def sample(_, decoder, args: Options, trace: ViewMem):
if trace.memory is None:
num_items = get_integer((1, 8))
else:
num_items = trace.memory[1]
z = torch.randn(num_items, args.dim_z).to(trace.device)
gms = decoder(z)
vs, splits = gm_utils.hierarchical_gm_sample(gms, trace.points_in_sample, False)
vs = vs.cpu().numpy()
splits = [s for s in splits]
palette = create_palettes(splits)
im, points = view([vs_ for vs_ in vs], splits, palette)
return True, (sample, num_items), im, (points, splits, palette)
def interpolate(encoder, decoder, args: Options, trace: ViewMem):
if trace.memory is None:
msg = '\tPlease choose number of interpolation: from 8 to 20\n\t'
num_interpulate = get_integer((7, 21), msg)
else:
num_interpulate = trace.memory[1]
z = get_z_by_id(encoder, args, 2, trace.last_idx, trace)
gms = decoder.interpulate(z, num_interpulate)
vs, splits = gm_utils.hierarchical_gm_sample(gms, trace.points_in_sample)
spread = [vs[i].cpu().numpy() for i in range(vs.shape[0])]
splits = [s for s in splits]
palette = create_palettes(splits)
im, points = view(spread, splits, palette) #, save_path=f'{cp_folder}/interpulate_{idx[0].item()}_{idx[1].item()}.png')
return True, (interpolate, (num_interpulate,)), im, (points, splits, palette)
def hgmms(encoder, decoder, args: Options, trace: ViewMem):
z = get_z_by_id(encoder, args, 1, None, trace)
gms = decoder(z)
num_gms = len(gms)
vs = []
splits = []
for i in range(num_gms):
gms_ = [gms[i] for i in range(i+1)]
vs_, splits_ = gm_utils.hierarchical_gm_sample(gms_, trace.points_in_sample)
vs.append(vs_.squeeze(0).cpu().numpy())
splits.append(splits_.squeeze(0).cpu().numpy())
palette = create_palettes(splits)
im, points = view(vs, splits, palette)
return True, (hgmms, ), im, (points, splits, palette)
def before_plot(self, data, eval_size=-1):
y, encoder_inp = self.arrange_data(data, eval_size)
z, _, _ = self.encoder(encoder_inp)
gms = self.decoder(z)
vs, splits = gm_utils.hierarchical_gm_sample(
gms, self.opt.partial_samples[0], self.opt.flatten_sigma)
# splits_inp = np.array([0, encoder_inp.shape[1]])
# splits_y = np.array([0, y[1]])
# transform back
if self.opt.recon or len(self.opt.transforms) > 0:
transforms = data[3:]
for i in range(len(transforms)):
transform = transforms[-(i + 1)][:vs.shape[0]].to(self.device)
if transform.dim() == 2:
t = lambda x: x - transform[:, None, :]
else:
t = lambda x: torch.einsum('bnd,brd->bnr', x, transform)
vs, vs_in, y = list(map(t, [vs, encoder_inp, y]))
out = list(
map(lambda x: x.data.cpu().numpy(), [vs, encoder_inp, y, splits]))
return out