Gitiles
Code Review Sign In
review.mlplatform.org / ml / armnn / a0f8b15d4ddb5075f380003ff31b271d389d3b66 / . / docs / 05_02_deserializer_serializer.dox
blob: 6cfaf29968b36e48fa2af411e21034f1e9c74133 [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
Matthew Sloyanb63a3112021-09-08 13:05:51 +0100[diff] [blame]33- Convolution3d
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]34- DepthToSpace
35- DepthwiseConvolution2d
36- Dequantize
37- DetectionPostProcess
38- Division
39- ElementwiseUnary
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]40- Fill
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]41- Floor
42- FullyConnected
43- Gather
Teresa Charlin6966bfa2022-04-25 17:14:50 +0100[diff] [blame]44- GatherNd
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]45- Input
46- InstanceNormalization
47- L2Normalization
James Ward590f3f72020-12-02 10:20:48 +0000[diff] [blame]48- Logical
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]49- LogSoftmax
50- Lstm
51- Maximum
52- Mean
53- Merge
54- Minimum
55- Multiplication
56- Normalization
57- Output
Matthew Sloyan2e5d0b22021-10-21 14:05:31 +0100[diff] [blame]58- Pad (Constant, Symmetric, Reflect)
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]59- Permute
60- Pooling2d
Tamas Nyirid998a1c2021-11-05 14:55:33 +0000[diff] [blame]61- Pooling3d
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]62- Prelu
James Ward590f3f72020-12-02 10:20:48 +0000[diff] [blame]63- QLstm
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]64- Quantize
65- QuantizedLstm
James Ward590f3f72020-12-02 10:20:48 +0000[diff] [blame]66- Rank
Jan Eilers1121c6b2021-02-19 09:38:03 +0000[diff] [blame]67- Reduce
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]68- Reshape
69- Resize
Keith Davis3ae3f972021-05-21 16:33:48 +0100[diff] [blame]70- Shape
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]71- Slice
72- Softmax
73- SpaceToBatchNd
74- SpaceToDepth
75- Splitter
76- Stack
77- StandIn
78- StridedSlice
79- Subtraction
80- Switch
James Ward590f3f72020-12-02 10:20:48 +0000[diff] [blame]81- Transpose
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]82- TransposeConvolution2d
Narumol Prangnawarata0162e12021-07-23 14:47:49 +0100[diff] [blame]83- UnidirectionalSequenceLstm
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]84
85More machine learning layers will be supported in future releases.
86
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]87@subsection serializersupportdeprecated Deprecated layers
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]88
89Some 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:
90
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]91- Abs will deserialize as ElementwiseUnary
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]92- Equal will deserialize as Comparison
93- Greater will deserialize as Comparison
94- Merger will deserialize as Concat
95- ResizeBilinear will deserialize as Resize
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]96- Rsqrt will deserialize as ElementwiseUnary
97<br/><br/><br/><br/>
98
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]99@page deserializer Deserializer
100@tableofcontents
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]101
102The `armnnDeserializer` is a library for loading neural networks defined by Arm NN FlatBuffers files
103into the Arm NN runtime.
104
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]105@section deserializersupport Supported Layers
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]106
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]107This 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]108
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]109@subsection deserializersupportfully Fully supported
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]110
111The Arm NN SDK Deserialize parser currently supports the following layers:
112
113- Abs
114- Activation
115- Addition
116- ArgMinMax
117- BatchToSpaceNd
118- BatchNormalization
mathad01b392e982021-04-07 12:07:30 +0100[diff] [blame]119- Cast
Simon Obute51f67772021-09-03 15:50:13 +0100[diff] [blame]120- ChannelShuffle
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]121- Concat
122- Comparison
123- Constant
124- Convolution2d
125- DepthToSpace
126- DepthwiseConvolution2d
127- Dequantize
128- DetectionPostProcess
129- Division
James Ward590f3f72020-12-02 10:20:48 +0000[diff] [blame]130- ElementwiseUnary
131- Fill
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]132- Floor
133- FullyConnected
134- Gather
Teresa Charlin6966bfa2022-04-25 17:14:50 +0100[diff] [blame]135- GatherNd
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]136- Input
137- InstanceNormalization
138- L2Normalization
James Ward590f3f72020-12-02 10:20:48 +0000[diff] [blame]139- Logical
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]140- LogSoftmax
141- Lstm
142- Maximum
143- Mean
144- Merge
145- Minimum
146- Multiplication
147- Normalization
148- Output
149- Pad
150- Permute
151- Pooling2d
Teresa Charlin6966bfa2022-04-25 17:14:50 +0100[diff] [blame]152- Pooling3d
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]153- Prelu
154- Quantize
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]155- QLstm
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]156- QuantizedLstm
James Ward590f3f72020-12-02 10:20:48 +0000[diff] [blame]157- Rank
Jan Eilers1121c6b2021-02-19 09:38:03 +0000[diff] [blame]158- Reduce
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]159- Reshape
James Ward590f3f72020-12-02 10:20:48 +0000[diff] [blame]160- Resize
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]161- ResizeBilinear
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]162- Slice
163- Softmax
164- SpaceToBatchNd
165- SpaceToDepth
166- Splitter
167- Stack
168- StandIn
169- StridedSlice
170- Subtraction
171- Switch
Jan Eilers1121c6b2021-02-19 09:38:03 +0000[diff] [blame]172- Transpose
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]173- TransposeConvolution2d
Narumol Prangnawarata0162e12021-07-23 14:47:49 +0100[diff] [blame]174- UnidirectionalSequenceLstm
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]175
176More machine learning layers will be supported in future releases.
177
Jan Eilers1ed388f2021-01-28 14:45:09 +0000[diff] [blame]178@subsection deserializersupportdeprecated Deprecated layers
Ryan OSheaf3a43232020-02-12 16:15:27 +0000[diff] [blame]179
180Some 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:
181
182- Equal will deserialize as Comparison
183- Merger will deserialize as Concat
184- Greater will deserialize as Comparison
185- ResizeBilinear will deserialize as Resize
186
187**/
188}