def test_device2(self):
print(f"{self.__class__.__name__}: test_device2 SparseTensor")
if not is_cuda_available():
return
coordinates = np.random.rand(8192,3) * 200
quant_coordinates, quant_features = sparse_quantize(coordinates, coordinates)
bcoords, bfeats = sparse_collate([quant_coordinates], [quant_features])
bcoords, bfeats = bcoords.cuda(), bfeats.cuda()
print(bcoords, bfeats)
SparseTensor(bfeats, bcoords)
def test_device(self):
if not is_cuda_available():
return
coords = torch.IntTensor([[0, 1], [0, 1], [0, 2], [0, 2], [1, 0],
[1, 0], [1, 1]])
feats = torch.FloatTensor([[0, 1, 2, 3, 5, 6, 7]]).T
SparseTensor(feats, coords)
SparseTensor(feats.to(0), coords.to(0))
feats = torch.FloatTensor([[0, 1, 2, 3, 5, 6, 7]]).T.to(0)
st = SparseTensor(feats, coords, device=feats.device)
print(st)
def test_device_unique(self):
print(f"{self.__class__.__name__}: test_device_unique SparseTensor")
if not is_cuda_available():
return
coords = torch.IntTensor(
[[0, 1], [0, 2], [0, 3], [0, 4], [1, 0], [1, 1], [1, 2]]
)
feats = torch.FloatTensor([[0, 1, 2, 3, 5, 6, 7]]).T
SparseTensor(feats.to(0), coords.to(0))
feats = torch.FloatTensor([[0, 1, 2, 3, 5, 6, 7]]).T.to(0)
st = SparseTensor(feats, coords, device=feats.device)
print(st)
def test_extraction(self):
print(f"{self.__class__.__name__}: test_extraction")
coords = torch.IntTensor([[0, 0], [0, 1], [0, 2], [2, 0], [2, 2]])
feats = torch.FloatTensor([[1.1, 2.1, 3.1, 4.1, 5.1]]).t()
X = SparseTensor(feats, coords)
C0 = X.coordinates_at(0)
F0 = X.features_at(0)
self.assertTrue(0 in C0)
self.assertTrue(1 in C0)
self.assertTrue(2 in C0)
self.assertTrue(1.1 in F0)
self.assertTrue(2.1 in F0)
self.assertTrue(3.1 in F0)
CC0, FC0 = X.coordinates_and_features_at(0)
self.assertTrue((C0 == CC0).all())
self.assertTrue((F0 == FC0).all())
coords, feats = X.decomposed_coordinates_and_features
for c, f in zip(coords, feats):
self.assertEqual(c.numel(), f.numel())
print(c, f)
self.assertEqual(len(coords[0]), 3)
self.assertEqual(len(coords[1]), 0)
self.assertEqual(len(coords[2]), 2)
if not is_cuda_available():
return
coords = torch.IntTensor([[0, 0], [0, 1], [0, 2], [2, 0], [2, 2]])
feats = torch.FloatTensor([[1.1, 2.1, 3.1, 4.1, 5.1]]).t()
X = SparseTensor(feats, coords, device=0)
coords, feats = X.decomposed_coordinates_and_features
for c, f in zip(coords, feats):
self.assertEqual(c.numel(), f.numel())
print(c, f)
self.assertEqual(len(coords[0]), 3)
self.assertEqual(len(coords[1]), 0)
self.assertEqual(len(coords[2]), 2)
