def call(self, x: tf.Tensor, *args, **kw_args) -> tf.Tensor:
size = x.size()
x = x.view(*size[:2], -1)
f, g, h = self.query(x), self.key(x), self.value(x)
# beta = F.softmax(torch.bmm(f.permute(0, 2, 1).contiguous(), g), dim=1)
# o = self.gamma * torch.bmm(h, beta) + x
return None # o.view(*size).contiguous()
def get_energies(self, y: tf.Tensor, weights_in_time: tf.Tensor):
weight_sum = tf.cond(tf.greater(weights_in_time.size(), 0),
lambda: tf.reduce_sum(weights_in_time, axis=0),
lambda: 0.0)
coverage = weight_sum / self.fertility * self.attention_mask
coverage_exp = tf.expand_dims(tf.expand_dims(coverage, -1), -1)
logits = tf.reduce_sum(
self.similarity_bias_vector *
tf.tanh(self.hidden_features + y +
self.coverage_weights * coverage_exp), [2, 3])
return logits
def get_energies(self, y: tf.Tensor, weights_in_time: tf.Tensor):
weight_sum = tf.cond(
tf.greater(weights_in_time.size(), 0),
lambda: tf.reduce_sum(weights_in_time, axis=0),
lambda: 0.0)
coverage = weight_sum / self.fertility * self.attention_mask
coverage_exp = tf.expand_dims(tf.expand_dims(coverage, -1), -1)
logits = tf.reduce_sum(
self.similarity_bias_vector * tf.tanh(
self.hidden_features + y
+ self.coverage_weights * coverage_exp),
[2, 3])
return logits
