7cebf32b75ef71ea8e367579c71fea84deaac91b,chainerrl/agents/dpp.py,AbstractDPP,_compute_y_and_t,#AbstractDPP#Any#Any#,48

Before Change

        xp = cuda.get_array_module(qout.greedy_actions.data)

        batch_actions = chainer.Variable(
            xp.asarray([elem["action"] for elem in experiences]))
        // Q(s_t,a_t)
        batch_q = F.reshape(qout.evaluate_actions(
            batch_actions), (batch_size, 1))

After Change

        batch_q = F.reshape(qout.evaluate_actions(
            batch_actions), (batch_size, 1))

        with chainer.no_backprop_mode():
            // Compute target values
            target_qout = self.target_q_function(batch_state, test=True)

            // Q"(s_t,a_t)
            target_q = F.reshape(target_qout.evaluate_actions(
                batch_actions), (batch_size, 1))

            // LQ"(s_t,a)
            target_q_expect = F.reshape(
                self._l_operator(target_qout), (batch_size, 1))

            // r + g * LQ"(s_{t+1},a)
            batch_q_target = F.reshape(
                self._compute_target_values(exp_batch, gamma), (batch_size, 1))

            // Q"(s_t,a_t) + r + g * LQ"(s_{t+1},a) - LQ"(s_t,a)
            t = target_q + batch_q_target - target_q_expect

        return batch_q, t


class DPP(AbstractDPP):

Instances