Gitiles
Code Review Sign In
review.mlplatform.org / ml / armnn / 51f67776a695c217a32596af806afeeb080f5528 / . / docs / 01_02_deserializer_serializer.dox
blob: 55259ddeb85a3008d8b2307a347b3b12c0ed0419 [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
mathad01b392e982021-04-07 12:07:30 +0100[diff] [blame]27- Cast
Simon Obute51f67772021-09-03 15:50:13 +0100[diff] [blame^]28- ChannelShuffle
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]29- Comparison
30- Concat
31- Constant
32- Convolution2d
33- DepthToSpace
34- DepthwiseConvolution2d
35- Dequantize
36- DetectionPostProcess
37- Division
38- ElementwiseUnary
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]39- Fill
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]40- Floor
41- FullyConnected
42- Gather
43- Input
44- InstanceNormalization
45- L2Normalization
James Ward590f3f72020-12-02 10:20:48 +0000[diff] [blame]46- Logical
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]47- LogSoftmax
48- Lstm
49- Maximum
50- Mean
51- Merge
52- Minimum
53- Multiplication
54- Normalization
55- Output
56- Pad
57- Permute
58- Pooling2d
59- Prelu
James Ward590f3f72020-12-02 10:20:48 +0000[diff] [blame]60- QLstm
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]61- Quantize
62- QuantizedLstm
James Ward590f3f72020-12-02 10:20:48 +0000[diff] [blame]63- Rank
Jan Eilers1121c6b2021-02-19 09:38:03 +0000[diff] [blame]64- Reduce
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]65- Reshape
66- Resize
Keith Davis3ae3f972021-05-21 16:33:48 +0100[diff] [blame]67- Shape
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]68- Slice
69- Softmax
70- SpaceToBatchNd
71- SpaceToDepth
72- Splitter
73- Stack
74- StandIn
75- StridedSlice
76- Subtraction
77- Switch
James Ward590f3f72020-12-02 10:20:48 +0000[diff] [blame]78- Transpose
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]79- TransposeConvolution2d
Narumol Prangnawarata0162e12021-07-23 14:47:49 +0100[diff] [blame]80- UnidirectionalSequenceLstm
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]81
82More machine learning layers will be supported in future releases.
83
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]84@subsection serializersupportdeprecated Deprecated layers
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]85
86Some 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:
87
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]88- Abs will deserialize as ElementwiseUnary
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]89- Equal will deserialize as Comparison
90- Greater will deserialize as Comparison
91- Merger will deserialize as Concat
92- ResizeBilinear will deserialize as Resize
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]93- Rsqrt will deserialize as ElementwiseUnary
94<br/><br/><br/><br/>
95
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]96@page deserializer Deserializer
97@tableofcontents
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]98
99The `armnnDeserializer` is a library for loading neural networks defined by Arm NN FlatBuffers files
100into the Arm NN runtime.
101
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]102@section deserializersupport Supported Layers
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]103
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]104This 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]105
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]106@subsection deserializersupportfully Fully supported
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]107
108The Arm NN SDK Deserialize parser currently supports the following layers:
109
110- Abs
111- Activation
112- Addition
113- ArgMinMax
114- BatchToSpaceNd
115- BatchNormalization
mathad01b392e982021-04-07 12:07:30 +0100[diff] [blame]116- Cast
Simon Obute51f67772021-09-03 15:50:13 +0100[diff] [blame^]117- ChannelShuffle
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]118- Concat
119- Comparison
120- Constant
121- Convolution2d
122- DepthToSpace
123- DepthwiseConvolution2d
124- Dequantize
125- DetectionPostProcess
126- Division
James Ward590f3f72020-12-02 10:20:48 +0000[diff] [blame]127- ElementwiseUnary
128- Fill
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]129- Floor
130- FullyConnected
131- Gather
132- Input
133- InstanceNormalization
134- L2Normalization
James Ward590f3f72020-12-02 10:20:48 +0000[diff] [blame]135- Logical
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]136- LogSoftmax
137- Lstm
138- Maximum
139- Mean
140- Merge
141- Minimum
142- Multiplication
143- Normalization
144- Output
145- Pad
146- Permute
147- Pooling2d
148- Prelu
149- Quantize
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]150- QLstm
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]151- QuantizedLstm
James Ward590f3f72020-12-02 10:20:48 +0000[diff] [blame]152- Rank
Jan Eilers1121c6b2021-02-19 09:38:03 +0000[diff] [blame]153- Reduce
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]154- Reshape
James Ward590f3f72020-12-02 10:20:48 +0000[diff] [blame]155- Resize
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]156- ResizeBilinear
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]157- Slice
158- Softmax
159- SpaceToBatchNd
160- SpaceToDepth
161- Splitter
162- Stack
163- StandIn
164- StridedSlice
165- Subtraction
166- Switch
Jan Eilers1121c6b2021-02-19 09:38:03 +0000[diff] [blame]167- Transpose
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]168- TransposeConvolution2d
Narumol Prangnawarata0162e12021-07-23 14:47:49 +0100[diff] [blame]169- UnidirectionalSequenceLstm
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]170
171More machine learning layers will be supported in future releases.
172
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]173@subsection deserializersupportdeprecated Deprecated layers
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]174
175Some 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:
176
177- Equal will deserialize as Comparison
178- Merger will deserialize as Concat
179- Greater will deserialize as Comparison
180- ResizeBilinear will deserialize as Resize
181
182**/
183}