forked from pytorch/pytorch
-
Notifications
You must be signed in to change notification settings - Fork 0
/
layer_parameter_sharing_test.py
233 lines (202 loc) · 9.09 KB
/
layer_parameter_sharing_test.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
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
from __future__ import unicode_literals
from caffe2.python import core, scope
from caffe2.python.modeling.parameter_sharing import (
ParameterSharing,
)
from caffe2.python.optimizer import AdagradOptimizer, AdamOptimizer
from caffe2.python.layer_test_util import LayersTestCase
import six
class ParameterSharingTest(LayersTestCase):
def test_layer_parameter_name(self):
output_dims = 2
with scope.NameScope('global_scope'):
fc1_output = self.model.FC(
self.model.input_feature_schema.float_features,
output_dims
)
self.assertEquals(self.model.layers[-1].w, 'global_scope/fc/w')
self.assertEquals(fc1_output(), 'global_scope/fc/output')
with scope.NameScope('nested_scope'):
fc2_output = self.model.FC(
fc1_output,
output_dims
)
self.assertEquals(self.model.layers[-1].w,
'global_scope/nested_scope/fc/w')
self.assertEquals(fc2_output(),
'global_scope/nested_scope/fc/output')
fc3_output = self.model.FC(
fc1_output,
output_dims
)
self.assertEquals(self.model.layers[-1].w,
'global_scope/nested_scope/fc_auto_0/w')
self.assertEquals(fc3_output(),
'global_scope/nested_scope/fc_auto_0/output')
def test_layer_shared_parameter_name_different_namescopes(self):
output_dims = 2
with scope.NameScope('global_scope'):
with ParameterSharing({'scope_1': 'scope_0'}):
with scope.NameScope('scope_0'):
fc1_output = self.model.FC(
self.model.input_feature_schema.float_features,
output_dims
)
self.assertEquals(self.model.layers[-1].w,
'global_scope/scope_0/fc/w')
self.assertEquals(fc1_output(),
'global_scope/scope_0/fc/output')
with scope.NameScope('scope_1'):
fc2_output = self.model.FC(
self.model.input_feature_schema.float_features,
output_dims
)
self.assertEquals(self.model.layers[-1].w,
'global_scope/scope_0/fc/w')
self.assertEquals(fc2_output(),
'global_scope/scope_1/fc/output')
def test_layer_shared_parameter_name_within_same_namescope(self):
output_dims = 2
with scope.NameScope('global_scope'):
with ParameterSharing({'fc_auto_0': 'fc'}):
self.model.FC(
self.model.input_feature_schema.float_features,
output_dims
)
self.assertEquals(self.model.layers[-1].w,
'global_scope/fc/w')
self.model.FC(
self.model.input_feature_schema.float_features,
output_dims
)
self.assertEquals(self.model.layers[-1].w,
'global_scope/fc/w')
def test_layer_shared_parameter_name_within_same_namescope_customized_name(self):
output_dims = 2
with scope.NameScope('global_scope'):
with ParameterSharing({'new_fc': 'shared_fc'}):
self.model.FC(
self.model.input_feature_schema.float_features,
output_dims,
name='shared_fc'
)
self.assertEquals(self.model.layers[-1].w,
'global_scope/shared_fc/w')
self.model.FC(
self.model.input_feature_schema.float_features,
output_dims,
name='new_fc'
)
self.assertEquals(self.model.layers[-1].w,
'global_scope/shared_fc/w')
def test_layer_shared_parameter_name_different_shapes(self):
output_dims = 2
with scope.NameScope('global_scope'):
with ParameterSharing({'fc_auto_0': 'fc'}):
self.model.FC(
self.model.input_feature_schema.float_features,
output_dims
)
self.assertEquals(self.model.layers[-1].w,
'global_scope/fc/w')
with six.assertRaisesRegex(self, ValueError, 'Got inconsistent shapes .*'):
self.model.FC(
self.model.input_feature_schema.float_features,
output_dims + 1
)
def test_layer_duplicated_parameter_init(self):
output_dims = 2
with scope.NameScope('global_scope'):
with ParameterSharing({'new_fc': 'shared_fc'}):
self.model.FC(
self.model.input_feature_schema.float_features,
output_dims,
name='shared_fc'
)
self.model.FC(
self.model.input_feature_schema.float_features,
output_dims,
name='new_fc'
)
train_init_net = core.Net('train_init_net')
train_net = core.Net('train_net')
for layer in self.model.layers:
layer.add_operators(train_net, train_init_net)
op_outputs = []
for op in train_init_net._net.op:
op_outputs.extend(op.output)
# only fill these parameter blobs once
self.assertEquals(
sorted(op_outputs),
['global_scope/shared_fc/b', 'global_scope/shared_fc/w']
)
def test_layer_shared_parameter_optim_validator(self):
"""
This test is to cover the _validate_param_optim function in
layer_model_helper class.
"""
output_dims = 2
adagrad_optim = AdagradOptimizer(
alpha=0.004,
epsilon=0.02,
)
self.model.default_optimizer = adagrad_optim
# the following covers the branch -- optim is None
with scope.NameScope('global_scope_0'):
with ParameterSharing({'scope_1': 'scope_0'}):
with scope.NameScope('scope_0'):
fc1_output = self.model.FC(
self.model.input_feature_schema.float_features,
output_dims,
weight_optim=self.model.NoOptim,
)
with scope.NameScope('scope_1'), self.assertRaises(Exception):
fc2_output = self.model.FC(
self.model.input_feature_schema.float_features,
output_dims
)
# the following covers the branch -- optim is NoOptim
with scope.NameScope('global_scope_1'):
with ParameterSharing({'scope_1': 'scope_0'}):
with scope.NameScope('scope_0'):
fc1_output = self.model.FC(
self.model.input_feature_schema.float_features,
output_dims,
weight_optim=None,
)
with scope.NameScope('scope_1'), self.assertRaises(Exception):
fc2_output = self.model.FC(
self.model.input_feature_schema.float_features,
output_dims,
weight_optim=self.model.NoOptim,
)
# the following covers the branch -- optim is an instance of Optimizer
adagrad_optim_2 = AdagradOptimizer(
alpha=0.005,
epsilon=0.02,
)
adam_optim = AdamOptimizer()
self.model.default_optimizer = adagrad_optim_2
with scope.NameScope('global_scope_2'):
with ParameterSharing({'scope_1': 'scope_0', 'scope_2': 'scope_0'}):
with scope.NameScope('scope_0'):
fc1_output = self.model.FC(
self.model.input_feature_schema.float_features,
output_dims,
weight_optim=None, # it will use adagrad_optim_2
)
with scope.NameScope('scope_1'), self.assertRaises(Exception):
fc2_output = self.model.FC(
self.model.input_feature_schema.float_features,
output_dims,
weight_optim=adagrad_optim,
)
with scope.NameScope('scope_2'), self.assertRaises(Exception):
fc2_output = self.model.FC(
self.model.input_feature_schema.float_features,
output_dims,
weight_optim=adam_optim,
)