d33c114816429c494e2a40625849d57ae2103f3f,softlearning/algorithms/sac.py,SAC,_update_target,#SAC#,399

Before Change

    def _update_target(self):
        source_params = self._V.get_weights()
        target_params = self._V_target.get_weights()

        self._V_target.set_weights([
            (1 - self._tau) * target + self._tau * source
            for target, source in zip(target_params, source_params)

After Change

        self._update_target()

    def _update_target(self):
        for Q, Q_target in zip(self._Qs, self._Q_targets):
            source_params = Q.get_weights()
            target_params = Q_target.get_weights()
            Q_target.set_weights([
                (1 - self._tau) * target + self._tau * source
                for target, source in zip(target_params, source_params)
            ])

    def _do_training(self, iteration, batch):
        Runs the operations for updating training and target ops.

        feed_dict = self._get_feed_dict(iteration, batch)

Instances