wehub-resource-sync
220 lines
7.7 KiB
Python
220 lines
7.7 KiB
Python
# Copyright (c) 2026 PaddlePaddle Authors. All Rights Reserved.
|
|
#
|
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
|
# you may not use this file except in compliance with the License.
|
|
# You may obtain a copy of the License at
|
|
#
|
|
# http://www.apache.org/licenses/LICENSE-2.0
|
|
#
|
|
# Unless required by applicable law or agreed to in writing, software
|
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
# See the License for the specific language governing permissions and
|
|
# limitations under the License.
|
|
|
|
"""
|
|
张量数学运算高级测试 / Advanced Tensor Math Operations Tests
|
|
|
|
测试目标 / Test Target:
|
|
paddle.tensor.math 高级数学函数
|
|
|
|
覆盖的模块 / Covered Modules:
|
|
- paddle.logsumexp: log-sum-exp
|
|
- paddle.logit: logit变换
|
|
- paddle.log1p: log(1+x)
|
|
- paddle.expm1: exp(x)-1
|
|
- paddle.i0/i1: 贝塞尔函数
|
|
- paddle.sinc: sinc函数
|
|
- paddle.round/ceil/floor/trunc: 取整函数
|
|
|
|
作用 / Purpose:
|
|
补充高级数学运算API的测试，提升覆盖率。
|
|
"""
|
|
|
|
import unittest
|
|
|
|
import numpy as np
|
|
|
|
import paddle
|
|
|
|
paddle.disable_static()
|
|
|
|
|
|
class TestLogarithmicOps(unittest.TestCase):
|
|
"""测试对数运算 / Test logarithmic operations"""
|
|
|
|
def test_logsumexp(self):
|
|
"""测试log-sum-exp / Test log-sum-exp"""
|
|
x = paddle.to_tensor([[1.0, 2.0, 3.0], [4.0, 5.0, 6.0]])
|
|
result = paddle.logsumexp(x, axis=1)
|
|
self.assertEqual(result.shape, [2])
|
|
# log-sum-exp of [1,2,3] should be ≈ 3.4076
|
|
np.testing.assert_allclose(
|
|
float(result[0].numpy()),
|
|
np.log(np.exp(1) + np.exp(2) + np.exp(3)),
|
|
rtol=1e-5,
|
|
)
|
|
|
|
def test_log1p(self):
|
|
"""测试log(1+x) / Test log1p"""
|
|
x = paddle.to_tensor([0.0, 1.0, 2.0])
|
|
result = paddle.log1p(x)
|
|
expected = np.log1p([0.0, 1.0, 2.0])
|
|
np.testing.assert_allclose(result.numpy(), expected, rtol=1e-5)
|
|
|
|
def test_expm1(self):
|
|
"""测试exp(x)-1 / Test expm1"""
|
|
x = paddle.to_tensor([0.0, 0.5, 1.0])
|
|
result = paddle.expm1(x)
|
|
expected = np.expm1([0.0, 0.5, 1.0])
|
|
np.testing.assert_allclose(result.numpy(), expected, rtol=1e-5)
|
|
|
|
def test_logit(self):
|
|
"""测试logit变换 / Test logit transform"""
|
|
x = paddle.to_tensor([0.1, 0.5, 0.9])
|
|
result = paddle.logit(x)
|
|
self.assertEqual(result.shape, [3])
|
|
|
|
|
|
class TestRoundingOps(unittest.TestCase):
|
|
"""测试取整运算 / Test rounding operations"""
|
|
|
|
def test_round(self):
|
|
"""测试round / Test round"""
|
|
x = paddle.to_tensor([1.4, 1.5, 2.6, -1.4])
|
|
result = paddle.round(x)
|
|
self.assertEqual(result.shape, [4])
|
|
|
|
def test_ceil(self):
|
|
"""测试ceil / Test ceiling"""
|
|
x = paddle.to_tensor([1.1, 1.9, 2.0, -1.1])
|
|
result = paddle.ceil(x)
|
|
expected = np.ceil([1.1, 1.9, 2.0, -1.1])
|
|
np.testing.assert_allclose(result.numpy(), expected)
|
|
|
|
def test_floor(self):
|
|
"""测试floor / Test floor"""
|
|
x = paddle.to_tensor([1.9, 2.0, 2.1, -1.1])
|
|
result = paddle.floor(x)
|
|
expected = np.floor([1.9, 2.0, 2.1, -1.1])
|
|
np.testing.assert_allclose(result.numpy(), expected)
|
|
|
|
def test_trunc(self):
|
|
"""测试截断 / Test truncation"""
|
|
x = paddle.to_tensor([1.7, -1.7, 2.3, -2.3])
|
|
result = paddle.trunc(x)
|
|
expected = np.trunc([1.7, -1.7, 2.3, -2.3])
|
|
np.testing.assert_allclose(result.numpy(), expected)
|
|
|
|
def test_frac(self):
|
|
"""测试小数部分 / Test fractional part"""
|
|
x = paddle.to_tensor([1.7, -1.7, 2.3])
|
|
result = paddle.frac(x)
|
|
self.assertEqual(result.shape, [3])
|
|
|
|
|
|
class TestTrigonometricOps(unittest.TestCase):
|
|
"""测试三角函数 / Test trigonometric operations"""
|
|
|
|
def test_sin_cos_tan(self):
|
|
"""测试基本三角函数 / Test basic trig functions"""
|
|
x = paddle.to_tensor([0.0, np.pi / 4, np.pi / 2, np.pi])
|
|
sin_result = paddle.sin(x)
|
|
cos_result = paddle.cos(x)
|
|
tan_result = paddle.tan(paddle.to_tensor([0.0, np.pi / 4]))
|
|
np.testing.assert_allclose(sin_result.numpy()[0], 0.0, atol=1e-6)
|
|
np.testing.assert_allclose(cos_result.numpy()[0], 1.0, atol=1e-6)
|
|
|
|
def test_asin_acos_atan(self):
|
|
"""测试反三角函数 / Test inverse trig functions"""
|
|
x = paddle.to_tensor([0.0, 0.5, 1.0])
|
|
asin_result = paddle.asin(x)
|
|
acos_result = paddle.acos(x)
|
|
atan_result = paddle.atan(x)
|
|
self.assertEqual(asin_result.shape, [3])
|
|
self.assertEqual(acos_result.shape, [3])
|
|
self.assertEqual(atan_result.shape, [3])
|
|
|
|
def test_atan2(self):
|
|
"""测试atan2 / Test atan2"""
|
|
y = paddle.to_tensor([1.0, -1.0, 1.0])
|
|
x = paddle.to_tensor([1.0, 1.0, -1.0])
|
|
result = paddle.atan2(y, x)
|
|
self.assertEqual(result.shape, [3])
|
|
|
|
def test_sinh_cosh_tanh(self):
|
|
"""测试双曲函数 / Test hyperbolic functions"""
|
|
x = paddle.to_tensor([0.0, 0.5, 1.0])
|
|
sinh_result = paddle.sinh(x)
|
|
cosh_result = paddle.cosh(x)
|
|
tanh_result = paddle.tanh(x)
|
|
np.testing.assert_allclose(
|
|
sinh_result.numpy(), np.sinh([0.0, 0.5, 1.0]), rtol=1e-5
|
|
)
|
|
|
|
|
|
class TestClampAndAbs(unittest.TestCase):
|
|
"""测试clip/abs运算 / Test clamp and abs operations"""
|
|
|
|
def test_clip(self):
|
|
"""测试clip / Test clip (clamp)"""
|
|
x = paddle.to_tensor([-2.0, -1.0, 0.0, 1.0, 2.0])
|
|
result = paddle.clip(x, min=-1.0, max=1.0)
|
|
expected = np.clip([-2.0, -1.0, 0.0, 1.0, 2.0], -1.0, 1.0)
|
|
np.testing.assert_allclose(result.numpy(), expected)
|
|
|
|
def test_abs(self):
|
|
"""测试绝对值 / Test absolute value"""
|
|
x = paddle.to_tensor([-3.0, -1.5, 0.0, 1.5, 3.0])
|
|
result = paddle.abs(x)
|
|
expected = np.abs([-3.0, -1.5, 0.0, 1.5, 3.0])
|
|
np.testing.assert_allclose(result.numpy(), expected)
|
|
|
|
def test_sign(self):
|
|
"""测试符号函数 / Test sign function"""
|
|
x = paddle.to_tensor([-3.0, 0.0, 3.0])
|
|
result = paddle.sign(x)
|
|
expected = np.sign([-3.0, 0.0, 3.0])
|
|
np.testing.assert_allclose(result.numpy(), expected)
|
|
|
|
def test_maximum_minimum(self):
|
|
"""测试逐元素最大/最小值 / Test element-wise max/min"""
|
|
a = paddle.to_tensor([1.0, 5.0, 3.0])
|
|
b = paddle.to_tensor([4.0, 2.0, 6.0])
|
|
max_result = paddle.maximum(a, b)
|
|
min_result = paddle.minimum(a, b)
|
|
np.testing.assert_allclose(max_result.numpy(), [4.0, 5.0, 6.0])
|
|
np.testing.assert_allclose(min_result.numpy(), [1.0, 2.0, 3.0])
|
|
|
|
|
|
class TestModAndDivOps(unittest.TestCase):
|
|
"""测试取模和整除运算 / Test modulo and floor divide operations"""
|
|
|
|
def test_mod(self):
|
|
"""测试取模 / Test modulo"""
|
|
x = paddle.to_tensor([10.0, 11.0, 12.0])
|
|
y = paddle.to_tensor([3.0, 4.0, 5.0])
|
|
result = paddle.mod(x, y)
|
|
expected = np.mod([10.0, 11.0, 12.0], [3.0, 4.0, 5.0])
|
|
np.testing.assert_allclose(result.numpy(), expected)
|
|
|
|
def test_floor_divide(self):
|
|
"""测试整除 / Test floor divide"""
|
|
x = paddle.to_tensor([10.0, 11.0, 12.0])
|
|
y = paddle.to_tensor([3.0, 4.0, 5.0])
|
|
result = paddle.floor_divide(x, y)
|
|
expected = np.floor_divide([10.0, 11.0, 12.0], [3.0, 4.0, 5.0])
|
|
np.testing.assert_allclose(result.numpy(), expected)
|
|
|
|
def test_pow(self):
|
|
"""测试幂运算 / Test power operation"""
|
|
x = paddle.to_tensor([2.0, 3.0, 4.0])
|
|
y = paddle.to_tensor([2.0, 3.0, 0.5])
|
|
result = paddle.pow(x, y)
|
|
expected = np.power([2.0, 3.0, 4.0], [2.0, 3.0, 0.5])
|
|
np.testing.assert_allclose(result.numpy(), expected, rtol=1e-5)
|
|
|
|
|
|
if __name__ == '__main__':
|
|
unittest.main()
|