in phasic_policy_gradient/ppo.py [0:0]
def compute_losses( model, ob, ac, first, logp, vtarg, adv, state_in, clip_param, vfcoef, entcoef, kl_penalty,