4586366c1018257dd448d62f1fc7534e3f5a4338,enjoy.py,,,#,49

Before Change

    cpu_actions = action.data.cpu().numpy()

    // Obser reward and next obs
    obs, _, done, _ = env.step(cpu_actions[0])

    if args.env_name.find("Bullet") > -1:
        if torsoId > -1:

After Change

    masks.fill_(0.0 if done else 1.0)

    if current_obs.dim() == 4:
        current_obs *= masks.unsqueeze(2).unsqueeze(2)
    else:
        current_obs *= masks
    update_current_obs(obs)

    if args.env_name.find("Bullet") > -1:
        if torsoId > -1:

Instances