ArmNN  NotReleased
src/armnnSerializer/SerializerSupport.md
Go to the documentation of this file.
1 # The layers that ArmNN SDK Serializer currently supports.
2 
3 This reference guide provides a list of layers which can be serialized currently by the Arm NN SDK.
4 
5 ## Fully supported
6 
7 The Arm NN SDK Serializer currently supports the following layers:
8 
9 * Activation
10 * Addition
11 * ArgMinMax
12 * BatchToSpaceNd
13 * BatchNormalization
14 * Comparison
15 * Concat
16 * Constant
17 * Convolution2d
18 * DepthToSpace
19 * DepthwiseConvolution2d
20 * Dequantize
21 * DetectionPostProcess
22 * Division
23 * ElementwiseUnary
24 * Floor
25 * FullyConnected
26 * Gather
27 * Input
28 * InstanceNormalization
29 * L2Normalization
30 * LogSoftmax
31 * Lstm
32 * Maximum
33 * Mean
34 * Merge
35 * Minimum
36 * Multiplication
37 * Normalization
38 * Output
39 * Pad
40 * Permute
41 * Pooling2d
42 * Prelu
43 * Quantize
44 * QuantizedLstm
45 * Reshape
46 * Resize
47 * ResizeBilinear
48 * Slice
49 * Softmax
50 * SpaceToBatchNd
51 * SpaceToDepth
52 * Splitter
53 * Stack
54 * StandIn
55 * StridedSlice
56 * Subtraction
57 * Switch
58 * TransposeConvolution2d
59 
60 More machine learning layers will be supported in future releases.
61 
62 ## Deprecated layers
63 
64 Some 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:
65 
66 * Equal will deserialize as Comparison
67 * Merger will deserialize as Concat
68 * Greater will deserialize as Comparison
69 * ResizeBilinear will deserialize as Resize
70 * Abs will deserialize as ElementwiseUnary
71 * Rsqrt will deserialize as ElementwiseUnary
72