-
Notifications
You must be signed in to change notification settings - Fork 5.6k
/
test_tensor_fill_diagonal_tensor.py
214 lines (192 loc) · 7.37 KB
/
test_tensor_fill_diagonal_tensor.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
# Copyright (c) 2019 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
import paddle
import paddle.nn.functional as F
from paddle import base
class TensorFillDiagTensor_Test(unittest.TestCase):
def setUp(self):
self.typelist = ['float32', 'float64', 'int32', 'int64']
self.places = [base.CPUPlace()]
if base.core.is_compiled_with_cuda():
self.places.append(base.CUDAPlace(0))
def test_dim2(self):
expected_np = np.array(
[[1, 2, 2], [2, 1, 2], [2, 2, 1], [2, 2, 2]]
).astype('float32')
expected_grad = np.array(
[[0, 1, 1], [1, 0, 1], [1, 1, 0], [1, 1, 1]]
).astype('float32')
for idx, p in enumerate(self.places):
if idx == 0:
paddle.set_device('cpu')
else:
paddle.set_device('gpu')
for dtype in self.typelist:
v = paddle.ones((3,), dtype=dtype)
var = np.random.random() + 1
x = paddle.ones((4, 3), dtype=dtype)
x.stop_gradient = False
y = x * 2
y.retain_grads()
ny = y.fill_diagonal_tensor(v, offset=0, dim1=0, dim2=1)
loss = ny.sum()
loss.backward()
self.assertEqual(
(ny.numpy().astype('float32') == expected_np).all(), True
)
self.assertEqual(
(y.grad.numpy().astype('float32') == expected_grad).all(),
True,
)
def test_dim2_offset_1(self):
expected_np = np.array(
[[2, 2, 2], [1, 2, 2], [2, 1, 2], [2, 2, 1]]
).astype('float32')
expected_grad = np.array(
[[1, 1, 1], [0, 1, 1], [1, 0, 1], [1, 1, 0]]
).astype('float32')
for idx, p in enumerate(self.places):
if idx == 0:
paddle.set_device('cpu')
else:
paddle.set_device('gpu')
for dtype in self.typelist:
v = paddle.ones((3,), dtype=dtype)
var = np.random.random() + 1
x = paddle.ones((4, 3), dtype=dtype)
x.stop_gradient = False
y = x * 2
y.retain_grads()
ny = y.fill_diagonal_tensor(v, offset=-1, dim1=0, dim2=1)
loss = ny.sum()
loss.backward()
self.assertEqual(
(ny.numpy().astype('float32') == expected_np).all(), True
)
self.assertEqual(
(y.grad.numpy().astype('float32') == expected_grad).all(),
True,
)
def test_dim2_offset1(self):
expected_np = np.array(
[[2, 1, 2], [2, 2, 1], [2, 2, 2], [2, 2, 2]]
).astype('float32')
expected_grad = np.array(
[[1, 0, 1], [1, 1, 0], [1, 1, 1], [1, 1, 1]]
).astype('float32')
for idx, p in enumerate(self.places):
if idx == 0:
paddle.set_device('cpu')
else:
paddle.set_device('gpu')
for dtype in self.typelist:
v = paddle.ones((2,), dtype=dtype)
var = np.random.random() + 1
x = paddle.ones((4, 3), dtype=dtype)
x.stop_gradient = False
y = x * 2
y.retain_grads()
ny = y.fill_diagonal_tensor(v, offset=1, dim1=0, dim2=1)
loss = ny.sum()
loss.backward()
self.assertEqual(
(ny.numpy().astype('float32') == expected_np).all(), True
)
self.assertEqual(
(y.grad.numpy().astype('float32') == expected_grad).all(),
True,
)
def test_dim4(self):
expected_np = np.array(
[
[
[[0, 3], [2, 2], [2, 2]],
[[2, 2], [1, 4], [2, 2]],
[[2, 2], [2, 2], [2, 5]],
[[2, 2], [2, 2], [2, 2]],
],
[
[[6, 9], [2, 2], [2, 2]],
[[2, 2], [7, 10], [2, 2]],
[[2, 2], [2, 2], [8, 11]],
[[2, 2], [2, 2], [2, 2]],
],
]
).astype('float32')
expected_grad = np.array(
[
[
[[0, 0], [1, 1], [1, 1]],
[[1, 1], [0, 0], [1, 1]],
[[1, 1], [1, 1], [0, 0]],
[[1, 1], [1, 1], [1, 1]],
],
[
[[0, 0], [1, 1], [1, 1]],
[[1, 1], [0, 0], [1, 1]],
[[1, 1], [1, 1], [0, 0]],
[[1, 1], [1, 1], [1, 1]],
],
]
).astype('float32')
for idx, p in enumerate(self.places):
if idx == 0:
paddle.set_device('cpu')
else:
paddle.set_device('gpu')
for dtype in self.typelist:
v = paddle.to_tensor(
np.arange(12).reshape(2, 2, 3), dtype=dtype
)
var = np.random.random() + 1
x = paddle.ones((2, 4, 3, 2), dtype=dtype)
x.stop_gradient = False
y = x * 2
y.retain_grads()
ny = y.fill_diagonal_tensor(v, offset=0, dim1=1, dim2=2)
loss = ny.sum()
loss.backward()
self.assertEqual(
(ny.numpy().astype('float32') == expected_np).all(), True
)
self.assertEqual(
(y.grad.numpy().astype('float32') == expected_grad).all(),
True,
)
def test_largedim(self):
if len(self.places) > 1:
bsdim = 1024
fsdim = 128
paddle.set_device('gpu')
for dtype in self.typelist:
v = paddle.arange(bsdim * fsdim, dtype=dtype).reshape(
(bsdim, fsdim)
)
y = paddle.ones((bsdim, fsdim, fsdim), dtype=dtype)
y.stop_gradient = False
y = y * 2
y.retain_grads()
ny = y.fill_diagonal_tensor(v, offset=0, dim1=1, dim2=2)
loss = ny.sum()
loss.backward()
expected_pred = v - 2
expected_pred = F.diag_embed(expected_pred) + 2
expected_grad = paddle.ones(v.shape, dtype=dtype) - 2
expected_grad = F.diag_embed(expected_grad) + 1
self.assertEqual((ny == expected_pred).all(), True)
self.assertEqual((y.grad == expected_grad).all(), True)
if __name__ == '__main__':
unittest.main()