Rockchip · AI & Linux Hub

AI inference

RKNN2 V2.5.0 · RV1103B/RV1106B：Basics and convolution

RKNN2 V2.5.0 · RV1103B/RV1106B：Basics and convolution · Complete operator parameters and constraints from the V2.5.0 source.

· ArmSoM Editorial · Rockchip source documentation · 7 min read

This RKNN2 V2.5.0 platform reference helps ArmSoM developers select and adapt models. It does not establish support on every ArmSoM board: verify the board SoC, actual Toolkit2, Runtime, NPU driver and BSP versions. The source specifies no universal Linux kernel version. No ArmSoM hardware validation was performed.

Source: the supplied Chinese and English RKNN2 Supported Operators and Constraints Reference V2.5.0, 374 and 384 pages respectively. Chinese follows the Chinese edition; English, Russian and German follow the English edition. Platform constraints remain separate. The English PDF itself contains Chinese I/O figures and output tables; their provenance is retained with editable localized tables. Source spellings such as ConvTranpose and the original LSTM/GRU equations are preserved; these equations are not corrections to SDK implementations.

Source pages: CN 317–324; EN 326–334.

Rockchip RKNN2

Operator reference series

RKNN2 V2.5.0 · RK3566/RK3568：Basics and convolution

RKNN2 V2.5.0 · RK3566/RK3568：Recurrent networks, arithmetic and pooling

RKNN2 V2.5.0 · RK3566/RK3568：Shapes, slicing and transpose

RKNN2 V2.5.0 · RK3566/RK3568：Activations, normalization and extensions

RKNN2 V2.5.0 · RK3588：Basics and convolution

RKNN2 V2.5.0 · RK3588：Recurrent networks, arithmetic and pooling

RKNN2 V2.5.0 · RK3588：Shapes, slicing and transpose

RKNN2 V2.5.0 · RK3588：Activations, normalization and extensions

RKNN2 V2.5.0 · RK3562：Basics and convolution

RKNN2 V2.5.0 · RK3562：Recurrent networks, arithmetic and pooling

RKNN2 V2.5.0 · RK3562：Shapes, slicing and transpose

RKNN2 V2.5.0 · RK3562：Activations, normalization and extensions

RKNN2 V2.5.0 · RV1103/RV1106：Basics and convolution

RKNN2 V2.5.0 · RV1103/RV1106：Recurrent networks, arithmetic and pooling

RKNN2 V2.5.0 · RV1103/RV1106：Shapes, slicing and transpose

RKNN2 V2.5.0 · RV1103/RV1106：Activations, normalization and extensions

RKNN2 V2.5.0 · RK3576/RV1126B：Basics and convolution

RKNN2 V2.5.0 · RK3576/RV1126B：Recurrent networks, arithmetic and pooling

RKNN2 V2.5.0 · RK3576/RV1126B：Shapes, slicing and transpose

RKNN2 V2.5.0 · RK3576/RV1126B：Activations, normalization and extensions

RKNN2 V2.5.0 · RK2118：Basics and convolution

RKNN2 V2.5.0 · RK2118：Recurrent networks, arithmetic and pooling

RKNN2 V2.5.0 · RK2118：Shapes, slicing and transpose

RKNN2 V2.5.0 · RK2118：Activations, normalization and extensions

RKNN2 V2.5.0 · RV1103B/RV1106B：Basics and convolution

RKNN2 V2.5.0 · RV1103B/RV1106B：Recurrent networks, arithmetic and pooling

RKNN2 V2.5.0 · RV1103B/RV1106B：Shapes, slicing and transpose

RKNN2 V2.5.0 · RV1103B/RV1106B：Activations, normalization and extensions

RKNN2 V2.5.0 · CPU operator support and constraints

RKNN2 V2.5.0 · GPU operator support and constraints

RKNN2 V2.5.0 · Model I/O, alignment and convolution limits

RV1103B/RV1106B

Add

Performs element-wise binary addition(C=A+B),supporting multi-directional NumPy-style broadcasting

Inputs

  • A:T1,T2
    • shape:[batch,channel,height,width]
    • Quantization support
      • per-layer
    • Broadcasting support
      • Supports broadcasting along arbitrary dimensions
      • Broadcasting constraints are the same as the Expand operator
  • B:T1,T2
    • shape:[batch,channel,height,width]
    • Quantization support
      • per-layer
    • Broadcasting support
      • Supports broadcasting along arbitrary dimensions
      • Broadcasting constraints are the same as the Expand operator

Outputs

  • C:T1
    • shape:[batch,channel,height,width]
    • Quantization support
      • per-layer

Data Type Constraints

  • T1:int8(Tensor)
  • T2:float(constTensor)

AddRelu

Add and Relu fused computation, C=Relu(A+B)

Inputs

  • Same as Add

Outputs

  • Same as Add

Other Support

  • Multi-core: Supported

BatchNormalization

Performs batch normalization on the input tensor according to the description in https://arxiv.org/abs/1502.03167

Inputs

  • input:T1
    • shape:[batch,channel,height,width]
    • Quantization support
      • per-layer

Outputs

  • output:T1
    • shape:[batch,channel,height,width]
    • Quantization support
      • per-layer

Data Type Constraints

  • T1:int8(Tensor)

Attributes

  • epsilon: float (default value = 1e-5)
    • A real number added when dividing by the standard deviation to prevent division by 0
    • epsilon∈(0,∞]
  • momentum:float
    • Sliding average parameters during training

Concat

Concatenate a sequence of vectors into a single vector along the axis specified by axis. All vectors must be the same size except in the axis dimension

Inputs

  • input_tensor:T
    • shape:[batch,channel,height,width]
    • Quantization support
      • per-layer

Outputs

  • output:T
    • shape:[batch,channel,height,width]
    • Quantization support
      • per-layer

Data Type Constraints

  • T:int8(Tensor)

Attributes

  • axisint64
    • axis∈{0,1,2,3}

Convolution

Convolution

Inputs

  • input_tensor:T1
    • shape:[batch,channel,height,width]
      • width: when dilated_kernel_h>1, width<16383. In addition, the first-layer input width is restricted; see the model input description for details.
    • Quantization support
      • per-layer
  • weight:T1
    • shape:[output_channel,input_channel,kernel_h,kernel_w]
      • kernel_h∈[1,31]
      • kernel_w∈[1,31]
    • Quantization support
      • per-layer
      • per-channel

Outputs

  • output:T1
    • shape:[batch,channel,height,width]
    • Quantization support
      • per-layer

Attributes

  • strides:int(strides_h,strides_w)
    • stride_h∈[1,7]
    • stride_w∈[1,7]
  • pads:int(pads_top,pads_left,pads_bottom,pads_right)
    • pads_top∈[0,15]
    • pads_left∈[0,15]
    • pads_bottom∈[0,15]
    • pads_right∈[0,15]
  • group:int
  • dilations:int(dilations_h,dilations_w)
    • dilations_h∈[1,31]
    • dilations_w∈[1,31]

Data Type Constraints

  • T1:int8(Tensor)

Depthwise Convolution

Depthwise separable convolution

Inputs

  • input_tensor:T1
    • shape:[batch,channel,height,width]
      • width: when dilated_kernel_h>1, width<16383. In addition, the first-layer input width is restricted; see the model input description for details.
    • Quantization support
      • per-layer
  • weight:T1
    • shape:[output_channel,input_channel,kernel_h,kernel_w]
      • kernel_h∈[1,8]
      • kernel_w∈[1,8]
    • Quantization support
      • per-layer
      • per-channel

Outputs

  • output:T1
    • shape:[batch,channel,height,width]
    • Quantization support
      • per-layer

Attributes

  • strides:int(strides_h,strides_w)
    • stride_h∈[1,7]
    • stride_w∈[1,7]
  • pads:int(pads_top,pads_left,pads_bottom,pads_right)
    • pads_top∈[0,15]
    • pads_left∈[0,15]
    • pads_bottom∈[0,15]
    • pads_right∈[0,15]
  • group:int
  • dilations:int(dilations_h,dilations_w)
    • dilations_h∈[1,32]
    • dilations_w∈[1,32]

Data Type Constraints

  • T1:int8(Tensor)

ConvolutionRelu

Convolution and Relu fused computation, Output=Relu(Conv(Input)). Description and specification constraints are the same as Convolution

ConvolutionClip

Convolution and Clip fused computation, Output=Clip(Conv(Input)). Description and specification constraints are the same as Convolution

ConvolutionPRelu/LeakyRelu

Convolution and PRelu/LeakyRelu fused computation, Output = PRelu(Conv(Input)) or Output = LeakyRelu(Conv(Input)). Description and specification constraints are the same as Convolution

ConvolutionAdd

Convolution and Add fused computation, Output=Add(Conv(Input0),Input1). Description and specification constraints are the same as Convolution

ConvolutionSigmoid

Convolution and Sigmoid fused computation, Output=Sigmoid(Conv(Input)). Description and specification constraints are the same as Convolution

ConvolutionTanh

Convolution and Tanh fused computation, Output=Tanh(Conv(Input)). Description and specification constraints are the same as Convolution

ConvolutionSoftplus

Convolution and Softplus fused computation, Output=Softplus(Conv(Input)). Description and specification constraints are the same as Convolution

ConvolutionHardSigmoid

Convolution and HardSigmoid fused computation, Output=HardSigmoid(Conv(Input)). Description and specification constraints are the same as Convolution

ConvolutionHardSwish

Convolution and HardSwish fused computation, Output=HardSwish(Conv(Input)). Description and specification constraints are the same as Convolution

ConvolutionElu

Convolution and Elu fused computation, Output=Elu(Conv(Input)). Description and specification constraints are the same as Convolution

ConvolutionSwish

Convolution and Swish fused computation, Output=Swish(Conv(Input)). Description and specification constraints are the same as Convolution

ConvolutionMish

Convolution and Mish fused computation, Output=Mish(Conv(Input)). Description and specification constraints are the same as Convolution

ConvTranspose

Transposed convolution

Inputs

  • input_tensor:T1
    • shape:[batch,channel,height,width]
      • width: when dilated_kernel_h>1, width<16383. In addition, the first-layer input width is restricted; see the model input description for details.
  • Quantization support
    • per-layer
  • weight:T1
    • shape:[output_channel,input_channel,kernel_h,kernel_w]
      • kernel_h∈[1,31]
      • kernel_w∈[1,31]
    • Quantization support
      • per-layer
      • per-channel

Outputs

  • output:T1
    • shape:[batch,channel,height,width]
    • Quantization support
      • per-layer

Attributes

  • strides:int(strides_h,strides_w)
    • stride_h∈[1,8]
    • stride_w∈[1,8]
  • pads:int(pads_top,pads_left,pads_bottom,pads_right) Note when setting pad: kernel_hdilations_h-dilations_h-pads_top<0 is not supported; kernel_wdilations_w- dilations_w-pads_left<0 is not supported; stride_h(height-1)-pads_top+1<output_h is not supported; stride_w(width-1)-pads_left+1<output_w is not supported
    • pads_top∈[0,15]
    • pads_left∈[0,15]
    • pads_bottom∈[0,15]
    • pads_right∈[0,15]
  • group:int
    • group: only 1 is supported; num_output is supported if and only if num_input=num_output
  • dilations:int(dilations_h,dilations_w)
    • dilations_h∈[1,32]
    • dilations_w∈[1,32]
    • When dilation is not equal to (1,1), only kernel=(1,1) is supported

Data Type Constraints

  • T1:int8(Tensor)

ConvTranposeRelu

ConvTranspose and Relu fused computation, Output=Relu(ConvTranspose(Input)). Description and specification constraints are the same as ConvTranspose

ConvTranposeClip

ConvTranspose and Clip fused computation, Output=Clip(ConvTranspose(Input)). Description and specification constraints are the same as ConvTranspose

ConvTranposePRelu/LeakyRelu

ConvTranspose and PRelu/LeakyRelu fused computation, Output=PRelu(ConvTranspose(Input)) or Output= LeakyRelu(ConvTranspose(Input)). Description and specification constraints are the same as ConvTranspose

Related articles

Practical guideHands-on

RK3572: 8nm, eight cores and 4 TOPS AI — ArmSoM plans a Zero-form-factor board

Rockchip RK3572 specifications and ArmSoM’s planned Raspberry Pi Zero-form-factor product, with clear limits on power figures and board compatibility.

Read guide →
Practical guideHands-on

RKNN-Toolkit2 Python API V2.5.0 — Runtime, operator registration and inference

RKNN-Toolkit2 Python API V2.5.0 — Runtime, operator registration and inference | Complete parameters, examples and platform constraints

Read guide →
Practical guideHands-on

RKNN-Toolkit2 Python API V2.5.0 — Requirements and object lifecycle

RKNN-Toolkit2 Python API V2.5.0 — Requirements and object lifecycle | Complete parameters, examples and platform constraints

Read guide →

Explore this topic → · Learning paths →