def forward(self, x, *, u=None, dequant_logd=None):
assert (u is None) == (dequant_logd is None)
if u is None:
u, dequant_logd = self.calc_dequant_noise(x)
assert u.shape == x.shape and dequant_logd.shape == (x.shape[0], )
assert (u >= 0).all() and (u <= 1).all()
z, main_logd = self.main_flow(x + u, aux=None, inverse=False)
z_logp = sumflat(standard_normal_logp(z))
total_logd = dequant_logd + main_logd + z_logp
assert z.shape[0] == x.shape[0] and z.numel() == x.numel()
assert main_logd.shape == dequant_logd.shape == total_logd.shape == z_logp.shape == (
x.shape[0], )
return {
'u': u,
'z': z,
'total_logd': total_logd,
'dequant_logd': dequant_logd,
}
def calc_dequant_noise(self, x):
eps = torch.randn_like(x)
u, dequant_logd = self.dequant_flow(eps=eps, aux=x, inverse=False)
assert u.shape == x.shape and dequant_logd.shape == (x.shape[0], )
return u, dequant_logd - sumflat(standard_normal_logp(eps))
def forward(self, input_, *, aux, inverse: bool):
assert isinstance(input_, tuple)
cf, ef = input_
ef_flow = self._make_elementwise_flow(cf=cf, aux=aux)
out_ef, logd = ef_flow(ef, aux=None, inverse=inverse)
return (cf, out_ef), sumflat(logd)