d5b0cd8e7960c247bb7c5b7c832358f8831780fb,ch15/03_train_trpo.py,,,#,90

Before Change

            trajectory.clear()
            writer.add_scalar("advantage", traj_adv_v.mean().data.cpu().numpy()[0], step_idx)
            writer.add_scalar("values", traj_ref_v.mean().data.cpu().numpy()[0], step_idx)
            writer.add_scalar("loss_value", loss_value_v.data.cpu().numpy()[0], step_idx)

After Change

            trajectory.clear()
            writer.add_scalar("advantage", traj_adv_v.mean().item(), step_idx)
            writer.add_scalar("values", traj_ref_v.mean().item(), step_idx)
            writer.add_scalar("loss_value", loss_value_v.item(), step_idx)

Instances