def create_shapenet_voxel_dataloader(dset_type_, data_base_dir_, batch_size_):
dataset = VoxelDataset(
dset_type=dset_type_,
data_base_dir=data_base_dir_,
transform=transforms.Compose([transforms.ToTensor()]))
dataloader = DataLoader(
dataset,
batch_size=batch_size_,
shuffle=True,
num_workers=4)
return dataloader
epoch_id = filename[61:-3]
sdf_net = SDFNet()
sdf_net.filename = filename[7:]
sdf_net.load()
sdf_net.eval()
clouds = sample_point_clouds(sdf_net,
50,
2048,
voxel_resolution=64,
latent_codes=latent_codes)
np.save('data/chairs/results/voxels_{:s}.npy'.format(epoch_id), clouds)
if 'dataset' in sys.argv:
from datasets import VoxelDataset
dataset = VoxelDataset.from_split('data/airplanes/voxels_64/{:s}.npy',
'data/airplanes/val.txt')
from torch.utils.data import DataLoader
voxels = next(iter(DataLoader(dataset, batch_size=100, shuffle=True)))
print(voxels.shape)
clouds = sample_from_voxels(voxels, 2048)
np.save('data/dataset_airplanes_point_cloud_sample.npy', clouds)
if 'test' in sys.argv:
import pyrender
data = np.load('data/dataset_point_cloud_sample.npy')
for i in range(data.shape[0]):
points = data[i, :, :]
scene = pyrender.Scene()
scene.add(pyrender.Mesh.from_points(points))
pyrender.Viewer(scene, use_raymond_lighting=True, point_size=8)
SURFACE_LEVEL = 0.011
FRAMES = SAMPLE_COUNT * TRANSITION_FRAMES
progress = np.arange(FRAMES, dtype=float) / TRANSITION_FRAMES
if USE_VAE:
from model.autoencoder import Autoencoder, LATENT_CODE_SIZE
vae = Autoencoder()
vae.load()
vae.eval()
print("Calculating latent codes...")
from datasets import VoxelDataset
from torch.utils.data import DataLoader
dataset = VoxelDataset.glob('data/chairs/voxels_32/**.npy')
dataloader = DataLoader(dataset, batch_size=1000, num_workers=8)
latent_codes = torch.zeros((len(dataset), LATENT_CODE_SIZE))
with torch.no_grad():
position = 0
for batch in tqdm(dataloader):
latent_codes[position:position + batch.shape[0], :] = vae.encode(
batch.to(device)).detach().cpu()
latent_codes = latent_codes.numpy()
else:
from model.sdf_net import SDFNet, LATENT_CODES_FILENAME
latent_codes = torch.load(LATENT_CODES_FILENAME).detach().cpu().numpy()
sdf_net = SDFNet()
def create_shapenet_voxel_dataloader(dset_type_,
data_base_dir_,
batch_size_,
subset_size_=None):
dataset_ = VoxelDataset(dset_type=dset_type_, data_base_dir=data_base_dir_)
return create_dataloader_from_dataset(dataset_, batch_size_, subset_size_)
ITERATION = int(get_parameter('iteration', 0))
# Continue with model parameters that were previously trained at the SAME iteration
# Otherwise, it will use the model parameters of the previous iteration or initialize randomly at iteration 0
CONTINUE = "continue" in sys.argv
FADE_IN_EPOCHS = 10
BATCH_SIZE = 16
GRADIENT_PENALTY_WEIGHT = 10
NUMBER_OF_EPOCHS = int(get_parameter('epochs', 250))
VOXEL_RESOLUTION = RESOLUTIONS[ITERATION]
dataset = VoxelDataset.from_split(
'data/chairs/voxels_{:d}/{{:s}}.npy'.format(VOXEL_RESOLUTION),
'data/chairs/train.txt')
data_loader = DataLoader(dataset,
batch_size=BATCH_SIZE,
shuffle=True,
num_workers=4)
def get_generator_filename(iteration):
return 'hybrid_progressive_gan_generator_{:d}.to'.format(iteration)
generator = SDFNet(device='cpu')
discriminator = Discriminator()
if not CONTINUE and ITERATION > 0:
generator.filename = get_generator_filename(ITERATION - 1)