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
|
/*
* Copyright (c) 2019 Samsung Electronics Co., Ltd. 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.
*/
#include "MaxPoolCanonicalizer.h"
#include <moco/IR/TFDialect.h>
#include <moco/Support/TFShapeInferenceHelper.h>
#include "CodecHelper.h"
#include <moco/Log.h>
namespace
{
bool canonicalize_maxpool2d(loco::Graph *graph, moco::TFMaxPool *node)
{
LOGGER(l);
/**
* @note This will replace TFMaxPool node with Canonical FeatureEncode +
* MaxPool2D + FeatureDecode
*
* Before
* A -- TFMaxPool -- C
*
* After
* +- TFMaxPool --
* |
* A -+- FeatureEncode -- MaxPool2D -- FeatureDecode -- C
*
* Where
* A : value of TFMaxPool
* C : a node that uses TFMaxPool as an input
* TFMaxPool is disconnected from other nodes
*/
auto data_layout = plier::tf::as_data_layout(node->data_layout());
auto feature_enc = graph->nodes()->create<loco::FeatureEncode>();
auto maxPool2d_node = graph->nodes()->create<loco::MaxPool2D>();
auto feature_dec = graph->nodes()->create<loco::FeatureDecode>();
set_feature_enc(feature_enc, data_layout);
set_feature_dec(feature_dec, data_layout);
// paddata to pad
auto input_shape = moco::node_shape(node->input());
assert(input_shape.domain() != loco::Domain::Unknown);
auto node_stride = moco::stride_of(node->strides(), node->data_layout());
auto node_window = moco::window_of(node->ksize(), node->data_layout());
moco::Padding2DInference infer_padding2d;
infer_padding2d.padding(node->padding());
infer_padding2d.stride(node_stride);
infer_padding2d.window(node_window);
auto input_feature_shape = moco::as_feature_shape(input_shape, node->data_layout());
auto input_plane_shape = moco::make_plane_shape(input_feature_shape);
*maxPool2d_node->pad() = infer_padding2d(input_plane_shape);
*maxPool2d_node->stride() = node_stride;
*maxPool2d_node->window() = node_window;
INFO(l) << "Canonicalize TFMaxPool pad = T " << maxPool2d_node->pad()->top() << ", L "
<< maxPool2d_node->pad()->left() << ", B " << maxPool2d_node->pad()->bottom() << ", R "
<< maxPool2d_node->pad()->right() << std::endl;
// update graph
auto node_A = node->input();
// update connections
feature_enc->input(node_A);
maxPool2d_node->ifm(feature_enc);
feature_dec->input(maxPool2d_node);
// replace node
replace(node).with(feature_dec);
return true;
}
} // namespace
namespace moco
{
namespace tf
{
bool MaxPoolCanonicalizer::transform(TFMaxPool *node) const
{
return canonicalize_maxpool2d(node->graph(), node);
}
} // namespace tf
} // namespace moco
|
