f57e1d9d32a84f60738a73fa35b407999394501e,chainerrl/agents/double_pal.py,DoublePAL,_compute_y_and_t,#DoublePAL#Any#,18

Before Change

            with state_kept(self.q_function):
                next_qout = self.q_function(batch_next_state)

            with state_kept(self.target_q_function):
                target_next_qout = self.target_q_function(
                    batch_next_state)
            next_q_max = F.reshape(target_next_qout.evaluate_actions(
                next_qout.greedy_actions), (batch_size,))

            batch_rewards = exp_batch["reward"]

After Change

class DoublePAL(pal.PAL):

    def _compute_y_and_t(self, exp_batch):

        batch_state = exp_batch["state"]
        batch_size = len(exp_batch["reward"])

        if self.recurrent:
            qout, _ = self.model.n_step_forward(
                batch_state, exp_batch["recurrent_state"],
                output_mode="concat")
        else:
            qout = self.model(batch_state)

        batch_actions = exp_batch["action"]
        batch_q = qout.evaluate_actions(batch_actions)

        // Compute target values

        with chainer.no_backprop_mode():
            batch_next_state = exp_batch["next_state"]
            if self.recurrent:
                next_qout, _ = self.model.n_step_forward(
                    batch_next_state, exp_batch["next_recurrent_state"],
                    output_mode="concat")
                target_qout, _ = self.target_model.n_step_forward(
                    batch_state, exp_batch["recurrent_state"],
                    output_mode="concat")
                target_next_qout, _ = self.target_model.n_step_forward(
                    batch_next_state, exp_batch["next_recurrent_state"],
                    output_mode="concat")
            else:
                next_qout = self.model(batch_next_state)
                target_qout = self.target_model(batch_state)
                target_next_qout = self.target_model(batch_next_state)

            next_q_max = F.reshape(target_next_qout.evaluate_actions(
                next_qout.greedy_actions), (batch_size,))

            batch_rewards = exp_batch["reward"]

Instances