Skip to content

P
pytensor
提交 d782630a authored 作者: Brandon T. Willard's avatar Brandon T. Willard 提交者： Brandon T. Willard
浏览文件
操作

Move Numba random tests to test_random

上级 df58bd22
隐藏空白字符变更
内嵌 并排
正在显示 包含 593 行增加 和 573 行删除
tests/link/numba/test_basic.py
浏览文件 @ d782630a
......@@ -5,14 +5,12 @@ from unittest import mock
import numba
import numpy as np
import pytest
import scipy.stats as stats
import aesara.scalar as aes
import aesara.scalar.math as aesm
import aesara.tensor as at
import aesara.tensor.basic as atb
import aesara.tensor.math as aem
import aesara.tensor.random.basic as aer
from aesara import config, shared
from aesara.compile.function import function
from aesara.compile.mode import Mode
......@@ -1217,577 +1215,6 @@ def test_shared():
np.testing.assert_allclose(numba_res, new_a_value * 2)
@pytest.mark.parametrize(
"rv_op, dist_args, size",
[
(
aer.normal,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
at.as_tensor([3, 2]),
),
(
aer.uniform,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
at.as_tensor([3, 2]),
),
(
aer.triangular,
[
set_test_value(
at.dscalar(),
np.array(-5.0, dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(5.0, dtype=np.float64),
),
],
at.as_tensor([3, 2]),
),
(
aer.lognormal,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
at.as_tensor([3, 2]),
),
pytest.param(
aer.pareto,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
],
at.as_tensor([3, 2]),
marks=pytest.mark.xfail(reason="Not implemented"),
),
(
aer.exponential,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
],
at.as_tensor([3, 2]),
),
(
aer.weibull,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
],
at.as_tensor([3, 2]),
),
(
aer.logistic,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
at.as_tensor([3, 2]),
),
(
aer.geometric,
[
set_test_value(
at.dvector(),
np.array([0.3, 0.4], dtype=np.float64),
),
],
at.as_tensor([3, 2]),
),
(
aer.hypergeometric,
[
set_test_value(
at.lscalar(),
np.array(7, dtype=np.int64),
),
set_test_value(
at.lscalar(),
np.array(8, dtype=np.int64),
),
set_test_value(
at.lscalar(),
np.array(15, dtype=np.int64),
),
],
at.as_tensor([3, 2]),
),
(
aer.wald,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
at.as_tensor([3, 2]),
),
(
aer.laplace,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
at.as_tensor([3, 2]),
),
(
aer.binomial,
[
set_test_value(
at.lvector(),
np.array([1, 2], dtype=np.int64),
),
set_test_value(
at.dscalar(),
np.array(0.9, dtype=np.float64),
),
],
at.as_tensor([3, 2]),
),
(
aer.normal,
[
set_test_value(
at.lvector(),
np.array([1, 2], dtype=np.int64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
at.as_tensor(tuple(set_test_value(at.lscalar(), v) for v in [3, 2])),
),
(
aer.poisson,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
],
None,
),
(
aer.halfnormal,
[
set_test_value(
at.lvector(),
np.array([1, 2], dtype=np.int64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
None,
),
(
aer.bernoulli,
[
set_test_value(
at.dvector(),
np.array([0.1, 0.9], dtype=np.float64),
),
],
None,
),
(
aer.randint,
[
set_test_value(
at.lscalar(),
np.array(0, dtype=np.int64),
),
set_test_value(
at.lscalar(),
np.array(5, dtype=np.int64),
),
],
at.as_tensor([3, 2]),
),
pytest.param(
aer.multivariate_normal,
[
set_test_value(
at.dmatrix(),
np.array([[1, 2], [3, 4]], dtype=np.float64),
),
set_test_value(
at.tensor("float64", [True, False, False]),
np.eye(2)[None, ...],
),
],
at.as_tensor(tuple(set_test_value(at.lscalar(), v) for v in [4, 3, 2])),
marks=pytest.mark.xfail(reason="Not implemented"),
),
],
ids=str,
)
def test_aligned_RandomVariable(rv_op, dist_args, size):
"""Tests for Numba samplers that are one-to-one with Aesara's/NumPy's samplers."""
rng = shared(np.random.RandomState(29402))
g = rv_op(*dist_args, size=size, rng=rng)
g_fg = FunctionGraph(outputs=[g])
compare_numba_and_py(
g_fg,
[
i.tag.test_value
for i in g_fg.inputs
if not isinstance(i, (SharedVariable, Constant))
],
)
@pytest.mark.parametrize(
"rv_op, dist_args, base_size, cdf_name, params_conv",
[
(
aer.beta,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
(2,),
"beta",
lambda *args: args,
),
(
aer.gamma,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
(2,),
"gamma",
lambda a, b: (a, 0.0, b),
),
(
aer.cauchy,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
(2,),
"cauchy",
lambda *args: args,
),
(
aer.chisquare,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
)
],
(2,),
"chi2",
lambda *args: args,
),
(
aer.gumbel,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
(2,),
"gumbel_r",
lambda *args: args,
),
(
aer.negative_binomial,
[
set_test_value(
at.lvector(),
np.array([100, 200], dtype=np.int64),
),
set_test_value(
at.dscalar(),
np.array(0.09, dtype=np.float64),
),
],
(2,),
"nbinom",
lambda *args: args,
),
pytest.param(
aer.vonmises,
[
set_test_value(
at.dvector(),
np.array([-0.5, 0.5], dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
(2,),
"vonmises_line",
lambda mu, kappa: (kappa, mu),
marks=pytest.mark.xfail(
reason=(
"Numba's parameterization of `vonmises` does not match NumPy's."
"See https://github.com/numba/numba/issues/7886"
)
),
),
],
)
def test_unaligned_RandomVariable(rv_op, dist_args, base_size, cdf_name, params_conv):
"""Tests for Numba samplers that are not one-to-one with Aesara's/NumPy's samplers."""
rng = shared(np.random.RandomState(29402))
g = rv_op(*dist_args, size=(2000,) + base_size, rng=rng)
g_fn = function(dist_args, g, mode=numba_mode)
samples = g_fn(
*[
i.tag.test_value
for i in g_fn.maker.fgraph.inputs
if not isinstance(i, (SharedVariable, Constant))
]
)
bcast_dist_args = np.broadcast_arrays(*[i.tag.test_value for i in dist_args])
for idx in np.ndindex(*base_size):
cdf_params = params_conv(*tuple(arg[idx] for arg in bcast_dist_args))
test_res = stats.cramervonmises(
samples[(Ellipsis,) + idx], cdf_name, args=cdf_params
)
assert test_res.pvalue > 0.1
@pytest.mark.parametrize(
"dist_args, size, cm",
[
pytest.param(
[
set_test_value(
at.dvector(),
np.array([100000, 1, 1], dtype=np.float64),
),
],
None,
contextlib.suppress(),
),
pytest.param(
[
set_test_value(
at.dmatrix(),
np.array(
[[100000, 1, 1], [1, 100000, 1], [1, 1, 100000]],
dtype=np.float64,
),
),
],
(10, 3),
contextlib.suppress(),
),
pytest.param(
[
set_test_value(
at.dmatrix(),
np.array(
[[100000, 1, 1]],
dtype=np.float64,
),
),
],
(5, 4, 3),
contextlib.suppress(),
),
pytest.param(
[
set_test_value(
at.dmatrix(),
np.array(
[[100000, 1, 1], [1, 100000, 1], [1, 1, 100000]],
dtype=np.float64,
),
),
],
(10, 4),
pytest.raises(
ValueError, match="objects cannot be broadcast to a single shape"
),
),
],
)
def test_CategoricalRV(dist_args, size, cm):
rng = shared(np.random.RandomState(29402))
g = aer.categorical(*dist_args, size=size, rng=rng)
g_fg = FunctionGraph(outputs=[g])
with cm:
compare_numba_and_py(
g_fg,
[
i.tag.test_value
for i in g_fg.inputs
if not isinstance(i, (SharedVariable, Constant))
],
)
@pytest.mark.parametrize(
"a, size, cm",
[
pytest.param(
set_test_value(
at.dvector(),
np.array([100000, 1, 1], dtype=np.float64),
),
None,
contextlib.suppress(),
),
pytest.param(
set_test_value(
at.dmatrix(),
np.array(
[[100000, 1, 1], [1, 100000, 1], [1, 1, 100000]],
dtype=np.float64,
),
),
(10, 3),
contextlib.suppress(),
),
pytest.param(
set_test_value(
at.dmatrix(),
np.array(
[[100000, 1, 1], [1, 100000, 1], [1, 1, 100000]],
dtype=np.float64,
),
),
(10, 4),
pytest.raises(ValueError, match="Parameters shape.*"),
),
],
)
def test_DirichletRV(a, size, cm):
rng = shared(np.random.RandomState(29402))
g = aer.dirichlet(a, size=size, rng=rng)
g_fn = function([a], g, mode=numba_mode)
with cm:
a_val = a.tag.test_value
# For coverage purposes only...
eval_python_only([a], FunctionGraph(outputs=[g], clone=False), [a_val])
all_samples = []
for i in range(1000):
samples = g_fn(a_val)
all_samples.append(samples)
exp_res = a_val / a_val.sum(-1)
res = np.mean(all_samples, axis=tuple(range(0, a_val.ndim - 1)))
assert np.allclose(res, exp_res, atol=1e-4)
def test_RandomState_updates():
rng = shared(np.random.RandomState(1))
rng_new = shared(np.random.RandomState(2))
x = at.random.normal(size=10, rng=rng)
res = function([], x, updates={rng: rng_new}, mode=numba_mode)()
ref = np.random.RandomState(2).normal(size=10)
assert np.allclose(res, ref)
def test_random_Generator():
rng = shared(np.random.default_rng(29402))
g = aer.normal(rng=rng)
g_fg = FunctionGraph(outputs=[g])
with pytest.raises(TypeError):
compare_numba_and_py(
g_fg,
[
i.tag.test_value
for i in g_fg.inputs
if not isinstance(i, (SharedVariable, Constant))
],
)
def test_scan_multiple_output():
"""Test a scan implementation of a SEIR model.
......
tests/link/numba/test_random.py 0 → 100644
浏览文件 @ d782630a
import contextlib
import numpy as np
import pytest
import scipy.stats as stats
import aesara.tensor as at
import aesara.tensor.random.basic as aer
from aesara import shared
from aesara.compile.function import function
from aesara.compile.sharedvalue import SharedVariable
from aesara.graph.basic import Constant
from aesara.graph.fg import FunctionGraph
from tests.link.numba.test_basic import (
compare_numba_and_py,
eval_python_only,
numba_mode,
set_test_value,
)
rng = np.random.default_rng(42849)
@pytest.mark.parametrize(
"rv_op, dist_args, size",
[
(
aer.normal,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
at.as_tensor([3, 2]),
),
(
aer.uniform,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
at.as_tensor([3, 2]),
),
(
aer.triangular,
[
set_test_value(
at.dscalar(),
np.array(-5.0, dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(5.0, dtype=np.float64),
),
],
at.as_tensor([3, 2]),
),
(
aer.lognormal,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
at.as_tensor([3, 2]),
),
pytest.param(
aer.pareto,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
],
at.as_tensor([3, 2]),
marks=pytest.mark.xfail(reason="Not implemented"),
),
(
aer.exponential,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
],
at.as_tensor([3, 2]),
),
(
aer.weibull,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
],
at.as_tensor([3, 2]),
),
(
aer.logistic,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
at.as_tensor([3, 2]),
),
(
aer.geometric,
[
set_test_value(
at.dvector(),
np.array([0.3, 0.4], dtype=np.float64),
),
],
at.as_tensor([3, 2]),
),
(
aer.hypergeometric,
[
set_test_value(
at.lscalar(),
np.array(7, dtype=np.int64),
),
set_test_value(
at.lscalar(),
np.array(8, dtype=np.int64),
),
set_test_value(
at.lscalar(),
np.array(15, dtype=np.int64),
),
],
at.as_tensor([3, 2]),
),
(
aer.wald,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
at.as_tensor([3, 2]),
),
(
aer.laplace,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
at.as_tensor([3, 2]),
),
(
aer.binomial,
[
set_test_value(
at.lvector(),
np.array([1, 2], dtype=np.int64),
),
set_test_value(
at.dscalar(),
np.array(0.9, dtype=np.float64),
),
],
at.as_tensor([3, 2]),
),
(
aer.normal,
[
set_test_value(
at.lvector(),
np.array([1, 2], dtype=np.int64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
at.as_tensor(tuple(set_test_value(at.lscalar(), v) for v in [3, 2])),
),
(
aer.poisson,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
],
None,
),
(
aer.halfnormal,
[
set_test_value(
at.lvector(),
np.array([1, 2], dtype=np.int64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
None,
),
(
aer.bernoulli,
[
set_test_value(
at.dvector(),
np.array([0.1, 0.9], dtype=np.float64),
),
],
None,
),
(
aer.randint,
[
set_test_value(
at.lscalar(),
np.array(0, dtype=np.int64),
),
set_test_value(
at.lscalar(),
np.array(5, dtype=np.int64),
),
],
at.as_tensor([3, 2]),
),
pytest.param(
aer.multivariate_normal,
[
set_test_value(
at.dmatrix(),
np.array([[1, 2], [3, 4]], dtype=np.float64),
),
set_test_value(
at.tensor("float64", [True, False, False]),
np.eye(2)[None, ...],
),
],
at.as_tensor(tuple(set_test_value(at.lscalar(), v) for v in [4, 3, 2])),
marks=pytest.mark.xfail(reason="Not implemented"),
),
],
ids=str,
)
def test_aligned_RandomVariable(rv_op, dist_args, size):
"""Tests for Numba samplers that are one-to-one with Aesara's/NumPy's samplers."""
rng = shared(np.random.RandomState(29402))
g = rv_op(*dist_args, size=size, rng=rng)
g_fg = FunctionGraph(outputs=[g])
compare_numba_and_py(
g_fg,
[
i.tag.test_value
for i in g_fg.inputs
if not isinstance(i, (SharedVariable, Constant))
],
)
@pytest.mark.parametrize(
"rv_op, dist_args, base_size, cdf_name, params_conv",
[
(
aer.beta,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
(2,),
"beta",
lambda *args: args,
),
(
aer.gamma,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
(2,),
"gamma",
lambda a, b: (a, 0.0, b),
),
(
aer.cauchy,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
(2,),
"cauchy",
lambda *args: args,
),
(
aer.chisquare,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
)
],
(2,),
"chi2",
lambda *args: args,
),
(
aer.gumbel,
[
set_test_value(
at.dvector(),
np.array([1.0, 2.0], dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
(2,),
"gumbel_r",
lambda *args: args,
),
(
aer.negative_binomial,
[
set_test_value(
at.lvector(),
np.array([100, 200], dtype=np.int64),
),
set_test_value(
at.dscalar(),
np.array(0.09, dtype=np.float64),
),
],
(2,),
"nbinom",
lambda *args: args,
),
pytest.param(
aer.vonmises,
[
set_test_value(
at.dvector(),
np.array([-0.5, 0.5], dtype=np.float64),
),
set_test_value(
at.dscalar(),
np.array(1.0, dtype=np.float64),
),
],
(2,),
"vonmises_line",
lambda mu, kappa: (kappa, mu),
marks=pytest.mark.xfail(
reason=(
"Numba's parameterization of `vonmises` does not match NumPy's."
"See https://github.com/numba/numba/issues/7886"
)
),
),
],
)
def test_unaligned_RandomVariable(rv_op, dist_args, base_size, cdf_name, params_conv):
"""Tests for Numba samplers that are not one-to-one with Aesara's/NumPy's samplers."""
rng = shared(np.random.RandomState(29402))
g = rv_op(*dist_args, size=(2000,) + base_size, rng=rng)
g_fn = function(dist_args, g, mode=numba_mode)
samples = g_fn(
*[
i.tag.test_value
for i in g_fn.maker.fgraph.inputs
if not isinstance(i, (SharedVariable, Constant))
]
)
bcast_dist_args = np.broadcast_arrays(*[i.tag.test_value for i in dist_args])
for idx in np.ndindex(*base_size):
cdf_params = params_conv(*tuple(arg[idx] for arg in bcast_dist_args))
test_res = stats.cramervonmises(
samples[(Ellipsis,) + idx], cdf_name, args=cdf_params
)
assert test_res.pvalue > 0.1
@pytest.mark.parametrize(
"dist_args, size, cm",
[
pytest.param(
[
set_test_value(
at.dvector(),
np.array([100000, 1, 1], dtype=np.float64),
),
],
None,
contextlib.suppress(),
),
pytest.param(
[
set_test_value(
at.dmatrix(),
np.array(
[[100000, 1, 1], [1, 100000, 1], [1, 1, 100000]],
dtype=np.float64,
),
),
],
(10, 3),
contextlib.suppress(),
),
pytest.param(
[
set_test_value(
at.dmatrix(),
np.array(
[[100000, 1, 1]],
dtype=np.float64,
),
),
],
(5, 4, 3),
contextlib.suppress(),
),
pytest.param(
[
set_test_value(
at.dmatrix(),
np.array(
[[100000, 1, 1], [1, 100000, 1], [1, 1, 100000]],
dtype=np.float64,
),
),
],
(10, 4),
pytest.raises(
ValueError, match="objects cannot be broadcast to a single shape"
),
),
],
)
def test_CategoricalRV(dist_args, size, cm):
rng = shared(np.random.RandomState(29402))
g = aer.categorical(*dist_args, size=size, rng=rng)
g_fg = FunctionGraph(outputs=[g])
with cm:
compare_numba_and_py(
g_fg,
[
i.tag.test_value
for i in g_fg.inputs
if not isinstance(i, (SharedVariable, Constant))
],
)
@pytest.mark.parametrize(
"a, size, cm",
[
pytest.param(
set_test_value(
at.dvector(),
np.array([100000, 1, 1], dtype=np.float64),
),
None,
contextlib.suppress(),
),
pytest.param(
set_test_value(
at.dmatrix(),
np.array(
[[100000, 1, 1], [1, 100000, 1], [1, 1, 100000]],
dtype=np.float64,
),
),
(10, 3),
contextlib.suppress(),
),
pytest.param(
set_test_value(
at.dmatrix(),
np.array(
[[100000, 1, 1], [1, 100000, 1], [1, 1, 100000]],
dtype=np.float64,
),
),
(10, 4),
pytest.raises(ValueError, match="Parameters shape.*"),
),
],
)
def test_DirichletRV(a, size, cm):
rng = shared(np.random.RandomState(29402))
g = aer.dirichlet(a, size=size, rng=rng)
g_fn = function([a], g, mode=numba_mode)
with cm:
a_val = a.tag.test_value
# For coverage purposes only...
eval_python_only([a], FunctionGraph(outputs=[g], clone=False), [a_val])
all_samples = []
for i in range(1000):
samples = g_fn(a_val)
all_samples.append(samples)
exp_res = a_val / a_val.sum(-1)
res = np.mean(all_samples, axis=tuple(range(0, a_val.ndim - 1)))
assert np.allclose(res, exp_res, atol=1e-4)
def test_RandomState_updates():
rng = shared(np.random.RandomState(1))
rng_new = shared(np.random.RandomState(2))
x = at.random.normal(size=10, rng=rng)
res = function([], x, updates={rng: rng_new}, mode=numba_mode)()
ref = np.random.RandomState(2).normal(size=10)
assert np.allclose(res, ref)
def test_random_Generator():
rng = shared(np.random.default_rng(29402))
g = aer.normal(rng=rng)
g_fg = FunctionGraph(outputs=[g])
with pytest.raises(TypeError):
compare_numba_and_py(
g_fg,
[
i.tag.test_value
for i in g_fg.inputs
if not isinstance(i, (SharedVariable, Constant))
],
)
Markdown 格式
0% 或
您添加了 0 人 到此讨论。请谨慎行事。
请先完成此评论的编辑！
请 注册 或者 后发表评论