about torch.distribution sample

zdist = get_zdist(config['z_dist']['type'], config['z_dist']['dim'],
                      device=device)
ntest = 8
ztest = zdist.sample((ntest,))

zdist is Normal(loc: torch.Size([256]), scale: torch.Size([256])). -> torch.distribution(Normal distribution)

I wonder about this part :

zdist.sample((ntest,))

My understanding is to extract ntest number of samples from zdist, right?

And what else is different from this?

zdist.sample(torch.Size([ntest]))

please help me.