wehub-resource-sync
147 lines
4.6 KiB
Python
147 lines
4.6 KiB
Python
# Copyright (c) 2024 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.
|
|
|
|
import unittest
|
|
|
|
import numpy as np
|
|
from op_test import get_devices
|
|
|
|
import paddle
|
|
|
|
|
|
def log_normal_mean(mean, std):
|
|
return np.exp(mean + np.power(std, 2) / 2.0)
|
|
|
|
|
|
def log_normal_var(mean, std):
|
|
var = np.power(std, 2)
|
|
return (np.exp(var) - 1.0) * np.exp(2 * mean + var)
|
|
|
|
|
|
class TestLogNormalRandomInplaceOpDtype(unittest.TestCase):
|
|
def setUp(self):
|
|
self.shape = (1000, 784)
|
|
|
|
def test_log_normal_inplace_op_dtype(self):
|
|
def test_fp32():
|
|
tensor_fp32 = paddle.ones(self.shape, dtype=paddle.float32)
|
|
tensor_fp32.log_normal_()
|
|
self.assertEqual(tensor_fp32.dtype, paddle.float32)
|
|
|
|
def test_fp64():
|
|
tensor_fp64 = paddle.ones(self.shape, paddle.float64)
|
|
tensor_fp64.log_normal_()
|
|
self.assertEqual(tensor_fp64.dtype, paddle.float64)
|
|
|
|
places = get_devices()
|
|
for place in places:
|
|
paddle.set_device(place)
|
|
test_fp32()
|
|
test_fp64()
|
|
|
|
|
|
class TestLogNormalRandomInplaceOpIsInplace(unittest.TestCase):
|
|
def setUp(self):
|
|
self.shape = (1000, 784)
|
|
|
|
def test_log_normal_inplace_op_is_inplace(self):
|
|
tensor_a = paddle.ones(self.shape)
|
|
tensor_b = tensor_a.log_normal_()
|
|
self.assertTrue(tensor_a is tensor_b)
|
|
|
|
|
|
class TestLogNormalRandomInplaceOpSeedIsZero(unittest.TestCase):
|
|
def setUp(self):
|
|
self.shape = (1000, 784)
|
|
|
|
def test_log_normal_inplace_op_not_equal(self):
|
|
tensor = paddle.ones(self.shape)
|
|
tensor.log_normal_()
|
|
tensor_data_first = tensor.numpy()
|
|
tensor.log_normal_()
|
|
tensor_data_second = tensor.numpy()
|
|
self.assertFalse((tensor_data_first == tensor_data_second).all())
|
|
|
|
|
|
class TestLogNormalRandomInplaceOpShape(unittest.TestCase):
|
|
def setUp(self):
|
|
self.shape = (1000, 784)
|
|
|
|
def test_log_normal_inplace_op_shape(self):
|
|
tensor = paddle.ones(self.shape)
|
|
tensor.log_normal_()
|
|
tensor_shape_np = np.array(tensor.shape)
|
|
origin_shape = np.array(self.shape)
|
|
self.assertTrue((tensor_shape_np == origin_shape).all())
|
|
|
|
|
|
class TestLogNormalRandomInplaceOpDistribution(unittest.TestCase):
|
|
def setUp(self):
|
|
self.shape = (1000, 784)
|
|
self.mean = -1
|
|
self.std = 1
|
|
|
|
def test_log_normal_inplace_op_distribution(self):
|
|
tensor = paddle.ones(self.shape)
|
|
tensor.log_normal_(self.mean, self.std)
|
|
mean = np.mean(tensor.numpy())
|
|
var = np.var(tensor.numpy())
|
|
mean_ref = log_normal_mean(self.mean, self.std)
|
|
var_ref = log_normal_var(self.mean, self.std)
|
|
np.testing.assert_allclose(mean_ref, mean, rtol=0.2, atol=0.2)
|
|
np.testing.assert_allclose(var_ref, var, rtol=0.2, atol=0.2)
|
|
|
|
|
|
class TestLogNormalRandomInplaceOpEmptyTensor(unittest.TestCase):
|
|
def test_log_normal_inplace_op_empty_tensor(self):
|
|
places = get_devices()
|
|
test_shapes = [(200, 0), (0, 200)]
|
|
for place in places:
|
|
paddle.set_device(place)
|
|
for test_shape in test_shapes:
|
|
tensor = paddle.empty(shape=test_shape)
|
|
tensor.log_normal_()
|
|
tensor_shape_np = np.array(tensor.shape)
|
|
origin_shape = np.array(test_shape)
|
|
self.assertTrue((tensor_shape_np == origin_shape).all())
|
|
|
|
|
|
class TestLogNormalRandomInplaceGrad(unittest.TestCase):
|
|
def setUp(self):
|
|
self.shape = (1000, 784)
|
|
|
|
def run_(self):
|
|
def test_grad():
|
|
tensor_a = paddle.ones(self.shape)
|
|
tensor_a.stop_gradient = False
|
|
tensor_b = tensor_a * 0.5
|
|
tensor_b.retain_grads()
|
|
tensor_b.log_normal_(mean=-2.0, std=2.0)
|
|
loss = tensor_b.sum()
|
|
loss.backward()
|
|
log_normal_grad = tensor_b.grad.numpy()
|
|
self.assertTrue((log_normal_grad == 0).all())
|
|
|
|
places = get_devices()
|
|
for place in places:
|
|
paddle.set_device(place)
|
|
test_grad()
|
|
|
|
def test_log_normal_inplace_grad(self):
|
|
self.run_()
|
|
|
|
|
|
if __name__ == '__main__':
|
|
unittest.main()
|