def test_construct_base_samples(self, cuda=False):
device = torch.device("cuda") if cuda else torch.device("cpu")
test_shapes = [
{
"batch": [2],
"output": [4, 3],
"sample": [5]
},
{
"batch": [1],
"output": [5, 3],
"sample": [5, 6]
},
{
"batch": [2, 3],
"output": [2, 3],
"sample": [5]
},
]
for tshape in test_shapes:
batch_shape = torch.Size(tshape["batch"])
output_shape = torch.Size(tshape["output"])
sample_shape = torch.Size(tshape["sample"])
expected_shape = sample_shape + batch_shape + output_shape
for dtype in (torch.float, torch.double):
for qmc in (False, True):
for seed in (None, 1234):
samples = construct_base_samples(
batch_shape=batch_shape,
output_shape=output_shape,
sample_shape=sample_shape,
qmc=qmc,
seed=seed,
device=device,
dtype=dtype,
)
self.assertEqual(samples.shape, expected_shape)
self.assertEqual(samples.device.type, device.type)
self.assertEqual(samples.dtype, dtype)
# check that warning is issued if dimensionality is too large
with warnings.catch_warnings(record=True) as w:
construct_base_samples(
batch_shape=torch.Size(),
output_shape=torch.Size([200, 6]),
sample_shape=torch.Size([1]),
qmc=True,
)
self.assertEqual(len(w), 1)
self.assertTrue(issubclass(w[-1].category, SamplingWarning))
exp_str = f"maximum supported by qmc ({SobolEngine.MAXDIM})"
self.assertTrue(exp_str in str(w[-1].message))
def test_construct_base_samples(self):
test_shapes = [
{
"batch": [2],
"output": [4, 3],
"sample": [5]
},
{
"batch": [1],
"output": [5, 3],
"sample": [5, 6]
},
{
"batch": [2, 3],
"output": [2, 3],
"sample": [5]
},
]
for tshape, qmc, seed, dtype in itertools.product(
test_shapes, (False, True), (None, 1234),
(torch.float, torch.double)):
batch_shape = torch.Size(tshape["batch"])
output_shape = torch.Size(tshape["output"])
sample_shape = torch.Size(tshape["sample"])
expected_shape = sample_shape + batch_shape + output_shape
samples = construct_base_samples(
batch_shape=batch_shape,
output_shape=output_shape,
sample_shape=sample_shape,
qmc=qmc,
seed=seed,
device=self.device,
dtype=dtype,
)
self.assertEqual(samples.shape, expected_shape)
self.assertEqual(samples.device.type, self.device.type)
self.assertEqual(samples.dtype, dtype)
# check that warning is issued if dimensionality is too large
with warnings.catch_warnings(record=True) as w, settings.debug(True):
construct_base_samples(
batch_shape=torch.Size(),
output_shape=torch.Size([2000,
11]), # 200 * 11 = 22000 > 21201
sample_shape=torch.Size([1]),
qmc=True,
)
self.assertEqual(len(w), 1)
self.assertTrue(issubclass(w[-1].category, SamplingWarning))
exp_str = f"maximum supported by qmc ({SobolEngine.MAXDIM})"
self.assertTrue(exp_str in str(w[-1].message))
def test_construct_base_samples(self, cuda=False):
device = torch.device("cuda") if cuda else torch.device("cpu")
test_shapes = [
{"batch": [2], "output": [4, 3], "sample": [5]},
{"batch": [1], "output": [5, 3], "sample": [5, 6]},
{"batch": [2, 3], "output": [2, 3], "sample": [5]},
]
for tshape in test_shapes:
batch_shape = torch.Size(tshape["batch"])
output_shape = torch.Size(tshape["output"])
sample_shape = torch.Size(tshape["sample"])
expected_shape = sample_shape + batch_shape + output_shape
for dtype in (torch.float, torch.double):
for qmc in (False, True):
for seed in (None, 1234):
samples = construct_base_samples(
batch_shape=batch_shape,
output_shape=output_shape,
sample_shape=sample_shape,
qmc=qmc,
seed=seed,
device=device,
dtype=dtype,
)
self.assertEqual(samples.shape, expected_shape)
self.assertEqual(samples.device.type, device.type)
self.assertEqual(samples.dtype, dtype)
# check that warning is issued if dimensionality is too large
with warnings.catch_warnings(record=True) as w:
construct_base_samples(
batch_shape=torch.Size(),
output_shape=torch.Size([200, 6]),
sample_shape=torch.Size([1]),
qmc=True,
)
self.assertEqual(len(w), 1)
self.assertTrue(issubclass(w[-1].category, SamplingWarning))
exp_str = f"maximum supported by qmc ({SobolEngine.MAXDIM})"
self.assertTrue(exp_str in str(w[-1].message))
