Gitiles
Code Review Sign In
review.mlplatform.org / ml / armnn / 31a7c890207f80afc866f706d797267d78d05a0d / . / docs / 01_02_deserializer_serializer.dox
blob: 047cb5dc8d6d7447a9bb9c7020da28ada707f284 [file] [log] [blame]
Jan Eilers31a7c892021-01-29 14:18:13 +0000[diff] [blame^]1/// Copyright (c) 2021 ARM Limited and Contributors. All rights reserved.
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]2///
3/// SPDX-License-Identifier: MIT
4///
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]5
6namespace armnn
7{
8/**
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]9@page serializer Serializer
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]10@tableofcontents
11
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]12The `armnnSerializer` is a library for serializing an Arm NN network to a stream.
13
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]14@section serializersupport Supported Layers
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]15
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]16This reference guide provides a list of layers which can be serialized by the Arm NN SDK.
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]17
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]18@subsection serializersupportflully Fully supported
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]19
20The Arm NN SDK Serializer currently supports the following layers:
21
22- Activation
23- Addition
24- ArgMinMax
25- BatchToSpaceNd
26- BatchNormalization
27- Comparison
28- Concat
29- Constant
30- Convolution2d
31- DepthToSpace
32- DepthwiseConvolution2d
33- Dequantize
34- DetectionPostProcess
35- Division
36- ElementwiseUnary
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]37- Fill
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]38- Floor
39- FullyConnected
40- Gather
41- Input
42- InstanceNormalization
43- L2Normalization
James Ward590f3f72020-12-02 10:20:48 +0000[diff] [blame]44- Logical
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]45- LogSoftmax
46- Lstm
47- Maximum
48- Mean
49- Merge
50- Minimum
51- Multiplication
52- Normalization
53- Output
54- Pad
55- Permute
56- Pooling2d
57- Prelu
James Ward590f3f72020-12-02 10:20:48 +0000[diff] [blame]58- QLstm
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]59- Quantize
60- QuantizedLstm
James Ward590f3f72020-12-02 10:20:48 +0000[diff] [blame]61- Rank
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]62- Reshape
63- Resize
64- Slice
65- Softmax
66- SpaceToBatchNd
67- SpaceToDepth
68- Splitter
69- Stack
70- StandIn
71- StridedSlice
72- Subtraction
73- Switch
James Ward590f3f72020-12-02 10:20:48 +0000[diff] [blame]74- Transpose
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]75- TransposeConvolution2d
76
77More machine learning layers will be supported in future releases.
78
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]79@subsection serializersupportdeprecated Deprecated layers
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]80
81Some layers have been deprecated and replaced by others layers. In order to maintain backward compatibility, serializations of these deprecated layers will deserialize to the layers that have replaced them, as follows:
82
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]83- Abs will deserialize as ElementwiseUnary
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]84- Equal will deserialize as Comparison
85- Greater will deserialize as Comparison
86- Merger will deserialize as Concat
87- ResizeBilinear will deserialize as Resize
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]88- Rsqrt will deserialize as ElementwiseUnary
89<br/><br/><br/><br/>
90
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]91@page deserializer Deserializer
92@tableofcontents
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]93
94The `armnnDeserializer` is a library for loading neural networks defined by Arm NN FlatBuffers files
95into the Arm NN runtime.
96
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]97@section deserializersupport Supported Layers
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]98
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]99This reference guide provides a list of layers which can be deserialized by the Arm NN SDK.
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]100
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]101@subsection deserializersupportfully Fully supported
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]102
103The Arm NN SDK Deserialize parser currently supports the following layers:
104
105- Abs
106- Activation
107- Addition
108- ArgMinMax
109- BatchToSpaceNd
110- BatchNormalization
111- Concat
112- Comparison
113- Constant
114- Convolution2d
115- DepthToSpace
116- DepthwiseConvolution2d
117- Dequantize
118- DetectionPostProcess
119- Division
James Ward590f3f72020-12-02 10:20:48 +0000[diff] [blame]120- ElementwiseUnary
121- Fill
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]122- Floor
123- FullyConnected
124- Gather
125- Input
126- InstanceNormalization
127- L2Normalization
James Ward590f3f72020-12-02 10:20:48 +0000[diff] [blame]128- Logical
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]129- LogSoftmax
130- Lstm
131- Maximum
132- Mean
133- Merge
134- Minimum
135- Multiplication
136- Normalization
137- Output
138- Pad
139- Permute
140- Pooling2d
141- Prelu
142- Quantize
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]143- QLstm
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]144- QuantizedLstm
James Ward590f3f72020-12-02 10:20:48 +0000[diff] [blame]145- Rank
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]146- Reshape
James Ward590f3f72020-12-02 10:20:48 +0000[diff] [blame]147- Resize
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]148- ResizeBilinear
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]149- Rsqrt
150- Slice
151- Softmax
152- SpaceToBatchNd
153- SpaceToDepth
154- Splitter
155- Stack
156- StandIn
157- StridedSlice
158- Subtraction
159- Switch
160- TransposeConvolution2d
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]161
162More machine learning layers will be supported in future releases.
163
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]164@subsection deserializersupportdeprecated Deprecated layers
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]165
166Some layers have been deprecated and replaced by others layers. In order to maintain backward compatibility, serializations of these deprecated layers will deserialize to the layers that have replaced them, as follows:
167
168- Equal will deserialize as Comparison
169- Merger will deserialize as Concat
170- Greater will deserialize as Comparison
171- ResizeBilinear will deserialize as Resize
172
173**/
174}