beta = 100
partial_input = True
data_completeness = 0.7
data_sparsity = 100
try:
volume = np.load(
'sdf' + save_fold +
'/sdf_{}_{}_{}.npy'.format(split, checkpoint, conditioned_ind))
except FileNotFoundError:
volume = None
if volume is None:
DATA_PATH = 'data/ShapeNet'
fields = {'inputs': dataset.PointCloudField('pointcloud.npz')}
category = ['02958343']
test_dataset = dataset.ShapenetDataset(
dataset_folder=DATA_PATH,
fields=fields,
categories=category,
split=split,
partial_input=partial_input,
data_completeness=data_completeness,
data_sparsity=data_sparsity,
evaluation=True)
# conditioned_input = test_dataset.__getitem__(conditioned_ind)['points'].unsqueeze(0)
ds_kitti = dataset.KITTI360Dataset('data/KITTI-360/data_3d_pointcloud',
'train',
'building',
for k, v in saved_model_state.items()
})
# set multi-gpu if available
if torch.cuda.device_count() > 1:
print("Let's use", torch.cuda.device_count(), "GPUs!")
net = torch.nn.DataParallel(net)
net.to(device)
num_params = sum(p.numel() for p in net.parameters() if p.requires_grad)
print('The number of parameters of model is', num_params)
# create dataloader
# ShapeNet
DATA_PATH = cfg['data']['path']
fields = {
'inputs': dataset.PointCloudField(cfg['data']['pointcloud_file'])
}
category = cfg['data']['classes']
shapenet_dataset = dataset.ShapenetDataset(
dataset_folder=DATA_PATH,
fields=fields,
categories=category,
split='train',
with_normals=use_normal,
points_batch=points_batch,
partial_input=partial_input,
data_completeness=data_completeness,
data_sparsity=data_sparsity)
shapenet_loader = torch.utils.data.DataLoader(
shapenet_dataset,
batch_size=batch_size_shapenet,