ad60498885593181f25f02cf01a72ef8f3f18167,dqn.py,DQN,_compute_loss,#DQN#Any#Any#Any#,124

Before Change

        batch_q = self.q_function(batch_state, test=False).q_values

        batch_q_target = self._compute_target_values(
            experiences, gamma, cuda.to_cpu(batch_q.data))

        if self.gpu >= 0:
            batch_q_target = cuda.to_gpu(batch_q_target, self.gpu)

        batch_q_target = Variable(batch_q_target)

        if errors_out is not None:
            del errors_out[:]
            delta = F.sum(

After Change

            batch_actions), (batch_size, 1))

        batch_q_target = F.reshape(
            self._compute_target_values(experiences, gamma), (batch_size, 1))

        if errors_out is not None:
            del errors_out[:]

Instances