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
|
# Copyright 2017 The TensorFlow 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.
# ==============================================================================
"""Tests for Adagrad."""
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
import numpy as np
from tensorflow.compiler.tests import xla_test
from tensorflow.python.framework import constant_op
from tensorflow.python.ops import resource_variable_ops
from tensorflow.python.ops import variables
from tensorflow.python.platform import test
from tensorflow.python.training import adagrad
class AdagradOptimizerTest(xla_test.XLATestCase):
def testBasic(self):
for dtype in self.float_types:
with self.cached_session(), self.test_scope():
var0 = resource_variable_ops.ResourceVariable([1.0, 2.0], dtype=dtype)
var1 = resource_variable_ops.ResourceVariable([3.0, 4.0], dtype=dtype)
grads0 = constant_op.constant([0.1, 0.1], dtype=dtype)
grads1 = constant_op.constant([0.01, 0.01], dtype=dtype)
ada_opt = adagrad.AdagradOptimizer(3.0, initial_accumulator_value=0.1)
ada_update = ada_opt.apply_gradients(
zip([grads0, grads1], [var0, var1]))
variables.global_variables_initializer().run()
# Fetch params to validate initial values
self.assertAllClose([1.0, 2.0], var0.eval())
self.assertAllClose([3.0, 4.0], var1.eval())
# Run 3 steps of adagrad
for _ in range(3):
ada_update.run()
# Validate updated params
self.assertAllCloseAccordingToType(
np.array([-1.6026098728179932, -0.6026098728179932]), var0.eval(),
float_rtol=1e-5)
self.assertAllCloseAccordingToType(
np.array([2.715679168701172, 3.715679168701172]), var1.eval(),
float_rtol=1e-5)
def testTensorLearningRate(self):
for dtype in self.float_types:
with self.cached_session(), self.test_scope():
var0 = resource_variable_ops.ResourceVariable([1.0, 2.0], dtype=dtype)
var1 = resource_variable_ops.ResourceVariable([3.0, 4.0], dtype=dtype)
grads0 = constant_op.constant([0.1, 0.1], dtype=dtype)
grads1 = constant_op.constant([0.01, 0.01], dtype=dtype)
ada_opt = adagrad.AdagradOptimizer(
constant_op.constant(3.0), initial_accumulator_value=0.1)
ada_update = ada_opt.apply_gradients(
zip([grads0, grads1], [var0, var1]))
variables.global_variables_initializer().run()
# Fetch params to validate initial values
self.assertAllClose([1.0, 2.0], var0.eval())
self.assertAllClose([3.0, 4.0], var1.eval())
# Run 3 steps of adagrad
for _ in range(3):
ada_update.run()
# Validate updated params
self.assertAllCloseAccordingToType(
np.array([-1.6026098728179932, -0.6026098728179932]), var0.eval(),
float_rtol=1e-5)
self.assertAllCloseAccordingToType(
np.array([2.715679168701172, 3.715679168701172]), var1.eval(),
float_rtol=1e-5)
def testSharing(self):
for dtype in self.float_types:
with self.cached_session(), self.test_scope():
var0 = resource_variable_ops.ResourceVariable([1.0, 2.0], dtype=dtype)
var1 = resource_variable_ops.ResourceVariable([3.0, 4.0], dtype=dtype)
grads0 = constant_op.constant([0.1, 0.1], dtype=dtype)
grads1 = constant_op.constant([0.01, 0.01], dtype=dtype)
ada_opt = adagrad.AdagradOptimizer(3.0)
# Apply the optimizer twice. Both applications will use
# the same accums.
ada_update1 = ada_opt.apply_gradients(
zip([grads0, grads1], [var0, var1]))
ada_update2 = ada_opt.apply_gradients(
zip([grads0, grads1], [var0, var1]))
self.assertEqual(["accumulator"], ada_opt.get_slot_names())
slot0 = ada_opt.get_slot(var0, "accumulator")
self.assertEquals(slot0.get_shape(), var0.get_shape())
slot1 = ada_opt.get_slot(var1, "accumulator")
self.assertEquals(slot1.get_shape(), var1.get_shape())
variables.global_variables_initializer().run()
# Fetch params to validate initial values.
self.assertAllClose([1.0, 2.0], var0.eval())
self.assertAllClose([3.0, 4.0], var1.eval())
# Mix the first and the second adagrad for 3 steps.
ada_update1.run()
ada_update2.run()
ada_update1.run()
# Validate updated params (the same as with only 1 Adagrad).
self.assertAllCloseAccordingToType(
np.array([-1.6026098728179932, -0.6026098728179932]), var0.eval(),
float_rtol=1e-5)
self.assertAllCloseAccordingToType(
np.array([2.715679168701172, 3.715679168701172]), var1.eval(),
float_rtol=1e-5)
if __name__ == "__main__":
test.main()
|
