def test_trivial_fancy_out_of_bounds(self):
a = zeros(5)
ind = ones(20).long()
ind[-1] = 10
self.assertRaises(RuntimeError, a.__getitem__, ind)
self.assertRaises(RuntimeError, a.__setitem__, ind, 0)
ind = ones(20).long()
ind[0] = 11
self.assertRaises(RuntimeError, a.__getitem__, ind)
self.assertRaises(RuntimeError, a.__setitem__, ind, 0)
def test_index_is_larger(self):
# Simple case of fancy index broadcasting of the index.
a = zeros((5, 5))
a[[[0], [1], [2]], [0, 1, 2]] = tensor([2, 3, 4])
self.assertTrue((a[:3, :3] == tensor([2, 3, 4])).all())
def test_broadcast_subspace(self):
a = zeros((100, 100))
v = Variable(torch.arange(0, 100))[:, None]
b = Variable(torch.arange(99, -1, -1).long())
a[b] = v
expected = b.double().unsqueeze(1).expand(100, 100)
self.assertEqual(a, expected)
if __name__ == '__main__':
run_tests()
x = torch.randn(4, 4)
y = torch.randn(4, 4)
z = cpp_extension.sigmoid_add(x, y)
self.assertEqual(z, x.sigmoid() + y.sigmoid())
def test_extension_module(self):
mm = cpp_extension.MatrixMultiplier(4, 8)
weights = torch.rand(8, 4)
expected = mm.get().mm(weights)
result = mm.forward(weights)
self.assertEqual(expected, result)
def test_jit_compile_extension(self):
module = torch.utils.cpp_extension.load(
name='jit_extension',
sources=[
'cpp_extensions/jit_extension.cpp',
'cpp_extensions/jit_extension2.cpp'
],
extra_include_paths=['cpp_extensions'],
extra_cflags=['-g'],
verbose=True)
x = torch.randn(4, 4)
y = torch.randn(4, 4)
z = module.tanh_add(x, y)
self.assertEqual(z, x.tanh() + y.tanh())
if __name__ == '__main__':
common.run_tests()
self.assertFalse(torch.sparse_coo_tensor(([0, 0],), (0., 0.), (1,)).is_nonzero())
self.assertFalse(torch.sparse_coo_tensor(([0, 0],), (-1., 1.), (1,)).is_nonzero())
# NB: We should test "scalar" sparse tensors, but they don't actually
# work at the moment (in principle, they should)
class TestUncoalescedSparse(TestSparse):
def setUp(self):
super(TestUncoalescedSparse, self).setUp()
self.is_uncoalesced = True
@unittest.skipIf(not TEST_CUDA, 'CUDA not available')
class TestCudaSparse(TestSparse):
def setUp(self):
super(TestCudaSparse, self).setUp()
self.is_cuda = True
self.IndexTensor = torch.cuda.LongTensor
self.ValueTensor = torch.cuda.DoubleTensor
self.SparseTensor = torch.cuda.sparse.DoubleTensor
@unittest.skipIf(not TEST_CUDA, 'CUDA not available')
class TestCudaUncoalescedSparse(TestCudaSparse):
def setUp(self):
super(TestCudaUncoalescedSparse, self).setUp()
self.is_uncoalesced = True
if __name__ == '__main__':
run_tests()
