deecf3c0d7452276a87eaeeee533fad8c4ba24b5,ch06/01_dqn_pong.py,,,#,146

Before Change

            losses.append(loss_v.data.cpu().numpy())
        epsilon *= 0.99
        epsilon = max(epsilon, 0.1)
        print("Loss %.6f, epsilon=%f" % (np.mean(losses), epsilon))

    pass

After Change

            losses.append(loss_v.data.cpu().numpy())
        epsilon *= 0.99
        epsilon = max(epsilon, 0.1)
        writer.add_scalar("epsilon", epsilon, iter_idx)
        writer.add_scalar("loss", np.mean(losses), iter_idx)
        if rewards:
            writer.add_scalar("reward", np.mean(rewards))
        iter_idx += 1

Instances