tracked(tf.GraphKeys.REGULARIZATION_LOSSES)
if other_losses:
loss = [loss] + other_losses
loss = tf.add_n(loss, name="TotalLoss")
self._total_loss = loss
return losses, loss
def _build_loss(self, results, features, labels):
"""Creates the loss operation
Returns:
tuple `(losses, loss)`:
`losses` are the per-batch losses.
`loss` is a single scalar tensor to minimize.
"""
losses, loss = getters.get_loss(
self.loss_config.module, results, labels, **self.loss_config.params)
self._loss = loss
self._losses = losses
other_losses = get_tracked(tf.GraphKeys.REGULARIZATION_LOSSES)
if other_losses:
loss = [loss] + other_losses
loss = tf.add_n(loss, name="TotalLoss")
self._total_loss = loss
return losses, loss
def _build_loss(self, results, features, labels, loss, **kwargs):
return getters.get_loss(loss.IDENTIFIER, results, features, **loss.to_dict())
def _build_loss(self, results, features, labels, loss_config, **kwargs):
return getters.get_loss(loss_config.module, results, features, **loss_config.params)
def _build_loss(self, incoming, results, loss_config, **kwargs):
return getters.get_loss(loss_config.module, results, incoming,
**loss_config.params)
def _build_loss(self, results, features, labels, loss_config, **kwargs):
return getters.get_loss(loss_config.module, results, features,
**loss_config.params)