rnncell.py 文件源码

def __call__(self, inputs, state, scope=None):
        """Gated recurrent unit (GRU) with nunits cells."""
        with tf.variable_scope(scope or type(self).__name__):  # "GRUCell"
            if self.pretanh:
                state = state[:, :self.num_units]
            with tf.variable_scope("Gates"):  # Reset gate and update gate.
                # We start with bias of 1.0 to not reset and not update.
                r, u = tf.split(1, 2, utils.linear([inputs, state], 2 * self.num_units, True, 1.0))
                r, u = tf.nn.sigmoid(r), tf.nn.sigmoid(u)
            with tf.variable_scope("Candidate"):
                preact = utils.linear([inputs, r * state], self.num_units, True)
                c = self.activation(preact)
            new_h = u * state + (1 - u) * c
        if self.pretanh:
            new_state = tf.concat(1, [new_h, preact])
        else:
            new_state = new_h
        return new_h, new_state