Overview ¶

This document lists the layers (or operators) which can be imported and converted. Supported operators allow to address a large range of classical topologies targeting a Mobile or IoT resource-constrained runtime environment: SqueezeNet, MobileNet V1 or V2, Inception, SSD MobileNet v1,…

Purpose of this document is to list the operators and their associated constraints or limitations, please refer to the original documentation for details on a given layer.

Tensorflow Lite is the format used to deploy a neural network model on mobile platforms. ST Edge AI Core imports and converts the .tflite files which is based on the flatbuffer technology. The official ‘ schema.fbs ’ definition ( tags v2.18.0) is used to import the models. A number of operators from the supported operator list are handled, including the quantized models and/or operators generated by the Quantization Aware Training or/and Post-training quantization processes.

This file was automatically generated.

ST Edge AI Core version : 4.0
102 operators found

Summary table ¶

Following table contains the list of the operators that can be imported, if the constraints or limitations are met.

supported optional fused activation (or non-linearity): gelu, linear, relu, quantized_relu, relu_n1_to_1, relu_0_to_1, leaky_relu, relu6, elu, selu, sigmoid, hard_sigmoid, hard_swish, exponential, tanh, softmax, softplus, softsign, abs, acos, acosh, asin, asinh, atan, atanh, ceil, clip, cos, cosh, erf, flexerf, exp, floor, identity, log, logistic, neg, logical_not, prelu, probit, reciprocal, relu_generic, relu_thresholded, round, sign, sin, sinh, softmax_zero, sqrt, swish, tan
supported optional fused integer activation (or non-linearity): prelu, relu, clip, lut, swish, identity, relu6
mixed data operations (i.e hybrid operator) are not natively supported, activations and weights should be quantized
if an operator is not supported in integer, floating point version is used. Converters are automatically added by the code generator.

operator	data types	constraints/limitations
ABS	float32	common
ADD	float32, int8, uint8	common
ADD_N	float32	common
ARG_MAX	float32, int32	common
ARG_MIN	float32, int32	common
ATAN2	float32	common
AVERAGE_POOL_2D	float32, int8, uint8	common, specific
BATCH_MATMUL	float32, int8, uint8	common, specific
BATCH_TO_SPACE_ND	float32, int8, uint8	common
BROADCAST_ARGS	float32, int8, uint8, int32	common
BROADCAST_TO	float32, int8, uint8	common, specific
CAST	bool, int8, uint8, float32	common, specific
CEIL	float32	common
CONCATENATION	float32, int8, uint8	common, specific
CONV_2D	float32, int8, uint8	common, specific
COS	float32	common
DEPTH_TO_SPACE	float32, int8, uint8	common, specific
DEPTHWISE_CONV_2D	float32, int8, uint8	common, specific
DEQUANTIZE	float32, int8, uint8	common
DIV	float32, int8, uint8	common
ELU	float32, int8, uint8	common
EQUAL	float32, bool	common
EXP	float32	common
EXPAND_DIMS	float32, int8, uint8	common
FILL	float32	common
FlexErf	float32	common
FLOOR	float32	common
FLOOR_DIV	float32	common
FLOOR_MOD	float32	common
FULLY_CONNECTED	float32, int8, uint8	common, specific
GATHER	float32, int8, uint8	common, specific
GATHER_ND	float32, int8, uint8	common, specific
GELU	float32	common
GREATER	float32, bool	common
GREATER_EQUAL	float32, bool	common
HARD_SWISH	float32	common
L2_NORMALIZATION	float32	common
LEAKY_RELU	float32, int8, uint8	common
LESS	float32, bool	common
LESS_EQUAL	float32, bool	common
LOCAL_RESPONSE_NORMALIZATION	float32	common
LOG	float32	common
LOG_SOFTMAX	float32	common, specific
LOGICAL_AND	bool	common
LOGICAL_NOT	float32	common
LOGICAL_OR	bool	common
LOGISTIC	float32	common
MAX_POOL_2D	float32, int8, uint8	common, specific
MAXIMUM	float32, int8, uint8	common
MEAN	float32	common
MINIMUM	float32, int8, uint8	common
MIRROR_PAD	float32	common, specific
MUL	float32, int8, uint8	common
NEG	float32	common
NOT_EQUAL	float32	common
PACK	float32, int8, uint8	common, specific
PAD	float32	common, specific
PADV2	float32	common, specific
POW	float32	common
PRELU	float32, int8, uint8	common
QUANTIZE	float32, int8, uint8	common
REDUCE_ANY	float32	common
REDUCE_MAX	float32	common
REDUCE_MIN	float32	common
REDUCE_PROD	float32	common
RELU	float32, int8, uint8	common
RELU6	float32, int8, uint8	common
RELU_0_TO_1	float32, int8, uint8	common
RELU_N1_TO_1	float32, int8, uint8	common
RESHAPE	float32, int8, uint8	common
RESIZE_BILINEAR	float32	common
RESIZE_NEAREST_NEIGHBOR	float32	common
REVERSE_V2	float32, int8, uint8	common, specific
ROUND	float32	common
RSQRT	float32	common
SCATTER_ND	float32, int8, uint8	common, specific
SELECT	float32, int8, uint8, int16, uint16, int32, uint32, bool	common
SELECT_V2	float32, int8, uint8, int16, uint16, int32, uint32, bool	common
SHAPE	float32, int8, uint8, int32	common
SIGN	float32	common
SIN	float32	common
SLICE	float32, int8, uint8	common
SOFTMAX	float32	common, specific
SPACE_TO_BATCH_ND	float32, int8, uint8	common
SPACE_TO_DEPTH	float32, int8, uint8	common
SPLIT	float32, int8, uint8	common, specific
SPLIT_V	float32, int8, uint8	common, specific
SQRT	float32	common
SQUARE	float32	common
SQUARED_DIFFERENCE	float32	common
SQUEEZE	float32, int8, uint8	common
STRIDED_SLICE	float32, int8, uint8	common
SUB	float32, int8, uint8	common
SUM	float32	common
TANH	float32	common
TILE	float32, int8, uint8	common, specific
TOPK_V2	float32, int8, uint8	common
TRANSPOSE	float32, int8, uint8	common, specific
TRANSPOSE_CONV	float32, int8, uint8	common, specific
UNIDIRECTIONAL_SEQUENCE_LSTM	float32, int8	common, specific
UNPACK	float32, int8, uint8	common
ZEROS_LIKE	float32	common

Common constraints ¶

input and output tensors must be not dynamic.
- variable-length batch dimension (i.e. (None,)) is considered as equal to 1
- must not be greater than 6D
- dimension must be in the range [0, 65536[
- batch dimension is partially supported in the axis/axes parameters
data type for the weights/activations tensors must be:
- float32, int8, uint8
- only int32 for the bias tensor is considered
- for some operators, bool and binary types are also supported
1D operator is mapped on the respective 2D operator by adding a singleton dimension on the input: (12,3) -> (12, 1, 3)

Operators ¶

ABS ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32
output data types: float32

ADD ¶

Performs element-wise operation

category: eltwise operator
input data types: float32, int8, uint8
output data types: float32, int8, uint8

ADD_N ¶

Performs element-wise operation

category: eltwise operator
input data types: float32
output data types: float32

ARG_MAX ¶

Computes the indices of the max elements of the input tensor’s element along the provided axis.

category: generic layer
input data types: float32
output data types: int32

ARG_MIN ¶

Computes the indices of the min elements of the input tensor’s element along the provided axis.

category: generic layer
input data types: float32
output data types: int32

ATAN2 ¶

Performs element-wise operation

category: eltwise operator
input data types: float32
output data types: float32

AVERAGE_POOL_2D ¶

Downsamples the input

category: pooling layer
input data types: float32, int8, uint8
output data types: float32, int8, uint8

Specific constraints/recommendations:

arbitrary strides, provided that they are smaller than the input size
arbitrary pool sizes, provided that they are smaller than the input size

BATCH_MATMUL ¶

Fully Connected operation

category: core layer
input data types: float32, int8, uint8
output data types: float32, int8, uint8
fused activations (if present): gelu, linear, relu, quantized_relu, relu_n1_to_1, relu_0_to_1, leaky_relu, relu6, elu, selu, sigmoid, hard_sigmoid, hard_swish, exponential, tanh, softmax, softplus, softsign
integer schemes: weights / activations
- Signed Symmetric / Signed Asymmetric (SSSA)
- Signed Symmetric per channel (or per-axis) / Signed Asymmetric (SSSA_CH)
- Signed Symmetric / Unsigned Asymmetric (SSUA)
- Signed Symmetric per channel (or per-axis) / Unsigned Asymmetric (SSUA_CH)

Specific constraints/recommendations:

Only up to 3D matrix multiplication is supported

BATCH_TO_SPACE_ND ¶

Reshape the batch dimension of a tensor

category: Reshaping operation
input data types: float32, int8, uint8
output data types: float32, int8, uint8

BROADCAST_ARGS ¶

Returns a tensor containing the shape of the input tensor

category: Reshaping operation
input data types: float32, int8, uint8
output data types: int32

BROADCAST_TO ¶

Constructs a tensor by tiling the input tensor

category: reshaping layer
input data types: float32, int8, uint8
output data types: float32, int8, uint8

Specific constraints/recommendations:

tiling on batch-dimension is not supported

CAST ¶

Cast elements of the input tensor to the specified output tensor data

category: conversion layer
input data types: bool, int8, uint8, float32
output data types: bool, int8, uint8, float32

Specific constraints/recommendations:

The attribute saturate different from the default value is not supported

CEIL ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32
output data types: float32

CONCATENATION ¶

Performs concatenation of a list of inputs

category: merge operator
input data types: float32, int8, uint8
output data types: float32, int8, uint8
fused activations (if present): gelu, linear, relu, quantized_relu, relu_n1_to_1, relu_0_to_1, leaky_relu, relu6, elu, selu, sigmoid, hard_sigmoid, hard_swish, exponential, tanh, softmax, softplus, softsign

Specific constraints/recommendations:

concatenating on the batch dimension is not supported

CONV_2D ¶

Performs convolution operation

category: convolutional layer
input data types: float32, int8, uint8
output data types: float32, int8, uint8
fused activations (if present): gelu, linear, relu, quantized_relu, relu_n1_to_1, relu_0_to_1, leaky_relu, relu6, elu, selu, sigmoid, hard_sigmoid, hard_swish, exponential, tanh, softmax, softplus, softsign
integer schemes: weights / activations
- Signed Symmetric / Signed Asymmetric (SSSA)
- Signed Symmetric per channel (or per-axis) / Signed Asymmetric (SSSA_CH)
- Signed Symmetric / Unsigned Asymmetric (SSUA)
- Signed Symmetric per channel (or per-axis) / Unsigned Asymmetric (SSUA_CH)

Specific constraints/recommendations:

arbitrary strides, provided that they are smaller than the input size
arbitrary filter kernel sizes, provided that they are smaller than the input size
dilation factors different from 1 are not supported for int8 model

COS ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32
output data types: float32

DEPTH_TO_SPACE ¶

Permutes the dimensions of the input according to a given pattern

category: reshaping layer
input data types: float32, int8, uint8
output data types: float32, int8, uint8

Specific constraints/recommendations:

transposing the batch dimension is not supported

DEPTHWISE_CONV_2D ¶

Performs convolution operation

category: convolutional layer
input data types: float32, int8, uint8
output data types: float32, int8, uint8
fused activations (if present): gelu, linear, relu, quantized_relu, relu_n1_to_1, relu_0_to_1, leaky_relu, relu6, elu, selu, sigmoid, hard_sigmoid, hard_swish, exponential, tanh, softmax, softplus, softsign
integer schemes: weights / activations
- Signed Symmetric / Signed Asymmetric (SSSA)
- Signed Symmetric per channel (or per-axis) / Signed Asymmetric (SSSA_CH)
- Signed Symmetric / Unsigned Asymmetric (SSUA)
- Signed Symmetric per channel (or per-axis) / Unsigned Asymmetric (SSUA_CH)

Specific constraints/recommendations:

arbitrary strides, provided that they are smaller than the input size
arbitrary filter kernel sizes, provided that they are smaller than the input size
dilation factors different from 1 are not supported for int8 model

DEQUANTIZE ¶

Computes element-wise data conversion low precision to full precision, based on the scale/zeropoint parameters

category: conversion layer
input data types: float32, int8, uint8
output data types: float32, int8, uint8

DIV ¶

Performs element-wise operation

category: eltwise operator
input data types: float32, int8, uint8
output data types: float32, int8, uint8

ELU ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32, int8, uint8
output data types: float32, int8, uint8

EQUAL ¶

Performs logical element-wise operation

category: eltwise operator
input data types: float32
output data types: bool

EXP ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32
output data types: float32

EXPAND_DIMS ¶

Reshapes a tensor

category: Reshaping operation
input data types: float32, int8, uint8
output data types: float32, int8, uint8

FILL ¶

Generates a tensor with given value and shape

category: constant layer
input data types: float32
output data types: float32

FlexErf ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32
output data types: float32

FLOOR ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32
output data types: float32

FLOOR_DIV ¶

Performs element-wise operation

category: eltwise operator
input data types: float32
output data types: float32

FLOOR_MOD ¶

Performs element-wise operation

category: eltwise operator
input data types: float32
output data types: float32

FULLY_CONNECTED ¶

Fully Connected operation

category: core layer
input data types: float32, int8, uint8
output data types: float32, int8, uint8
fused activations (if present): gelu, linear, relu, quantized_relu, relu_n1_to_1, relu_0_to_1, leaky_relu, relu6, elu, selu, sigmoid, hard_sigmoid, hard_swish, exponential, tanh, softmax, softplus, softsign
integer schemes: weights / activations
- Signed Symmetric / Signed Asymmetric (SSSA)
- Signed Symmetric per channel (or per-axis) / Signed Asymmetric (SSSA_CH)
- Signed Symmetric / Unsigned Asymmetric (SSUA)
- Signed Symmetric per channel (or per-axis) / Unsigned Asymmetric (SSUA_CH)

Specific constraints/recommendations:

Only up to 3D matrix multiplication is supported

GATHER ¶

Gathers values along a specified axis

category: activation function
input data types: float32, int8, uint8
output data types: float32, int8, uint8

Specific constraints/recommendations:

Gather is not supported with indices dimensions > 2 (Batch is not considered), axis > 3and axis = 0, batch_dims attribute is not handled

GATHER_ND ¶

Gathers slices from input tensor into an output tensor with shape specified by indices

category: activation function
input data types: float32, int8, uint8
output data types: float32, int8, uint8

Specific constraints/recommendations:

Batch_dims attribute is not handled. 4D or more indices (BATCH is not included) are not implementedLast dimension of indices > 4 with 2D inputs tensor case is not handled

GELU ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32
output data types: float32

GREATER ¶

Performs logical element-wise operation

category: eltwise operator
input data types: float32
output data types: bool

GREATER_EQUAL ¶

Performs logical element-wise operation

category: eltwise operator
input data types: float32
output data types: bool

HARD_SWISH ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32
output data types: float32

L2_NORMALIZATION ¶

Apply Lp-normalization along the provided axis

category: normalization function
input data types: float32
output data types: float32

LEAKY_RELU ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32, int8, uint8
output data types: float32, int8, uint8

LESS ¶

Performs logical element-wise operation

category: eltwise operator
input data types: float32
output data types: bool

LESS_EQUAL ¶

Performs logical element-wise operation

category: eltwise operator
input data types: float32
output data types: bool

LOCAL_RESPONSE_NORMALIZATION ¶

Apply Local Response Normalization over local input regions

category: normalization function
input data types: float32
output data types: float32

LOG ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32
output data types: float32

LOG_SOFTMAX ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32
output data types: float32

Specific constraints/recommendations:

It is supported only for 1D tensor and only on the channel dimension

LOGICAL_AND ¶

Performs boolean element-wise operation

category: eltwise operator
input data types: bool
output data types: bool

LOGICAL_NOT ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32
output data types: float32

LOGICAL_OR ¶

Performs boolean element-wise operation

category: eltwise operator
input data types: bool
output data types: bool

LOGISTIC ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32
output data types: float32

MAX_POOL_2D ¶

Downsamples the input

category: pooling layer
input data types: float32, int8, uint8
output data types: float32, int8, uint8

Specific constraints/recommendations:

arbitrary strides, provided that they are smaller than the input size
arbitrary pool sizes, provided that they are smaller than the input size

MAXIMUM ¶

Computes the maximum (element-wise) a list of inputs

category: eltwise operator
input data types: float32, int8, uint8
output data types: float32, int8, uint8

MEAN ¶

Computes the Mean of the input tensor’s element along the provided axes

category: reduction operation
input data types: float32
output data types: float32

MINIMUM ¶

Computes the minimum (element-wise) a list of inputs

category: eltwise operator
input data types: float32, int8, uint8
output data types: float32, int8, uint8

MIRROR_PAD ¶

Pads an input tensor

category: Reshaping layer
input data types: float32
output data types: float32

Specific constraints/recommendations:

padding ‘edge’ is not supported

MUL ¶

Performs element-wise operation

category: eltwise operator
input data types: float32, int8, uint8
output data types: float32, int8, uint8

NEG ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32
output data types: float32

NOT_EQUAL ¶

Performs element-wise operation

category: eltwise operator
input data types: float32
output data types: float32

PACK ¶

Packs a list of tensors into a tensor along a specified axis

category: merge operator
input data types: float32, int8, uint8
output data types: float32, int8, uint8

Specific constraints/recommendations:

related TF operator: tf.stack

PAD ¶

Pads an input tensor

category: Reshaping layer
input data types: float32
output data types: float32

Specific constraints/recommendations:

padding ‘edge’ is not supported

PADV2 ¶

Pads an input tensor

category: Reshaping layer
input data types: float32
output data types: float32

Specific constraints/recommendations:

padding ‘edge’ is not supported

POW ¶

Performs element-wise operation

category: eltwise operator
input data types: float32
output data types: float32

PRELU ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32, int8, uint8
output data types: float32, int8, uint8

QUANTIZE ¶

Computes element-wise data conversion full precision to low precision, based on the scale/zeropoint parameters

category: conversion layer
input data types: float32, int8, uint8
output data types: float32, int8, uint8

REDUCE_ANY ¶

Computes the logical ‘or’ of elements across dimensions of a tensor

category: reduction operation
input data types: float32
output data types: float32

REDUCE_MAX ¶

Computes the Max of the input tensor’s element along the provided axes

category: reduction operation
input data types: float32
output data types: float32

REDUCE_MIN ¶

Computes the Min of the input tensor’s element along the provided axes

category: reduction operation
input data types: float32
output data types: float32

REDUCE_PROD ¶

Computes the Product of the input tensor’s element along the provided axes

category: reduction operation
input data types: float32
output data types: float32

RELU ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32, int8, uint8
output data types: float32, int8, uint8

RELU6 ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32, int8, uint8
output data types: float32, int8, uint8

RELU_0_TO_1 ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32, int8, uint8
output data types: float32, int8, uint8

RELU_N1_TO_1 ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32, int8, uint8
output data types: float32, int8, uint8

RESHAPE ¶

Reshapes a tensor

category: Reshaping operation
input data types: float32, int8, uint8
output data types: float32, int8, uint8

RESIZE_BILINEAR ¶

Resize input tensor using bilinear interpolation

category: resizing operation
input data types: float32
output data types: float32

RESIZE_NEAREST_NEIGHBOR ¶

Resize input tensor using nearest interpolation mode

category: resizing operation
input data types: float32
output data types: float32

REVERSE_V2 ¶

Reverses specific dimensions of a tensor

category: Reshape layer
input data types: float32, int8, uint8
output data types: float32, int8, uint8

Specific constraints/recommendations:

reversing the batch dimension is not supported
support more than 1 axis is not supported

ROUND ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32
output data types: float32

RSQRT ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32
output data types: float32

SCATTER_ND ¶

Scatters updates into a tensor of shape equal to shape attribute according indices (TFLite),The output of the ScatterND layer is produced by creating a copy of the input data, and then updating its values to values specified by updates at specific index positions specified by indices (ONNX)

category: activation function
input data types: float32, int8, uint8
output data types: float32, int8, uint8

Specific constraints/recommendations:

5D or more indices (BATCH is not included) are not implemented, 5D or more data (BATCH is not included) are not implemented

SELECT ¶

Where layer

category: generic layer
input data types: float32, int8, uint8, int16, uint16, int32, uint32, bool
output data types: float32, int8, uint8, int16, uint16, int32, uint32, bool

SELECT_V2 ¶

Where layer

category: generic layer
input data types: float32, int8, uint8, int16, uint16, int32, uint32, bool
output data types: float32, int8, uint8, int16, uint16, int32, uint32, bool

SHAPE ¶

Returns a tensor containing the shape of the input tensor

category: Reshaping operation
input data types: float32, int8, uint8
output data types: int32

SIGN ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32
output data types: float32

SIN ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32
output data types: float32

SLICE ¶

Crops the input

category: reshaping layer
input data types: float32, int8, uint8
output data types: float32, int8, uint8

SOFTMAX ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32
output data types: float32

Specific constraints/recommendations:

It is supported only for 1D tensor and only on the channel dimension

SPACE_TO_BATCH_ND ¶

Divides spatial dimensions

category: Reshaping operation
input data types: float32, int8, uint8
output data types: float32, int8, uint8

SPACE_TO_DEPTH ¶

Rearranges blocks of spatial data into depth

category: reshaping layer
input data types: float32, int8, uint8
output data types: float32, int8, uint8

SPLIT ¶

Splits a tensor into a list of sub tensors

category: split operator
input data types: float32, int8, uint8
output data types: float32, int8, uint8

Specific constraints/recommendations:

Only supported if the number of splits is equal to the size of the splitting dimension

SPLIT_V ¶

Splits a tensor into a list of sub tensors

category: split operator
input data types: float32, int8, uint8
output data types: float32, int8, uint8

Specific constraints/recommendations:

Only supported if the number of splits is equal to the size of the splitting dimension

SQRT ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32
output data types: float32

SQUARE ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32
output data types: float32

SQUARED_DIFFERENCE ¶

Compute (x-y)(x-y)

category: eltwise operator
input data types: float32
output data types: float32

SQUEEZE ¶

Reshapes a tensor

category: Reshaping operation
input data types: float32, int8, uint8
output data types: float32, int8, uint8

STRIDED_SLICE ¶

Crops the input

category: reshaping layer
input data types: float32, int8, uint8
output data types: float32, int8, uint8

SUB ¶

Performs element-wise operation

category: eltwise operator
input data types: float32, int8, uint8
output data types: float32, int8, uint8

SUM ¶

Computes the Sum of the input tensor’s element along the provided axes

category: reduction operation
input data types: float32
output data types: float32

TANH ¶

Applies an activation function to the input tensor

category: activation layer
input data types: float32
output data types: float32

TILE ¶

Constructs a tensor by tiling the input tensor

category: reshaping layer
input data types: float32, int8, uint8
output data types: float32, int8, uint8

Specific constraints/recommendations:

tiling on batch-dimension is not supported

TOPK_V2 ¶

Retrieve the top-K largest or smallest elements along a specified axis

category: topK operator
input data types: float32, int8, uint8
output data types: float32, int8, uint8

TRANSPOSE ¶

Permutes the dimensions of the input according to a given pattern

category: reshaping layer
input data types: float32, int8, uint8
output data types: float32, int8, uint8

Specific constraints/recommendations:

transposing the batch dimension is not supported

TRANSPOSE_CONV ¶

Transposed convolutional layer

category: convolutional layer
input data types: float32, int8, uint8
output data types: float32, int8, uint8

Specific constraints/recommendations:

arbitrary strides, provided that they are smaller than the input size
arbitrary filter kernel sizes, provided that they are smaller than the input size

UNIDIRECTIONAL_SEQUENCE_LSTM ¶

Computes a multi-layer long short-term memory (LSTM) RNN to an input sequence (batch=1, timesteps, features)

category: recurrent layer
input data types: float32, int8
output data types: float32, int8

Specific constraints/recommendations:

stateless mode support only
fused activation: sigmoid
fused recurrent activation: sigmoid
return_state not supported
time_major not supported

UNPACK ¶

Unpacks num tensors from values along specified axis

category: split operator
input data types: float32, int8, uint8
output data types: float32, int8, uint8

ZEROS_LIKE ¶

Generates a tensor with values 0

category: zeros like
input data types: float32
output data types: float32