373ae159f7ae1cabaf87228d1ae0fb6acd1c6363,ch14/lib/common.py,,unpack_batch_a2c,#,8

Before Change

    // handle rewards
    rewards_np = np.array(rewards, dtype=np.float32)
    if not_done_idx:
        last_states_v = ptan.agent.float32_preprocessor(last_states, cuda=cuda)
        last_vals_v = net(last_states_v)[2]
        last_vals_np = last_vals_v.data.cpu().numpy()[:, 0]
        rewards_np[not_done_idx] += last_val_gamma * last_vals_np

After Change

    // handle rewards
    rewards_np = np.array(rewards, dtype=np.float32)
    if not_done_idx:
        last_states_v = ptan.agent.float32_preprocessor(last_states).to(device)
        last_vals_v = net(last_states_v)[2]
        last_vals_np = last_vals_v.data.cpu().numpy()[:, 0]
        rewards_np[not_done_idx] += last_val_gamma * last_vals_np

Instances