['--alg', 'sac', '--env', 'HalfCheetah-v2', '--learn', 'behavior', '--traj', 'medium', '--seed', '0', '--data', '10000', '--sub']
5053.504169503249
episode: 0 training return: tensor(-112.6611, device='cuda:0')
episode: 1 training return: tensor(-132.6227, device='cuda:0')
episode: 2 training return: tensor(-33.3462, device='cuda:0')
